#Building a NAV/COM radio
1 messages · Page 2 of 1
4 layer isn't much different from 2 layer. The process is the same.
The above mentioned 4layer stack-up just allows you to completely forget about VCC and GND, because VCC and GND is just EVERYWHERE. Instead you've got much more real-estate for signal traces...
Definitely worth giving a try at some point☺️
I have plans for another project maybe next year.
Does the company you order the acrylic panels from do the painting and laser etching?
Yes. I order from hispapanels.com in Spain, who have panels for 737 and a320, but I know there are some other sites with a larger portfolio, too... just with customs and shipping Spain works best for me...
So to shorten a long story: they have the panels I need ready and I build my project around that. They're not built to match my design, nor are they 100% replicas of the real thing. They're bloody close in most cases, tho
I'm loving this thread! Feels like a couple of guys each tinkering with their own project while chatting about all sorts of semi-related things...
I just hope @river river isn't annoyed by me hijacking his thread over and over...!😅🙈
@mattlockr No sweat. Every little detour is just another learning opportunity for me.
yeah this has turned into a fun chat of its own
Just don't let perfect become the enemy of good enough. Often it is better to get something done when you need it, rather than spending ages on optimizations
You should tell me that few years ago when I started to build first device 🤣
Took me a while to develop that as well.
it helps to have no idea what you are doing at first, you dont know what your design is missing 😄
True, but then you start watching Robert Ferenec and John Teel videos on YouTube because, well, the algorithm... and all of the sudden you start worrying about sh*t you didn't even know existed only a few hours ago like signal to noise ratios, EMI, power consumption, ESR, trace inductance etc etc etc 😅
All because you googled whether to use a switching regulator or an LDO for your 3.3V supply🤣🤣🤣
yeah 🙂
I have a friend who does super tiny stuff on pcbs and knows a lot about those things
good way to learn
he taught me to use the "single layer view" on kicad
so you can see better how to route, sometimes its harder to see with both layers visible
You mean dimming the inactive layers?
yeah, there is the mode
Yes, and it's the best thing in the world, especially when laying out boards with more than 2 layers!
like here in hindsight...
might have routed that one trace on top layer and avoided that giant multi-lane motorway curve
but they are just button pins
so no big deal
it just looks silly 🙂
but this basically goes around the RP2040 layout I used as a starting point
those moments are worst when you realize the whole thing would work better if you rotated the microcontroller chip 180 degrees ....
but a lot of the pcb designs seem to be 1000 compromises laid out on a piece of circuit board
100%. Even te reference designs have compromises
Rpi foundation mentions them
I haven't yet laid out an RP2040 design on 2 layers... but I imagine it being a bit restrictive.
Yes, the wide trace on the top layer could have gone else where and that would have avoided all this vias, but I've seen worse (some of it in my own, early work!).
Something that really helped my layout strategies was the fact I made my first PCBs myself. Drilling 1.76billion vias and riveting then all was not desired so I spent a lot of time trying to avoid them where possible😅
I agree. A lot of the time you really only have the least worse option to work with, but not the ideal one...
And then comes Tuomas with info about 2000 missing tiny details and the entire roof falls down 😃
Here is technically NOT an electronic PCB board having to deal with "signal to noise ratios, EMI, power consumption, ESR, trace inductance etc etc etc" but merely a board to connect an MCU to a few switches and a couple of displays without wires.
The more difficult routing was between the MAX7219 pins and the 5-digit displays. So many criss-crossing paths for the digits to connect to the associated pin on the Max.
To begin with, there were approximately 250 Freerouting runs. Then once I started to understand manual routing then came the rabbit hole of the what-ifs until I just had to settle on what I have.
The DRC runs are now only giving the silkscreen warnings and now on to the breadboard prototype to see if it actually works.
Here is technically NOT an electronic PCB board having to deal with "signal to noise ratios, EMI, power consumption, ESR, trace inductance etc etc etc" but merely a board to connect an MCU to a few switches and a couple of displays without wires.
This is what most of our boards really are
this is pretty darn nice first impression, didnt look too deeply into the connections etc
good to also verify with the 3d view (and admire how nice that usually looks!) to see that everything is on the correct, intended side
This!
I have to agree... the routing looks perfectly fine. I think there isn't much that could've been done better. You've even intuitively chosen arduino pins for your encoders that make routing straight forward and darn good looking! Nice work!
looks like this has been a good thread also for @river river 🙂
It was a fun challenge for a noobie and my architectural design experience rather than the small amount of electronics knowledge had some influence.
THese are the 3D views showing the required locations of the front components based on the existing faceplate holes.
Then the back is where are the other stuff needed to be fit between all the front stuff.
I am only hoping this is going to work as intended when I get the prototype finished.
This thread has been educational and the help received here from all willing to contribute was invaluable.
It is still not over yet until the board comes back from the fabricators and connects as it should.
I never could find a model of the 5-digit displays to add to this.
you are doing the pushbuttons with smd version on purpose?
as long as you check that the pinout of the 5 digit module matches your wiring
sometimes they might differ a bit by manufacturer
SMD buttons will be my first attempt at SMT soldering.
Gotta start somewhere.
Hopefully the info for these symbols in the schematic is correct.
And that I got them all connected per the diagram.
The Arduino pin connections were cut and pasted in the schematic many, many times.
Thanks for the compliments.
The breadboard connections are completed for the COM2 side, both Active and Standby Freqs. and trying now to connect to MF. No issues with the Dual encoder and the swap button, but the explanation for connecting Active and Standby displays in MF doesn't exist. Also, the instructions for a single display is specific to MSFS and the XPlane options for the freqs. are confusing.
I am not sure where to even begin to explain where I am failing in understanding the config. setup.
This is what I have done...
Created the Radio Config under modules and uploaded.
Assigned DIN, CS and CLK to the Nano pins (Selected 2 under NUM which is unclear to me)
Uploaded config
Created COM2 Radio Output Config.
Edited the Config, SIM Variable set to Xplane Dataref (not sure what Preset code to use yet)
--- MF showed MSFS preset as COM Active Frequency:Index---
Xplane Dateref preset does not have the same named preset.
Skipped over <Modify> and went to <Display> tab.
Set <Output Device> --- Mobiflight Nano / SN-d6a-7fc_ Display Module for "use Type of".
Name/Number set to (COM2 Radio / 2)
And now it's at Values <Number of Digits>
I know that it will be 5 digits for each of the displays.
As I get into this should there be two separate Configs, one for the Active Display and one for the Standby display?
But with them being daisy chained that doesnt make sense either.
Where am I getting confused?
Yes, two output configs bound to different set of digits.
In the Max7219 hardware config, Num refers to how many Max chips are daisy chained.
So I got the num 2 correct....two max chips daisy chained but each one has it's own display and config.
So back to the drawing board...COM2 Active and COM2 Standy configs.
But I am confused about the pins.
There is one set of pins for the daisy chain but each separate config needs a set of unique pins.
I can't get my head wrapped around this.
So under the MF Modules page there is one Radio config with the 3 pins and number of max chips assigned?
Then create two Output configs (Active and Standby)...
If that part is correct, then the next part about selecting the Frequency presets is throwing me.
Xplane has freqs separated into the MHz and KHz presets.
I do not see how the MHz freq. is assigned to the left 3 digits and the KHz is assigned to the right two digits.
yea and in the output configs is a selector for the max module
you can first just set all digits on and turn on all checkboxes and test that you get numbers
it takes a bit of testing to understand how it all works
but basically use the Test toolbar on the bottom of the display tab to set for example 12345678
and see what digit gets what
and how it all works
and whether you get numbers
so you know the wiring is correct
Trial and error
then once you have the test working, you add a simvar, and it basically just supplies your value instead of the "test" field
what comes from simulator is always in "Flight Sim Value", then you use Modify tab and modifiers to format the number if needed
and then it is sent to Display
and the Output value in MF COnnector window shows what MF is sending to the display
you will figure it out, it is a horrible firehose of information at first right now
As a recap...to make sure I have the configs correct.
Active config assigned to COM2 Radio as Display 1
Standby config assigned to COM2 Radio as Display 2
yeah it depends which way they are on the pcb
but that is easy to swap on the config
try to show just 12345678 on the digits first
wiht the test function
10-4
hm?
In the hardware config, only one device config for each daisy chain. Each additional Max chip in the daisy chain just increases the Num parameter. No additional hardware config nor pin assignments are needed.
It means ok...in radio talk.
oh!
I think that is what I did....the first Config under the Mobiflight module "Mobiflight Nano" is the hardware config.
Then there are two Output configs as subsets of the Hardware config.
One of those is the Active Freq. display and one is the Standby Freq. Display.
That is making sense but it's a bit murky beyond that when it comes to assigning presets for Xplane.
I am trying to figure out how to test the wiring without assigning anything to Xplane.
so you have the displays with their pins done?
I do
so you add an output config,
Each display connected to the max chip per my schematic. 5v external power is running to the max chips. The two caps are in place as well as the resistor.
I am ignoring it
should have something like this
'
so check all checkboxes,
set number of digits to 8
and type 12345678 on the Test value on the bottom
and press "Test"
it should give you numbers on the displays
and then we figure, if not, why not.
I see that...and tried it with just 5 digits.
I am not getting any power to the displays. Nothing lights up. Going to go back and troubleshoot all my connections again.
hmm
if you have a multimeter check the voltages etc on the MAX7219
and that stuff is on the correct pins
do you have them both connected or just one?
oh wait hey
I actually configured mine wrong
doimg voltage test now
They are both daisy chained with the power connected to each in parallel..
I am using one set of pins for the two that are daisy chained
Yes one configured with the 3 pins and num set to 2
this then selects which one
so try both options there
but it should just select either one
if both are breadboarded, but if only one, then this would be significant
I have that set to 1 right now for now...will stry display num 2
That is what I assume....I want to try lighting up individual leds in the display to make sure I am getting power.
Not correct. If the Max modules are daisy chained, like I said, only one device config per daisy chain.
Ok
As mentioned, the first Max chip is Num 1 and the second is Num 2
I now that...but I had two Output configs
ok, never mind what I said, then
I need to go back through my pin connections first. Power is good
you could display 8.8.8.8.8.8.8.8.
and measure multimeter between the DIGIT and segment pins
because they are just leds
also disconnect and reconnect the arduino at times, if the `MAX gets bad connection or somehow screwed up, it needs a reboot
so replugging the usb will help
then stop and test again
Only up to two Max chips should be powered from Arduino. If more, external supply is needed.
Measure continuity between a digit pin and a segment pin?
yeah, try also with just one max and set the brightness in the pin config to minimum
digit pin and segment pin should give you voltage because there is a led
I am powering the two max chips with external PSU..
if the segment is displaying something
I am getting nothing on any segments..
Yep
can you take a photo of the setup? maybe we spot something
All grounds are connected.
How would I illuminate a single segemnt without blowing it out.
I think I will start over by just connecting the first display....start simple.
Gotta rewire and disconnect the daisy chaining
I disconneted the second daisy chained display. Took the CS and CLK from the first max chip directly to the arduino instead of the second max chip.
Now I am getting 12345 on the first display with the DP between 3 and 4 as it should be.
I think I know what my mistake was with the daisy chaining. Need to try something
Question....With the PSU power I get a good bright display . But when I turn off the PSU I expect the display to go dark.
However it was still lit but dimmer. I do not have any 5v from the arduino connected .
There are a positive and negative 5v lead pins from my PSU breakout board to power the breadboard rails.
After I turned off the PSU breakout board, the display was still lit but dim.
Then to test something I removed the power leads from the breadboard and the display then brightened up for a few seconds then went black.
Now when I plug it all back in and turn on the PSU breakout controller ....nothing. The display does not light at all now.
I switch to the Arduino 5v since it is only one display and the display still did not light.
So I changed out the 5-digit display module. The second display module is now working off the arduino 5v power.
Is it possible that when I pulled the PSU power pins from the breadboard and teh display came on brighter then went black that I possibly killed all of the segment LEDs?
so the max is always powered
you can blank display two ways in config
either write " " spaces
or use the "brightness reference" variable
so for example set CIRCUIT GENERAL PANEL ON
or alternatively if there is a simvar for the COM1 power for example
use that
and use a modifier to set it to a desired value
you can also mix in the radio backlight dimmer value from the simulator
but lets not make it too complicated right now, lets just say you can simulate everything the real thing has, as long as we can get data from the aircraft
always reconnect the arduino if the max acts up
it gets confused easily if power is interrupted etc
so either stop and start MF or just reconnect it
but this is encouraging that you get digits
I am getting digits on the second swapped out display.....but the first one wasnt coming back on immediately with either power supply.
You can test this with a multimeter set to continuity... it should measure around 2V between any given segment anode and any give digit cathode... Alternatively use a 220ohm resistor to connect a cathode of your choice to ground. Then poke around the anodes with a jumper wire connected to 5V...
I'm not sure how the LEDs would still be on if you've removed the only power supply... I'm worried there might have been a short somewhere.
Where was the power coming from? Vusb?
One thing I forgot to do when I disconnected the wires to the daisy chained second display was to also remove the second set of caps and resistor from the power and ground rails of the PSU power. Would having the other two caps and resistor in that circuit connected to the GND and 5V of the PSU supply cause damage to the display leds?
The power was first coming from the PSU to Display one that was no longer daisy chained. Worked fine until I pulled the PSU power pins fromthe breadboard.
I don't think so, no
When plugging the PSU power back into the breadboard, the display wold no longer light
Now that I changed the display module and am using the Nano 5v it works.
Does the suspicious display work with nano 5V?
I would generally recommend powering down everything before changing any connections on breadboards... especially the MAX ICs seem to be pretty finicky
I was about to try that again....but now the second one suddenly does not work under the test mode.
And I touched nothing,,,,
I will shut it all down and start over.
Unplug the nano and the power supply, give it a minute and try again
I know "have you tried unplugging it" sounds lame, but it helps more often than one would think🤣
I will keep it simple for now and unplug the nano since I am using the 5V from that for the single display
A reboot of the Nano and of MF did the trick. And that is on the first display module that I thought I damaged.
I will give it some time to see if it is stable.
Then I will go back to the PSU power.
You cannot just cut power to Max7219 chips. An Arduino reboot is necessary after that because the Max7219 needs to be initialized again.
makes sense.
But is the only way to cut power to the max chips is to unplug the USB cable from the arduino.
it might be the breadboard wires are a bit fiddly
yeah
so yeah, if the panel is plugged in, it will be powered. I think there is some sleep mode in the arduinos if they are idle for long, Ralf would know
but the displays will not be lit if Mobiflight is stopped
it should turn off all digits and leds
True about mobiflight and when I stop the test.
So after about 15 minutes the display went blank. Restarting MF app brought it back.
yeah, what version of mobiflight do you have?
this was an issue at some point, it will not happen if the value changes I think
Its rather consistent that after about 15 min. the test stops and I have to restart MF app, to restart the display test.
I think this is already fixed in the latest beta
there was no keep alive message being generated when in Test mode.
The test is not so important as long as it doesn't do that when I get it running with my flight sim.
Now it's time to switch back to PSU power and then reconnect the daisy chain.
I, too, have seen this mystery of the max staying lit after cutting power. Here's my previous discussion if it helps: https://discord.com/channels/608690978081210392/1304412511168237619
And the problem with test mode as well: https://discord.com/channels/608690978081210392/1295371715047587850
Thanks for the links to the discussions.Those same issues have also been my experience with MF running in test mode and the disconnection of the power.
It sounds like an anomaly with no solution..yet.
My understanding is that they are possibly leaching power from the signal pins when the main power supply is disconnected, because of the way the ESD protection circuits work. I now make it a point to power down my Arduino's before turning off the external PSUs to avoid any damage.
Are there ESD protection diodes in the arduino or the max chips?
Likely in the max, but I'm not really sure. Way beyond my current skill level. 😁
The MAX7219 datasheet I have pulled up doesn't mention any ESD protection diodes. The Atmega 328p used in Arduino Nanos (as well as the 328pb used on many clones) has protection diodes on each I/O on to both VCC and GND.
Interesting. That was one of the theories in the other thread. Oh well, bottom line is that the max is getting enough power from somewhere to keep driving the display if you kill only its power supply and keep the Arduino running. Figuring out where will likely need more tools than I currently have.
you cannot really do that anyway, because the max wont suddenly start to work if you power it back up
because it needs to be properly started by the arduino lib in order to work
if you power it down, I would think the internal state of the shift registers is lost
and you will need to stop and start MF at least to make it work again
you should then rather get a usb hub with power cut off switch
Yeah, the intent at that time wasn't to do a quick restart of the max. It was simply a moment of "How is this thing still on if I'm not supplying power? Have I messed up my circuit somewhere? Is something shorting?" 😅
yeah
on my dual MAX7219 radio I am just powering it from USB
it has worked fine
and I have 47kOhm for iSET
After some more testing with a single display before daisy chaining the 2nd display I did a couple of experiments.
Just for grins I added a 470k bleed resistor parallel to the electrolytic cap by putting the leads into the same 5v and GND holes that the cap is plugged into.
Powering off the PSU before powering off my Nano with the USB hub switch, the display stays lit but at about half the brightness. The display powered down after about 15 min. and not immediately as expected with the extra resistor.
It does not appear that the bleed resistor works unless I installed it incorrectly.
There is another mystery in that my PSU breakout board has lights to indicate the status of the connections.
One of the lights stays lit as does the display when powering off the PSU prior to shutting down the arduino. See photo
It is odd that its a different power rail than the one being used to power the breadboard.
What may be a factor is that I was using that connector yesterday when I inadvertently shorted the GND and 5v leads on that one. Perhaps that is why that particular breakout board light stays on.
But I found that as long as the USB switch is powered down first, then the PSU is turned off all is ok.
If the USB powers on before the PSU, then MF app does not need to restart to run the display test.
I finally got the daisy chained displays sorted out.
But what I learned when setting pins in mobiflight is that pins D0, D1 and D13 are off limits on a UNO and a NANO and while MF will allow A6 and A7 to be selected in the config, I read that they will not work.
I am sure about the D0 and D1 pins but is it true that A6 and A7 are not usable?
oh yeah
I mean you shouldnt be able to select D0 and D1
those are the usb serial port that is used for communication between the arduino and MF Connector
and nano has special hardware limitation on A6 and A7 indeed
what is the deal with nano D13?
I just read through google that D13 is reserved
A6 and A7 have no digital functionality on a nano, only analog. They had some extra pins on the smd version and I guess the only spare contacts the silicon had were ADC channels
D13 has the LED, which can cause issues with pull-up resistors depending on the specific board you are using
I am avoiding those 5 pins now.
But if I want to add the two volume pots would I be able use A6 and A7 for those?
The amp was introduced with the Uno R3, boards based on older designs or optimized for compactness just drive the LED straight from the pin. Even the official nano does not have the amp.
Sure, that is why they exist
If DO and D1 are unavailable and the A6 and A7 pins are reserved for analog, then I am short by 4 pins for one of the encoders plus my two switches.
Oops sorry....I misread
D13 should be fine for outputs, btw
Like an LED?
Just avoid straight switches, encoders and multiplexers on that pin
Seems like I have just enough pins.
How does one test the function of a button config?
Its a simple momentary switch as a button config uploaded to the Nano module.
The correct arduino pin D6 is set and the switch has gnd on one side and signal to D6 pin on other side.
The tactile switch is confirmed to be working with multimeter.
Should I see something in the bottom window when I press the switch?
When I select run I get a line towards the bottom - Mobiflight Nano =>Button=>RELEASE=>executing "Button"
Then pressing the button there is nothing...
With Mobiflight logging on, you should see both press and release events when you press and release the button switch.
That is what I assume....but that is not happening
What is happening exactly?
How did you wire it up?
Is it a two or three terminal button switch?
Ground to switch then Wire from switch to D6.
When I connect my multimeter from ground to pin on arduino I get tone when I press the button
Its just a simple 2 pole switch....in and out.
You mean Mobiflight only reports the release event?
The circuit is closing but not translating into mobiflight except for the release function (normally open) when I start run in MF.
Yes....and only when I start run....it reads the normally open position and that is all.
When you start run? Without pressing I guess?
That would be auto Retrigger working
What happens when you press and release?
Yes....starting run....no pressing of the button
That is normal behavior
Nothing at all happens when I press and release
Nothing on the log?
Test your button with the multimeter, do the contacts show continuity when you press?
As I said I get tone on my multimeter when I press the button witch the probes in my ground rail and touching the arduino pin D6
This should be so simple. Its these little problems that makes me want to send this flight sim to the dump.
Tried a different switch....same thing.
What am I not understanding?
If the multimeter triggers, arduino issues seem more likely than switch issues
It was a grounding issue....
Its always something simple that just gets missed
It has to do with the human mind always making assumptions based on prior experiences
but it did a retrigger release and never pressed
is it a four leg switch?
see which legs connect when you press
often it is 2 -> 2
like one direction, and 2 are always connected
so its like ]- -[
or I guess it worked out already
but yeah, this is the joy and horror of this stuff
" always something" 😄
its just a lot of them when you are new to it all
it'll make sense bit by bit
I swapped to a four legged switch. The closed position is pretty standard and it works now.
The dual encoders are working now also.
Next thing is to get is all connected to Xplane.
sweet
some datarefs to try:
- sim/cockpit2/radios/actuators/com1_power
- sim/cockpit2/radios/actuators/com1_frequency_hz
- sim/cockpit2/radios/actuators/com2_frequency_hz
- sim/cockpit2/radios/actuators/com1_standby_frequency_hz
- sim/cockpit2/radios/actuators/com1_standby_frequency_hz
- sim/cockpit2/radios/actuators/com2_standby_frequency_hz
those exist as presets too
so if you paste those into the search box it should find them, or you can also paste them into the [x] Show preset code box
the frequencies are in Hz I think so you need to do a Transform modifier of $/1000 for example to get the decimal point (or you can also add a fixed decimal point in the display config)
but this is a lot at once, so lets see if you have questions, and we can sort them out
Some of that I am familiar with. The transform modifier will come later.
What I have built so far on my breadboard is the Com2 radio with active and standby displays.
Now its time to get the Nav2 radio wired up....now that I sort of know what I am doing.
yeah
that will be helpful then, its similar stuff but just differerent datarefs
now I am tempted to 3d model the case
and we could have the pcb so we could have as many radios as we want 😄
Can you expand on what you mean about the case and the PCB? I am not sure I follow you.
I mean if we 3d model the origianl radio panel we could 3d print a new one, and create new radios without the original panel even
I've been thinking of making and offering "generic" radio panels with proceeds going to MF. I just haven't come to a satisfying conclusion in my internal debate regarding power supply... screw terminals? Barrel jack? Molex MiniFit Jr.? 12V? 7-15V? 5V & 12V?
Also, Comm & Nav or Comm & Nav & ADF?
A lot of these control panels are very similar across various planes, at least in function: active and stdby frequency, dual rotary encoder, TFR button... any model specific inputs (adf tone and mode switch, comm radio selection etc) could be handled by buttons on the bottom?
Seems like everyone is flying different planes, so model specific devices don't make sense to me
I did the 737 radio on mf shop, and I have also a new version, plus trying to finally get the GNC255 out
its just so many iterations and tests to get satisfactory quality
but yeah, MF shop pcbs use JST-XH generally
OLED actually
The KX 155 radio I am trying to build is very model specific. In fact it is specific to my pilot friend's 1970 Piper Arrow II. In fact my entire sim is based on his aircraft only because I get to do some IRL piloting in it.
I cant imagine anyone else wanting such an antiquated radio. If I recall, the only other default aircraft in Xplane that has that radio is the King Air. Cant tell you if it even exists in any MSFS A/C since I rarely use that sim.
Fancy☺️
How is it interfaced?
I think I'll have to look at the source code for that at some point before proceeding with my cdu idea...
Hardware SPI. Ralf and Tuomas did the work on that, I am just copying their homework for my own generic OLED panel
It's pretty cool you get to fly the plane you sim around irl!
Wish I could say that about the 737-800 🤣🤣🤣
Unfortunately, friends with airplanes aren't a thing for me... the local flight school offers trials but they're around 500€/hour I believe
Oh I definitely need to look at that then!
Not sure they make graphical OLEDs in CDU size though. Also not sure if the custom device framework is that great for CDU levels of content...
Love those dangling toggles
@autumn thorn so I am finally trying to push the gnc255 or well, "MF255" to the mobiflight shop, its just a lot of reifining to get the consistent results with everything
yeah this was being serviced
when I took the photo back then
we had painted the panel matt black
and removed the switches since those two were on the plastic directly
that is the aerobat, it has the door eject handles
sadly this got one too many nosegear collapses and was scrapped 🙁
for parts
The custom device framework has been a limitation. I will need A LOT of output configs!
On the other hand, I'm working with an RA8875 chip to do the heavy lifting, using the RP2040 to initialize it and send strings and locations to print.
The screen I'm using is zj050na
this thread has turned into "pcb and electronics chat" but I dont mind it, as long as @river river gets the answers needed for the pcb work
@river river a lot of people do like the old classic avionics though
flying with raw needles in bad weather in the sim is fun, even though one is very happy about glass cockpit situational awareness in real life
sadly there are not that many VORs left
Especially not ADF stations.
Seems like Germany is decommissioning a lot of VORs to make room for windmills...
On the other hand, apparently the EU is working on preserving a network of VORs after the recent spike in GPS interference...
But it does feel like the NAV and ADF panels are being used less and less...
six pack avionics flying always seemed like real flying.
My friend and I have flown a few approaches to minimums in his Arrow. It was always fun.
our VORs are just DME now
so airliners can triangulate distances
but not so fun with simming
I have heard that GA flying in the EU is not so common now. Is that true?
I guess it has never been as common as in the US
It has never been nearly as wide spread as in the US and Canada. Norway and Sweden are an exception, but there it's just more practical rather than recreational.
I remember as a kid in Canada, every Saturday and Sunday the skies would be buzzing with Skyhawks, often on floats, as people would fly to the west coast of Vancouver Island for a weekend away. Here in Germany, people tend to take the train or drive😅 might be due to taxes on fuel.
The most popular form of recreational GA in my area is definitely gliders!
Wide open land outside of urban areas and low fuel cost probably contribute a lot to interest in various motorized activities, aviation among that.
That is true! ATVs, Motocross bikes and 4x4s exist here, but they're rarely used to their full potential. I think the dense population is certainly a factor.
@kind tulip NOW the conversation has most certainly digressed🤣🤣🤣
europe has quite a lot of ultralights though
but we have now plenty of crossovers and SUVs ...in the cities 🤣
They're not 4x4s🤣
Houston we have a problem...
I got my breadboard all wired up nice and neat and then being really stupid connected the power wires to my external power supply backwards. I smelled my mistake as soon as I hit the PSU switch.
Did I fry all 4 of the max chips I have wired in parallel?
I am assuming I need to get another set of max chips now.
That was such a bonehead move.
Fortunately they are not too expensive.
I would suspect so, yes.
Possibly also electrolytic capacitors, as they tend to not cope with reverse polarity very well.
Bonehead or not, I take comfort in assuming this has happened to the best of us...
Yes...not paying close attention and making ASSumptions.
In checking for replacements I can find the CNG+ at Digikey for $17.00 each, however on Amazon there is a set of 5 CNGs for $10.00.
Does it matter? Does the price necessarily indicate better quality?
I would personally buy the cheaper option after my experiences.
Statistically (MY statistics) this goes wrong every so often but the money saved in the meantime more than makes up for the rare losses.
I bought my MAX7219CWG chips on aliexpress (50 pcs for 20some € + shipping) and they've been fine.
I know this is annoying. I'm guessing you made time this weekend to complete your testing which will now be delayed. Don't beat yourself up too bad, though. 😉
Yea...was hoping to finish it today. Should have double checked the wiring.
This sim has been a 4 year going on 5 year project.
Whats another few days.
😄 Don't I know the feeling
I'd reckon that even if genuine chips are better than generics, they are not 10x the price better
The cheap amazon chips have been ordered.
Another question...
I set up a mini display breadboard to test my max chips.
Using a 4 digit 7 seg display I have it working with the max7219 on an Arduino micro.
However, I am still trying to understand the connection between the MF output config and what is shown on the display.
To display the numerals on the 4 digit display reading 1234 left to right requires the numbers in the output config's " test current settings " window to be typed as 4321.
I set the number of digits to 4 in the config.
How is this config test mode supposed to work?
Is there some thing that needs to be done to reverse the digits in the config, or is it a matter of swapping the connections between the max and the display?
@river riverthere is Reverse digits option in Display tab.
Right below the Values where you set amount of digits.
I tried that a dozen times but the display would not light at all.
So I decided to check the pinouts again and googled the display model for the pinouts which was identical to the first diagram I found.
I went through all the connections to make sure they were good and then.....I plugged in my Arduino micro and nothing.
Now the arduino does not even register in my device manager let alone MF.
Its very frustrating how fragile this stuff is. One day it works, you move a few things around and then nothing.
Sheesh........
Sometimes its the ghost in the machine. A reboot of my desktop brought all my arduinos back online. The display config was gone and had to be recreated.
Back where it was and still the reverse digits (use for custom PCBs only) does not reverse the digits on the display. There is no output at all.
so set this:
its good to set number of digits to 8
that setting is something that does not really make much sense
its more confusing than useful
just set number of digits to 8 for the display, and then check those "use display" checkboxes for the digits you want to use
but first check them all, and set "12345678" as a test value
so you see which digit is which
if your breadboard has space, you can also wire the other 4 digits
so then you see what your digits are
even with the current just 4, you will see which digits they are
whether they are [----5678]
or [1234----]
or if they indeed are [8765----] or [----4321]
I swapped the pinouts from the display to the max which worked.
I will move them back to the original location and try setting the digits to 8 in the config.
the checkboxes mean you can have two configs use the same MAX, this is quite confusing at first for sure. But this is so you can have for example heading and alt in the same max like this:
[HHH--ALT]
That level of complexity will come later. For now I only wanted a way to test my Max chips. For that it doesn't really matter which way the digits display but by switching the pinouts like this I get the output to have the same order as the config test input.
The reverse digits check box in the config just does not work with the display and max combo for some reason.
Just for grins, now that I have a 4 digit circuit working with a new max chip, I replaced the new max with the first max of my "official" 4 display prototype on which I reversed the polarity and found that the first of 4 chips is working.
There is a display on the 4 digit mini prototype board using a chip that I thought might have been damaged.
I will check the other 3 max chips later.
Happy to hear the IC is still working. Wonder where the smell you described came from, tho.
I have no idea what it was. Maybe it was the smoke coming out of my ears when I realized the mistake.
I recommend wiring the digits in reverse order with respect to the Max chip number, so reverse digits option is not used.
See here
#1139182169063956523 message
I think that is what I did per the image above to get the display to match the test number order.
And that reverse order was set up in the 5 digit display pinout I have been using for my radio build.
Oh I remember those. Those modules numbering the digits from right to left are the only exception I know of. All the rest number from left to right.
yeah, so on that one you should reverse the digit connections I guess
it generally works both ways, just having the checkbox for reverse if needed
The reverse digits checkbox hasnt worked for me. Dozens of tries, even uploading config after selecting.
When the test button is selected nothing displays.
Something about my setup does not allow for the reverse digit function to work.
hm, and you had 8 digits selected and 8 checkboxes on?
then it should display at least numbers, even if in reverse order if you test
Yes, after selecting 8 in the number of digits box and then selecting only the 1234 checkboxes, the reverse function worked. And since I have the wiring reversed, the 1234 numbers in the test box came up on the display as 4321. So it seems like I can set this thing up either way.
yeah
it works either way. the most common way is to have the hardware digits start at zero from the right side
hardware-wise both work, its just a matter of a convention, with a reverse option possible too
I think I got it now....
cool
its a bit complicated to set up those at times because each MAX can have each digit as a "display"
ie, you can use a digit or digits for something and the other digits for something different
hence the checkboxes
Now that I am connecting all of the components of my NAV / COM radio on my breadboard circuit to MF, I am getting confused about how to separate and control these four displays.
There are two displays for the COM radio (active and Standby) and two for the NAV radio (active and standby).
1...if all of these are daisy chained and all 4 are running off only (3) digital pins, doesn't that mean I have only one Device under the Nano Module and is a single device connected to those three pins to drive all 4 displays?
2...If there is only one Nano module device for all 4 daisy chained displays, then will I have a separate Output config for each of the 4 displays (Active, Standby, Active, Standby)?
3...How do I control each individual display because right now when I run the test both the COM and NAV active displays light up.
Once that is cleared up I will have additional questions.
Ok....I think I am figuring it out.
The Device is set to Num 4...displays.
Then the output configs are set to 1, 2, 3, 4 respectively...
I usually have to ask the questions which helps me to figure it out on my own.
It looks like all the MAX chips are working well. Once the numbering was set properly the readouts are good on each of the 4 displays.
Is there any significance to the Flight Sim Value and Output Value showing as 0, 0 for the first config and blank on the other 3 configs?
yes, you have one display device, and there then is a selector of which one in chain you want to display to
so "com standby" could be for example "COM2/NAV2 RADIO" -> Num 1
and whichever the other frequency in the chain is is Num 2 etc
Dont you love it when a plan comes together.?
All the displays are working and connected to Xplane with the proper datarefs. When changing the freqs in xplane I get the readouts on my breadboard displays.
There are still some encoders and switches to connect to MF but those will be easy.
After the switches then its time to send the files for PCB fabrication.
Now I see what those columns are for.
The columns are a before and after modifiers applied
this is an awesome photo btw!
yeah like Jaime said, the "flight sim value" is the raw value received from sim -> modifiers -> Display tab -> Output value is sent to display
@river river I can imagine the feeling of accomplishment so far, though it is a lot fo work of course
but imagingin you can just imagine, design and build a fully 100% custom sim panel yourself is what makes all this so great
by the way, the use of those angled pin headers on the prototyping board is genius
and in general the breadboard test is amazing, I am blown away how neat it is
this was me in 2021
I discovered that idea online about soldering two regular male pins into an L- shape to keep the jumpers flat. Took a little time to solder all the pins but it made it easier for me to see what was going on.
And I was able to move the group of pins and jumper wires around much more easily
Problem with the typical L-shaped pins is that one leg is short and doesn't hold the dupont pins too well nor inserts into the breadboard as far
ah true!
That breadboard looks so good! Mine usually look like something you might find in a Wikipedia article about "rat's nest" 😆
Its difficult for me to see what is what when the jumpers are sticking straight up. To help keep things easier to understand, I used the same wire colors for each set of max to display connections. It can get confusing with so many wires.
yes, this is very nicely done
All encoders and switches are connected to Xplane and working perfectly. All components are connected per my schematic and pcbnew in kicad, so I think I can order that PCB board now.
All wiring has been triple and quadrupled checked.
I have to say that MF has been much easier to work with than Air Manager. It was a struggle in AM for the first few years with all of the scripting that is a challenge to learn.
you can also export the pcb from kicad as pdf and print it and cut it and see that everything fits on it
and that it fits on the panel
I usually 3D print my PCB exporting it as STEP from kicad, but not sure you have a printer
just to see if it fits
I do have a 3d printer.
I am not too worried about the width and height of the board, but want to make sure I get the depth of the board relative to the front of the face plate correct to allow for the protruding encoders and the location of the displays relative to the front lens.
I will 3d print that board without the encoders and rotary power switches and with the mounting holes for everything.
yeah, you can test fit stuff
it might sound silly but at least for me its super helpful to have the stuff in my hands to check everything feels and fits right
you could also model and print extra radios now that you have the pcb 🙂
@river river you and I are doing almost the exact project at the same time, looking forward to seeing your progress
So am I...😅
Is your project based on a particular radio?
brilliant! I'll have to make some of those!
Having seen your meticulous breadboard work I don't see any reason why the PCB shouldn't work.
I don't believe I've ever gone to anywhere near as much trouble prototyping a circuit. Hats off, that is truely clean work. Very nice looking!
Probably sounds silly, but I've been staring at your breadboard assembly for the last 5 minutes just enjoying the view!
Thanks for that. Sometimes I may go too far and take more time than really necessary.
I like the ATX PSU adapter, that looks very fancy. Could you share a link?
I don't see why not, if you're enjoying yourself... I do the same when it comes to PCB layout... I should probably spend more time prototyping 😄
The PSU breakout board is one of a few I found on Amazon. It just looked like it would be best for me.
https://www.amazon.com/dp/B099RJ7572?ref=ppx_pop_dt_b_product_details&th=1
Very nice. I like the fact it has both indicators and fuses on board as well as some USB power connectors.
Seemed like a useful device. It has the 12 volt output so I can use the power supply for my landing gear lamps instead of the Meanwell power unit that is dedicated for those.
At first I was planning a 737 com radio. But after some thought, and because I fly a couple different types, I’m now planning a generic radio panel. 2x 6 digit led but with buttons to select com1/2/nav/adf/transponder.
Who among you are good with Fusion 360? I have had the software for 5 years but never had a lot of time to learn how to do more than some basic design.
Would these photos be helpful for anyone to give me suggestions on what might be a good approach for getting this faceplate modeled?
It has some complicated faces, slopes, chamfers etc. and I am not sure what might be the best strategy.
Granted, photos might not be the best way to get an adequate understanding of this object.
not fusion, but freecad a bit
I think they painted their white plastic panels black and then sanded off the faces of the text labels just like @hushed wind and many others have done 3d printed panels too
pretty cool
I have 3 sets of digital calipers. Measuring is not an issue.
Its the Fusion process and where to start, what with sketches, bodies and components and the best work planes to begin with that make sense with the shapes and protrusions on this faceplate.
I will DM, once I figure out how to do that.
answered to DM
I think I would do make a sketch of the from face, leaving out all the white protrusions. Just the front 2D view. Extrude the basic shape and then make another sketch on the front face with all the details.
I'm away for work this week, but if you are still looking for input next weekend I'd be up for a video call if that helps.
That would be very helpful if you have the time. I dont want to put you out but any advice would be very helpful.
Thanks.
I am thinking that the work plane for the front face should be on the z axis where the faceplate would basically be upright as it would be in a panel.
That doesn't really matter... when it comes time to assemble things in CAD you can flip the panel around how ever you need
Thats true
What time zone are you in?
Eastern Time Zone
Ok, so a couple hours behind CET... If you still need help on Friday, lmk... we'll work something out😊
Sounds good to me. Will let you know. Thanks
My PCB has some mounting holes that are going through the ground plane filled areas.
Should there be a keep out zone around those holes to keep the copper from extending right up to the edge of the holes?
There are videos online about keep out zones and keep out layers in Kicad 5.0 but where is that ability in Kicad 9.0?
I can't seem to find that option in any of the 9.0 menus.
it depends if your holes have a ground pad
I do se grounded holes myself
I guess its a matter of preference, I could do either way
figured if I ever do metal casings the grounded holes would be useful to ground the case too
but most are plastic so it doesnt really matter I think
So its okay to let the copper extend right up to the edge of the hole, even if I might have a metal standoff touching the copper?
I can use my nylon standoffs
my reasoning was nothing more than "oh these look cool" 😄
and I kinda made up that other reason
The copper will not extend all the way to the hole by default. Zones (pours) keep a certain distance from board edges so the inside perimeter of the hole will not be conductive (the distance is determined in the zone properties and in the board settings). If the hole is just a hole, the copper around it will be covered by soldermask and therefore be completely isolated.
If you are using mounting hole footprints with a connection (blank copper and all) they can either be left unconnected or connected to ground. I wouldn't recommend connecting a metal housing to several 0V mounting points, however, as it may turn into an unintended current return path causing noise issues. Metal housings are best connected directly to an earth wire rather than to the 0V rail (aka GND)
Thats good info. Learning something new everyday.
I asked because I created mounting holes just by drawing circles on the edge cuts layer and refilled the ground zone. The shading of the ground plane copper seems to go right to the edge of the holes. I guess I was expecting to see a blank space between the shading and the circle.
I tried the mounting hole option from the library that does not have an annular and redid the zones. After that I moved the mounting hole and saw that the ground zone has a hole in it based on the inner cyan circle.
The components of this mounting hole are a mystery to me. There is a Courtyard layer which I understand, but the light purple circle is on the User. Comments layer and that I don't understand what it's purpose is.
press "b" to recalculate the zone if you move components
it does not do it automatically to avoid computing all the time
oh wait maybe you didnt ask that, sorry let me grab my morning coffee 😄
it might be a visual indicator of the commonly used standoff diameter maybe?
That circle on the edgecuts layer should have copper free area around it. 🤔
Open the properties of the pour (highlight the pour by clicking on it, the hit the E key) and edit the clearances. If there isn't an option there concerning copper to board edge clearance see if you can find it in the board setup (second icon from the left in the upper icon bar I believe)
I find it very risky having (potentially exposed) copper go all the way to the board edge and would recommend setting a clearance automatically kept by KiCad
From an earlier post of yours...🙈😅 this is what the clearance should ideally automatically look like.
Couldn't find any other pictures on my phone to demonstrate... my PC is a couple hundred miles away
Disregard the orange circle, that was relevant back when the screenshot was originally taken
I think the defaults have a clearance
I just have a big oversized rectangle for my fill
like this
I tend to do the same
It took awhile to find it <Board Setup> <DesignRules> <Constraints> <Copper to Edge Clearance>
0.25mm should be fine.
I guess it might have just not been rendering correctly.
It was set at 0. I changed the value from 0 to .25. I can increase that to .3 but beyond that I get a DRC error at another location on my board.
0.25 is plenty
What DRC error is that? Missing connection?
However, if copper to edge clearance is throwing drc errors at 0.3mm, check that your footprints are far enough away from the edge. I suspect one of your pads/tht might be marginally close to the edge of the board.
If there's not enough clearance, I believe the fab might reject your design. It may also be difficult to mount your completed board with components so close to the edge
The errors are happening where I have this notch which is very close to the pads for the MCU.
At 3.0mm clearance I dont get the errors. At 3.1mm the errors pop up.
I think I can get by with the 3mm clearance at the holes.
Yeah, you should be ok
However, the inner corners in that notch may be rejected. Add a radius to streamline the process. I've never had issues at 1.5mm corner radius so I know that will work, smaller might also be possible... I couldn't find any info on the jlcpcb website
Disclaimer: AI replies MUST be taken with a grain or twenty of salt... but this is the closets to a real answer I found
yeah or you need to do dog ears if you need exact fit of a corner
maybe kicad can do that in gerbers?
So if I read that correctly I need to have a min. .65mm radius for the corners in my notch due to the drill bits the fabricators can use.
These fillets came out to be about .89 so good to go?
looks like it
I mean I think I have done just straight edges and it has worked out
but I think in practice it is not a straight corner
it will be cut with a milling bit that is round
Yes, .65mm minimum is my understanding.
I didn't realize how small that notch is. I suspect .89mm is the largest you can fit due to constraints in the geometry of your panel?
the design rule there is just "don't expect a perfectly sharp corner, no refunds for puzzle pieces that don't fit together because you failed to take out end mills into account", not "we'll reject it if it is drawn with sharp corners"
I would think you're good to go, tho. Here's me overthinking on how in earth they're going to mill out such small fillets🙈😅
Sounds plausible
so in practice it will result in a cutout like that regardless of whether you draw it like that or not
like @quartz walrus said
so it will work fine
Interesting, and good to know. I assumed they would either charge their "engineering fee" or send the design back for review.
I think I never looked
they sometimes ask if it is critical
but maybe not about this stuff
but I mean it does not hurt to round them
so your design is what you also get in reality
got a reject once where I was trying to route traces through V-cuts with PTH abuse and did not specify it (and also had economy PCBA selected, which does not support V-cuts)
rejected board was refunded, my other boards still got made and shipped under the agreed-upon terms
oh the snap-off parts?
the jlcpcb kicad library has symbols and footprints for jlcpcb "mousebites" for breakaway parts
yeah, I did some mouse bites on a later board
The sparkfun guys are the ones who gave me my ideas for creative mouse bite abuse
What do you mean by creative abuse?
This is the second time I have seen the term mousebite....gotta look that one up.
That notch is not much wider than the teeth on a carbide saw blade
And I have no idea how they cut PCBs
There is a little metal protusion on the bottom of the faceplate. I might be able to grind that off instead of using a PCB notch.
In fact...I think I will just do that. Modify the faceplate instead of the PCB.
I wouldn't worry about grinding down the faceplate.
Since you've excluded ground pour around the notch on the PCB, you can use something like a nail file or a key file to clean up the notch without unearthing your copper if necessary
Basically, mouse bites are just a row of holes. If placed correctly, they allow you to easily break a PCB in two. They're commonly used for mass ordered small PCBs, where a bunch of them are all attached to each other as a panel. Makes it easier to handle than individual 1"x0.5" boards for example... once the boards are all populated, they're broken apart for individual use... the edges of each individual board then have what look like mouse nibbles...
It appears @quartz walrus has some more interesting use cases as well, though...
don't really want to hijack this thread with it
I can cut that notch after I get the board. Easy enough to do that also.
I have plenty of those PCBs I ordered that snap into smaller sizes.
we can start a pcb chatter thread on #1265357893889294429 too
#1435377709998936266 message
but to me its fine to also chat here on the stuff that relates to the radio design
but it might be useful to have a channel like this even when the radio pcb is finished 🙂
Since I am not an EU resident, I am planning on using PCBWAY to manufacture this board.
Is it normal that a quote only needs the size of the board, the thickness and accepting all the typical defaults for color, finishes, track spacing and hole sizes without providing a gerber file of the board design?
So no matter the complexity of the board layout it doesn't figure into the price?
No experience with PCBWay, but at JLC, most design features are either options you need to select, or they do not affect the price in sane quantities
Just like a copy shop charges per page and not per letter on each page
I just selected all the defaults which look fine to me. So a one trace board will be the same as a board with 100 traces.
Traces do not affect manufavturing cost in measurable ways. If anything, areas without copper would, but that is still so marginal that no hobbyist-friendly fab would charge you
I think some fabs have fees for ridiculous amounts of vias, but usuqlly that is for manufacture at commercial scales
Interesting
The process is the same no matter how many features... except vias, as @quartz walrus pointed out.
Vias and complex outlines dull the drill/mill bits.
I'm not sure about PCBWay, but JLC start charging extra once a certain average number of vias per area is exceeded on a board. It does take A LOT of vias to get there tho. Like 100 holes per square inch including tht holes or something like that, averaged across the entire board.
The extra cost is negligible, though... € 0.47 per 10000/m² holes😅😅😅
a few drill bits worth yeah
How's your CAD design progressing?
I was incorrect when I said I was going to use PCBWay. The fab shop I was actually thinking about was OSHPark since they are US based. But their basic 1.6mm boards are $5 per square inch which for my board would be 65 USD for (3) with $0 for USPS basic shipping. The basic quote I received from PCBWay before uploading my gerbers was about 25 USD for (5) but the shipping is almost the same cost as the board. Makes the PCBWay board about 50 USD. So I will think it over.
I created the typical 2d engineering views of the faceplate and saved it out as a DXF file.
Then I imported that into a Fusion file that I started some weeks ago but didnt get too far. I wanted to get the images into the fusion file and all I could figure out was to extrude the dxf import and place the images on that surface. No doubt not the best way I am sure.
Can you use these to help me get started on how to model what is to me a rather complicated shape?
Have you looked at jlcpcb also?
I did
Knowing nothing about you off the top of my head I would guess you work in metal fabrication?
Your methods here would suggest it anyway. Not terrible at all, but in additive manufacturing we get the luxury of doing things a little easier.
I work in architecture, but have been drafting for 50 years.
If you upload the DXF file for me, either here or as a direct message I'll take a quick peak and start a project for you so I can explain what I mean 🙂
Ok, that also makes sense.
Please forgive me if I came across as a bit of a dick there.
No prob. I will get that DXF and the two referenced images into a .zip file and upload that . It will have to be later today.
Since JLCPCB cost is less than 8 USD + 24 USD shipping, I suppose it might be best to go with them in case my design has some undetected problems.
I think this is one of the cool things in this type of hobbies
you get people with vastly different experience so its always new things to learn from others
Cool!
If you link me by name my phone will notify me and I'll get right on it.
I haven't ever used discord before MF but I absolutely love what I've seen here.
The kindness, the intriguing questions, the time people take for each other... you'd wish the world was a bit more like the MF discord server ❤️
Is that express shipping or idiocracy-tax? 😮
It's DHL express...
Or more accurately DHL Express (DDP) but that was even before I gave my address.
right, phew 😄
If you have a week's worth of patience... and budget shipping is available... I can recommend it. My boards have always shown up within 14 calender days of ordering with the cheapest option, usually for around 3-5€ shipping.
How much did that change the price? 😬
THere is the Global Direct for 3.30 USD..
I have not yet actually submitted anything and ordered. All of this is just from the basic quoting going to the USA.
I am waiting one basic pcb now, got the cheapest option for shipping to finland this time
to EU the taxes (VAT) is also depending on the shipping price
so the tax is calculated from goods + shipping
so it kinda adds up if you use the fast option
Quick mock-up
I'm not sure ANY of the dimensions or in fact the face's shape is anywhere close to the original, but I can walk you through the steps of creating this.
I started with a very simple sketch with only the rough dimensions of the face, along with a rectangle depicting the thickness of the outer wall
Next I extruded the face to 2 mm thickness
I then extruded the wall backward, making sure to select the "JOIN" operation
After clicking on the front of the face, I clicked "create sketch". This basically put all the features of the selected surface in the new sketch to work with.
I drew a circle slightly below center of the left encoder opening.
That circle was then extruded 1mm deep into the 2mm thick face of the panel
4mm radius fillets were added to the upper and lower edge of the extrusion
After again selecting the front face, I "create sketch"ed again, drawing out the outlines of the white margins around the displays etc.
I chamfered the edges
Then extruded the margins around the displays etc, making sure to extrude "Two sides" due to the extrusion and fillets around the left encoder hole
Using the sketch I'd created for the extrusion around the left encoder hole, I extruded into the margins.
Then added fillets to round things off
Another sketch created on the front face has all the letters on it as well as the protruding squares around some of the letters
I extruded the squares
Then the letters, making sure to select the "JOIN" operation
Left click on the body, then appearance allowed me to apply "plastic glossy black" and "acetal resin white" appearances to individual surfaces. This step is not strictly necessary, just some eyecandy
This method is likely not the only and probably not the best one out there. Others certainly have other workflows for something like this.
The way I've gone about this ensures I have a single, fully closed body that fusion360 can export as an STL for 3D-printing.
All features are mathematically baked into the design. Because everything is done one thing after the other, and because the later sketches are referenced to the faces they're on, if I go back and change pretty much any dimension in one of the sketches or extrusions the rest that came after SHOULD(!!!!!) follow suit and the whole thing SHOULD(!!!!!) still math out. (Sometimes fusion gets confusioned if the changes are too extreme 😄 )
If you'd like to, mark me in any questions you have 🙂
Sorry for spamming 🙈
I wonder how many of those comments in the thread are mine 😄
It does!
And straight forward enough to build from scratch even without the original panel 🙂
I have two of those audio panel faces and switches with one connected to MF and Xplane using a raspberry pico and a hand built pcb.
The marker beacon original bulbs also work off 5 volts.
Is this getting to be an unusually long thread?
I don't think there's a rule stipulating a limit on thread length. 🤷♂️ If you're not upset with the number of comments...
I think the topics in this thread are a little more diverse than usual, tho😅
No problem with me. I dont mind the tangents. Every detour is a learning opportunity.
we should just start creating threads in #1265359709880451134 but things dont always work out how you plan
this started from a project related question, and went off the rails a long ago, but hey, it looks like we'll get a pcb in the end
so I dont complain
its possible for a moderator to forward a comment to another thread, so maybe it works to "branch off" different sidequests into their own threads that way
maybe even anyone can forward a message
The PCB is ready to be fabricated, so once that is built this thread should be winding down.
I believe so
I can confirm
What does "mark me" mean as related to Discord?
I downloaded your Fusion model. It looks good. I assume you got all the dims from the .jpg image since I did not upload the DXF.
No need for me to upload the DXF then?
One question....the circular recesses on the left encoder hole are not required and I cannot figure out how to delete them.
@river river this is "marking" or "tagging
@river river not sure if "mark" is the right term
Thank you💪🏻
I think I figured it out...just went through the history at the bottom and found the sketches / bodies to delete.
The dimensions are completely made up! Just guestimates based on the shapes in the jpg.
Looks just like mine
Wasn't sure whether I was looking at discoloration or a recess around the comm encoder
If you're going to actually use the design I uploaded (feel free to btw!) I would suggest scrolling the history marker at the bottom to the very beginning and editing the sketch first of all, then walking through the complete history step by step, editing everything... I doubt there's a single dimension that's spot on 😅
This only links me to the WordPress website
oh weirtd
stuff like this, someone does retrofit service
looks like our sim panel projects 😄
That font is hideous😅
yes
the site is weird too, the guy talks about crypto payments only etc
just a bit sketchy 😄
Very 2025😅😅😅
That explains what the nubbins on the back of the die-cast faceplate are for💡
yeah
that could be laser cut from orange acrylic
A faulty aircraft radio model Bendix King KX-155 recently came to my hands.
As a private pilot of light General Aviation aircraft for about 2 years, I have used this model of radio in almost all the planes I have flown but never really looked at what goes on behind the scenes.
I decided to take this opportunity to do a teardown of the radio to i...
this is so interesting to look how these were constructed
So many trim pots
This video shows how to install a new LED Display in your Bendix King KX 155 KX165 radios yourself. There is absolutely no modification needed on your Bendix King KX 155 KX 165 radio. All you have to do is pull out the old "glass" display and slide the LED Display into its socket and connect two power wires - a red long wire and a short black wi...
this has some internal construction view on the front panel and knobs
this is some 7segment retrofit kit
those displays are not available as spares anymore
and I guess there were so many of these that there was a market
Thats exactly what they are for. I couldnt afford the entire radio so I only bought the faceplate and the lens inserts
I need those white buttons...
you can print the caps
I think I would just 3d print them.
You can make them fit your buttons
I have plenty of white PLA.
Recess the arrows a bit, maybe 0.5mm. Paint them in with a sharpie and then sand down the front face of the button to clean up any paint bleed
Easy enough...
If you're feeling fancy you could even design them to have the OEM spec travel, I reckon
I will give the Fusion file dim editing a try, but if you want, here is the DXF file and referenced photos of the front and back of the faceplate.
I am not sure this large file will be accepted. I dont know the limitations of Discord.
Once I get the SMT switches mounted on the PCB, I can then figure out the button size.
That hole was for mounting the tray in the aircraft dash. I don't need it.
It would fit with my compulsive perfectionism, but its not absolutely necessary.
😄 😄 😄
Thought we could branch off the front panel discussion into a new thread
Might be easier to find and follow for other users.
Sounds like a plan to branch off.
That fusion work would have taken me at least a week. I do appreciate your efforts on this.
Now how do I get my PCB board step file and the faceplate into the same file?
I couldn't find an import step file option for the Kicad export file. All I could do is open up the exported file in Fusion.
But now I have two different files. I think I need the PCB step file to be a component.
And then one or the other needs to be rotated.
I think I can find a youtube video on that
I think I know now....but it's too late for my brain to sort it all out. It will have to wait until tomorrow.
Thanks a lot for the faceplate. Now I can test it with my PCB layout.
https://discord.com/channels/608690978081210392/1436498790625447986
I've described the process here.
Yeah, its late. Time for some Zzzz
It has taken a while to find time to send out the PCB to a fabricator but now is the moment of truth.
The PCB has been uploaded to OSHPark for fabrication and hopefully the boards will be delivered in maybe 10 days.
Fingers crossed 🤞
Hadn't heard of OSHPark before. Keen to hear about your experience with a non-chinese fab.
It's more expensive but includes the shipping. I feel more comfortable with a US based company.
I understand that. I've been scoping out German fabricators recently, but haven't tried any of them yet (haven't ordered any PCBs recently). At first glance they are significantly more expensive and the list of options seems to be shorter, too... I have yet to compare capabilities like minimum copper width and clearances.
At first glance they look significantly pricier, even after adjusted price of shipping and customs. I definitely want to compare finished products, tho!
Do they actually fab on-shore, or do they just collect, panelize and sort orders domestically?
The ones I've looked at advertised Made in Germany
I haven't looked further into how much of the "making" is actually done here, tho.
Was thinking of OSHPark for the Americans.
I feel like Aisler is the most popular choice for fabbing with a German company these days, and their price range is rather different from JLC
oshpark was popular before chinese manufacturers, and they can also host your design for others to fabricate if you want
they are in the US so it wasnt a good choice for me shipping wise so havent tried but I know a lot of open hardware projects have used it
They definitely predate jlc and pcbway
yeah, they exist to provide economical pcbs for open source projects, their idea was to pool up designs to fill up pcb material sheets from several projects to avoid waste and reduce cost
so they take a bit longer to make the boards, depending on demand but of course jlcpcb is an alternative nowadays but the geopolitics can flip things around
Ah, just saw a statement that they have run on domestic fabs since the dorkbotpdx days
From what I could tell, OSHPARK offer 2 layer boards only?
The basic 2 layer is 5.00 USD per sq. In.
They have 4 layer @ 10.00 per sq. In and
6 layer @ 15.00 per sq. In.
Good lord, that'll add up quick!!!😬 Is that for a single board or like a run of five?
3 copies only....welcome to USA capitalism.
I dont think oshpark is exactly the most capitalistic thing
"You can order as many copies as you want, as long as they’re in multiples of three."
maybe it just is a nice number in practice to lay out the big sheets of pcb raw material with designs
I would assume they drill 3 large panels simultaneously, with one large gantry moving the three drills synchronously. JLCPCB do 5 at once, hence the multiples of 5 there...
The 2 layer boards are reasonable.
https://youtu.be/LRiNIUZUAa0?si=Y9F5nXiowbntRk2j
An interesting insight into how this stuff is made with economies of scale in mind
#printedcircuitboard #pcbfactory #pcbmanufacturer
Established in 2004, Ucreate is a highly professional and experienced PCB and PCB Assembly Manufacturer. As an integrated collaborative factory, its business ranges from PCB Fabrication, PCB Assembly, Electronic Components Sourcing and other areas to meet various requirements among the market.
...
4 gives you SOOOOO much more freedom. I haven't done any 2 layer boards in a while. No need to worry about return paths, power supply traces, etc.
so GND and VCC in middle layers and traces on top and bottom?
What types of PCB components would require 4 layers?
High voltage caps, transformers, high speed transistors....
Or boards with a lot of stuff
I mostly go 2 layers because they suffice for most purposes and having all layers on the surface helps with rework
Also, I rarely find myself in situations where I run out of routing space, unless I drastically try to optimize RF performance
And on 2 layers, you don't pay extra for blue/black/white, and pay less extra for oversize
A circa 1960s multi layered PCB in a KR 85 ADF receiver.
I wonder hwat are those red/blue/black square metal cubes
No clue yet.
adjustable inductors and transformers according to google
searched for "TOKO RCL"
btw I just realized that now with the multimaterial printers
one could create those mechanical frequency dials
4 digits, 24hour-notation (military time) mechanical digital clock controlled by only one servo.
Time adjustment from PC is possible via bluetooth.
3D model (STL files, Blender source) and micro:bit codes are available online at
https://www.thingiverse.com/thing:4613624
By modifying the program, it can show the time in 12-hour notation, or the...
I did a 4 layer for the overhead panel, it was the only way I could obtain an evenly distribution of the 9V needed for backlighting and an evenly distributed GND layer. I spent A LOT though... about 100 euro per 5 panels (but it was big). Talking about jlc here.
I generally use 4 layer boards with signals on the outer layer and power/gnd on the inner layers. Bottom layer has most smd components while things like buttons, encoders, LEDs (7seg and indication) are on the top side. Any "high speed" (ekhhhm) signals like SPI (to MAX7219 for example) are routed on bottom layer while pwm is on top. Inner layer 1 (top inner) is used for power supply traces with throughly stitched ground pour while inner layer 2 (bottom inner) is a solid, uninterrupted ground plane for best possible signal integrity.
This stack-up is super useful when working with multiple voltages (5Vload for MAX7219 and shift register outputs, 5Vlogic for level shifters, 3.3Vlogic for uC and multiplexers, 3.3VAvdd for uC and potentiometers) needed at different locations on the board. The continuous ground plane on L3 enables unobstructed current return on signal traces and ideal decoupling of finicky ICs. I have yet to encounter noise issues from switch mode power supplies intense enough to disturb the function of my circuits.
Having two layers for signals and two just for power distributing and ground really helps with layout, as I tend to end up with fairly dense layouts mainly due to future-proofing. I might want to write custom code for these uCs one day, and with hardware debouncing it would be easier... I might want to switch to a different power supply setup one day, and with 12V -> 5V and 5V -> 3.3V buck regulators I can without needing a new PCB. Electricity might become super expensive one day, so SMPSs are more desirable than LDO linear regulators.
Granted, I might be overengineering (some of?😅) my PCBs given the nature of the kind of PCBs we need for a sim cockpit. I feel more comfortable that way though.
I'm more than happy to admit that half of the components in my PCBs are overkill or will never be used and that the resulting complexity is an unnecessary use of resources.
However I'm also adamant that diving down this rabbit hole has taught me a lot about the electronics world to which I previously had very little connection.
I could probably save a lot of time, money, PCB layers and gray matter if I opted to use uC dev-boards jammed in pin sockets rather than laying out my own microcontroller circuits.
I'm not advocating for everyone to follow my lead, tho! I think I started down the 4layer path out of interest mainly rather than using 4 layers to solve an issue I couldn't solve on 2 layer boards.
Actually, I was perfectly happy with 2 layers + an arduino until @kind tulip introduced me to the wonderful world of RP2040 uCs and then it all spiraled out of control😅😅😅
Thank you, Tuomas ❤️😅
I mean why would we live, if it wasnt to learn new things anyway
you have also shared a lot of awesome infromation, this is one of those threads that is also cool, that it seems to have resulted in the radio panel pcb also in the end ! 🙂
but the branching into rabbit holes has been great for a chance
I'm really looking forward to seeing @river river 's finished panel😊
yes
Those must have been HUGE!!!
Did you have the entire OHP connected to one board?
I have always had the philosophy to have one rp2040 per panel, so each panel is self contained
so I was able to rearrange them when my setup changed and I first had a pedestal and then also a desktop frame
Feel like that depends on your definition of "panel". Like, shift registers only get you so far for massive units like a full-scale overhead. But they are divided into smaller units, some of which are the perfect size for a small controller, some even so small that dedicating a controller to them feels like a waste.
yeah, I do feel bad to put a RP2040 into a panel and only use 6 pins
but its 0.7€, but that is probably a neurodivergent thing to "emphatise" the poor MCU wasting its potential 😄
Yes, not the entire OVH obviously, but signs, elec, air cond, gpws, adirs, apu and fuel.
when it could do so much more 😄
that's one huge pcb inside
but the basic button + led stuff is way most tested, and usb is also pretty robust. Even though all devices should work, I do not feel adventurous enough to create a huge panel from a single nano and a ton of shift registers. I mean, sure it probably works fine, but for me a 20 port usb hub has done the job well
But in the next project i will follow @kind tulip way and have separated pcb. I don't know yet if they will share a single mega2560 or separate rps with a usb hub
but certainly they will have independent pcbs
Which is why the pro micro feels so awkward to me. So many 32u4 pins not broken out
yeah
oh wait how many gpio does it have? because its mostly pro micro that I have used ever
as far as the 'u4 is concerned
Pro micro has 17. Leonardo / full micro have 20. U4 has 26.
One of them is the boot mode selector though
so its like if a rp2040 was trapped in a nano suit
My radio build compared to what some of you are building is like comparing a bicycle with training wheels to a Ducati.
@river river this was your first pcb!!! 🤣🤣🤣
Only a lunatic would start (and most likely immediately end) their riding career on a Ducati!!!!
I'm sure that you've already got ideas what all you might do differently on the next panel's PCB so you're done with the bicycle, next project will be a Honda Super-Cub...☺️
Not to be confused with a Piper Super Cub
because everybody in here knows that after a panel's PCB, there is always the next panel's PCB.
Similar era😅
had to look it up
hah, this is quite similar to the very iconic Finnish "old man" light motorcycle from the brand Tunturi
It looks like that panel of Stefano_F78 is 3D printed. What printer was used?
I have a new Bambu Lab P2S that will hopefully give me some good 2 color prints.
I own a Bambulab P1S and yes, that panel is entirely printed, it is composed of various elements and the case is also composed by various pieced tied toghether with nuts and bolts. If you look around in showcase you'll find some pic of the inside of the model also. It has been printed using 0.2 nozzle on a smooth plate
I have the default 0.4 and have been considering if I should get the smaller nozzle.
I will check out the showcase to see if there are some ideas for attaching my panels pieces.
In my experience, when you're working with text smaller than 4 or 5 mm, 0.2 nozzle is the way to achieve the best results. You can have 2x 0.1mm layers for white colour with a line width of 0.22mm, using arachne for wall generation you will get a ver fine line around the text and a good infill.
Keep in mind that I used a very strong light now, I will be using 0805 leds in the 592nm range to get the backlighting effect
Is this what I need for the 0.2 mm swap?
I don't own a P2S so I can't be sure. I'd go with the original hotends though. Bambulab makes excellent products. Not cheap, but good.
I agree...stick with OEM products.
Found it.....
I guess i have to get back to the drawing board (freecad) because there is too much bleeding here. Backlighting color is spot on though (those leds cost an arm and a leg on digikey!)
How do you create the backlit lettering?
The writings have the first two layers in white, then transparent layers all the way behind. Have a look here https://discord.com/channels/608690978081210392/1435748072331280434 where I put the screenshot of the model I used to create this panel
Are you saying those are all 3D printed layers printed with your P1S? Grey, white and clear filaments?
yessir! 🙂
I am just starting the journey down the multicolor printing rabbit hole within Bambu Studio.
So far it's been a head exploder.
Have a look at @round gale 's work here: https://discord.com/channels/608690978081210392/1278230568341078027 . He explains very well how to design a multilayer color printing. Also have a look at The Cockpit Printer youtube channel
Thanks for the resources....there is much to learn.
All of that is very helpful information.
I’m in the same boat “head exploding”
Yeah you know, coming from an Anycubic Kobra, lots of headaches have already been addressed. Bambulab printers are a breeze on this regard.
no doubt
multicolor / multimaterial printing is quite a game changer in what can be done from a design
In the method that you and @round gale used, I can see how modeling the separate layers can help with color assignments of each layer. I assume that it's relatively easy to align each .stl file afterward.
But I see that Bambu Studio has the ability to paint using height range. It seems that the entire model could be brought into BS and then each layer color could be defined with it's own height range.
It's just supposition on my part for now until I can get more into the process.
Do you print a layer of black on the gray face layers before printing transparent? I'm thinking light barriers might be possible with black between individual backlit features?
Bottom two layer: gray and white.
Next two layers: black on gray, transparent on white features.
After that all transparent except for black "walls" between individually backlit features (around the TFR button for example and the inner circumference of the toggle switch hole). Possible lining the walls of any openings for buttons with black will reduce bleed around the buttons?
The colors look incredible, though!
you do not have to align anything because they will be placed exatcly as the design.
in my design I could not use a black layer as @round gale does but I managed to reduce bleeding with some black contouring and 100% infill on the first 2 mm of the print
Another 8 hours to go (this solution is taking almost 18 hours to print, and this is about halfway through). I surrounded the flat top leds with a black circle that should drastically reduce light bleeding inside while keeping lettering and lines well backlit. Let's hope it's good, next week I'll also have the Biqu glacier plate and I will print it again for matte finish on the face.
take a look to the Discor of The Cockpit Printer, i have shared some of my A320 Panels for fully print there.
Fully printed
@round gale it would be great if you shared those in MF repo so it doesn't feel lonely 😉
Our wish is to build one place to go for great designs
I am glad that a Piper Arrow panel for mounting the navigation instruments is so much simpler and without any backlighting.
that's on another level mate. Did you use, like Cockpit Printer, led strips for backlighting? I've tried them on my FCU but got lots of bleeding even with the black layer, how did you achieve that good result?
I am using 12v led stripes with many led‘s. The thing is that the layer thickness is calculated for 0.14 layer height. I am also using am Bambulab Ams with 0.2 nozzle. I have made many tests to find out how many layers I need to block the illumination on the panel and not blocking the text illumination. Additionally the transparent layer is needed for better diffusion of illumination.
Hi, thats a really good idea, but I am using the Korry design of The Cockpit Printer so I have to ask first. How do i upload it to the Mobiflight repo?
@kind tulip might be more helpful here. I fial to understand how github works
my constraint was the design I used for the rmp, my FCU is made using the same layers you used (0.14) but for the rmp that was not entirely viable. I will post pics of my last print as soon as I get back home
Great way to get your first taste! I think in general the Piper Arrow's is a lovely cockpit to build.
Are you planning on building a full main instrument panel?
There are a few things getting in the way of doing that....space, money, time. Other than that, I would.
Pretty much all that is missing now is the instrument six pack. For those instruments that require settings via knobs, I have added encoders for the two VORS, altitude, and HSI. The goal is to make it so that I do not need the use of a mouse or keyboard to do anything in the sim.
I hear you! A full cockpit build does occupy a lot of... well, life. 😅
Makes sense. I always felt using a mouse and keyboard really slowed things down. I don't know the Piper arrow at all, but in the 738 there are a lot of things to do and doing it all by mouse, changing views etc really made a 15 minute cold and dark start start take more like 30 minutes...😕
The first image is the Arrow in which I received my complex endorsement. At that time they had the standard 6 pack gauges.
The second image is the Arrow which I get to fly right seat from time to time.
Everything within the red line is what I have built including the brake handle, the flap lever, the trim wheel between the seats and the fuel selector.
Also the rudder pedals from Slaw.
The flight instruments are just displayed on my computer screen.
I can't take a pic of the sim because it is tucked out of the way for now.
@river river have you seen the Mobiflight (share) Arrow that @supple wigeon flies? These are nice airplanes
No, I havent seen it.
Very nice with the 3 blade prop.
The one in the pic I posted is a 1970s model that came with the 3 blade, but it was changed to a 2 blade recently which added a couple knots to the cruise speed.
Oooh I remember! ❤️
@lofty valley may I share a nice picture of you in said arrow?
or wait, was it in that
yeah
sure
Sometimes the proper solution is just a low tech one. Some tape around the Sel "hole"' and some thermoplastic plus a small pla disc on the bottom of the led did the trick. No bleeding at all. I'm currently printing thw fifth and I hope last iteration of the panel to test it the complete assembly but I've already tested the light and now they are absolutely spot on.
yeah if you can print multicolor, just have the inside walls where you dont want light bleed to be 1 or 2 perimeters of black
or if just single color printing, make drop in single wall pieces
I tried that, but it didn't work so well, some bleeding was still present because the problem was the butt of the led, not the walls. This, however, completely blocked any refracted light.
oh yeah
What is the type of LED in that photo? Is that one of the expensive ones you previously mentioned?.
nope, the expensive ones are smd 0805, because I wanted them with an exact wavelength (592nm). Those are cheap flat top 3mm led for the green light above the buttons, bought a couple hundreds of them on aliexpress
That's a brilliant idea, @hybrid sedge . Heatshrink... some things would be so easy if you just thought of them 😄
Thanks. The idea came by looking at the original RMP module from this post here: https://discord.com/channels/608690978081210392/1255588990778740746 , in particular if you look at the picture here #1255588990778740746 message you will notice that black ring below the led's end. That led me google and bit to find the led spacers and When I thought of a way to keep them close to the end of the led, I thought of the heat shrinking tubes
So the heat shrink is just meant to hold the black disk? Would bending the leads in an S -shape take care of this? I imagine all the heat shrinking on small parts like this would have your fingers start smelling like crispy bacon 🔥🤣
No, not only for this reason, it also eliminates any possible lateral bleeding. I heated the plastic when the led still had the full lenght of their legs so there was enough distance to save my fingers 😄
Holding the black disk was a nice extra result
yeah, light bleeding has ben definitely sorted out. 592nm wavelength color is perfect.
That looks fantastic!
Have you ever tried dark tinted film on the 7 seg displays? I find it to be a nice finishing touch
They will have an orange acrylic sheet in front. I'm still waiting for their pcb to arrive from jlc to finish the 7 segment modules (the will be like the original, two mountable modules connected via flat ribbon cable
11 days after ordering them, the PCBs have arrived.
Lets hope they work.
Ii isn't in MF blue color - NOT APPROVED! 😉
You get what you pay for...
isn't it the same price for both colors? Asking as I didn't pay attention.
It may have been...I just accepted the default. Never really gave it much thought
Just to be clear - I was joking about being approved 😄
I know that....tongue in cheek.
can't wait to see it assembled
Not as much as me....😅
but hey, this is awesome! can't wait to hear if it works
Yes they were from OSHPark.
I need to run continuity tests on one of them tomorrow.
they look amazing. Happy soldering! 😁
Crossing the fingers that they perform as they should.
Been there. Let's hope you didn't mix up some tracks (i did it multiple times!)
That would be an unfortunate setback and a real patience tester.
did I mention that the 7-segment pcb I'm still waiting from jlc is the second one? Guess why 😆
Got your signals crossed no doubt.
The soldering challenge I will have are the (2) SMD buttons. This type of soldering will be a first for me....any advice on how to go about doing this?
Perhaps tin one of the pads then solder down the corresponding button lug while moving the button with tweezers? Then solder the rest of the lugs?
Is there any best temperature to set the iron for this type of work?
exactly that. Place some tin on one of the pads, then gently press with the solder tip the button over the tin. Then solder the rest. I usually work at 320C using leaded tin (non lead free), and using some flux
The solder I have is Pb40 noted on the spool.
so it's 40% lead 60% tin. It will be perfect, you just need the usual precautions using leaded tin, meaning ventilation is required for safety.
I have my fan for that.
My Hakko solder station has a number of tip sizes.
Is it best to use a small tip?
For smd, usually a small flat top one is good, grab a size that is about the same size of the pad. Not too small or you'll have and hard time heating the pins, not too big or you'll be risking bridging
I usually use a 1.5 or 1.8 mm tip (being that the width)
I think I have one about that size.
Remember to use generous amounts of flux when soldering SMD. Better heat transfer, better flow and any small beads of solder get suspended in the flux. Isopropyl alcohol can then be used to clean off the flux and any beads suspended therein. An old toothbrush can be used to agitate...
yeah tiny bit of tin on one pin and like mentioned, hold switch on while gently applying pressure, heat up the pad so it melts and "sinks in"
the rest you can heat on one side and tap a bit of solder on the other side so it flows in
Soldering smd resistor (390 ohms) Hand soldering Techniques #JLCPCB #resistor #soldering #short
JLCPCB PCB+PCBA from $2. Click here to sign up. Get $54 new user coupons at https://jlcpcb.com/?from=SMTEB 350K+ In-stock Genuine Components. Global Parts Sourcing Reaching Up to 1M+ Authentic parts.
(mute it, be warned :D)
unless you enjoy elevator music
Makes it look easy but that comes from experience.
I bought some soldering practice kit on aliexpress and practiced a lot soldering down to 0603. Now 1206 are a breeze, 0805 are just a bit more challenging but still easy. It's just a matter of practice. I got used to soldering mega2560 with 0.5 pitch...
I found a practice kit on Amazon that will be helpful. Thanks for the advice on the kit.
This practice kit has 0805 caps, 0603 resistors, 3216 diodes and some other SOT and SO components.
I tried a couple of the 0805 caps while losing 4 of them before I even started.
Its difficult to even see them with my eyesight and the less than ideal soldering station with a pretty much worthless magnifying glass.
So what options are there to be able to magnify these minuscule components?
I think I need some sort of jewelers headset with high power magnifying glasses and integral LED lighting.
Any suggestions on a setup that can allow me to see what I am working on.
Those are my instruments, paired with my reading glasses (I'm almost 50 😅). The digital microscope is used to double check soldering specially on those 0.5 pitch chips like the mega2560.
do not look at the mess on the desk please 😅
I wish I was almost 50....23 years ago I was saying that.
I think I need a digital microscope.
This hobby doesn't have an age limit, I sincerely hope I'll be still playing with these toys 23 years from now. The digital microscope is very handy but it's not very practical to do the actual soldering. I think it would be better a magnifing visor like this one (just an example, i don't actually own the product so I can't speak about it): https://www.amazon.it/dp/B07WFLNVPH/?coliid=I1WLZQXKFLWF6G&colid=3I0PZK1LSUZTR&ref_=list_c_wl_lv_ov_lig_dp_it&th=1
I find myself at ease enough with the magnifying lens in combination with my normal reading glasses, but here everyone's mileage may vary a lot.
It might be best to find a better magnifying lens setup.
The one I have is not adequate.
yeah, for smd soldering a good magnifying lens is mandatory
and good tweezers. Even as a relatively young mid-30s person, heck, even in my 20s at uni, I found that being able to grab a part precisely and not having it jump off is far more important than how fancy your soldering iron is.
That's right, I have many different tweezers depending on the chip I have to solder. I use bent tip for resistors and capacitors, straight ones for ic and so on
All types of Tweezers are plentiful here. But once the component jumps off its gone for good.
for less than 30 bucks you can pick up a (fairly low resolution) digital microscope for use with a phone or tablet. Pretty handy for the samllest components. They're never gonna provide the image quality seen on youtube videos but it's not as bad as one is led to believe.
Something to look into
Images taken from 10" and 2" distance respectively, with a piece of 1mm thick solder shown for reference
these are a little fancier than the one I got a few years back...
A google search for "usb microscope" should spit out something useful
its pretty easy to solder 1206 size
anything smaller is kinda yeah maybe doable with tools etc but then I just let jlcpcb do smd for me if I need that (like there is no way I would do a RP2040 panel with hand soldering
and the button should be easy peasy once you practice with a few 1206 resistors
3216 I think is 1206
1206 is imperial size
with the right magnifying lens and tools you can go down to 0805. Any smaller than that and it's almost impossible to do manually. And I agree, if I had to solder a Rp2040 I would use JLC service. There is no way one can manually solder that chip, considering it has the pins hidden below the chip itself and it requires 0402 caps very very very close to the chip. As long as I'll continue to use the mega250 i can do it manually with relative ease built on practice.
The SMD switches I have for my build have larger lugs than anything that came with this kit.
Most of the components are caps and resistors.
So not really applicable to the switches I need to solder.
However, it will be a fun exercise to try my hand at learning to solder those things once I can see them.
For me, I'd say 1206 is easy, 0805 is fairly OK, 0603 is a challenge and 0402 is testing the limits of possibility
I usually use 0805 for hand soldering and would always defer to machine assembly for 0402 and below
I did two 0805 caps with a less than optimal magnifying lens...not very attractive. The tip didnt seem too large for them but I want to try something smaller.
I think the button will be fine, tin one pad, hold switch in place and press with iron tip so it sinks in (you can see and feel it drop in place)
then its easy to just heat and dab solder on the remaining pads
only one leg at first to get it aligned properlyu
because then it is easy to heat and adjust if needed
I usually solder multi-legged parts in a cross pattern
Solder one corner, then the opposie corner, then the other corners, then down the sides
same here, both for switches and larger ICs. First one pin then the opposite, then the rest.
Hot air and/or a hotplate!
I populate all of my boards myself. RP2040 is a little more challenging, but perfectly achievable even without a stencil and paste. In fact the flash chip is the more difficult. 0402 capacitors are a nightmare because they tend to stick to tweezers, but with a little patience they too can be hand soldered
Feel I might need a hot air station some day. Paint stripping guns just have too much airflow and no temperature adjustment, they just burn your parts
I generally try to use the largest tip I can physically fit between my components without melting things I don't want to.
The smaller the tip, the higher you need to turn the temperature which I try to avoid. At 400°C flux just tends to disappear while the joint is still cold. More thermal mass in the tip allows for lower temperatures
I would certainly recommend getting one. Doesn't have to be a fancy name brand one! 30€ on AliExpress should get you something usable. If the temperature and air flow is adjustable you're all set! The temperature readout might not be 100% accurate, but you'll get a feel for it very quickly!
A few years ago I bought a hotair station and have used it once. Its time to pull it out and use it with the practice kit. I will need to make sure that I have some solder paste still.
I keep telling myself I will try to solder one some day. But can't get decided to do it yet. 😄
I generally use pb-solder wire to tin my pads with lots of extra flux and a soldering iron, then place my components (relying on the sticky flux to hold them in place). I then use the hot air to reflow everything with small circular movements...
I have a YiHua 853D, soldering and hot air station, bought on amazon a couple years ago for about 100€. Air flow is variable, it has also a voltage tester/supply to test your components. So far, it has been a good purchase.
Find it harder and harder to get snpb
I made some stock because I also found it difficult to find now. I have a kg or so 0.8 and 1/2kg of 0.5 (that I use onl for 0.5 pitch to avoid excessive bridging). It should last a good while
Yeah, I've got a few spools sitting in shelves... if I need new spools I'll probably look on AliExpress (Amazon also have several offers, but I've come to dislike their pricing policy... )
Reputable electronics suppliers in the EU hardly sell SnPb solder now. I haven't seen anything on reichelt.de or conrad.de in recent years...
Alternatively, the low temperature lead free solder (SnBi or SnBiAg) works fairly well... I'm just unsure about the mechanical strength, so I avoid it when soldering surface mount components that'll be exposed to repeated mechanical stress (buttons, connectors etc. )
Eheemmm... right, just looked on Ali out of curiosity... apparently the Chinese are boarding the RoHS train...😅
I want expecting to find so very few SnPb products!
All SnAg or SnCu!
A 500g spool of 63/37 is offered at 33€ 😅
Apparently Amazon isn't the most expensive place to buy SnPb anymore😅😅😅
Is the Sn/Ag/Cu solder being used more extensively in the EU due to the regulations against lead based solder imposed more than 20 years ago?
Yes, it's almost impossible to find SnPb here nowadays unless you buy it from China - which is rapidly transitioning to RoHS as @autumn thorn just said
Seems like RoHS compliant soldering is a bit trickier to do from what I was reading....higher temps required and lower flow.
Yeah, there are some issues with it.
For one, using lead-free solder should be planned for before ordering PCBs. Leaded HASL finish on pads apparently doesn't play nice with lead-free solder. If you're ordering ENIG (gold plate) it's less of an issue.
Lead-free solder in my own experience doesn't flow nearly as nice. The hotplate helps, but it is a pain in the bits and requires very thorough inspection, especially under quad flat no-lead (QFN) packaged chips like the RP2040 or Atmega 328P-MUR found on some Arduino clones.
Loads of flux certainly helps, but doesn't entirely mitigate RoHs compliant solders' shortcomings.
I mean, I usually outsource QFNs and such
Have not had issues with lead-free solder wire on leaded HASL pads myself yet, but I am sort of in the "apply generously and remove later" camp when it comes to solder joints
i only use lead-free solder for the last ten years and i hadnt had any issues whatsoever
in EU you cannot get leaded solder anymore, but I have some stannol with silver mix that is pretty good. My friends who are into soldering all swear by lead but if you do anything for sale it has to be RoHS in EU
Maybe this dilutes the lead to a point where it doesn't cause issues?
I haven't played around with lead-free a lot. The times I did I found it didn't work the way I wanted (the way I was used to with SnPb)... maybe I should give it another try and change my expectations🙃
It is still available on Amazon, although not as widely as before. I'm not up to date on current regulation... my last stance was SnPb was still allowed for personal use while commercial products have to be RoHS compliant? Has this changed?
I think the main reason for me to prefer lead based solder might be BigClive on youtube😅🤷♂️
yeah but they won't ship to eu
🤷♂️ shipping address is my home in Germany
Since lead based solder seems to be relatively unobtainable these days I reckon I'll get some experience with lead-free on my next projects
Just did some reading... I was wrong!
I was unaware that selling lead based solder to private individuals is banned in the EU. Thank you for making me aware... I'll be looking at the alternatives from now on, regardless of availability on the internet.
That is just crazy. How can you ban an individual from buying it. I have some lead free solder and I had a heck of a time using it.
I mean, it is (technically) poison, and most poisons are only sold to people who can demonstrate a practical need and the ability to handle it safely. Which is usually professionals in their line of work.
well you ban leaded gasoline from cars because lead is nasty stuff
you also ban selling dynamite to individuals without permit, or guns (in europe)
lead free works but you want 420 degrees C
it will just make a mess <400C
even though it melts, it just wont flow properly unless you heat it well
then it is fine. It is more brittle over time, but for non vibration stuff it is fine. It is just annoyingly harder to solder, leaded is so easy
May be inconvenient, so much fun stuff is too toxic to be sold to the public. But also, when you see how some people use that stuff, something needs to be done, and it is both easier and less invasive to police distribution than actual use.
And as for whether we agree with it... ultimately matters very little considering what influence a few individuals have.
this is mine
but there are lots of others, not really the only right one
its more or less the same - Sn 96'ish% and 3% silver (Ag)
but 420C or more
That is such a German answer😅❤️
Thank you for the links, @kind tulip
My next question was going to be regarding solders that you guys have had good experiences with.
Have you tried Aliexpress ? Just curious if it would work from there.
I wouldn't risk, solder is stuff you buy a roll and use years
don't cheap on it. and any leaded likely is 10x easier to work with, but with lead free those I have had success with
and flux if things turn out difficult
Do you mean ordering SnPb from China and whether it gets delivered?
I haven't tried, but I think I'll be going with the times (better late...) and switching to lead-free...
No I meant the 60/40
Jesus, Maria und Joseph!
70$US for a 500g spool!!!
I don't believe I've ever spent that much on solder! 😄
Oh my god.
@kind tulip what's "Germany" in Finnish? 😄
Yeah, the 75$ one is the newer listing
This one is listed at >50€ for 250g 😄