Ikoka Nano / Stick / Handheld

[north flicker]

Hello, this thread is for discussion, support and development of a series of designs I am working on that begin with the name "Ikoka" (“行こか?” or “行こか!” is Kansai dialect for “shall we go?” or “let’s go!”).

What they have in common: Seeed Studio XIAO nRF52840 based, with EBYTE E22-900M30S 1-watt LoRa (SX1262) modules (pin-compatible with other E22 variants, but watch out for your Tx power setting)

If you found my designs useful, do consider sponsoring me on GitHub or tipping me a coffee on Ko-Fi

• Ikoka Nano - Smallest of the series
• Ikoka Stick - Longer form factor, but uses a 21700 cylindrical Li-ion battery (18650 should work too, with a magnet/stretch of the spring)
• Ikoka Handheld - A middle ground between the Nano and Stick

If you have any issues, feedback or improvements, please feel free to raise an issue or PR against the respective GitHub repos (preferred, for tracking reasons) or drop a message here.
If you need help with JLCPCB settings or building the nodes, do drop a message here.

[velvet scarab]

The github is so detailed. Great job and thank you

[north flicker]

Published Ikoka Handheld to GitHub. I've just ordered it for myself, haven't built it yet, order at your own risk! Rotations may not be correct in the CPL export from KiCad, so check the PCBA positions before ordering!

[north flicker]

Right now my biggest worry is that I had to relocate the OLED circuitry quite far from the OLED connector due to lack of space and wanting to keep the layout single-sided, so I don't know if this will affect stability of the OLED

[merry snow]

@north flicker Your fix from last night my time boots and sees the I2C devices on my Nano.

[north flicker]

Cool, could you note that in the PR and I’ll take it out of draft

[merry snow]

OK, will do!

Channel 1 on the INA is reporting strange values however. Hmmm, I will make a note in the PR. never mind, it's off batttery LOL.

[north flicker]

Ok, the PR is ready for review

[rain spire]

Now we just need a case for the Stick. I love this thing!

[merry snow]

I’ve asked the designer if they’d like to share the design.

It’s posted here https://discord.com/channels/867578229534359593/1383128367779414056 (thanks @silent haven

[north flicker]

TP4056 gets pretty hot at 1A, but i think the fast charging is nice to have

[merry snow]

What the Seeed really needs is a way to wire in an separate reset and user switch. This format is EXCELLENT but if used in a mobile node, access to reset and a user button would really be helpful.

I love the size of this node!

[north flicker]

but definitely don’t go smaller than a 1000mAh battery or you’ll charge above 1C which is above what most batteries are designed for

it’s paralleled with the onboard charger so it’s more like 1.25A charging i think

[north flicker]

haven’t yet cleaned the flux off but

added RST

[north flicker]

Unfortunately, the OLED flex pinout is mirrored. I’ll have to fix the design. The Kicad footprint is intended for the flex to fold onto the same side rather than to the opposite side of the board.

[rain spire]

Definitely love how thin I was able to build this one. Now I just need to commission someone to design a case for it lol

[north flicker]

Oh wow! Love to see it

[north flicker]

I've fixed this up in the github repo, but have not ordered PCBs yet, kinda busy with work lately

And also spent about as much as I can on prototypes this month, do sponsor me on ko-fi/github if you have some spare cash 🙂

I'm harvesting E22 modules from old boards to resell locally, but any extra money will help to restock XIAO boards at least

Probably order plain PCBs and reflow them myself transferring parts off the old PCBs for now

[merry snow]

Done!

[rain spire]

honestly, i've never done PCBA for any of the boards of yours that i've built. I also appreciate that you don't really go below 0603 on any of the components. Helps me maintain my sanity during builds.

[fossil stratus]

oooo what's that like 802595 lipo? and what's the capacity?

[north flicker]

Hmm, when I’m designing I’m actually designing for PCBA, which is why there’s no reference designators.

Tbh the ikoka handheld is getting quite compact and some of the pads are getting hard to reach 🥲

Well to be accurate i’m designing for reflow soldering, so if you have a hot plate you’ll be fine, but if you’re trying to use a soldering iron tip you’re going to have a bad time

[north flicker]

Appreciate the support. If you have ideas about recognition, do let me know too!

[rain spire]

103395 3700mAh

[fossil stratus]

Oooooo very nice

[north flicker]

Hello everyone who is using the Ikoka Nano, the building and flashing instructions for the xiao_ble variant are very out of date, I am working on fixing it upstream.

The GitHub issue on the Ikoka Nano repo is https://github.com/ndoo/ikoka-nano-meshtastic-device/issues/10 and thanks to tremmert81-tx for reporting it, if you are here.

If anyone knows where/if an alternative web flasher exists that does have the xiao_ble binaries, please let me know, I will add them to the flashing instructions, to make it easier for those who are not comfortable with building their own firmware.

@obsidian jackal sorry for the lack of replies via DM, I have been kinda working on the PR directly, it's still pending https://github.com/meshtastic/firmware/pull/7105

[obsidian jackal]

No worries whatsoever, I greatly appreciate your replies, it’s so great to see your work getting the attention it deserves. Looking for forward to PR getting approved!

[north flicker]

if you are familiar with git, you can pull my branch and build from there

if you are not, you can download this zip file, then open the folder in vscode and the environments picker will have these variants (assuming you are using the stick/handheld/something you DIY'ed using the wio-sx1262 pinout)

if you are still using the ikoka nano (or wiring according to the old xiao_ble variant), then the xiao_ble variant from mainline meshtastic firmware will do.

[round holly]

Ho @north flicker looking at ordering some of these as I really like the look of these boards.. will be able to use up some spares I have here too. Nice work!

Should I wait for this PR before ordering?

[obsidian jackal]

Courtesy of @inland blaze:

https://mrekin.duckdns.org/flasher/

I use the "Xiao BLE" option for the Ikoka Nano.

[north flicker]

Which board?

Over the weekend, I plan to update the 2 older designs (stick, nano) to match the values i’m using on the Handheld

As well as to address some bugs etc.

[round holly]

Its the handheld

[north flicker]

Not functional bugs sorry. More accurately improvements.

[north flicker]

I haven’t tested the OLED FPC footprint yet, so i would say consider the design alpha 🙂

[round holly]

ok - had not realised there was scope for a screen!

I had quite a job finding the jlpcb parts. Some were out of stock etc

I think I have got them down to this...

Comment,Designator,Footprint,JLCPCB Part #
1k,"R10,R4",R_0603_1608Metric,C21190
1uF,"C8,C9,C5",C_0603_1608Metric,C15849
100k,R3,R_0603_1608Metric,C25803
47uF,"C3,C4",C_0805_2012Metric,C15850
10uF,"C11,C12",C_0603_1608Metric,C19702
FTC303015D220MBCA,L1,L_Wuerth_MAPI-3015,C18724424
4.7k,"R7,R8,R6",R_0603_1608Metric,C23162
DSK26,"D3,D1",D_SOD-123F,C50553
39k,R2,R_0603_1608Metric,C23144
KT-0603R,D4,LED_0603_1608Metric,C72038
2.2uF,"C10,C6,C7",C_0603_1608Metric,C23630
22uF,"C2,C1",C_0603_1608Metric,C59461
300k,R1,R_0603_1608Metric,C22975
MT3608,U1,SOT-23-6,C84817
TP4056-42-ESOP8,U4,SOIC-8-1EP_3.9x4.9mm_P1.27mm_EP2.41x3.3mm_ThermalVias,C16581
SW_Push,"SW2,SW1",TS-1088-AR02016,C720477
AO3401A,Q1,SOT-23,C20917
270k,R5,R_0603_1608Metric,C25874
E22-900M30S,U2,XCVR_E22-900M30S,
Conn_Coaxial,J1,SMA_Samtec_SMA-J-P-H-ST-EM1_EdgeMount,
1N4148WT,D2,D_SOD-523,C2128
1.25-2PWB,J4,Molex_PicoBlade_53261-0271_1x02-1MP_P1.25mm_Horizontal,C2905009
1.2k,R9,R_0603_1608Metric,C107068

I don't fancy soldering all that with an iron - using seed and ebyte I already have (thats enough soldering for me)

[north flicker]

Oh dear, actually, if you install this, https://github.com/Bouni/kicad-jlcpcb-tools

All the mappings are already there

You can add the repo to Kicad's PCM and install it from there

[round holly]

great - thanks! First time in kicad 😉

[north flicker]

Sorry, maybe the instructions in the README are not clear/incomplete, I should clean it up, the Handheld is really not release quality yet

Some of the parts have incorrect rotation in JLCPCB PCBA, there is a page where you can visually correct them, essentially you just click once (don't drag!) on components that need fixing, then hit space until they are oriented correctly

[round holly]

I noticed this too!

it's pretty easy to match them to the correct pin where they are obviously wrong. I was more concerned if parts such as D3 and D1 were the correct way around

[round holly]

Do you have plans to test the LCD soon?

[north flicker]

I am trying to batch together a few different things i want from JLCPCB because fast shipping is getting pricey. Tomorrow I will try soldering the OLED on my problematic boards reversed, if that works then the new design should work since all I did was reverse the pinout

[north flicker]

https://github.com/ndoo/ikoka-handheld-meshtastic-device/commits/main/

the PCB in that photo is from commit 30e5063

[round holly]

I see the charge current was reduced - was it getting too hot?

[north flicker]

the OLED solders to this. the part number you want is probably this N096-2896KSWEG01-H30

https://lcsc.com/product-detail/OLED-Display_Newvisio-N096-2864KSWEG01-H30_C2890596.html

The flex connector goes through the slot and solders to this

Taping part of the connector down with kapton tape (the yellow tape JLCPCB uses to tape your PCBs together, just re-use it) helps with the soldering

Or you can gently fold the flex to shape it which will help stop it from trying to free itself (obviously don't press down hard enough to crease it and break the traces)

these 4 fixes are kinda important. haha. i noticed my lipo getting a little bloated 😛 but it's an old lipo i rescued out of a thinkpad

[north flicker]

The diodes are thankfully correct. Eventually I will get around to adding custom rotations so that the generated files are correct, but haven't had the time to do that yet

[north flicker]

I updated the boost circuit on the Ikoka Nano, pushing it in a bit

[north flicker]

the diodes are good to 2A and the inductor is good to 1.35A, should give some breathing room for the E22 especially the 33S model

I know @obsidian jackal managed to blow the diodes on the current design... 😅 so some improvements are long overdue.

the power path switching also always has the boost running, so that tx power is as high as on battery power when on usb power

the connectors at the bottom are also now flush against the pcb edge.

[silent haven]

didn't see anything about this yet (or maybe i missed it). but is anyone having charge issues with the Ikoka on the Xiao? mine only gets up to about 3.8V on a 2Ah LiPo before it stops taking charge. curious if anyone noticed similar on the Xiao

[soft shuttle]

2 batteries ?

[merry snow]

Maybe I’ll pull my attic test (ikoka) down and try to reproduce this here this weekend.

[north flicker]

Is that how it measures on a multimeter too? I was wondering if it was a voltage divider thing (i.e. measurement is wrong, not voltage is wrong)

[silent haven]

correct, the divider appears fine as it measures the same in the multimeter. it behaves the same between two batteries i’ve tested

[fossil stratus]

just fired up my nano v3, had 2.7.0 and was reporting 4.22V when on USB, 99%/4.19V when I disconnected and paired with phone to see batt w/o USB. then upgraded to xiao_ble-2.7.3-5ae8021a.uf2 from bender's buildbot and seeing the same #s with a Makerfocus 1100mAh LiPo

[merry snow]

That’s good but you are starting with a full battery. @silent haven is reporting the inability to charge UP beyond 3.8V.

[fossil stratus]

hmmm true - i've def drained it down but on 2.6 - i'll leave it on over night and see how it does recharging from < 50%

[north flicker]

The actual battery voltage is 3.8V or 4.2V?

[silent haven]

it’s a 4.2V lipo battery. i think it’s a brand new”pkcell” from Adafruit

[north flicker]

I mean, measuring it when the charging stops with a multimeter, does it read 4.2 or 3.8

i have an issue with maybe 2 of my 8+ xiao nrf boards too where it stops at 3.8, but didn't think much of it, since i have reflowed those so many times that i assumed i had damaged them lol

[ionic cipher]

Might be worth adding an ibom export:
https://github.com/openscopeproject/InteractiveHtmlBom

For the firmware compiles, I need to have another look, but GitHub actions are now essentially free, so less need for vscode.

[merry snow]

@silent haven I've thrown a 3Ah Makerfocus on an ikoka Nano, it's at 3.93V currently. I'll let it burn down to 3.7 or so and then plug it in to USB power. Would that test help?

I don't have a smaller battery with the right connector currently.

[silent haven]

maybe? if your is getting over 3.8V then i think that is already different than what i am seeing. unless it gets to about 3.93 and stops there.

[silent haven]

yes, it is 3.8V on the multimeter. other boards (like a RAK) will charge it up to 4.2V

[merry snow]

Wasn't yours an issue charging? I threw a 75% full battery on it, so I need to drain it down to do the charge test.

I'm heading out for the afternoon, but it's going to take a while to do. Maybe I should put a crappy antenna on it, then put it full power router late and let it blast my house with 1W of ikoka love

OK, done. That should speed things up quite a bit.

[silent haven]

it is not a voltage reporting issue. it appears to be reporting what the battery actually is (the multimeter agrees) so it seems to be a charging issue. it's also the same on multiple 4.2V LiPo cells so it isn't the battery (they also seem to be going up to 4.2V on my test RAK19003 as well)

it was unclear to me if there is some charger influence on the Xiao based on GPIO pin state (i couldn't really grok the schematic)

i've got my hands on a stick now so i'll tesst the same thing on the stick. it could just be this xiao board not good

[fossil stratus]

I'm still waiting on my battery to drain down low enough, we're testing MediumSlow now so there's not a lot of LF traffic to help deplete it

[merry snow]

@silent haven my batttery is down to 64%/3.84V so I'll let it go overnight and whereever it lands, I'll charge it up and report back here.

[north flicker]

i can't think of why this is happening, the schematic looks simple enough

[silent haven]

wonder if possible p0.14 or p0.31 get stuck open and they actually act as a sink during charging?

[north flicker]

This is a good point. I will have to read the code to be sure. I did make some change regarding those pins because Meshtastic implemented proper functions to toggle an ADC pin because of Heltec T114 (and possibly other variants) so I changed the variant.cpp code while doing some cleanup.

this is the specific change
https://github.com/meshtastic/firmware/pull/6930/commits/9a15101f6e183cc0eee83e9a9cacc7c3ba83f0be

i'll have to read the code to make sure the pins are getting set correctly using this new way of doing things

[north flicker]

So, I re-soldered the oled on my old PCBs, but this time in the correct orientation for the footprint i was using, but it still got stuck on init, i may have damaged the flex from how many times I had desoldered it

[north flicker]

it was not so much the heat but just that it was over 1A of charge current which is more than 1C for 1000mAh batteries, most batteries the size of the pcb are around 900~1200mAh or so.

[merry snow]

@silent haven it's down to 3.7V, I'm going to plug it into USB and see how it charges.

[fossil stratus]

haha i just plugged myne in too, finally dropped to 3.64V overnight - sitting at 3.74V on the charger now

[merry snow]

i put mine on full power and router late (with a poor antenna, sitting in my office) in order to burn up the battery faster 🙂

[fossil stratus]

i got 23hrs off this same battery in testing (also had an INA3221) but that was when we weren't testing MediumSlow (and have too many 1W on MediumSlow right now to also run the Ikoka at higher power so left it on LF)

[fossil stratus]

welp i've broken 3.80V (3.81V currently with one 3.82V so still climbing) - could be ADC slightly off but seems like this one isn't affected at least

[north flicker]

Liking this form factor, but still no idea how to get the OLED working, need to order a bunch from LCSC…

Trying to squeeze in a larger inductor, this one gets burning hot during Tx but then rapidly cools back down lol. Or maybe I will just stitch vias through the board to help it cool down…

[ionic cipher]

What value inductor are you using?

There's some nice flat ones on JLC that I've been enjoying for the micro MPPT.

[north flicker]

22uH but rated for exactly 1.2A

or maybe 1.3 i can’t remember. 3x3mm footprint. going to go up to 4x4mm but have to squeeze things a bit more.

but that gives me 1.4A which gives some breathing room and the slightly larger footprint probably better for reducing heat density

[merry snow]

And mine has climbed through 3.84V and is still climbing.

[ionic cipher]

I think you need to double check the difference between switching current and output current (and ripple current)

I'll dig out the inductor on my micro-mppt later. On paper it was great.

I was trying to work out why you've got traces in places, but that slot...

You could save some space by making the holes into half-holes. The hole type you've picked can be used with economic assembly without having to have edge -plating

I don't know what resistor values you're using, but would resistor -arrays save space?

[north flicker]

yeah the heat rating current is just barely enough so

[north flicker]

my pcb qualifies for economic no problem

[north flicker]

unfortunately i don’t have four of the same value anywhere

wait which holes?

[ionic cipher]

The ones with the vias at each corner

You can make the hole unplated, and just use the vias to connect top & bottom

then turn them into half-holes

[north flicker]

Oh, how does that save space?

Or do you mean to have them just be part of the edge of the board

[ionic cipher]

yes

[north flicker]

like a corner cutout for the screws to fix on to. Hmm

[ionic cipher]

locating notches

[north flicker]

How does that work with an enclosure? I’d love to see an example

though i guess if my screws are countersunk they can absolutely be at the corners

I would just get rid of the screw holes entirely and just have it slot into a case, one end secured by screws and the other by the sma locking nut 😂

Kept a 1mm clearance around the edge of the board in case i decided to go that way lol

[north flicker]

Oh!

[north flicker]

Looks like 22uH is way too high of an inductor value for the input and output voltage. i need to pick a lower value. vibe'd it, looks like i'll pick a 2.2uH https://chatgpt.com/share/68766370-cf40-8009-813d-bf85f6b7753f

[ionic cipher]

Nah, I'll do some calcs later

Double check 😉

[north flicker]

Sure, for now I will switch to 2.2uH, seems like i can avoid that burning hot inductor without changing the footprint too much, if the heat output calculations are correct

[ionic cipher]

The reasoning is probably correct - core saturation from either ripple current or the peaks during tx.

What inductor are you using now?

[north flicker]

https://lcsc.com/product-detail/power-inductors_cjiang-changjiang-microelectronics-tech-ftc303015d220mbca_C18724424.html

[ionic cipher]

And what's your current target?

[north flicker]

is these 😛

https://lcsc.com/product-detail/Power-Inductors_cjiang-Changjiang-Microelectronics-Tech-FNR3015S2R2MT_C167747.html

[north flicker]

i'm also gonna change the oled schematic values to match the meshtiny 🥲 just in case

[ionic cipher]

https://www.powerelectronicsnews.com/the-dc-dc-boost-converter-power-supply-design-tutorial-section-5-1/

Inductor selection gives a couple of formulas halfway down

Switch frequency is 1.2mhz, and duty cycle will be whatever the output power Vs rating is.

[ionic cipher]

ok, 1/10th the DC resistance, and 2A Isat is better than the first one

If you can find something with a little bit more Isat...

[north flicker]

tagged releases on github, ordering PCBs to verify

[north flicker]

wow, they don’t close at night eh.

[ionic cipher]

JLC engineer saw your order, and made the call. "Ndoo ordered some more prototypes - put on an extra shift."

[rain spire]

Does this also impact the stick? Aka do I need to swap caps lol

[rain spire]

Does this look a little too detonator-esque to travel through an airport with safely? Granted, it will be packed in a back pack and not taken out. Just worried about TSA scanners lol

[ionic cipher]

Add a fake mouthpiece and speaker

[merry snow]

I haven’t been asked about mine across several flights. They scan it and see the battery. That’s all they care about out.

[north flicker]

yep, swap inductors not caps. well the stick’s inductor is quite large so it will happily work at the slightly lower efficiency

[north flicker]

Is mrekin here? I was wondering if it would be straightforward to add the xiao nrf52840 e22 variants to his flasher

rather than the full variant definitions

oh right looks like i just need to submit PR against https://github.com/mrekin/MeshtasticCustomBoards/blob/main/build_list.yaml

[ionic cipher]

@inland blaze

[ionic cipher]

What's different?

Power limit?

[inland blaze]

Here)
yaml file just lists variants to build (used by jenkins)
what do you want to add?

[north flicker]

the 2 variants i submitted
seeed_xiao_nrf52840_e22_900m30s
seeed_xiao_nrf52840_e22_900m33s

available since Meshtastic Firmware 2.7.3.cf574c7 Alpha

[inland blaze]

do you need daily builds or releases only?

Added with release builds and one daily builded fot testing

[north flicker]

Just releases should be fine since it’s in 2.7.3 but daily builds might help someone else.

[north flicker]

Power limit and no GPS - I2C by default

UwU

The OLED reset pulse and OLED charge pump isn’t going to like that… I’ll salvage caps from other boards meanwhile

I imagine they must have had to allocate a bunch of money on coupons for this, if they’ve really put a full 4000 cap reel of 100pF caps where customers needed 1uF

[north flicker]

@obsidian jackal if you are using the new pinout style you can get firmware there 🙂

https://mrekin.duckdns.org/flasher/

the nano too but with “xiao ble” instead.

[obsidian jackal]

Thank you and @inland blaze for adding those options, I can confirm that it works!

[north flicker]

handheld still getting stuck trying to detect oled... don't know why

[north flicker]

It’s a fresh OLED, don’t know why it’s hanging the i2c bus. pinout is correct too

[north flicker]

my worst nightmares came true, hahaha. the 30 pin pinout numbering on the ssd1306 is totally different than the FPC 30 pin numbering.

need to do a new PCB 🥲

or is it... i am trying to find the correct pins for i2c, they seem different from flushable to my schematic

it seems correct... :/

[north flicker]

sigh

wHY?????

[merry snow]

@north flicker : in the Stick there are two grounded holes for mounting on one end , and closer to the USB there is a single hole that’s a different diameter and very close to the Seeed pins. What was the design thought around the use of this hole? Was it for mounting? I’m finding that to be safe for using it with a screw, I’ll need to use a plastic washer to avoid touching the MCU pins.

[round holly]

Congratulations 👏

Is a PCB design change required or just remove resistors?

[ionic cipher]

On existing boards, it's "just" snipping the resistors off and making sure there's some pullup on the i2c - he's on with that now.

Well, he will tomorrow 😉

[fossil stratus]

i think the holes on the xiao/usb side are to fit the positioning bumps on the bottom of a 21700 holder

I ended up using the "tab" holders because none of my NiteCore 21700s fit right into the 21700 holder without tearing the plastic wrap (and risk exposing the ground)

[north flicker]

If you use the Beilamoo 21700 holder (seems to be the most common SMT 21700 holder on Aliexpress etc.) there is a polarity pin in the plastic housing that goes in that hole (it won’t sit flush if you have the polarity wrong)

[north flicker]

Just leave the resistors off. I’ll circle them on a pic in a bit.

I didn’t use the Keystone 21700 holders because between those and generic chinese ones it made more sense for me to get a bunch of the generic chinese ones since neither had a KiCad footprint available

If the Keystone had the footprint on kicad i’d have used that, they’re much better than the chinese ones in terms of the mounting frame sturdiness

[ionic cipher]

Are they glass-filled?

[north flicker]

As much as you would trust a chinese oem to.... I don't think they are but let me check. I think just plain nylon or PBT.

[north flicker]

if your PCB looks like this, then these resistors need to be removed to un-stick the I2C bus.

[north flicker]

High Temp Nylon 25% Glass Filled,UL 94V-0

https://www.beilamoo.com/sdm/1074412/4/pd-5180803/21061576-2986071/ONE_21700_Battery_Holder_with_Surface_Mount_SMT.html

[ionic cipher]

Excellent.

I still got it 😉

[soft shuttle]

Did you ever lose it in the first place ?

[ionic cipher]

Can't lose what you never had.

[soft shuttle]

No, but you cal always find it.. ;P

I find sanity sometimes.. But it's a slippery lil f*cker..

[north flicker]

ok, the i2c seems to freeze from time to time without external pull-ups, gonna put them back in

the inductor runs pretty cool now using a 2.2uH inductor, but i might go for a higher spec 4.7uH on the ikoka handheld just because the mt3608b datasheet says to pick from 4.7~22uH and 4.7uH inductors still have similar DCR and heat current ratings as 2.2

[ionic cipher]

It needs 4.7 minimum - see the spreadsheet I made and the ti application note.

I can't remember the limiting factor in the spreadsheet was.

[ionic cipher]

ok, I did another bit of twiddling - at I_in_min = 0.01A = 10mA, and Vin=4.2, Vout=5, the duty cycle is 16% and the inductor current is 14mA

Which, if it uses fixed frequency switching, means that at 1MHz, L_min will be 28uH

if you assume the schottky diode is 500mV, Dmin jumps up to 23%, and L_min goes to 40uH

I don't think its quite that bad, but I am no expert 😄

[north flicker]

10mA because you’re optimizing for idle right? not Tx. actually it might be worthwhile using the Tx switching to turn off the boost converter when not Txing

i would be a bit worried about the soft start meaning the tx PA isn’t getting full power soon enough though

all that need for optimization and the rise time means i think it’s better to just waste some power in idle

It’ll be running in PFM mode anyway, it can’t be that bad

In the earlier versions of the Ikoka Nano I did use the SX1262 RST/EN line to control the boost converter

that would save power and prevent (rapidly) draining the LiPo when Meshtastic shuts down.

[ionic cipher]

There's some really fast ones with like <0.5ms startup times

[north flicker]

for pcb routing simplicity i removed that trace but I should really squeeze it back in

[ionic cipher]

There's also some that basically allow the current to go negative to handle overshoot

i.e. they can handle a very low duty cycle.

[north flicker]

idk how much i wanna optimize idle, power users are generally on busy meshes

just wanna minimize blowing up or inductors desoldering themselves :p

[ionic cipher]

This is also a worthy goal

Just use the inductor they say

Do you have a schottky diode on there?

[north flicker]

yes, it was a DSK26 but i have changed it to a DSK34. 700+mV Vf to 550mV Vf and 2A -> 3A continuous

[silent haven]

anyone have a ebyte radio fail? my nano no longer finds the radio:

INFO  | ??:??:?? 2 Final Tx power: 22 dBm
INFO  | ??:??:?? 22 SX126x init result -2
WARN  | ??:??:?? 22 No SX1262 radio
ERROR | ??:??:?? 22 NOTE! Record critical error 3 at src/main.cpp:1412
INFO  | ??:??:?? 22 PowerFSM init, USB power=1
DEBUG | ??:??:?? 22 State: BOOT
DEBUG | ??:??:?? 22 Node status update: 1 online, 1 total

[fossil stratus]

oh no... guess it was random and not after a firmware update (e.g. pointing to sw not hw)

[silent haven]

i hadn't done any change to the radio (SW or HW). i did try a wipe and update to latest alpha. but didn't seem to change. nothing obvious stands out

[ionic cipher]

Is it hand made or PCBa?

[rain spire]

this new firmware structure's got me a little confused as to which variant i should be building for the Stick. Can anyone point me in the right direction?

[fossil stratus]

I had a variant.h i'd drop into variants/diy/seeed-xiao-nrf52840-wio-sx1262/ but you should be able to use seeed_xiao_nrf52840_e22_900m30s as the target as ndoo added 1 and 2W variants (note: i haven't tried building for xiao since the reshuffle and that target was added)

[fossil stratus]

i've seen the 30S/33S happily get past that during boot but then shortly thereafter get into a busytx/crash state when there's not sufficient current for the radio and it's trying to xmit so i'd lean towards checking the signal pins but not really how sure it'd die so suddenly like that, even if the LNA burned out from an antenna that came lose I think it'd still be able to talk SPI (hypothetically/theoretically)

[north flicker]

i did have an e22 fail, just stuck hot, but it was quite abused (moved between multiple boards and reflowed back down)

hot apparently means the PA is stuck on all the time, which, if the module is fine, means the tx_en pin got stuck high for some reason

[silent haven]

this one is hand soldered (based on visual) but it doesn't seem to get hot. i'll try reflowing the joints

[ionic cipher]

Bought from someone?

[silent haven]

correct

visually, D1 may have been nicked' during soldering of the Xiao

[ionic cipher]

Can you share photos?

And have you asked the person who sold it?

@north flicker is your design allowed to be sold?

[north flicker]

Yes I used CERN OHL-P which is the most permissive OHL license

[ionic cipher]

Cool cool cool.

[north flicker]

here is a very rudimentary enclosure for the ikoka handheld. use M2x10 to M2x15 screws. SHCS, pan head or countersunk will work. The holes are designed for SHCS but pan head will work. Countersunk will fit weird but still be fine.

i'll put it on the git repo along with my readme update when i find the time... 😦

ALIGN THE OLED WELL BEFORE PUSHING DOWN ON/TIGHTENING THE LID. I broke my OLED because I didn't.

The bottom edge of the OLED glass MUST line up with the bottom of the E22 module. I'll take pictures after replacing the cracked OLED.

If the OLED is aligned, it will be held in place by the ridges on the inside of the lid. If you are worried about the tolerances of your printer/tape causing excessive pressure on the OLED, don't use my enclosure!

as for what LiPos will fit, here are the ID

sorry, i just wanted something to chuck the device in my bag, so i did not spend time to make it pretty.

[merry snow]

@north flicker for your next design, think about USB on the side so one can charge with the node (and screen!) vertical, sitting on a desk.

[north flicker]

Yeah, i tried, doesn’t fit in the current pcb size

but i think it can be a bit taller haha.

The other option is getting rid of the lower 2 screw holes and using a clip/latch on the bottom of the enclosure

[merry snow]

Oh I just meant as you are considering future projects. I have an Ikoka Nano enclosure (designed by @silent haven ) that turns the Nano so the USB is on the side of the enclosure. The Nano lacks a screen so it’s a simple design decision.

Anyhow, we’re all fans of your work.

[obsidian jackal]

FYI: you can use the Qwiic connector to attach an i2c OLED.

[merry snow]

Thank you. I knew that.

@north flicker keeps evolving his designs to address different use cases. That’s what I’m giving feedback on. Why did he include an OLED header on the Stick if not to make it easier for users?

🤦‍♂️

[north flicker]

I’ll try!

[north flicker]

Haha you make it sound like I made BaseUI happen

Tbh that was a nice coincidence timing-wise, the regular UI wasn’t really worth the work but baseUI sure is

You know, the nano is square because I couldn’t decide on an orientation 😭

[merry snow]

And that’s what makes it perfect in that way. 😃

[fathom yacht]

hello guys, I having a weird issue with my xiao ble . I was trying to setup xiao ble + E22 module. I build meshtastic firmware for xiao_ble and upload to xiao. I able to connect over bluetooth but when I try configure it, it stucks at 50% reading from radio. upload firmware again and start working for a short period of time. now I cannot make it working again. tried reflash several times with no lucky

[merry snow]

Which Ikoka are you building for? nano, stick or handheld?

[fathom yacht]

I think I ask under wrong chanel. I'm building a Xiao BLE (diy/xiao_ble)

[merry snow]

I think of them synonymously from a format perspective.

Have you tried a prebuilt firmware for it? I’ve used this many times for community supported variants https://mrekin.duckdns.org/flasher/

[fathom yacht]

Tried now and no success. it´s not finding SX1262. I wired it following diy/xiao_ble

[merry snow]

That first line is telling me that the radio is dead or not connected. Has this node worked previously?

[fathom yacht]

yesterday It was working

[merry snow]

Did you also hook up VCC and Ground on the radio? Those are left out if those instructions but naturally are required. Maybe others. I’d have to go load up hr ikoka schematic on KiCad to see.

[merry snow]

Oh, did you run the radio without an antenna perhaps?

[fathom yacht]

tried now reflash with fw I build for xiao_ble . SX1262 was founded and I able to connect over bluetooth. I set region and it´s reboot, after reboot, appear again to add region, but now cannot read from radio, stucks at 0%

[ionic cipher]

Can you post a photo of it? There might be some clues there.

[fathom yacht]

no, always with antenna connected and VCC and ground came from XIAO

[fossil stratus]

have you tried powering the E22 with a dedicated 5V -> Vcc that can deliver upto 1A? it's very power hungry when it xmits and while it boots ok using 5V off the Xiao I don't believe it has enough current to support E2290030S (or 33S) at full power which is ~650mA on the 30S IIRC

probably also help to see the serial console/logs from right before/during/after a reboot vs just the initial boot

[ionic cipher]

easy enough to test this by turning it down to 11

[ionic cipher]

https://tenor.com/view/oh-my-star-trek-gif-19129737

[fossil stratus]

https://tenor.com/tIES.gif

[ionic cipher]

Very frustrating when folk don't get this.

It's one louder, ok?

[fathom yacht]

I was able to get it working after connect at other power supply, I configured again and everything works fine untill I disconnect from wall power supply and connect in to powerbank, afert that, start not reading from rádio, I disconnect from powerbank and connect at same wall power supply and doesn't work anymore.

[fathom yacht]

Is it possible to monitor serial using an wall power supply?

[ionic cipher]

yes

you can either make a data-only cable, or make sure that the everything comes from a source with a common ground

[fossil stratus]

hrmmm, could be the powerbank is seeing too low of a draw from the idle nRF52840 + radio and powering off? almost all of them behave that way, and once the firmware crashes from trying to xmit at 22dB with insufficient Vcc it'll stay freaked out until reboot even if you swapped back to the wall supply.
try 1: setting the xmit power very low, like 5 (there's still +8dB in the LNA so you'll be 13dB)
try 2: ONLY powering off the wall supply for like an hour and confirm it's stable which will point to the powerbank shutting itself off because the idle draw is to low

[fathom yacht]

I'll follow yours tips to deep test. One don't, could an under voltage power supply breaks meshtastic firmware?

[fossil stratus]

it gets unhappy when the radio isn't working as expected, and stays unhappy until rebooted even if enough power was applied because it's already in the "something very wrong" state

[fathom yacht]

got it. first attempt to diy with meshtastic (I´m very comfortable with esp32 under my home automation) but I found meshtastic is a little different.

[fossil stratus]

ahh nice! there's a lot of overlap but when you start playing with the 1W radios insufficient power is a common problem since they want an extra 500mA or so of power to transmit at full power and things act weird when there's not enough power

[fathom yacht]

I reflash to start over again. format file system using nrf_erase_sd7_3.uf2 and upload xiao_ble-v2.6.11.60ec05e.uf2 . after upload completes, I unplug from PC and plug it at wall power supply, pair over bluetooth, set region and xmit power to 5, click save and reboot, after that, appear again to set region and not able to communicate with radio, stucks at 0% 🙁

INFO_UF2.TXT content

UF2 Bootloader 0.7.0-22-g277a0c8 lib/nrfx (v2.0.0) lib/tinyusb (0.12.0-145-g9775e7691) lib/uf2 (remotes/origin/configupdate-9-gadbb8c7)
Model: Seeed XIAO nRF52840
Board-ID: nRF52840-SeeedXiao-v1
Date: Jun 14 2023
SoftDevice: S140 7.3.0

[fathom yacht]

A little advance, I replace VCC and gnd Dupont cable from E22 module, and now system is up and running with xmit at 5 powering from wall power supply. Let's see what happens next

[north flicker]

you must be able to supply 1.25A or so reliably at the module (i.e. have a capacitor near the module) otherwise things will go very poorly.

datasheet says to have around 100uF of capacitance as close to the module supply pins as possible. wiring it flying like that is a bad idea

at that point you should use the non amplified module (m22s) or just any other generic sx1262 module

[fathom yacht]

I wiring like that way only to test and see if everything works fine, and after that, I´m planning to buy an pcb as well. could I solder 100uf capacitor directly at E22 module (I can see an spare vcc and gnd at board) . I have others SX1262 (heltec v3) but my goal is to have an 30dBm node

It's possible solder like that ? I saw I have at hands only 220uf or 470uf capacitors, will it works or need to match 100uf ?

I have replaced all Dupont cables to better ones, but still have system unstable. I saw an note for 100uf capacitor, but don't give it a attention.

[ionic cipher]

You can

Whether or not you should...

It will work

I.e. add the capacitance you need

I wouldn't power it up and down too often, though

[fathom yacht]

I'm trying figure out where is issue. I have haplace all Dupont cables to better ones and system works but still having issues. Let's try capacitor

[ionic cipher]

I mean, if you're able to get a suitable PCB, that would be better

But spaghetti is fine too.

[fathom yacht]

Spaghetti (lol) is only for test propose. I'm planing use an PCB as soon have it working stable

[merry snow]

Question for @north flicker or anyone else who knows. Is the ikoka nano limited to 100mA charging via the solar or USB-C inputs? On other NRF variants I see that it might use 100mA at high states of charge but at lower states of charge it can charge at upwards of 400 mA. I know the variants all have slightly different charge circuits.

[ionic cipher]

As far as I can see, it's limited to the charge rate of the Xiao, which is supposed to be 100mA

[merry snow]

Ah, ok. And that makes sense.

[ionic cipher]

In fact, all the chargers I've used (admittedly only reputable and high(er) end) have preconditioning for low state-of-charge batteries, i.e. 10% charge rates

Some have a bypass to use with supercaps

But those are the more expensive ones. The cheap ones without preconditioning are relatively sought after for supercaps.

[merry snow]

Thanks

[north flicker]

my newer ikoka handheld has a tp4056 separately set at 800ish mA (forgot what value i used, it's in the schematic i think) for a total of just under 1A with the xiao

the rest, yeah, limited by the xiao

[merry snow]

Time for me to try that one out then!

[ionic cipher]

Oops! Missed that...

Oh, handheld! I've not looked.

[fossil stratus]

there's a note in seeed's wiki about the charging needing a pin set high just to get the 100mA (which the variant.h handle when I've checked) but it is annoying the onboard is anemic vs the ~250mA I see on nice!nano promicro clones - and super awesome you bypassed that silliness on the handheld 😄

[fathom yacht]

Capacitor in place and system is up for 21m with 22 xmits

[fathom yacht]

System is up for 1h

[fossil stratus]

sounds like you're ready to move to PCB! also thank you for running "repeater" with the correct CLIENT role (unlike some of the people we have here that didn't read the manual, don't have amazing position and ruined LF/slot20 for the rest of us)

[ionic cipher]

I'll be honest, in the early days the xiao was looking interesting, but that low charging current and my lack of knowledge of external chargers meant the promicro was the best choice

Now I'm a bit more adventurous, the lack of Io is holding me off using the xiao

[fossil stratus]

I have a few Xiao Plus but the extra pins haven't been needed (and the relocated battery pads means won't work w/ndoo's designs)

[fathom yacht]

Thank for all support. Like I said, my intention is to use an 30dbm node for airsoft Milsim operations and this is why I rename node as repeater and client role, will be strategic positioned between operation base and airsoft operators.

[fathom yacht]

My goal now is power system from battery instead of wall charger, let's see how it works

I'll think in a PCB and maybe an GPS to be able track node in the field

[fathom yacht]

15min system running from 5v ups Aliexpress board with two 18650 in parallel

[silent haven]

i have a ikoka mini also no longer working. i'd like to test some power options.

were you able to find a good way to test external power with the existing E22 module while still soldered to the board? in the photos it looks like maybe you are testing with a fresh E22 module

[fathom yacht]

5h system is running stable over 2x18650 parallel battery. In my case, needs to add capacitor for E22 VCC and gnd. It's only to test propose, I'll add those to an PCB

This is power source

Thinking to buy this one to use 6x18650 to be able haldle gps as well at least 24h

[north flicker]

yes 100mA is way too conservative. but I expect they had to do it because of how hot the tiny chip would have gotten if they set it any higher.

[north flicker]

Better to continue your dev this the main nRF channel #nrf52 or #1194757507013427250 so more people see it

[north flicker]

did the diodes blow up? my diodes were definitely undersized. i just built one with the newest pcbs from the git repo

the sod-123F can definitely take a lot more current. the flat leads help to dissipate the heat back into the traces.

@obsidian jackal had his come unsoldered from the heat of constant Tx which is not good. <-- this was the previous (current?) version with the SOD-323 diode which is much smaller

the decrease of inductor from 22uH to 2.2uH also greatly decreased the heat output in general on the dc-dc.

[silent haven]

just a fyi, i had a second ikoka nano stop working. nothing visibly wrong on the boards. this failure is slightly different: it stills sees the radio. but the radio does not appear to be able to send or receive anything (amplifier maybe burned out or something?)

[north flicker]

this is the early version with the small diodes? do you measure 5.2V at the module?

starting to wonder if the e22 doesn’t like 5.2V over 5V

[silent haven]

can you point me where to check?

i am trying the stick right now but if maybe the USB voltage is too high i'll stop using it before it burns up 😬

[fossil stratus]

is the stick v.0.2.x or v.0.3.0? bender had both in the last batch he sent me

[north flicker]

I’m in the jungle and away from my personal PC for a while. But you should see 5.2V on pin 9/10

[amber spade]

Quick question — does anyone know if there’s already a fork of the Nano and Handheld with ready-to-use production files for JLCPCB (Gerbers, BOM, CPL)?

If not, I’d really appreciate if anyone who has already generated them could share their files.

Thanks a lot in advance! 🙏

[north flicker]

I’ll add Gerbers + production files to the release page over the weekend

[amber spade]

Many thanks!!

[midnight bolt]

@north flicker hi, i was goign to get several ikoka sticks made, but is Hardware 0.3.0 the latest or recomended version?

[north flicker]

Sorry I’m out of town until the weekend without access to my laptop so I can’t give you a certain answer

[fossil stratus]

yeah v0.3.0 is the latest, at least unless there's some rev ndoo hasn't pushed to the repo in months but can confirm v0.3.0 works as-is (myne does at least)

[midnight bolt]

Yeah I've been using 0.3 but was going to make another set

[north flicker]

Sorry, I owe you guys updates to the release pages with gerbers and manufacturing files, but I will get around to it soon, was finishing up some Ikoka Pro test units for the local mesh

I’ll add the info to the top message after I sort out the release files for the existing designs. There’s already some improvements I had to/want to make.

For the pro, I plan to move the BLE module to the right side (top on the pcb orientation) to free up front panel space for a power switch.

And yeah I plan to figure out power switches for everything else too, more than one person has asked me about it

[spring wraith]

Ikoka Stick enclosure by @frosty rose it fits like a glove!

[north flicker]

just curious, is this using the tall OLED 2.54" headers or some custom/shorter ones that i should put in my BOM instead of the full height ones?

happy to change the BoM if there's a good short header part i can swap to, i just didn't have the patience to dig around jlcpcb/lcsc's piss poor parametric search

[spring wraith]

this is the one we use: https://www.amazon.com/dp/B0CN373JF4

since the display is a little floppy when inserted into the board, the enclosure has a spacer that you install between the display and the top of the E22 shield and there are guides on both side of the front case to keep the display in place.

[north flicker]

i see, but the headers are just the full height 2.54mm/.1" smt female headers originally specced?

because i realized the short ones also work, the oled pins just go all the way down and poke the pcb a little, but it's harmless, saves a tiny bit of z-height

[spring wraith]

I think it is full height?

[north flicker]

ah i mean on the board side. you just ordered according to the bom i guess

okay cool. just making sure it's compatible because someone (sorry, username slips my mind) built a super slim one

[spring wraith]

Yes, I am pretty sure it was built to the bom. There is usually advantage to make a radio as slim as possible. But for the tall Ikoka stick design, it is okay if it is a little thicker to balance out its length and make it possible to sit stably vertically on a flat surface. it is a good match to Alley cat's Beam series of enclosure design.

Would be great if you have a more squarish Ikoka design later

[north flicker]

ikoka handheld? but the usb port faces down which is annoying. i need to make it a bit larger. it is a bit too small now

washtastic works well too

definitely too small 🤣

[ionic cipher]

too big.

and my phone has a port at the bottom

[whole jolt]

@north flicker I was just directed to your nodes as a solution for a 1w lightweight battery-backed node. I was also told that there was an issue with ADC voltage divider readings at the moment -- are those solved, or will reading the battery still be inaccurate?

[spring wraith]

Yes, would be great to have the USB port on the side. The T-Echo is a great example, charge while standing up.

But can be worked around with a 90 degrees bendable antenna. There EBYTE one that comes with station g2 is very good. Same low SWR straight or bend

[north flicker]

Hmm which device has ADC divider issues? I think maybe the Pro but that is relatively early hardware, I just built 5 of my first prototype run ("v0.9.0" and already have a v1.0.0 im gonna order in a bit

for the rest, they are all using the internal divider on the XIAO module, I have seen some reports of bad readings from those, but i suspect they were either DOA from Seeedstudio or got dislodged somehow during hotplate reflow

[whole jolt]

Not sure which, I'll ask my friend

I am interested in building some sticks, however

[spring wraith]

@north flicker can you charge the 21700 battery through the USB-C port on the Xiao?

[north flicker]

Yes, but the charge is only 100mA so it will take some time

[spring wraith]

That matches what I see on my USB power meter. I feel better now that I know my unit is not broken. Previously I inserted a 21700 battery in reverse and I smelled magic smoke. I quickly removed the battery. I replaced the very hot Xiao with a new one. But I didn't know it is normal for it to charge at 100 mV.

[fossil stratus]

100mA is actually the Xiao nRF52840's "boosted" mode (set by flipping a pin in firmware), default is only 50mA - it's definitely anemic to get a 21700 up to full but once deployed with a panel that gets enough sun for most of the day it should be able to top itself up (previous testing with E22-900M30S + Xiao nRF on a Ikoka Nano v3 got me 23hrs of runtime in NYC's busy mesh off an 1100mAh battery that's been through a bunch of discharge/recharge cycles)

[fluid cloak]

What’s the biggest 21700 you can get? 3000? How long would one of these last on that

[hasty glen]

you can get 5000mah

https://liionwholesale.com/pages/search-results-page?page=1&rb_snize_facet5=21700&rb_snize_facet2=Flat top&tab=products&sort_by=price&sort_order=asc

one of these claims 6000mah, i've had success with this retailer before

I think you could hook up an external charging circuit to the second battery connector, not sure if that will blow up the xiao

[fluid cloak]

Cool! How long do those last?

[north flicker]

i'm using a 5000mAh made by EVE. it lasts around 3-4 days

[fluid cloak]

Really? I thought the boards where NRF like a RAK will last about a week on a 1000. Does the more power just draw a ton more I thought the Lora chip isn’t on if it’s not transmitting but I guess it’s always receiving

[ionic cipher]

The radio is using 6x as much power during tx as the one on a RAK, and the RAK module is lower power than the seeeeeeed one generally, too.

There's a boost circuit to account for as well (sorry @north flicker ) which is probably around 85% efficient.

[fossil stratus]

also depends on how active the local mesh is but little over 1k/day is about right for NYC

I have a Nitecore 5300mAh and yet to run it all the way down but I don't use that node much and pop it out instead of letting the nrf drain it down

[north flicker]

The LNA is always on when receiving so there’s quite a bit more idle draw too

[midnight bolt]

I built the boards to the BOM, that's what I shipped alleycat

@north flicker what's the latest on the pro? And on the stick what changes are underway or you are planning? I'm loving the stick and have got the Seattle region quite excited. I got some in circulation that sparked interest and @hasty glen and I are doing 2 different large buys to get more out in the region

[round holly]

LNA on the ebytes can be removed. @paper maple has done exactly this.. would only recommend it in very high noise environments.

[north flicker]

at that point maybe a g-nicerf

[paper maple]

Yep. Has sorted out my issue here. But I would only recommend if you are comfortable with doing so. Some more info here (i'm messords on reddit) https://www.reddit.com/r/meshtastic/comments/1n1z2lw/e22_900m30s_is_deaf/

[spring wraith]

@north flicker FYI, Alley Cat has released her Ikoka enclosure STLs on her discord server's beta channel. https://discord.gg/XDKDJyRZYK

[fossil stratus]

unfortunately it only fits if you built to the exact BOM, including socketed Xiao and even then my sockets seems to be 1mm too tall which need to try and dremel down so the back can close

def some cleverness in how to get the PCB inside though!

[frosty rose]

I can make small tweaks super easy, just put feedback in my beta channel.

[rain spire]

I already advocated (and sent measurements) for those of us that solder without reservation and cringe at the sight of a socket... the flatmounters 🤣

[frosty rose]

Wasn’t sure if it was the same setup!

[rain spire]

haha I think @fossil stratus did something similar to what I did, but I doubt its socketed. I'm just messing around. Even with the Faketec's I've always tried to get them physically as flat as humanly possible.

[frosty rose]

I can make a case for this with the long lipo on the side, then the OLED will be the decider of the overall thickness. 😅

That’s a whole new case though, so let me finish these 5 or so other projects first!

[fossil stratus]

haha yeah my original one was just pins, so I built another but with whatever random sockets i had lying on the workbench that ended up being 2mm too tall so the xiao side of the PCB was titled up/button was too far recessed/whole case didn't close... so i shaved the sockets down and now it all fits 🤌

[rain spire]

No rush, I built a stick specifically to fit in the case you made already haha

[spring wraith]

@north flicker, @frosty rose released her Ikoka enclosure on printables https://www.printables.com/model/1428538-acos-ikoka-stick-beam-case

[fluid cloak]

For the boards how do we know which board is which power DB?

[ionic cipher]

E22-900M22S = 22dbm output
E22-900M30S = 30dbm output
E22-900M33S = 33dbm output

The latter two are the right size for the Ikoka boards

[fluid cloak]

Thanks

[amber spade]

Hi @north flicker ! Would you be able to update it with the gerber and manufacturing files? Apprieciate your work!

[north flicker]

Sorry, I still don’t have access to a laptop/desktop as i’m overseas in a plantation

I bought a laptop recently though so hopefully i can get more of this done while on the road in future

[fluid cloak]

Are we able to use the ipec or the SMA on the Ikoka stick or only sma?

[ionic cipher]

Either/or. It's better if you leave the sma off if you're using the ipex

[fluid cloak]

So if I’ve got the sma on I should only use the sma really

[ionic cipher]

Theoretically or practically? 😉

[round holly]

Assembly of my first ikoka stick board completed today. Very nice job. Thank you.

[round holly]

Sharing this here. Thanks @west jolt

https://makerworld.com/en/models/1861469-meshtastic-meshcore-case-for-ikoka-stick#profileId-1991002

[spring wraith]

@north flicker how much power does a 2W ikoka need from the Xiao's USB-C port? It seems the USB-A port on a Pi5 is not enough to power it. Would a 1W ikoka work better with a Pi5?

[north flicker]

Ebytes DS wants 1.5 amps so you definitely need a way to inject much more than the USB port can supply e.g. a splitter cable

[spring wraith]

would a splitter cable be connected to the battery holder terminals?

[north flicker]

a dedicated 5V supply supplanting/replacing the usb 5V
gnd - to pi usb-a AND external 5V gnd
usbd+ - to pi usb-a
usbd- - to pi usb-a
5v - to another 5V supply or the pi 5V pin (if i’m not wrong this will bypass the overcurrent/fuse on the pi usb-a)

[rain spire]

@north flicker - is there another Inductor that's suitable for the Handheld if we're using a E22-900M30S? The FTC303020D4R7MBCA appears to be very hard to come by outside of ordering from LCSC.

[fossil stratus]

Heh yeah I ordered myne from LCSC but have some extras if you’re feeling impatient

[rain spire]

I'm feeling VERY impatient. Did you also get the DSK34's or Did you misread the labeling and figure a SS34's a suitable replacement like some of us?

[ionic cipher]

Coilcraft have an equivalent

[rain spire]

XGL3020-452 or the XEL3520-471MEC?

[ionic cipher]

Trying to find it now...

Cjiang basically have the entire coilcraft catalogue covered and then some

And may actually be the OEM for them, if the spec sheets match up...

[rain spire]

I also came across a digi-key friendly one that looks like it's in the same ballpark: MPS MPL-AY3020-4R7

https://www.digikey.com/en/products/detail/monolithic-power-systems-inc/MPL-AL5030-4R7/12418266

[ionic cipher]

Basically it's a 3x3mm footprint. Can't see all the bits on my phone

[ionic cipher]

No, too big

Resistance is 60 typ, 73max, isat is 4.8A typ, 4.0A max

[ionic cipher]

I've found 2 potential candidates from wurth on digikey, but they both have worse dcr

[fossil stratus]

LCSC had a 20 quantity minimum so I have spares, including some that are now in the USPS system enroute to a.li3n 😄

[ionic cipher]

The cjiang 1008 and 1212 ones are amazing.

I genuinely can't believe they exist, except I've actually got them on my designs and seen them working.

[ionic cipher]

20 of something that costs $0.16 is not bad 😉

[fossil stratus]

Hehe a bit more with tariffs but yeah I’ll take 20 and then when I lose one or two or ten during assembly I still have spares

[ionic cipher]

My assembly skills are slightly better with ceramic tweezers. I presume it's because they're stiffer, or maybe they don't get as sticky from the flux.

[fossil stratus]

hah I use ceramic for adding a little pressure on the PCB while on the hotplate, but they're chipped so not as useful for tiny components

[fossil stratus]

Anyone else checked their Ikoka's output on 2.7.11/2.7.12? I had the Handheld hooked up last night at while lora.tx_power was set to 22 I was only seeing 20.5dBm on RF Meter (there's an RF Filter + adaptors so ~1.5dB loss makes sense). Had to set app to 30 to get "expected" 28.5dBm reading

[ionic cipher]

That sounds like the correct behaviour?

Ndoo added proper offsets for the radio modules now.

[fossil stratus]

For E22-900M30s it used to be capped at 22 and the LNA took care of the extra 8, but I was curious after the new patches to address Heltec v4's power what it was actually putting out

[ionic cipher]

And surprised that it is correct? 😜

[fossil stratus]

surprised I needed to adjust, I haven't looked at the PRs closely

[rain spire]

@north flicker - just so you're aware, the latest gerbers for the handheld on your GitHub are the broken ones where the screen doesn't work. Some people that aren't paying attention, like myself, might order them because the gerbers already generated instead of generating gerbers from Kicad

[sage sparrow]

As a recipient of an Ikoka stick who just tried installing a screen with no luck, how do I check to see if the board(s) I have fall in this bucket?

[rain spire]

Sorry for the confusion, my comment was regarding the Handheld and only the handheld. I've built 3 sticks and they all turned out fine.

[round holly]

On mine I think vcc and gnd are the wrong way around for the screen. Have you checked this?

[supple tapir]

You can buy screens on Ali or Amazon with Vcc & GND swapped round.

[ionic cipher]

Swapped which way? There's technically no "right" way, I think

[rain spire]

The issue with the screen that I had mentioned was regarding the old design having pull downs instead of a pull up attached to it.

[ionic cipher]

Oh, on the i2c lines? That's not good...

[fossil stratus]

The Stick is Vcc/GND/SCL/SDA IIRC, make sure it's sitting on top of the radio for orientation, also make sure you're using the right firmware variant as the seeed nRF52840 kit firmware works but that's setup for UART instead of i2c so screen won't work, need to use the seeed_xiao_nrf52840_e22_900m30s variant

[fossil stratus]

GND/Vcc/SCL/SDA is far more common when ordering random ones off amazon or ali but yeah, no standard just like JSTs - sometimes you get lucky

[sage sparrow]

Completely hypothetically, what happens if one plugs in the wrong screen? 😬

[ionic cipher]

Magic smoke ensues 😶‍🌫️

[ionic cipher]

What's the cure? Snip snip snip or just external pull-ups?

[fossil stratus]

IIRC it was remove the pullups on that rev but don't quote me on that (or anything elsE)

or I guess pull-downs in a.li3n's case, also sending out "current" PCB to him in a bit

[rain spire]

Print the correct version of the board 🤣

[ionic cipher]

Oh, that old chestnut!

Only 2 rules:

1. do it right
2. don't do it wrong

[fossil stratus]

3> ???
4> profit!

[sage sparrow]

Smokes the board or smokes the screen? Or both?

[fluid cloak]

Anyone know how bad it is to use the ipec over the sma? There hooked up to the antenna pad so it should be fine right?

[ionic cipher]

It's fine.

[fluid cloak]

Really? Some people tell me No don't do it and some say its fine

[ionic cipher]

Who do you trust more?

🤨

[ionic cipher]

Noone here is going to tell you something that will damage your device.

The worst that can happen is you get worse signal.

This is fine.

[rain spire]

When you say DIY stick, did you build a Ikoka stick PCB yourself? Or did you use a Breadboard and truly make a DIY stick?

[frosty rose]

https://tenor.com/view/bravest-warriors-gas-powered-stick-cat-bug-bug-cat-cartoon-network-gif-10832364482942016648

[rain spire]

thats a TOTALLY different kind of stick haha

[rain spire]

Is there a ikoka handheld case floating around anywhere?

[fluid cloak]

There is alley cats case

This one https://www.printables.com/model/1428538-acos-ikoka-stick-beam-case

[ionic cipher]

He's looking for a case for this:
https://github.com/ndoo/ikoka-handheld-meshtastic-device

[rain spire]

The Stick, Handheld, Nano and Pro are all different boards, but thank you.

[fluid cloak]

Ohhhh

[ionic cipher]

Yeah, Ndoo has a habit of actually publishing his designs...

[rain spire]

@north flicker - for those select few of us that prefer DIY to PCBA: is there any chance we can beg you for a handheld variant with i2c pads, or possibly a ribbon connector for the display instead of the solder-to-board pads that I and at least one other person ( @fossil stratus ) have spent wayyyy too much time trying to get working but were ultimately unsuccessful?

I LOVE this little device. Seriously, it's my favorite one of yours to date, despite mine looking like Dr. Frankenstein assembled it:

[fossil stratus]

I'd even say a proper connector for the ribbon cable would work

[rain spire]

That's what I meant by the connector instead of pads lol

[fossil stratus]

oh LOL ... i was half awake when* I read that 🙃

[rain spire]

aside from @north flicker (and PCBA services) has anyone attached an OLED to the pads on the back successfully?
And by successfully, I mean it's attached and working as intended.

[fluid cloak]

Anyone have an average power draw for the 30db version of the ikoka stick?

[fluid cloak]

.15w when idle with a peak of 1.15w when transmitting

[round holly]

Depends which radio. I have a 33S. Will look later at usage and feedback.

[merry snow]

Average depends on how often it's transmitting. So it could be very low or very very high.

[fluid cloak]

Yeah so im guessing around .5w average

[merry snow]

Average what? If you are in the center of a busy mesh, and it's running as a router, it's going to use a lot of power. If you are using it on the edge of a mesh and in CLIENT_MUTE mode, it's not going to use much more than a RAK4631/19007 as the Seeed at that point is the main power consumption and the radio is silent.

There is no "average".

[fluid cloak]

Im making an average

[merry snow]

ok, your car is going to use 5 gallons of gas this week. I'm making an average for you.

[fluid cloak]

Its a router in PDX so busy so im getting if its 1.1w part of the day and .2w the other half and then theres is an average

[fluid cloak]

more like 50 but yes

[merry snow]

You are going to have to see how busy your node is in the real world to come up with your power consumption. I assume this is so you can plan around solar and battery... well, just using .5W might get you a number, but it's not based on anything.

I'm not trying to be grumpy or nasty here.... you need real world numbers in order to plan for anything.

[fluid cloak]

Yep thats what ill be doing

[merry snow]

So, I apologize if I came off a little snippity above (maybe I did not, maybe I did... on average it's a maybe 😃 )

[fluid cloak]

so a 10w pannel should be more and enough because it draws 12wd and my panel puts out 40wd

wd = watts per day

[merry snow]

No real numbers, but I had an ikoka nano as a client_mute node in a Harbor Breeze and it was doing great on power overnight, charging up during the day, etc. Then I put it into Router_late mode and moved it up the hill where it could get more coverage, and the battery ran down twice as far that night. I don't have the numbers committed to memory, but I was concerned that running it in router_lote move was going to require a bigger panel and battery setup than the Harbor Breeze allowed.

Ikoka Nano is basically the same power consumption as a Stick with no screen.

[fluid cloak]

Yeah the harbor breezes really need a xiao board or something a lot less

[merry snow]

Not if your part of the mesh isn’t doing much. That’s my point above. I’ve been running an ikoka nano harbor breeze for several months on MT and for over a month on MC.

[fluid cloak]

Oh wow, Ive had issues running a harbor breeze node on mc with just a xiao in it but the mesh here is pretty big

[merry snow]

Are you using the node charging or just letting the solar charge the battery and connecting the battery to the node battery in?

My ikoka nano charges at 100ma. So I let the solar panel directly charge the battery and then connect directly to the battery for node power.

[fluid cloak]

The solar charger is charging the battery and I run the xiao off it

[merry snow]

As we head into winter I will find out if my HB ikoka survives.

[ionic cipher]

@north flicker when you return from the wilderness, do you think it's worth putting the GPS pins onto the NFC pins of the i2c variant?

I've been giving it some thought

[fluid cloak]

Has anyone tried adding a battery protection chip to make a variant of the Ikoka stick?

[ionic cipher]

We've been chatting about some stuff in the background. Not sure Ndoo has much spare time right now, tho.

[fluid cloak]

Cool!

[uncut path]

I am looking at the Stick too. Could you sub in an E22P?

I am trying to get caught up here on the difference between the standard E22 and the P.

[rain spire]

certain pins aren't the same (one or two, I don't recall what ones they were), so you'd either be cutting traces and running jumpers, or you'd need to wait for an updated board.

[uncut path]

Ouch, bummer. Good to know though. At least the schematics are around to rewire things

[fossil stratus]

yeah you'd need to cut the trace for RXEN and jumper that to something 3v3 as on the E22P it's not RXEN but RFEN but otherwise should just work. probably.

[fluid cloak]

Can anyone verify with me that the 2w Ikoka sticks draw around 4w when transmitting?

For the 33dbm

[unique matrix]

From the E22-900M33S datasheet:

[rain spire]

Is this project still being maintained?

[merry snow]

Do you need a fix or improvement?

[ionic cipher]

I think people are expecting/wanting those, yes. I also think Ndoo will provide them as and when they can.

[merry snow]

I was responding to @rain spire because his question seems a little like it could be asking if ndoo has abandoned his projects (plural)

“This” is not a project. It’s a discussion that includes at least 3 different projects. Each project is hosted on GitHub and I just looked at two of them and some of the open issues are answered but not yet marked closed. Some are firmware questions about Meshtastic.

Anyhow, my real question should have been “what issue or fix are you interested in and have you opened an issue on the GH page”.

I don’t think @north flicker has signed up for continual improvements for his older designs like the Nano. But I don’t speak for him.

[ionic cipher]

Fair point(s)

[rain spire]

While I realize that this is the conversation for 3 different models, I just meant the Ikoka as a whole. I was just curious if @north flicker had any future plans of improvements/changes or if they had committed as much as they're going to for these projects (which is totally fine, since they've given us 3 fully working models).
My hope was that they were considering adding support for the E22P's, but I'd understand if that just isn't in the cards.
I should have phrased my question differently, focusing on future development vs. simple maintenance/upkeep.