off-topic-tech

[stuck nebula]

Youtube has a better streaming service but it's a terrible platform for new streamers

as long as they dont create a new algorithm solely for livestreaming

[twin dew]

And that the MBA doesn't need to understand anything about what his/her department does.

[night girder]

Twitch lost focus. On almost everything. It was a gaming platform to stream.

[stuck nebula]

what gaming platform

[twin dew]

At same time: "You know best" and "You don't need to know or understand anything under you".

[night girder]

But their hilarious back and forths with the rules. Banning their biggest streamers for shady reasons (not being transparant) allowing hot tub girls selling farts.

Yeah, I say they digged their own grave.

[stuck nebula]

they will be fine

as long as Amazon dosn't decide to abandon them

[night girder]

Yes they will now.

They closed South Korea.

And fired 1/3 of employees at Twitch.

But define "fine".

People said here that Twitter was doomed to fail once Elon Musk took over. I said it will be fine. But what is fine? Twitter is now X.

[stuck nebula]

i mean that's what they wanted for and what they get

[twin dew]

And still ongoing train wreck moving with the mass it had before Elon took over.

[stuck nebula]

it's really stupid for a ceo to earn millions by just signing some stuff

and the best part is, when they are hugely fucked up, they never held accountable of their stuff

[stuck nebula]

dont even get started man, twitter users are just another species

im happy it exists and i hope it never closes

so those people stay there

Reddit also started to become really similar to it

[night girder]

I don't know. They are implementing new features. Sharing job opportunities for example.

Ofc, only for premium in US.

[stuck nebula]

Musk is trying to the X like a wechat mega app

that will not work in western world

[night girder]

or X to compete with Linkedin.

[stuck nebula]

at best he will milk already existing Twitter users

that's all

[night girder]

is he even still in charge over X? Didn't he appoint someone else?

[twin dew]

And at same time the "Free speech absolutism" is allowing people and topic back in that is driving the advertisers away, causing lawsuits, including big EU one etc.
And at same time Elon is also trying his best to supress any critique of him on the platform, or topics he doesn't like like Pro-Ukraine content, so so much for the "Free speech absolutism" too.

[twin dew]

In theory.
Officially there is now separate CEO, but that doesn't really matter in actuality.

[night girder]

we don't know for sure. Can be a puppet. Can be a good person in the lead.

[twin dew]

Point is that the company is running out of money.

[stuck nebula]

"Free speech my ass" we've seen how he ran to Israel after the AD companies backlash

[twin dew]

When the CEO serves at leisure of the owner(s), so Musk.

[night girder]

That's a puppet.

But do you know that for sure?

[twin dew]

And Musk is still very much poking at the lowest levels of Twitters employee stack.

Like with those anti-musk comments just losing upvotes constantly.
Some people getting shadow-banned.
And so on.

[stuck nebula]

That is normal though, it's like buying wechat and starting anti chinese propaganda there

most of twitter's users are already western alligned people, atleast in ideologies

[night girder]

So I heard, but I haven't seen any data to back it up. And I did google for it a few times. It's not that X is being transparent (for obvious reasons)

[twin dew]

The advertising has crashed.
And the company needs to now pay more in interest payments each year (about 1B) than it ever created profit before.

And the per person "Twitter Blue" revenue is small drop compared.

Elon in lawsuit preparations has said the company is now half to tenth the worth he paid for it because of the advertizer loss.

Trying to blame that on some organizations, and not the moderation decision changes.

[stuck nebula]

so what isnt that his problem? If AD company dosn't like the current stand the X takes, don't they ahve all their rights to pull back their support,unless there is a contract between the said companies?

wether it's ethical or not, that's never the problem in business world

[twin dew]

Trying to blame some non-government organizations that tracked the antisemitism etc. blowing up next to ads from companies, and providing the data to those companies.

Which then stopped putting up new campaigns.

Elon I mean.

[stuck nebula]

well, the winners are here the old Twitter board who tricked him to pay that much for Twitter

XD

[night girder]

I don't know, it's a sensitive topic. The whole world has their opinion about Musk.

and only like 10 people or somethign actually know the man. His family, wife, kids, etc.

So the rest (including me) bases their opinion on facts. Those facts are the news. And the news is biased AF in this story.

[edgy hazel]

don't most of his kids distance themselves from him as much as possible?

[night girder]

Can be 😄 What did he call one of his daughters again? Angel? Luck? Was a weird name.

[twin dew]

That Musk completely changed how Twitter was moderated, and companies have stopped advertising because of that?
And Musk has started to prepare lawsuits related to that?

[stuck nebula]

i mean, even his current succesful companies aren't his own creation anyways

[night girder]

O2?

[stuck nebula]

there are many accusations of him forcibly "stealing" those companies

[twin dew]

First is fact, last if fact.
And the companies have stated why they stopped advertising too?

[night girder]

wait.

he has 11 children!?1

[stuck nebula]

no idea how many exactly, but he has many

[twin dew]

With several different women over the years.

[stuck nebula]

and one of his child is named X if i remmeber right

[night girder]

Oh, one of his kids is named "X Æ A-12 "

[stuck nebula]

poor child

[stuck nebula]

yeah that one i guess lmao

[night girder]

Then Jeff Bezos is also evil. Diaper story. Roomba story.

[stuck nebula]

i'm not calling anyone evil or not, but just saying those accusations exist to point that he is not that "genius" as he seems

[twin dew]

SpaceX was in reality created by him.
Paypal and Tesla weren't.

[night girder]

well Amazon kept underselling their own diapers to make another diaper company to go bankrupt.

Just because that company didn't want to sell to Amazon.

[stuck nebula]

im not surprised though

[twin dew]

And with SpaceX and Tesla, there were lot of filters in place to separate the good Musk ideas from the bad ones, because Musk himself doesn't seem to see any difference.

[stuck nebula]

yeah, that makes sense

atleast he has some good engineering team that does really good stuff

[twin dew]

And with both of those, he also knew he didn't know, so was willing to listen and learn.

[stuck nebula]

that's a reality

[night girder]

Lately, he seems to become more and more unhinged. Mentally.

[stuck nebula]

his biggest mistake was getting into politics and twitter

[twin dew]

But "he knows programming".
And yes, he went full alt-right insane during COVID.

[stuck nebula]

no sane person can stay sane there

[night girder]

The interview where he just said "Fuck em" really surprised me.

[twin dew]

Because of the limitations forced during it.

[night girder]

Well, Elon is a powerful man. Wether you like it or not. He holds some cards, and thus has power. Aka his satellites.

[charred pewter]

its nice to have money and be able to actually speak your mind in public without caring what others will whine about

[stuck nebula]

well this is a big debate, i really dont understand people turned into anti vaccine out of nowhere

[charred pewter]

or i should say "it must be nice"

[edgy hazel]

especially since the "fuck them" part was the only good part. The rest of the interview was an absolute shitshow

[stuck nebula]

i understand the complications of vaccines that had to be developed and had short cycles of test

[charred pewter]

the anti-vaccine stuff was from a specific person follower group who listend to idiocity

[stuck nebula]

but we got out of pandemic in just mere 2 years

thanks to that...

[twin dew]

Mainly because the closing of Tesla factories etc.

[night girder]

I didn't watch the rest, but the way he said fuck em. The cringe in the room.

Sometimes you think he is a comedian trying to have fun with journalists.

[twin dew]

That was the trigger.
He was against all limitations from start, but that pushed him completely over in my eyes.

[edgy hazel]

don't. It's bad and just hurts

[charred pewter]

heh... people gasping ... fuck marketters? what? how dare someone speak truths!

[stuck nebula]

then.. he would vent crazy if he lived in china

people werent even allowed to leave their apartments in Shanghai for months

and they were Vaccined

[twin dew]

He was somewhat unstable forever, but those Tesla factory closured in California pushed him into the extreme.

[stuck nebula]

in EU, i was able to go anywhere i want as long as i had my covid vaccine certificate

[twin dew]

Before that he wasn't that interested in politics etc.
And even stopped working with Trump.
To turning into semi-trumper after.

[charred pewter]

being a toxic walking spreader of virus is americas 'god given right' ... lol

[twin dew]

And now he really is very afraid that Russia will start WW3 with nukes if we don't give Putin Ukraine.

Which drives lot of current things.

[night girder]

Not only you Baldur. There are documentaries that said the same: covid really made Musk go mental. Because suddenly his factories had to close and he didn't agree with that. He felt that the world was against him.

[stuck nebula]

Russia is more restrictive in terms of Nukes compared to US

[twin dew]

Yup.

[stuck nebula]

even if putin said send nukes, that would not happen

[twin dew]

Very probably not, doesn't matter to Musk.

[night girder]

Eh, not to the outside world. They are very ambiguous.

[charred pewter]

i dont think putin wants to end his life, because the only reaction to sending nukes, is being nuked yourself

[stuck nebula]

exactly and even that is not the case

someone else from his circle will murder him the moment he goes crazy about nukes

[twin dew]

Only allowed if Russia is attacked itself as first step.
And then more limitations on top of that.

[stuck nebula]

because oligarchs has ALOT to lose

[night girder]

don't forget, elections are comming in US, so positioning you more central as a CEO is wise for upcomming president.

also Russia is a market for Musk. So many things that are playing here.

The one thing I do realize, is that lending starlink to Ukraine, was fully for his own benefit. To hold one over the american president.

[twin dew]

Elon was originally very pro-Ukraine, until Russia started saber rattling about escalations from west and nuke fear mongering.

[night girder]

He is what people want to hear at that point. Just like politicians.

[twin dew]

Complete turn after that, to this day. To pro-Russia.

Very quick one.

[stuck nebula]

the Whole Ukraine conflict is just sad, if you for a second think, that we are seeing a modern ww2 warfare were many young people die for nothing

[night girder]

I don't know happy conflicts.

[night girder]

I follow the story closely. I don't think he was pro ukrain for real. He saw an opportunity to get in the news with his Starlink.

[stuck nebula]

it just feels... dystopian

[night girder]

He used starlink the blackmail the world (remember he wanted to shut it down).

[stuck nebula]

i have my chill life, had everything in my life, meanwhile people are getting bombed and dying for nothing

[night girder]

And suddenly, the next day, after a few calls from the president. He agreed to keep starlink up and running.

[sterile plinth]

Y'all I have an issue

[twin dew]

From delivering free starlink service (on top of the bought ones), and anti-jamming changes to prevent russians from jamming the connections.
To dropping all that and limiting even the paid service etc.

[night girder]

speak. and be helped. maybe.

[sterile plinth]

So it is a known fact that in rooted phones banking apps refuse to work for safety reasons.

But 1) my phone is not rooted, it just had an unlocked bootloader

2)I am still running the stock Xiaomi OS even with an unlocked bootloader

3. not only my banking apps, but any NFC functionality at all is disabled

[night girder]

One last thing: I think Mark Zuckerberg is happy that he isn't the boogyman anymore. Been a while since I read something bad about him.

[sterile plinth]

No file transfer, no scanning, no anything

[twin dew]

They don't check for just rooting.
Unlocked bootloader is also often checked etc.

[sterile plinth]

My issue is that NFC itself doesn't work

[twin dew]

Because if you have rooted phone, just the rooting detection can be circumvented.

[twin dew]

Have you checked with anything but banking stuff?
And for the other stuff, do you have any suitable other end-point for the specific software?

Or doesn't allow to even enable it at all?

[stuck nebula]

im a total boomer when it comes to mobile stuff

[sterile plinth]

Again, when I say NFC functionality is disabled, I cannot even press the button

I press the NFC button but it doesn't turn it on

[twin dew]

Then contact the Xiaomi support?

[sterile plinth]

The thing is the banking app was saying some stuff about security, not about NFC not working

Could those 2 be completely seperate problems

The banking app telling me about security stuff and the NFC not working in general

[twin dew]

Or try what the HW diagnostic app (behind code) says about NFC HW functionality?

*#0*#

Into dialer

[sterile plinth]

That's a new one

[night girder]

Play Integrity is an API that is used by applications to determine device compatibility and security state. It has replaced SafetyNet for the most part, with a deadline of January 2025, when Google's SafetyNet servers will go offline. Apps that continue to exclusively depend on SafetyNet will no longer work once this happens. Most developers have already migrated to Play Integrity.​

To check Play Integrity status, you can do so through the Play Store.​
​
Tap the Profile icon in the upper right, go to Settings > About > Tap Play Store version 8 times. This unlocks developer mode in the Play Store.​
Now go to Settings > General > Developer options > Check integrity.​

It seems apps can use an API to check security on device, if that returns the bootloader state, apps can act upon it.

xdaforums.com/t/how-can-the-bootloader-unlock-cause-the-nfc-to-stop-working-and-can-i-mess-around-by-re-locking-bootloader.4646888/

[twin dew]

And at least Sony IIRC deleted some parts of the base OS stack when you unlock the bootloader, so that they cannot be copied out via that.
Even if you don't flash new OS.

[sterile plinth]

Seems to be broken

[sterile plinth]

I will check on that I guess

[twin dew]

Seems to not be there anymore when I check now.
Just got old results before.

[night girder]

Make sure to read the links badabing2003 posted. They seem pretty solid.

Oh fuck. I think my mac just bricked 😒

It was updating. I come back and now it wont turn on anymore 😒

Oh wait. It's comming back to life.

Holy shit. It was black screen for 5-10 minuts.

What is this for an update process? turning off the Macbook for so long.

My ethernet adapter didn't have a light anymore. Same for my iPhone. So 99% sure it was off. But why for so long.

Please hardware Gods 🙏 Can you stop fucking me over daily? 🙏

[charred pewter]

musta been firmware update

[night girder]

That long. It just gave me a heart attack.

I have my second monitor connected to my macbook pro. Sitting next to me. Whenever my second monitor goes to black, my heart skips a beat.

I am very paranoid/jumpy with black screens atm.

[charred pewter]

for me incredibly so because my mbp black-screened itself twice after i got it... seems apple finally fixed it, as its been fine since

but yeah, those software updates where it goes blackscreen without any growbar... is REALLY UNNERVING

[jagged snow]

It's almost like that used to be standard 😉

[night girder]

So I am used to the screen going black for a bit. But the thing was, the screen didn't go to black. My peripherals all were out of power too! (So the macbook was shut down). It stayed like that for 5-10 minutes. Then I got Apple logo, it continued to update for 25 more minutes. A lot of restarts. a 1GB update, just didn't expect it.

Atleast the netflix bug for going to fullscreen is fixed it seems. Cropped 3/4 of the movie when going fullscreen in Safari. People said it was a Netflix bug. But now I updated the OS and bug is gone. Mhh.

[nimble cargo]

my heart stops when my screen goes black. i'm starting to think the smoke smell and the reboots are connected. I mean, maybe my neighbour is outside at 5.30am smoking in the heavy rain and it's coming in my open window, but...

[night girder]

annoying right?

[nimble cargo]

very very. been running tests all night to see if something is failing.

[night girder]

my test is a fun one, gaming 😄

[nimble cargo]

true that. but i'm getting a bit over the whole screens go black and reboot thing. >.<

at first it was just satisfactory, now it's happening in other games

[night girder]

PC goes into instant black screen and reboots without BSOD and the eventviewer just shows a 41?

[nimble cargo]

lemme double check ev, but yeah i think so

[night girder]

Criticals will be event 41: PC didn't shut down properly.

[nimble cargo]

yeah, and zero beforehand

[night girder]

AMD CPU?

[nimble cargo]

5 3600

[night girder]

mhm, not sure on AM4. If I understand correctly from Baldur, who knows far more on this subject, only certain hardware errors will make PC shut down hard.

[twin dew]

AM4 AMD CPUs have in cases have had problems that they needed positive CO offset to be stable with U8, and some other games.

Just couple in here, but at least three that needed +10 all-core CO to stabilize.

By default many MBs are swallowing corrected errors.

[night girder]

"corrected" 😂

[twin dew]

And then just rebooting on uncorrected ones.

[twin dew]

Most ECC codes can now recover from at least one bit errors.

[night girder]

Well, it does correct the error. By shutting down the system 😄

Oh like that. Didn't know.

[twin dew]

Those are the corrected errors.
And then uncorrected errors cause reboots.

[nimble cargo]

Before AMD released a fix, I had all my usb devices uninstall/reinstall whenever I took a screenshot in EVE. I find it funny now, but... ugh.

[twin dew]

When 2 or more bits are corrupted.

[night girder]

I am still not sure how bits and ECC is connected the MHz (frequencies).

How the wrong frequency can cause bits to be wrong.

I though the frequency was the wave that can deliver the package. And the package itself are bits/data?

[twin dew]

Below certain frequency, you don't get any errors, above point you start to get errors, and above another point it doesn't work at all.

And at some point inside that "has errors sometimes" has "has uncorrectable errors sometimes" too.

[night girder]

This is the reason.

[twin dew]

Not related.

[night girder]

The purple, isn't that a frequency?

and the red dot the signal we send (the bits)

for DDR, two signals (two packages) can be send for one wave or clock cycle?

[twin dew]

In a way.
But it goes into analog electronics.
If the signal waveform has enough time to rise and fall in the time allotted by the frequency.
With the voltage, drive strenght, line capacitance etc. and reflections.
So that the other end gets the right bit or not.

SDR has sample once per clock, DDR has twice per clock.

But if the bit is get right depends on the actual data signal, not the clock signal.

[night girder]

Still a bit confused. Seems I need to learn way more. I though FLCK was clock signal. I just got confused how frequency can mess up bits.

[twin dew]

Stuff like this:

[night girder]

Oh I see. The real practical side of signals. Not a diagram.

[twin dew]

The left and right red spots are where the signal is changing, halfway between sample points.
And if that center black area isn't clean enough, it means the signal didn't get into reliable state before sampling is done at the other end.

And can cause corrupted transfer of a bit.

More frequency, less sideways space, so the signals have to rise and fall faster.

Digital signals are analog at the lowest level, they just are sampled at specific points and interpreted as 1 or 0.

And if the actual analog signal wasn't in the right state yet, because of too high frequency, too low voltage, etc. then you get data corruption.

[night girder]

Thanks Baldur, going to check for a YT video to watch 😄

[twin dew]

Called "Data Eye" for that pic type.

The clock signal just tells when to sample from that actual data signal.

And with DDR, that data signal is changing at twice the speed of the clock signal.

And the actual frequency signal is analog like that too.

There are no real square wave anywhere.

[night girder]

I don't know why they use squared wavelengths.

Easier to visualise?

[twin dew]

Just simplification.
Shows the theory.

And is enough for most uses.

[night girder]

and the jitter, is that caused by the reflection of the wave?

Sorry for the bad drawing.

[twin dew]

No.
And not really relevant to this.
Just something that the specific source was talking about.
Same for the SNR.

[night girder]

Ah ok.

[twin dew]

No, it is time going from left to right, with signal voltage on other direction.

Over many passes.

[night girder]

never had physics tbh. Kinda regret it now.

[twin dew]

With the clock signal changing from one state to the other at the center top-down line, and that marking the sample point.

So the data signal and the clock signal jitter a little compared to each other, and both don't always follow the exact same path (same voltage level at same repeated point in time).

Spreading the combined waveform over thousands of cycles.

The color tells how often the data signal voltage was at that level, at the specific time compared to the clock signal hitting specific voltage level.

Just oscilloscope picture with multicycle data collection on, and trigger set to the clock signal to sync the various takes.

[night girder]

interesting

[twin dew]

So added the red bars to show where one repeat is.

So those points are halfway between the clock signal trigger points.
And the pic will repeat after each in identical form if you widened the time window.

[night girder]

I am now staring at the wikipedia. But now I am more confused lol.

but this image atleast debunks my stupid idea that the other wave was comming back haha 😂

[twin dew]

That is the reflections part, but it just changes the voltage level in specific point in time.

[night girder]

What I am not understand is why there are two lines, binary 0 and binary 1. Because, a computer only needs to know 1 right? If it's 1, then it's not 0. And if it's 0 then it's not 1.

[twin dew]

On the analog side is is often that of the voltage range, lowest 1/3rd is 1, middle 1/3rd is unknown and highest 1/3rd is 0.

But the cutoff-levels are implementation dependant.

For old TTL (Transistor-Transistor Logic) the levels were 0.0-0.8V for "low" (often 1), and then 2.0V-5.0V for "high" (often 0).

And unknown between 0.8 and 2.0.

The receiver might interpret it as either depending on various things.

[night girder]

I understand.

going out for some food and then look on youtube for a video to understand analog waves better.

I want a better understanding 😄

[twin dew]

Main point is that behind all the digital signaling, there is actually analog layer at bottom that is very messy.
And gets harder to get right higher the frequency.

[night girder]

Yeah, that I understand. Higher frequency, less "snr" (the black eye area between two signals) and higher chance of errors.

[twin dew]

Mainly less time for the actual signal to rise and fall, and so the black area will be smaller.
And if it gets too small, you get errors.
More voltage helps to some extent to make the signal rise and fall faster.

[night girder]

There are some really scary images on wikipedia. Eye pattern of a 1.25 Gbit/s NRZ signal with a four-inch stub:

Context for image:

... an additional three inches of cable is added to the end of the same stub. The same "step" is present but is now four times as long, producing reflections at about 1280 ps or 1.6 UI. This produces extreme ISI (since the reflection of each UI arrives during the subsequent UI) which completely closes the eye.

[night girder]

Just watched a video, which said the same as you Baldur. But in different words. And I get it now. Atleast the basic. And it makes sense.

The bigger the difference in voltage (optinmal sampling instant) the easier it is for the receiver to differentiate between those voltages. The bigger the voltage differences, the bigger the eye?

[twin dew]

Higher feed voltage, higher current to same line, so higher change speed.

[night girder]

Big eye = good and efficient. And it makes a whole lotta sense.

[twin dew]

The feed voltage and the signal differentiation levels are not always synced.

[night girder]

Looking on amazon how much a Oscilloscope costs 😂

Looks like something fun 😛

[twin dew]

From hundred or so for 10-50MHz max ones to tens of thousands for ones that could look at GHz level signals.

[night girder]

https://www.youtube.com/watch?v=cL7QsELuv_M (giving this one a view, so interesting).

w2aew (youtuber) makes the correlation between clock speed and data signal.

[twin dew]

SDR signal with sample on rising clock and change on falling clock.

[night girder]

yes 😄

[twin dew]

Some ringing to be seen for the signal.

The wobbling after the change.

[sand saddle]

I think smth like this would work: www.amazon.de/dp/B08214QW3X
It says 40A max, 20A long operating. It honestly just needs to last a few minutes at a time.
You think we could mod it to work with a raspberry pi?

[night girder]

Ok, that last video really hit the nail on the head for me. Taking samples at the rising clock. Overlapping those samples and you get an eye.

non of this will fix my PC though. But I learned. I hope.

[twin dew]

One with just two or three wire adjustment would be more certain.
This one has five wires to the "pot", so it might be something more.
Still probably usable.
But usually the 40A would be in sub-second bursts, and 20A the normal.
Not sure how this chinese company uses them.

So I would really suggest looking for one that is 40A continous and more as the max.

[night girder]

Is it always on the rise of a clock cycle that signal changes happen? Or is that only in specific types of signal?

[twin dew]

No, depends on implementation.

For example SPI bus has 4 modes, that specify if the sampling is rising edge or falling edge, and which is the "idle" state of the clock.

www.analog.com/en/analog-dialogue/articles/introduction-to-spi-interface.html

Very simple serial bus.

Used in current PC:s for TPMs etc. instead of the older more complex LPC.

And in lot of other stuff.

Same that most BIOS, VBIOS etc. flash chips are SPI connected.

Two or three shared signals, one select signal per chip.
In some cases the communication is just one-way and then there is just MOSI or MISO between the chips.

Not both.

[night girder]

but they can do both? if I read corectly, you have main node with clock speed and chip selection, it can select one or more subnodes. And between main and subnodes are MOSI and MISO. It's up to the user how to implement it.

[twin dew]

One master, and infinite number of slaves, just limited by the bus signal integrity and chip select outputs on the master.

SCLK is the clock signal, MOSI = Master Out, Slave In, so master talking to all slaves (only the selected ones will listen), and MISO is Master In, Slave Out, where selected slave transmits data back to master.

[night girder]

mosi (main out, subnode1 in) - miso (master in, subnode out) 😄

[twin dew]

That is the new "corrected" naming.

[night girder]

Oh, right forgot about that. Let's not open that can of worms haha 😄

[twin dew]

Mainly meant that there are chips that never take any instructions, just output when given chip select and clock.
And some that you only ever command, that never reply.

And in those cases only one of MOSI or MISO is wired.

Most SPI parts need both.

First to get command on what to do, and then to reply.

[night girder]

Yeah was about to say that. It would feel ineffective to only command a chip and never receive feedback from it etc.

two way communication opens a lot of possibilities.

and if you only need one way communication, is SPI even the best option? Aren't there better alternatives for it then?

[twin dew]

It is semi-standardized version of the most basic digital bus you can implement without self-clocking.

For just one-to-one, the CS isn't needed for example.

Can just be hard-wired to always be on.

So for one direction, just clock and data lines.
With second data line for the other direction if needed.
Then CS lines for multiple chips on same data lanes.

You can get less wires, but the actual communication protocol etc. will get much more complex.

SPI is just bit banging.

[night girder]

what about 1-wire? (https://en.wikipedia.org/wiki/1-Wire)

[twin dew]

Like I said, much more complex.

[night girder]

it looks more straightforward to me 😂

[twin dew]

On the signaling level.

[night girder]

Yeah, probably.

[jagged snow]

Just like comparing i2c to spi

[twin dew]

For starters for 1-Wire:

The master starts a transmission with a reset pulse, which pulls the wire to 0 volts for at least 480 μs. This resets every slave device on the bus. After that, any slave device, if present, shows that it exists with a "presence" pulse: it holds the bus low for at least 60 μs after the master releases the bus.

Where as with SPI the bus doesn't care about the frequency or how stable it is most of the time.
As long as the frequency is low enough for all the chips.

[night girder]

Yeah, and 1-wire still has two connections it seems.

[sand saddle]

A bit of googling tells me that the 2 extra wires might just be used for physical on/off
So it's still a 3 wire knob.

[twin dew]

Probably, but wasn't sure.

[twin dew]

Still on same wire, but the timing determines who can speak.

[night girder]

while I was looking for a standard with only one way communication.

[sand saddle]

This has plenty amps: www.amazon.de/dp/B08214NZMX

[twin dew]

But for SPI, most of clients don't care about frequency or if the clock stops during transfer for a time etc.

[night girder]

wasn't the clock important to keep everything in sync?

[twin dew]

You can talk to the same chip with 1kHz or 10MHz signal, and you can have the clock frequency constantly changing etc.

[sand saddle]

Even found a solid review, 60A continuous isn't doable without active cooling, but I don't need 60A

[twin dew]

That is the whole point, nothing happens except during the clock signal changes.
So the timing between the changes doesn't matter on SPI bus.

Which makes it so easy to bit bang.
You can talk to SPI device with two manual switches if you want to, and single led for return data, manually.

[night girder]

I just misunderstood you 😄

[twin dew]

Change the clock signal to idle state, set the command signal, set the clock to active (transmit a bit), change the clock back to idle, set command, and so on.
The timing of the changes will be very variable, but it will work, as only the changes matter.
Not when they happen related to each other.

Not so on 1-wire for example, timing is critical.

[night girder]

mhm, yeah then SPI is superior.

[twin dew]

Different uses.
SPI is the "dumbest possible" thing.
I2C, 1-Wire etc. are newer and "better", but much more complex.

[night girder]

more complex = more than can go wrong, unless they have safeguards for everything that can go wrong. But that makes it most of the time even more complex.

Sorry GS, you had quetions. I'll 🤐

[twin dew]

And it actually specifies that it is just normal potentiometer with 270 degree throw and switch at one end for off.

So looks good for most part.

[sand saddle]

But I don't understand how I could bypass the potentiometer and wire it directly into a raspberry pi 4.

I'd like digital speed control

[twin dew]

Pi DAC output to the yellow/middle pin, or Pi PWM output with some filtering, with the black connected to Pi ground.

Just need to check if the red is same level as the Pi would output, if not, need extra thing in between.

And with the switch red being connected to digital pin on Pi for on-off control.

Or one of those Pi specific low power PWM control boards used instead.

If the voltage used by that board for the red power wires for the on-off and pot don't match to what the Pi can do.

So the Pi controls the low power PWM board, which controls that one instead of the potentiometer.

But probably uses 3.3V or 5V for the pot.

With 3.3V more probable at this point.

[sand saddle]

DAC?

You mean the 3.5mm header?

[twin dew]

Not really.
Just variable voltage level output.
But you get the same with the PWM outputs with some filtering caps after to average the output.

Some microcontrollers have internal DACs and ADCs, but Pi doesn't seem to have?
Just didn't remember right.

But the potentiometer in the original just acts as voltage divider, between 0V and voltage on the red wire.
With pot in one end outputting 0V to the yellow middle line, and at other end outputting the red wire voltage.
And that voltage level is what controls the thingie.
And you can just use anything else to feed that voltage.

[night girder]

but the knob, it's for manual adjusting of motor speed. You cannot program the knob afaik 😒

[twin dew]

So if the red wire is 3.3V, then you just need to cut the potentiometer off, and feed 0.0V-3.3V to the yellow wire.

[night girder]

you need to be able to adjust resistance programmatically

[twin dew]

While the black is tied to your Pi:s ground.

[sand saddle]

Since my brain isn't computing, is this hard to do? The whole digital speed control.
And do I need to order anything else?

[night girder]

www.circuitbasics.com/using-potentiometers-with-raspberry-pi/

[twin dew]

And then with PWM output with filtering, 0% would be 0V (original knob in one end), and 100% would be 3.3V (original knob at other end).
And 50% would be like knob halfway.

[twin dew]

That is in the other direction.

We are replacing a potentiometer on motor controller with output of a Pi.

[night girder]

DS1807 from Dallas Semiconductor is a digital driven potentiometer

[twin dew]

Yes, there are chips to do the thing.

Which you control with SPI or like.

But not needed if the voltage levels match.

[night girder]

simply put: more resistance = slower motor. So all you need to do is control resistance right?

[twin dew]

Not resistance, voltage on the yellow wire.

The potentiometer is just acting as configurable voltage divider.

[night girder]

huh, wiki says it's a resistor.

[twin dew]

That is how it works.
But in 3-pole use it is voltage divider.

Like that says.

You use it to select voltage level between the outer terminals.
As long as the input resistance of the thing reading the signal is high enough compared to the potentiometer resistance.

As in the middle terminal (wiper one) will have voltage that is somewhere between the outer terminal voltages, depending on the position.

[night girder]

I don't see how you can program a manual knob 😂

[twin dew]

You cut the potentiometer off.
And connect the wires to something that you can direct.

Because the thing doesn't care about the potentiometer, it only cares about voltage on the yellow wire.

The pot is just manual way to set that voltage.

[sharp oasis]

Let's go

New case new cooler

For $100

Coolermaster TD500 and a larger deepcool aio.

The LS720

Used from my freind.

I trust him enough

[night girder]

Is all this needed Baldur? I just read a tutorial and they connect PWM to motor.

[sand saddle]

Nice deal

[twin dew]

Only question is if that PWM controller uses 3.3V for the red wires to the pot.
In which case it can be directly connected to the Pi.
Or 5V or higher in which case it needs very low power additional voltage controller or programmable resistor/digital potentiometer chip between.

[night girder]

They can adjust speed in the code.

[sharp oasis]

Now I will need a new mobo badly else mine will look goofy as all hell. But that can be dealt with later.

[night girder]

Set the motor speed
GPIO.output(7, GPIO.HIGH) # Set PWMA
Motor B:
GPIO.output(18, GPIO.HIGH) # Set PWMB

[twin dew]

The red isn't connected in this case.
And only one PWM line is needed.
And the clock line etc. isn't used.

[sharp oasis]

I will eternally dislike a Matx mobo in an ATX case

[night girder]

red wire is 4.5V for this case.

[sharp oasis]

Ik there's nothing wrong with it it just looks bad. To me for some reason

[twin dew]

And one solution would be to take that board, and connect the pot black and yellow wires to one of the motor outputs of that addon-board, with extra capacitor for filtering.

[sharp oasis]

I'll shush now and let you guys talk

[night girder]

www.amazon.com/Adafruit-TB6612-Stepper-Driver-Breakout/dp/B00VY32VU4/?tag=howchoous-20 (that's the Arduino board they used in between)

[twin dew]

If the 60A motor controller board doesn't use 3.3V.

For the pot.

[night girder]

Here GS: howchoo.com/pi/controlling-dc-motors-using-your-raspberry-pi/ (not sure if it can help)

[twin dew]

But that is for the specific small power add-on board.
Idea is the same, use the PWM output on the Pi to control the voltage the 60A motor controller sees on the yellow pot line.

But next step is to get the 60A board and measure what voltage is on the red wires to the pot.

[night girder]

Yeah, just trying to give GS ideas. The more options they have, the better they can choose what fits their case 🙂

[languid gulch]

so who wants the $8500 Kohler toilet they're showing at CES

[night girder]

I want those new gaming chairs that vibrate your butt.

haptic feedback, but in your seat. Not joking. (Was also on CES)

[languid gulch]

for "gaming"

[night girder]

Yeah, so if you crash your car in GTA, your butt goes brrrrrrr.

[languid gulch]

i guess that's one way to get your gf into gaming

[night girder]

I just eat beans to get that effect.

Let me check what company it was.

[languid gulch]

hitachi?

[night girder]

Razer. it's not a chair. It's a cushion.

[twin dew]

But the point is that the Pi will output PWM, that gets converted or filtered to be voltage signal based on the PWM duty cycle, that gets input to the 60A motor controller yellow line, to control it instead of the potentiometer.

If there needs to be anything extra but single capacitor depends on the voltage the 60A motor controller board has on the red lines to the potentiometer.

[sand saddle]

The main problem is that there aren't really digital motor drivers that can have 18V 40A or so going through them.

[twin dew]

3.3V or less: No.
More than that and you need extra stuff between the Pi and the 60A board.

Like one of those low power motor controllers.

[night girder]

Oh right forgot.

[twin dew]

Or some SPI DAC, or digital potentiometer etc.

Or even just a MOSFET that gets driven by the Pi:s PWM output and is connected to the pot wires instead of the pot.

[night girder]

www.digikey.com/en/products/detail/MCP4151-103E%2FP/MCP4151-103E%2FP-ND/1874217 what about this one?

nvm. You need 18.

[twin dew]

There are lot of chips and ready solutions if the direct connection isn't possible.

But there is very high chance we only need single almost random value capacitor.
If the 60A board uses 3.3V or less for the pot.

Which can be measured with any multimeter or voltage meter between the red and black wires.

[night girder]

how does a digital potentiometer work? Does it need 1.8 - 5.5V to function or can it regulate between those two values?

[twin dew]

It has multiple inputs.
One for the logic, another for the pot part.

[night girder]

https://www.youtube.com/watch?v=qqIXkUIC2Mg

That's where I got that potentiometer from. Tutorial is in the description.

[sand saddle]

Could I order anything rn or do I need to measure the pot voltage first?
Cuz I'll only make a single amazon order and local shipping is like 3€ so if the caps are cheap, I'd just order a few and hope for the best.

[twin dew]

The logic part works with 1.8-5.5V input for power.
The pot part is 12.5V max

[twin dew]

If you can ask again tomorrow, I should have been in bed 3 hours ago.

[sand saddle]

No hurry

Good night! Sleep well!

[twin dew]

Was going to look at ready made boards between the Pi and the motor controller.

Just that going with digital potentiometer is probably the last thing you want to go for, once nothing else works.

[sand saddle]

Plan B is an on/off relay

The speed control is more to flex with, not required

[twin dew]

I meant that there are many better ways to give voltage signal to the motor controller instead of the physical pot.

[night girder]

then imo the digital potentiometer seems better. If I had to choose. And if one fits the specifications.

[twin dew]

As in PWM output, DAC or low power motor controller first, only then to digital pot if the range isn't suitable otherwise.

Because the digital potentiometers are for replacing real ones in the two-pin usage mostly.
Or if pure voltage control isn't ok and replacing real potentiometer with digital controlled one is needed 1:1.

They are very specialist devices.

And in this case just pure voltage signal is needed to high-impedance input.

And using PWM output on the Pi is easiest to do for the programming part.

Digital pot chip or DAC would need using SPI or I2C bus.

[sand saddle]

You can go to sleep

We're not holding you here 😅

[night girder]

As said earlier, we cannot draw more than 15mA from any GPIO pins and DC motor draws more than 15mA, so the PWM generated by Raspberry Pi cannot be fed to the DC motor directly. So if we connect the motor directly to PI for speed control, the board might get damaged permanently.

So we are going to use an NPN transistor (2N2222) as a switching device. This transistor here drives the high power DC motor by taking PWM signal from PI. Here one should pay attention that wrongly connecting the transistor might load the board heavily.

The motor is an induction and so while switching the motor, we experience inductive spiking. This spiking will heat up the transistor heavily, so we will be using Diode (1N4007) to provide protection to transistor against Inductive Spiking.

In order to reduce the voltage fluctuations, we will be connecting a 1000uF capacitor across the power supply as shown in the Circuit Diagram.

[maiden coyote]

well I'm working on my ebay special w7100 again. I've flashed a bios onto it I downloaded from techpowerup. about to put it back togeather and see if it'll still work

[night girder]

This person doesn't even use all of that. 2N2222 Transistor, Capacitor- 1000uF, 220Ω or 1KΩresistor (3)

[twin dew]

And in this case the PWM output would be driving the CMOS chip input on the motor controller board.
Very little power usage.

[maiden coyote]

if you want to run a motor, a arduino is a much much better choice than a pi

[sand saddle]

Pi is owned

[maiden coyote]

unless precise control doesn't matter.

[night girder]

How so?

[sand saddle]

It doesn't

[maiden coyote]

like if you need on or off the pi is fine. but if yu need to control the speed of the motor based on another variable arduino.

[twin dew]

Arduino is real microcontroller, with DAC outputs etc.
And can use and output 5V.

[night girder]

oh that's good to know.

Wanted to buy a PI eventually.

[jagged snow]

Arduino is better for anything that doesn't need to drive a GUI(with exceptions)

[twin dew]

Not really.
Pi can do PWM, or have DAC or like as accessory.

[sand saddle]

The programmer in my group would've liked an Arduino as well, but we got the Pi 4 for free basically

[maiden coyote]

a pi is a computer, it has to do operating system computer things. those computer things can and will get in the way if you need precise timings.

[jagged snow]

In reference to making gadgets anyways
Pi is more compute focused, Arduino is I/o focused

[twin dew]

It is mainly that Pi is overkill.
And yes, possible timing problems if very fine control is needed.

[night girder]

so arduino has just a terminal as GUI?

[maiden coyote]

the arduino has nothing

[jagged snow]

Arduino has no built in ui

[maiden coyote]

you write and upload your code to it, then when you power it on it excutes that code

[twin dew]

Arduino is raw CPU, no OS or UI etc.

[night girder]

dang, that's cool.

[twin dew]

Just basic microcontroller on a board.

[maiden coyote]

you can program a arduino to have a UI for it's task if you attach a LCD to it and buttons.

[twin dew]

You can buy the same controllers for 3-5e each as separate parts.

[night girder]

So you gotta write the OS yourself 😂 Good practice.

[maiden coyote]

but the arduino itself is a atmel microcontroller.. it can do alot of shit

[jagged snow]

Yup
Arduino also has a few extras and a pretty powerful ecosystem around it, just like pi

[jagged snow]

No, it doesn't run an os

Just very simple bootloader straight to whatever code you've written afaik

With some very powerful software libraries made avaliable

[maiden coyote]

control motors, generate a precise pwm signal to control model trains ( see DCCex ) it can act as the ECU for a gasoline engine.. control rgb leds... and the list goes on

[night girder]

I mean, you can write your "own" as in:

just a little interface 😄

[jagged snow]

Oh yes

[maiden coyote]

you can. we just mean stock.. I believe it comes with a bootloader burned on the chip and that's it

[maiden coyote]

I'll be damned.. my firepro might have drivers now

[maiden coyote]

it works. yay. downloaded a random(ish) bios from techpowerup.. padded it with zeros until it was 512kb, desoldered the vbios from the card and used a programmer to write it.. resoldered it back. fun.

[jagged snow]

Huh, cool

[maiden coyote]

moral of this is.. don't buy workstation cards on ebay.. there's a possibility they were used in a custom server build where the normal drivers provided by amd or nivida will not work

and because they're custom, you can't use software to change it to a normal bios. at least the amd bios tool would not let me do it

so the only way to write a different bios is to manually do it with a programmer.. so those firepro or quadro cards going for cheap.. beware

[sharp oasis]

I assume it was possibly a mining bios

Or something

[maiden coyote]

custom bios for a custom server xd this sorta follows along the lines of the issue https://www.techpowerup.com/forums/threads/subsystem-id-mismatch-barco-mxrt-7600-8gb.288136/

I gave up on the amd flasher tool after trying to force, and whatever that unlock command is supposed to do.

[sharp oasis]

Holy hell

Did I get a good 5600?

4.65ghz on 1.21v

[narrow rover]

Good to all, I have a problem when I try to play with friends and when they join the game, the loading screen satisfactory and then returns to the main menu and with it appears the following:

[twin dew]

Well, the bootloader is just that the CPU will read specific fixed address from the program flash to get an address into the flash that it will then start executing from...

[dire igloo]

#multiplayer-troubleshooting then #1038092680493801533

[twin dew]

@sand saddle For example:
www.amazon.de/-/en/stronerliou-Signal-Voltage-Converter-Digital/dp/B08PBKJ89S/
Set the PWM level selection to 5V position

Connect the Pi to PWM In + and - to the Pi
Connect the 18V fixed power supply you use for the motor to the Power Supply + and -
Connect the black wire that originally went to the potentiometer to the Voltage Output -
Connect the yellow wire that originally went to the potentiometer to the Voltage Output +
And tune the output voltage adjustment potentiometer so that on 100% PWM signal from the Pi, the Voltage output is same as the voltage between the red and black wires going to the pot (needs to be done before the yellow wire is connected).

DIY version would be something like this, followed by RC filter:
Where the +10V would be connected to the red wire, output to the yellow wire through the RC filter, and bottom ground to the black wire.
And then the left 60Hz would be PWM output from the Pi.
But would need to select suitable MOSFET, and RC filter values.

That specific circuit is inverting the the PWM output.
So needs another MOSFET stage to reinvert, or the the duty cycle will invert.
So that 100% PWM from the PI will mean stopped motor, and 0% PWM from the PI will mean maximum speed.

[twin dew]

Response time of about 200ms on PWM signal speed change with that config.
Didn't really calculate the filter, just tested in the sim.

So all the resistor and the capacitor values could be tuned more.

200ms response time as in from when the PWM signal changes, it takes about 200ms for the output to reach new steady state voltage.

Need to change the capacitor to 10uF for 1kHz PWM signal, and the response time goes up to 500ms or so.

The larger cap works for 10kHz and above too, just that the response time is always the same.

[twin dew]

Based on what I can find, the HW PWM outputs on Pi can be set from 1Hz to 19MHz.

But those absolutely lowest frequencies aren't really usable, as they severely limit the possible duty cycle values.
With that 1Hz just having 0%, 50% or 100% or near that.

And you want highest the specific MOSFETs are still efficient at anyways.

[twin dew]

@sand saddle But the first ready solution works if the motor controller uses 1-10V as the red wire voltage.
And you have some minimum 12V input supply, which you should have with the 18V motor input feed before the PWM motor controller.

[night girder]

So I just double checked. And Asrock is setting the 1.015V CPU VDDCR_SOC from itself. Everything is on default in the bios besides two settings.

1. is the auto driver installer disabled.
2. is the FLCK from auto to 1733MHz. So that seemed to fixed itself after we manual set it.

But before the restore to default settings wasn't doing that. So I am not really trusting this "default settings" in the BIOS.

[twin dew]

1.015V VSOC is normal AMD stock.
And the 1733MHz FCLK is the AMD stock.

Just that the Restore Defaults on that specific MB is doing weird shit.

[night girder]

Yes, I am 100% sure I did a restore all default before. And it wasn't correcting all the values I set manual to auto.

As if the "restore defaults" feature wasn't aware or able to override those weird settings. And only once I manual set em all to auto was alright with it.

If the manual settings were the default. And I set all those settings back to "auto" then the restore default feature would set them from auto back to numbers. But that didn't happen. I am so confused what happened.

That's what I had to do manual. But, I actually expected the "restore default" to do this for me. Which it didn't.

But now that everything is default, the "restore default" isn't setting them back either. So... super confused.

[twin dew]

Yup, that was what I said being completely weird about it on that BIOS version.

[night girder]

But just wanted to let you know that only the FCLK is being a annoying (can't put it on auto)

[pure karma]

i just cant help but laugh at this

[night girder]

how many tabs?

[edgy hazel]

Only a week until keyboard

[charred pewter]

i need something to cheer me up 😦

[edgy hazel]

uhhh uhhhhmmm... my first dnd game went super well and I had a blast?

[charred pewter]

first gaming experiences 😄

[jagged snow]

Oh neat, no idea it was that simple

[twin dew]

For many microcontrollers.

Arduinos come then with the FLASH already containing another stage bootloader with basic USB functionality etc. that the base Atmels don't come with.
So you don't need special programmer with SPI or JTAG to flash the controller with your own stuff.

[twin dew]

But almost anything starts that simple.
The more complex computer systems just chain more stages after the very simple ones.

[wanton orchid]

I'm happy cpubenchmark.net finally show more consistently cpu cache size details

[sand saddle]

Thank you! Looks great!
Still amazed by how much time you're willing to dedicate to helping others for free.

[twin dew]

Was weird to not find anything ready for the second type solution...

Everything available used SPI or I2C bus communication, and usually topped at 5V output that could be fed by the PI or like.

And I didn't use that much time.
Might have found something simpler as ready solution otherwise.

[twin dew]

Moment, checking if I can identify the type of one chip on the actual motor controller PCB...
The one that takes the pot input...

[visual tree]

Imagine being falsely charged for false accounting due to a software error and having your career ruined because the management was trying to cover it's ass https://en.wikipedia.org/wiki/British_Post_Office_scandal

I would sue the company for millions if If was one of the prosecuted victims

[dire igloo]

that's US law vs UK law

US law is basically "we don't have protection, but you can sue anytime and if you win you make bank"
UK and EU law is more like "there's safety measures in place, if those aren't violated you're on your own"

[visual tree]

What's even worse is that the management figured out at some point there is something wrong with the software but still decided to cover it up and blame the subpostmasters....

[twin dew]

@sand saddle So once you get the motor controller, need to know the markings on this chip to confirm it will work:
Trying to locate something that would seem relevant with anywhere near pinout currently.

[night girder]

So it's not only Apple that dropped the ball with their updates. KB5034441 has a bug which leads to the restore partition not being big enough. Resulting in 0x80070643 ERROR_INSTALL_FAILURE. As a work around users need to increase the size of the restore partition.

[twin dew]

Doesn't seem to always matter even if the restore partition is large.
My laptop has 600MB one and still fails.

[night girder]

And it is a important update. Fixing this: msrc.microsoft.com/update-guide/vulnerability/CVE-2024-20666

A successful attacker could bypass the BitLocker Device Encryption feature on the system storage device. An attacker with physical access to the target could exploit this vulnerability to gain access to encrypted data.

[twin dew]

Yes, but was waiting if MS fixes it before trying to force it in more.

[night girder]

Hey, person who said that perfect security is possible. Forgot who it was. But this is another example that perfect security is a fallacy.

[twin dew]

Because the danger of getting the laptop stolen by someone who would use that exploit is pretty slim right now.

[night girder]

Don't increase partition size yourself. Seems it can brick your PC.

Microsoft released two scripts that can do this.

support.microsoft.com/en-us/topic/kb5034957-updating-the-winre-partition-on-deployed-devices-to-address-security-vulnerabilities-in-cve-2024-20666-0190331b-1ca3-42d8-8a55-7fc406910c10

[night girder]

The chances are as high as you doxing yourself and someone using that to fuck you over 😉

[twin dew]

Something with small chance happening in next few weeks.
Vs. something with probably small chance happening during rest of my life.

And which cannot be pulled back if the probability later goes up for some reason.

[night girder]

My opinion still remains the same. If someone targets you. You probably screwed anyway and they hack you. But you have to have "value" to be targeted. Or someone really hating you.

but that doesn't mean you should leave the doors open to your PC either 😄

Will always be interesting to find the balance between security/easy-of-use.

[twin dew]

@sand saddle So looked more, and now unsure if it will work.
So wait on ordering.
If the chip is actually 555-timer.

[night girder]

Also Baldur, why you still on W10? 😒

[sand saddle]

I think you got the wrong product. We were looking at 60A(100A max):

[night girder]

Laptop doesn't support W11?

Or W11 doesn't support laptop 😉

[twin dew]

Ah, ok, moment...

[sand saddle]

Which chip do you need?

[twin dew]

Can you give the url to that?

[sand saddle]

This is smth above the white connectors

[twin dew]

Yeah, 555 timer.
But then depends on how it is connected to the pot.

[sand saddle]

There's also this:

[twin dew]

That is just 5V source for the low power electronics.

[sand saddle]

To the right of white connectors

[twin dew]

But it is that 555 that takes the pot input and converts it into kind-of-PWM.

[sand saddle]

www.amazon.de/dp/B08214NZMX

[twin dew]

So if you have the product page url, or some nice source for such PCB pics, I would like to take a look at the traces.

[sand saddle]

I need my dad to find the exact motor I'll be using, would be funny if he finds a brushless motor and all of this is irrelevant

[twin dew]

What is the multileg chip that drives the 7-segment screen?
If you can take shot of the markings for that?

[sand saddle]

No name

No markings

[night girder]

And in the specifications? Doesn't raspberry PI have overview of the chips in manual?

[twin dew]

Ok, moment.

[sand saddle]

This isn't Pi, it's this:

[night girder]

same logic applies? no manual? some websites put the manual with product if lucky

[twin dew]

But in this case, one pot leg is connected to the actual power-.
But it seems to be the red wire for some reason?

[twin dew]

Chinese random production.
All documentation you get is the advertising material on the site.

Do you have multimeter available?
Can you check the resistance between the red wire on the 3-pin and the Power- pole?

If you have that thing already.

[sand saddle]

No, https://www.youtube.com/watch?v=UzCmynbpAIk

[twin dew]

Ah, ok, moment.

[sand saddle]

I could make a comment on his video and maybe he answers?

He didn't test the pot voltage like that specifically

[twin dew]

I'll look at specific spots of the video.
And the same guy had videos for some other ones where he modified them for arduino control.

[sand saddle]

Indeed

[twin dew]

But in this case it seems to have 0V and +5V connected to the pot, just wrong way around.
So that the red wire is the 0V and the black wire is the +5V.
And the pot is voltage divider with the yellow going to the markingless multileg chip.
Which then controls the 555.

And not pot directly connected to the 555. with all three legs.

The potentiometer output is going straight to the bottom left leg of the mystery chip.
And in that case, the intended solution should work fine.

Just that will need to check that +5V and 0V thingie with the wires.

And in this case the voltage used is 5V.
Just wrong wire colors, as the whole backside is Power- fill at the bottom, and the away from display pin is connected to that fill.
Then the middle pin goes towards the display, and goes to the top side of the PCB, and enters the mystery chip.
And then trace from the +5V regulator output enters behind the white connector, and probably connects to the last leg.

But should work, just need to check the red and black pins on the 3-pin, to confirm which is +5V and which is 0V.

[sand saddle]

I still need the 3.3V PWM signal to 0-10V or could I just use 5V from Pi?

[twin dew]

Pi doesn't output 5V.
Would need minimum of one mosfet to raise the 3.3V to 5V, and then that RC filter to convert it into voltage instead of PWM.

[twin dew]

This thing, but one or both top feeds as 5V.

Pi just has two +5V supply pins that are directly connected to the USB input power plane.

In the header.

[sand saddle]

Top right, raspberry pi 4 pinout

[twin dew]

Like I just said, always on power pins.

[sand saddle]

Ahh, gotcha

[twin dew]

All the controllable outputs are 3.3V

And you don't have direct analog outputs at all.
So need level conversion from 3.3V PWM to 5.0V PWM, then conversion to non-PWM stable voltage.
Or some other 5V voltage output solution in between, DAC or like.

So that other thingie I linked is complex, but it does the job and is pretty cheap.

And should work fine in this case once the max output level is tuned with that trimmer.

There would be DAC boards etc. ready too for Pi, but those need I2C or SPI to direct and using the Pi native HW PWM outputs is much simpler in the code.

[jagged snow]

There are some very well done libraries for doing I2C or SPI with pi, but it's still a lot more work than pwm

[twin dew]

Not insurmountable, but not worth it to do it via "better" solution when the price would be the same.

[sand saddle]

Baldur, if you don't mind my interest, what's your education?

Like an academic degree

[twin dew]

Seems they are using the 555 as just a transistor/fet in that thing.
With Trigger and Threshold pins always active.
And just use the Reset pin to control the state, from the mystery chip.

Well, they are active on the reviewers sample, but the PCB on the other source pics is little different and they might be floating.
Hopefully the marketing pics are from older revision and not newer.

Yes, the PCB identification number on the reviewers sample is higher.

BLH-80432 on the marketing pics, and BLH-89287 on the reviewers sample.
And the BLH-80432 with the 555 shouldn't be reliable and would probably start to do weird things once it has been on for some time.

(Been filling my own curiosity on how the fuck does the thing work...)

[night girder]

🙂

Honestly, I would try to build it myself. Isn't that the fun part of raspberry PI and arduino's? Doing it all yourself and learning at the meantime?

[twin dew]

@sand saddle So check that PCB model marking when you get yours.

[night girder]

stuff like this:

But what I really want to achieve is more this:

Ground moisture meter for plants 😛

I even think an Arduino might be a better candidate to do this.

[maiden coyote]

It is. But a pi works too

Esp32 is might be even better, you can integrate it with home assistant and have notification pushed to your phone

[dreamy pelican]

how do i remove facist/christian ornaments from power poles?

[jagged snow]

You can just take them off... but I'm pretty sure you're trolling since fascim and christianity are rather different

[dusk rivet]

youtube done broke my sponsor blocker

so rude

[cyan crescent]

Seeing that wifi 7(is that the newest one?) will be able to use multiple bands at once is a god send. Why wasnt that a thing before?

[dusk rivet]

not experienced in this, but if i had to guess it just wasn't feasible with the technology of the time

[stray badger]

Im gueasing it was regulations

[twin dew]

Not enough bandwidth available to make it worth it, as it needs more complex transceivers and logic.
And having it as one continous chunk was possible in most cases when the max was 80MHz or less.
With 160MHz channel width as max it became a problem with the regulatory bandwidth allocations available.

An the 160MHz channels only came in Wifi 6, and most HW for that didn't support them.

On client side, so the AP support in the more expensive APs was useless.

[night girder]

It's not YT. It's the blocker. They already confirmed it.

[wide void]

I love winter

(the server)

[twin dew]

https://www.youtube.com/watch?v=0e1nDf0GYZU

[wide void]

thats been the most amusing difference between Steve and Jay, Steve is doing a nice deep dive thermals/pressures etc, Jay was more 'Mmmmm, shiny!'

[twin dew]

And then there is the LTT take back when.

Where they used incompatible GPU and complained about horrible temperatures.

And doubled down afterwards, then auctioned the block that was supposed to be returned.

[wide void]

I didn't watch ltt even before the 'drama', didn't really appeal to me

are they still running?

[twin dew]

Haven't watched either.
And yes.

[languid gulch]

i like it as a concept

it's at least different from what everyone else is doing

[twin dew]

Specialist product for the people with exact wants.

And no budget limit.

[languid gulch]

lol true

$1000

[wide void]

not messed with water cooling really since the bad old days. I run an AIO on the games rig but that doesn't count, lol

[languid gulch]

now, the next thing i wanna see is a combined air cooler that can handle 1200W

[twin dew]

Look into almost any 4+ GPU server?

😜

[languid gulch]

oh i mean for "normal" cases

[wide void]

that doesn't sound like a Eurofighter taking off...

someone here will know, what is the big chonker 120 x 240 air cooler called... I can never remember

[minor pawn]

Annoying question but would this be a good pc? https://be.pcpartpicker.com/list/7wGsjH

[dusk rivet]

ah perfect

[twin dew]

*Phew*
The beeping at home when I returned from store was just the UPS crying that the battery needs replacing...

[dusk rivet]

what’re you gonna do with the pc?

mostly gaming or productivity, or a mix of the two

[edgy hazel]

don't need no coffee now

[twin dew]

But anyways, got the UPS in 2012, replaced the battery first time in 2016 and now need to do it again.
But it wasn't used for few years in meantime.

Non-APC branded original is half the price of the APC branded one...

APC RBC110 would be the official replacement.
APC RBC17 would be larger capacity one that would also fit.
And seems I'm actually going to get CSB HR1234W.

And the HRL1234W long life version is even cheaper than the standard life HR1234W?=

[twin dew]

And seems I might be able to get the HRL1234W locally.

[twin dew]

Hmm... 68e via local store with few days delivery time.
Or from another country for 50e including delivery but with 3+ week delivery time...

[narrow folio]

I customised my old APC with bigger external batteries to have a bit more time to shutdown everything, but after 2 replaced sets over the years without ever needing the ups once... I omitted the thing

and as it was an oldie it drew around 30W idle when the batteries where charged and without anything on the outlets... that made the choice even easier

[edgy hazel]

my keyboard is in europe

hyped hyped

[twin dew]

Just same physical size battery with same connectors.
Just newer type cells that cram more power into same volume.
And then longer life industrial battery model vs. consumer model.

And the fiber transceivers finally arrived, and the thing works \o/

[narrow folio]

I went from 12Ah to 20Ah, no chance to fit those in the original case

[twin dew]

7.2AH to 9.0AH.

[narrow folio]

changed connectors and made a new case for them

[narrow folio]

this is a nice boost for the same size

[twin dew]

Original is APC RBC110.
Which should actually be just branded CSB GP1272: www.csb-battery.com.tw/english/01_product/02_detail.php?fid=5&pid=11

Then some other APC models use RBC17.
Which should actually be just branded CSB HR1234W: www.csb-battery.com.tw/english/01_product/02_detail.php?fid=7&pid=28

And that has long life version:
CSB HRL1234W, which is what I ordered: www.csb-battery.com.tw/english/01_product/02_detail.php?fid=8&pid=34

APCRBC110: www.apc.com/us/en/product/APCRBC110/apc-replacement-battery-cartridge-vrla-battery-7ah-12vdc-2year-warranty/

But anyways, nice that the TP-Link transceivers seem to work.

From 7.2Ah to 9.0Ah with 3 more years life between replacements rated by the manufacturer, even more by standards body.

[narrow folio]

a nice upgrade for lower price 👍
we need more like that 😁

[rancid coral]

Hey guys, I got an issue maybe you can help with. I'm looking for my dedicated server savefile, But it's not showing up in appdata like it should. I open game and look at saves it shows 4 different files I can load but can't find them where they are supposed to be. Any thoughts?

[twin dew]

#dedicated-servers
This is one of the off-topic channels, so for things NOT related to Satisfactory at all.

[twin dew]

And I had completely missed that Broadcom had bought VMWare for very high price, and are now both cutting costs and trying to raise revenue...

[night girder]

I am amazed how quickly the news spread that YT is slowing down videos for addblock.

quicker than the news of the developers behind the addblock, saying it's their fault 😂

Shows the distrust people have for YT.

[languid gulch]

it's also something that google would 100% do if they thought they could get away with it

which is why it'd be so easy to believe it was them, even if it wasn't

[night girder]

Yes, I agree with that. I wonder if they can get away with it though

[narrow folio]

Why can't they just subscribe and stop complaining. Why are those plebs so angry about a few ads here and there? /sarcasm

[charred pewter]

i wonder on average how many 'subscriptions' people have and dont really know where all their money is going

[narrow folio]

Seriously, I pay to not see ads and I'm fine with that. I use an ad-based service and don't want to see the ads so I pay the price.
If you want to avoid this by using ad blockers you can't be too surprised that yt will try to do something about you circumventing their business model.

In this case it's funny that the ad blocker is the culprit

[charred pewter]

ive got three "yearly recurring" ones .... which i tend to forget about =p

and oddly, i STILL have not seen any of these issues people are bitching about on youtube, and all i got installed on that one chrome profile is ublock origin. all my other chrome profiles dont have adblockers installed on them (and i only watch youtube junk on the profile that has ublock)

[twin dew]

One: Spotify.
Probably need to get another soon, for firewall & HIPS software.

[narrow folio]

I know people who still gave their print subscription from the time before the internet.
a few bucks a month. not reading it anyway, but can't be bothered to cancel

[charred pewter]

ive got Schedules Direct, Amazon Prime, AAA ... those are my three yearly ones. then the monthly one i have is Google Workplace for my personal domain email boxes

i kinda dont include "domain registrations" even though thats basically a subscription =p

i chalk those up to like insurance payments, and car title renewals lol

but when does a subscription go from 'being a frill subscription' to 'being a utility payment' ? ... i mean, water, power, gas, internet are all 'utility payments' ..... but they are just subscriptions too

[narrow folio]

Spotify, YouTube, Amazon, Disney, Google Drive, some domains with mail servers, a few ppl on patreon

[charred pewter]

for me, Amazon Prime is quickly becoming a 'utility payment' lol

given how much random shit i order on amazon

[narrow folio]

Prime is nice for faster delivery, the movie stuff and twitch deals are just a nice add-on

[charred pewter]

and ublock origin blosk Amazon Prime Videos 'free with ads' ... so i dont get commercials on amazon prime video

[narrow folio]

Oh, that works?

[charred pewter]

but you have to watch with like a browser, cant use those 'smartv apps' for amazon

[narrow folio]

would need a PC next to the TV for this to use

[charred pewter]

but thats fine, because my HTPC is my TV box Movie Box

it records OTA HD channels for me, which i replay using SageTV

then iwatch movies/shows on amazon through Chrome on it... and then movies using VLC

[narrow folio]

If I can't stream it I don't watch it. Too much stuff to watch anyway

[charred pewter]

i record some stuff from the local broadcast channels, nbc, cbs, fox, cw, pbs ....

[narrow folio]

and since covid I rather play some games than watch series/movie

[charred pewter]

sadly though my recordings OTA (over the air) for "Ask This Old House" the past two weeks... have had the video of the show, but the AUDIO coming from a different show of some guy painting

i cant tell if the broadcast is fubar'd, or something is getting crossed on the receiver

[narrow folio]

that's weird... I hope it was Bob Ross ❤️

[charred pewter]

googling it comes up with no one mentioning it

it sounded like bob ross

thing is, Ask This Old House is the ONLY one it has screwed up like this ... and only that show, not "This Old House" ... the other one that records

so odd

[narrow folio]

maybe multiple audio channels (like stereo and surround or different languages) getting mixed up?

[twin dew]

With current digital TV solutions, there are multiple channels bundled in each "stream" anyways.

[narrow folio]

I don't have any experience with recording... stopped recording stuff when my VHS died.

[twin dew]

So might have been local bitflip that caused mismatch in the video and audio stream-inside-the-stream decoding.

Can be from 4-10 channels in each bunch.

And even more on technical level, just that the bunch doesn't have enough bandwidth to provide usable quality bitrate for all.

[charred pewter]

yeah, the PBS channel is about 5 channels ....

all lower bitrates (not that great of quality visually)

I'll have my HTPC do a re-scan and sync with the channels... maybe its just 'off' enough on that one to cause a conflict

i'll brush the cobwebs off the HD antenna in the attic too 😉

[narrow folio]

What firewall do you plan to subscribe to?

[twin dew]

To Be Determined.
Need to test the couple possible ones on the new HW in trial mode first.

I will comment once I have made a choice.

[narrow folio]

Had to Google hips... my first thought was high impact polystyrene

so an IPS running on the client?

[twin dew]

Basically very intrusive antivirus type thing hooking into almost all Windows APIs and doing per executable level access right checking for HIPS.
Pain to train if you don't allow for "allow this code signer to have blanked rights".

[narrow folio]

Would you like to allow explorer.exe to open this file?

[twin dew]

Missing some rights from the middle of the list, didn't want to take third screenshot.

Just as example.

[narrow folio]

I used Sophos for firewall, proxy, IPS. But never those client side stuff

[visual tree]

R.I.P. Self-checkout. You will not be missed bbc.com/worklife/article/20240111-it-hasnt-delivered-the-spectacular-failure-of-self-checkout-technology

I only had problems with those

[twin dew]

Seems rest of the world cheaped out or something.
Works here except for user errors sometimes.

[visual tree]

I had so many problems scanning take-out food you buy at the grocery store. At one point, the guy standing near the machines helping customers started swearing loudly when people asked for help lol

[twin dew]

Mainly the user errors are that age restricted thing, which is just that people then have to repick their items and the transaction is cancelled from the employees station.

And they head to normal line.

[visual tree]

Worst part is when you buy a pen or something small. You have to input barcode number manually because the barcode scanner can't scan small barcodes or barcodes that are not on a flat surface

[minor pawn]

A mix sorry for the late reply btw

[twin dew]

Well, that is just then barcode put in wrong way on the item...

[visual tree]

But that problem also exists on regular checkout

[dusk rivet]

looks fine, cpu might be a bit overkill tho

[minor pawn]

A friend of mine said it doesn't qualify as a gaming pc cause its less then 2k

[onyx sable]

I am wanting to upgrade my video card and ram. I am wanting ddr4 4x32 gb and am wanting a 4K series graphic card, what are the community opinions around the GeForce RTX 4K Series?

[dire igloo]

Ah yes, gatekeeping gaming by setting up arbitrary monetary boundaries

I bet they'll still get wrecked by a 12yo kid on Fortnite that uses the family laptop

[serene charm]

Probably not the tech talk that is usually here but got my first pair of programming socks and holy are they cozy

[night girder]

I watched fortnite 2023 championship yesterday. These gamers have mad reflexes and skills.

[night girder]

programming socks?

[serene charm]

Pink and white thigh highs, its somewhere on know your meme

[night girder]

So hey, want to develop an app?

[visual tree]

I am not as dumb as Jerry

[onyx sable]

What kind of app is it?

[serene charm]

These are cool but they do not go up to the thighs

[charred pewter]

interestingly, the socks work on either foot

[visual tree]

But they will at least piss someone off

[charred pewter]

if my cat could read

[visual tree]

A dating app called Lovefinderrz

[onyx sable]

Oooh fancy

[serene charm]

I have removed the https because idk if know your meme is blocked

knowyourmeme.com/memes/programming-socks

[night girder]

oh right.

[glass panther]

self checkout works perfectly fine, it's the shoppers that are the problem

[visual tree]

Depends on what you buy here. When you buy items you picked up from the shelves, there are no problems. If you buy something like takeout food where price is calculated depending on the weight, things get more complicated.

[glass panther]

scan code weigh thing, problem solved

[visual tree]

The people who programmed this part messed something up

Wish It was as simple as scanning a barcode 😅

[glass panther]

the managers who managed the programmers who programmed that part messed up

[visual tree]

In one grocery store, it works perfectly. In the other one, there are always issues

[glass panther]

experience of working at a large software company has shown me it's product managers & the people writing product requirements that can't do shit properly

[visual tree]

When self-checkout was used for the first time, a store employee had to manually "unlock" (or whatever they were doing) your machine before scanning takeout food. They somewhat fixed it though

[glass panther]

in one of the local stores they can do that remotely

unless it's totally knackered

[visual tree]

I sometimes abused the self-checkout machines in order to exchange banknotes into a lot of coins lol

[night girder]

sometimes there is also a disconnect between departments.

[glass panther]

programmers: reality
product managers: fantasy land

[night girder]

😂

[visual tree]

The worst part is when companies are trying to get rid of experienced employees in order to hire inexperienced ones who are going to be paid less in order to reduce costs. This is one of the biggest fallacies I've seen at work and it ends up in disaster

They don't understand experienced workers add more value and only look at the cost while completely ignoring everything else

[sharp oasis]

It's sooo big. (7800XT I bought off a freind)

[dusk rivet]

hand reveal

[languid gulch]

yea, i went from a single fan 1660, to a dual fan 6600XT, to a triple fan 6800XT, all in the same case. CHONK

[sharp oasis]

the fucking 7800XT FLYS

Holy hell

30fps with PT on. Without really moddinbg fsr settings and overclocking

[languid gulch]

there really is more of a jump between what counts as midrange vs high end than i was expecting

i got the 1660 release week, & the jump between it & the 6600XT was noticeable, but not anything amazing. but WOW going to the 6800XT was one of those "ok, i'm good for about the next 4 years"

[sharp oasis]

the jump from the 6700XT to the 7800XT IS HUGE

my psu is probably struggling a bit

But its not overloaded

[languid gulch]

🤣

i'd start looking for a psu upgrade if it's getting close to like 90% of its max

[sharp oasis]

Tho Full load Im pulling Easly 300-400w. with cpu and gpu

i have 650w

[languid gulch]

just to protect from power spikes

[sharp oasis]

so im at like 60 ish% psu load

[languid gulch]

seeing those wild numbers from GN a couple years ago was a bit eye-opening when it came to what's expected vs what should be prepared for

[sharp oasis]

its pulling like 280w full load

[languid gulch]

sounds about right

[sharp oasis]

and cpu is doing 120w

so exactly 400w even

[languid gulch]

my 6800XT sits at about 275W stock

i didn't even bother OCing it, when i tried it jump to 325W & gave me a whopping like 3% extra. for the heat & noise i undid it

[sharp oasis]

this thing is so quiet

The fans are OFF wile im rendering video

[languid gulch]

which model did you go with?

[sharp oasis]

Power color Fighter 7800Xt

[languid gulch]

nice

[sharp oasis]

I'm so not used to 3slot gpus

It looks fucking massive in my case

[languid gulch]

hopefully you've already added in any NVMe drive that they like to hide underneath them, cuz it's a pain to deal with

[sharp oasis]

I only have one nvme slot

[languid gulch]

i bled last time i had to install one 🤣

[languid gulch]

big fat air coolers are not kind

[sharp oasis]

It's so big in my tiny case

[languid gulch]

looks good tho

[sharp oasis]

Quite

[languid gulch]

did you use separate PCIe power for each plug, or did you use a pigtail

[sharp oasis]

Separate power per plug

[languid gulch]

kk

i know it's probably still technically within spec, but i still don't like using the pigtailed connectors

[sharp oasis]

I cut the pigtail off lol

And rewired the connector

[languid gulch]

oh i'm not that brave 🤣 i just cable managed mine out of the way

[sharp oasis]

I hated seeing it

So got rid of it

[languid gulch]

i just keep all the RGB off so nobody can see it

once the glass is on everything's almost invisible

[sharp oasis]

460w peak powerdraw

[languid gulch]

yea you're fine for now

just don't grab a 14900K

[sharp oasis]

Yeah lol

I'd say closer to 470w bc fans and drives

I did buy a new case

[languid gulch]

i like how AMD used to be the company that slammed power thru their CPUs & GPUs to push performance, & now it's their competitors that do it

[sharp oasis]

From a freind

[languid gulch]

cases are one of those things i'd have 0 issues with buying secondhand

[sharp oasis]

It will be bigger

[languid gulch]

that & an air cooler

[sharp oasis]

So Def nice.

[languid gulch]

check GN reviews to see what its performance is like & if they have any fan position recommendations

[sharp oasis]

I already did

[languid gulch]

cheater 😛

[sharp oasis]

Ik what I'll prolly be doing

Rad top 3 140mm fans front

[sharp oasis]

I did get a new cooler as well

@languid gulch minor oc gets 310fps in rendering. The render engine on this gpu is insane

[sharp oasis]

Jesusssss

Stray with rt

[sharp oasis]

very stuttery

[twin dew]

In here there are weighting machines in produce etc. areas, where you put the item on the weighting machine, and enter the code for the product (2 numbers), and it prints scan code sticker that you attach to the product or bag.
And that is then scanned at checkout, with possible weight cross-check in some stores.

Was that way even before self-checkout was a thing, for long time.

The product codes are in the product price information signs.

And fuck... I have been running stability tests for time, and AMD CBS settings had reset to all auto...
So the results aren't valid and I have to use week more to run another pass.

Because the corrected ECC error reporting settings were in "swallow the shit" mode and not "report them to OS" mode.

[twin dew]

But today is for some network reconfiguration etc.

[sharp oasis]

@twin dew voltage adjustment for the 7800XT I have is completely locked off. For msi afterburner. Even after reinstall.

[twin dew]

And how about in the AMD Settings app overclocking section?

Because third party apps might just not have support for the specific card or card type.

[sharp oasis]

Amd adrenaline works fine

Apparently I have a odd model

It cannot be found on tpu

[twin dew]

And I would expect your card isn't MSI made.
So MSI Afterburner not working on Non-MSI, non-Nvidia card isn't that suprising.

[night girder]

afaik, afterburner isn't event developed inhouse by MSI. Alexey Nicolaychuk is the developer getting paid by MSI.

[twin dew]

At least up to the current war.
For about a year time it wasn't developed at all, as MSI didn't (couldn't legally?) pay the guy.

No idea how that was later "fixed" so they could go around the sactions.

As the guy is Russian living in Russia.
Which would already be reason for me to uninstall the program at this point.

[night girder]

But I would find it very interesting, to know that MSI (big company) put all it's eggs in one basket (one developer). That seems very dangerous.

if it's one developer, he can pretty much blackmail them at this point.

and all the code knowledge sits with one person.

[twin dew]

The software was his, MSI just bought right to name it.

Not the software as such AFAIK,.

More like sponsorship.

[night girder]

Still, my point stands.

From a development standpoint, this has so many flaws.

even in bigger teams I noticed it, one person gets sick, and some knowledge they didn't share and we need, prevented us from doing our job.

"Does anyone know how this endpoint works or where the repository is?" ... "No, XXX knows that, but they are sick".

You don't want to be dependent on one person for a program that has so many downloads/usage.

[twin dew]

Because MSI didn't buy the software?
Just "rights" to name and distribute it?
And the dev probably wasn't willing to sell?

Sponsorship of the dev, not ownership of the program.

[night girder]

yet it's on their website. Yet, they put MSI in front of it everywhere they can. One major leak in afterburner, who do you think people would blame?

Sponsorship, ok. I understand. That doesn't mean there are no flaws with this whole concept.

[twin dew]

Yes? In a way it is bad deal.
But the choice might have been to not have it at all.
And try to have inhouse team develop something similar.

[night girder]

I would never, ever, encourage this. Or give my go for this plan if I was MSI employee.

I wonder how many people even know it's not dveloped by MSI inhouse. Once I caught news of this, it was gone of my system (the old system, didn't even land on this new one).

[twin dew]

At some point Asus GPU Tweak was lisenced too to same base source.
And EVGA Precision X too.
EVGA isn't anymore, no idea about Asus.

[night girder]

Further investigation, MSI afterburner is actually RivaTuner. Developer by Alexey Nicolaychuk in 1997.

In 2014, EVGA terminated its contract with Alexey and was subsequently discovered to have engaged in source code theft after releasing a clone of RivaTuner that was claimed to be an in-house design.

[twin dew]

Yes. You didn't know?

[night girder]

No.

[twin dew]

Guy developed RivaTuner, which then got paid sponsorships/lisence deals, and the freeware direct release was dropped.

MSI Afterburner is the most known, and possibly only current one.

I wasn't aware of the other lisencees before I checked that same wikipedia article.

Trying to find out when RivaTuner turned into MSI Afterburner.

[night girder]

same

RivaTuner was last updated in 2009. It has since been licensed by computer hardware vendors and integrated into various overclocking utilities, including MSI Afterburner,

I guess it got "integrated" into MSI afterburner. Which is being developed by the same guy.

[twin dew]

While the right side says that last release was in 2012 or 2013

[night girder]

Yeah, noticed that too.

Well it says "last updated", that's not the same as stable release 😉

[twin dew]

And the RTSS part is still released direct too.

[night girder]

but I am still confused.

[twin dew]

The deal is confusing.

[night girder]

The more I read, the less I want to touch that piece of code again 😂

[twin dew]

Because most of the deal information isn't public.

[night girder]

which I can understand, more or less.

well, atleast it's a senior developer.

[night girder]

new disclaimer for google chrome icognito. They had to add that they still collect data.

[night girder]

Just watched another video, where someone connected two dots and the conclussion was: buy Nvidia.

The dots were this: AMD + Shaders in minecraft = "weird frame spikes".

Then a comment down below, read by YT content creator says the following:

The framespikes are being caused purely because of how S*** the optimization is in 1.18 and above.

So it has nothing to do with it being an AMD product, yet, the content creator still said: "Well I guess if you want to play minecraft with shaders you gotta buy Nvidia"

[stray badger]

There are some shaders that dont work with amd for some reason, but amd works just as well as nvidia for minecraft

Mainly because minecraft is so ancient that anything will work with it

[languid gulch]

it's 90% branding. nvidia named & advertised the fuck outta Ray Tracing & gave their GPUs a new prefix, & AMD didn't

& yea, minecraft is also old at this point, but also has those "please melt my system" settings that'll still tank any system

[night girder]

the shaders are pretty "new" no?

I played it at release and I can't remember this was an option though:

So even if minecraft is old, doesn't mean the modern version of it runs well 😄

[twin dew]

Fiber cable installed and working \o/

[charred pewter]

\o\

/o/

goto 10

https://tenor.com/view/shut-it-down-dark-city-talking-shut-it-down-forever-gif-17544075

[twin dew]

No, I'm not going to shut it down.
I even got the extra cable nicely stashed away:

[dire igloo]

Nothing's better for a nice relaxing day at work than having the modem of your WAN connection fail disconnecting 500 employees from whatever they are doing

And tomorrow we'll be offline again - planned downtime for a Netscaler update cuz new vulnerability just dropped

[twin dew]

I would have thought something like that would have been better to schedule for today after normal working hours...

[dire igloo]

Thing is: after normal working hours means after normal working hours for our team too

The more important issue at hand however is checking if the modem survives the server reboot tonight

So we don't want to change too much stuff at once

Also, we scheduled it for lunch break time - and if there's complications, we don't have to stay up until after midnight (which one of our admins had to do after doing a server update recently)

[dusk rivet]

sponsor blocker is working again, thankfully

[night girder]

yes, patch went live today 😊

[dusk rivet]

Love it, i got so used to no sponsors that hearing sponsored stuff in my videos felt wrong

[night girder]

fyi, there are other types of blockers too. Ones that go a bit further and have more control. Like https:// pi-hole.net

[dusk rivet]

network wide ad blocking?

now this is something i gotta look into

[night girder]

and I believe I read somewhere, that if companies are going to "scan" your PC for addblockers, this might go against x regulations and laws. Atleast in EU.

[dire igloo]

pihole can't skip ads that are part of a video's content