(10-Sep-2023, 03:01 AM)mstc Wrote: …You should have your soldering iron tip next to it in the photo, that would give some scale to the freakishly small items you are working on.

If you want a proper idea of scale…
This is a Hakko T18-BR02 bent tip 0.2mm:


Here is that tip next to that damn bga mcu, with the board it came off of above it.


Ridiculously tiny thing. 49 damn solder points on it.

I’ve decided that this board is officially a donor board, and a practice bga chip board.

I’m gonna get this bga thing down, eventually- but this is one of those things that requires the proper tools.

Someone kindly gifted me a Jesverty 30v 10A variable DC power supply.


I haven't dragged out my oscilloscope to measure ripple or anything, but I'm pretty impressed with it so far.
A quick check with a multimeter shows that changes to voltage or amperage happen in less than half a second, with no spikes.
I also find it really handy that when I turn it on, it doesn't put out anything until I hit the output button- so I can change my voltage or amperage to what I need. It saves my last setting, but it doesn't blast it out until I tell it to.
Amazon had this bugger on sale for $39.99, but the price bumped up to $55 today. There are comparable ones for $45-ish on the Amazon. Most of these are the same thing with a different brand name, so if you’re in the market, it’s really just down to which one has the biggest discount currently available.

If any of y'all have a use for the 0.5v to 30v 4A DC to DC buck-boost converter I was using, you're welcome to it at no cost. It is not stock- I direct soldered a low esr 35v 1000uf capacitor to the output, but honestly, that is something that you would want to do- so we'll call it better than stock. 


Anyhow. On the bench tonight, we have the JHEMCU GHF134AIO-BGA board #2.

Alright, folks, what are we going to do first?
If you answered, "Visual inspection under magnification!", you get a point. I would say you get a cookie, but, you're probably too far away from me to get a cookie that isn't gross and stale. But, if you really want one, I'll send you a stale cookie and a mystery box of random spare bits from my parts shelves. PM me, I enjoy sending and receiving random boxes of useless nifty spiffy shit. You don't get to request anything though- you might get a random frame or a set of motors, or props, or maybe just some nuts/bolts and shrink tubing, maybe just an antenna, or something completely useless and random like a bag of glitter- but you will also get a random cookie. The only requirement is that you have to make a post showing what you did with whatever I send you, even if it's just dropping it in a trashcan. 

Visual inspection shows a big ol' mess here:


As well as here on the main ground pad:


It's pretty clear to me that these messes happened when the previous owner desoldered everything from the board. When I'm pissed off at a thing, I don't carefully desolder it either.

Let's scrub this thing down with 99% isopropyl alcohol, and then clean all of that up.
Cleaned up that main ground pad. This wasn't really necessary because the side of that row of 3 caps next to the ground pad all go to ground anyway:


Cleaned up all pads including motor pads, and re-tinned most except for 3 that might be an issue:


I'm concerned about these 3 pads:


From the bottom up, we've got:
Vbat
Ground
5v

We can't be having what certainly looks like shorts between any of those.
Testing for continuity on our donor board, here's what we have:


Red squiggle continuity from 5v pad to the side of that cap closest to it.
Green squiggle continuity from ground pad to the side of that larger cap that the green squiggle goes to.
Vbat continuity to the other side of that cap (yellow squiggle).

After some real careful scraping, that's what we now have:


But, I'm concerned about soldering anything to those pads as is, because the exposed bits are real damn close to each other.

So, he're what we're gonna do. We're gonna scrub them down with isopropyl alcohol again. We're going to test for continuity again and make sure that things are connected to what they should be, but not shorted to anything they shouldn't be.

Scrubbed and tested, we're good.
Now, we're gonna put some solder mask over those bits that are uncomfortably close. That should protect those bits and keep them from bridging when I solder something to those pads.
Solder mask curing under my ridiculously overpowered UV light:


Alright! May not be as clean as I would like it to be, but I'm a twitchy MF-er.
I was able to tin all three pads, and we don't have a short between any of them:


Now that we're done with all of that, let's check for shorts on anything else.
No shorts! Awesome.
Now we can get down to actually testing this thing.

First up, connecting via usb.

Connected via usb just fine- damn good sign. That means our mcu and 3.3v regulator is good. Reset to dfu mode via the "bl" cli command, and flashed it with bf v4.4.2. Calibrated accelerometer, gyro is responding appropriately. Cool beans, gyro is good too.

Alright. Time to try to power this thing via the main battery pads. We're gonna solder on a buzzer real quick so that we can hear tones.
If we get none, we've got a serious problem. If we only get 3, we've got a problem with our 5v regulator or 5v rail. If we get all 5 sweet beeps, we're golden.


Well, I mean, it's a little confusing. We get a string of quick beeps, followed by 3 beeps. It's not the usual tones I'm used to. But, on power up, the red led flashes a couple times and stops and then the blue led starts flashing- which is the same as it does when powered via usb.
We're getting 5v on 5v pads, and more importantly we're geting 3.3v on 3.3v pads- our 3.3v regulators are downstream from our 5v, so, I think it's safe to say that our 5v regulator is okay. I *think* we're okay and this board is good to go, and the problem was bad manufacturing- those three pads (vbat, ground, and 5v) should have had a thicker coating between them to keep them from shorting.

I'm not 100% sure that we're good yet, though.
Let's reflash the esc's and see if we can get the proper tones we're looking for.
Flashed the esc's- we were on blheli_s 16.7, now we're on bluejay 0.19.
Same behavior- string of beeps followed by 3 beeps, same led behavior.

I want to say that this thing is good to go, but, we need to know for sure that the mcu is properly booting up via battery power. I've had a few, and at the moment, I can't think of a good way to test that short of slapping some motors and an rx on it, and seeing if it arms.

So for the moment, we're going to say that this board is undetermined, but potentially fixed.

Next up on the bench tonight is my damn laptop. This thing has slowed to a crawl, so, it got its fan cleaned out and some fresh thermal paste on the cpu, the junk on it was a bit dried out and powdery:


Seems to be running better now, maybe I can actually run a sim again.

For you fellow drinkers, grape brandy and Pepsi do *not* make for a tasty mix. I don't recommend it.

Very nice, I am getting good education and entertainment for the cost of a stamp 
Definitely sending you another batch when I've collected enough of them...

That JHEMCU was from years back, before I paid much attention to soldering. Maybe its my faulty memory, but I recall back then everything just went together and worked (or did after some rework), it was really rare to break a board on the bench. My first 16x16 board was built without any issues and don't think I even paid attention to be extra careful (guess I didn't know any better), but with recent boards even though I think I am being a lot more careful... oh well.

I’m always down to get new dead bits to play with.

I think quality control went way downhill when the chip shortage started, and the quality control doesn’t seem to have come back.

I don't ever sell anything that I repair. While I'm confident enough in my abilities to risk my own gear, I'm not willing to risk anyone else's.

Nobby sent me a Walksnail 1s v1 vtx and a Walksnail v1 full size vtx- with the agreement that if I can repair them both, I will send the 1s WS vtx back to him and I'll keep the fullsize.
I also recently sent a care package of some repaired boards to one of the igow4 finalists to help keep him in the air through the end of the competition, with the understanding that he can give away the boards with a warning, they can not ever be sold.

So I guess I am okay with some of my repairs going out into the world, as long as the user understands and accepts the risk, and as long they never hit the resale market.
I certainly wouldn't be comfortable with anything that I repaired floating around and being sold. As any of y'all that have followed this thread know, I am not and do not claim to be an expert or a professional. I just enjoy the challenge of trying to resurrect dead or misbehaving electronics.

Anyhow. First up on the bench tonight we have a Walksnail 1s v1 vtx:


According to Nobby, this board was on a moblite7HD, and stopped working after a crash.
That tells me that most likely we have a problem with the power system- either the voltage regulator, the Dialog DA9062 chip, or missing component that is keeping power from being delivered where it should be- ie: resistor, etc.
I could be wrong, but that's where we're going to focus after our usual visual inspection and short testing.
Something that worked until it crashed, that wasn't installed on a quad big enough to cause catastrophic dissasembly on impact points to either a component being knocked off a board, or a component that burned out perhaps while the owner was retreiving it.

Maybe. We'll certainly see, and our short testing could point us in a completly different direction.

First up, visual inspection. Checking this board for missing components is harder than it sounds. There's a whole bunch of pads with solder that do not have components. Thankfully, Mads Tech is awesome, and we've got some clear high resolution photos of these boards on repair.wiki

Visual inspections shows a couple potential issues:

We might be missing a capacitor and a resistor. Circled in red. If you look at the pads that I circled in yellow, these are pads that were tinned at the factory but didn't have a component soldered to them. If you look at the pads circled in red- they look different. They don't have the dome blob of solder that a tinned pad without a component would:


This pic of the board at an angle might better show that the top of those pads are flat like something used to be attached there:


Here's how that same section of the board should look according to repair.wiki, note the missing capacitor and resistor compared to the above pics:


It also looks like we might have a damaged mipi jack:


That could be a problem- I haven't determined what mipi jacks these boards use, and I certainly don't have any replacements on hand. Hopefully that jack isn't damaged beyond repair.

Alright, what's our next step? We're going to remove the black adhesive from the mipi connector and see if theings are still usable, then scrub the board down with isopropyl alcohol.

Nobby, if I get this board back in the air- You're going to want to put some liquid electrical tape on the ends of the mipi jack so that it looks like this:


And I'd suggest not repeatedly removing that cable from the jack. If you need to change it, do it from the camera side of the mipi cable unless the mipi cable is damaged.
We managed to clean up that mipi jack and make it functional, but, we did lose any grip on the side that was warped:


So, once again, Nobby- make sure you add some liquid electrical tape to the mipi connector.

It's time to scrub this board down and then check it for shorts.
I mean, we kinda already know what the likely problem is- the missing cap and resistor. But, let's check for shorts, and potentially apply power to it.

Okay. Scrubbed down. We don't have a short between our main power in- awesome. We don't have a short between the output of our vreg and ground- super awesome, that means our DA9062 chip is likely still good, as is our vreg.

Our issue is probably just those missing bits. Thankfuly, the full size v1 vtx and the 1s v1 vtx have almost the same board/layout. The v1 fullsize has that missing cap and resistor, so let's measure them and replace them.

There should be a photo here, but I can’t seem to find it… so maybe ignore the paragraph below.

Alright. Reisitor (right, yellow line) is measuring somewhere around 0.3 to .4 ohms. This is not an accurate measurement- my probes have about 0.3 ohm resistance. So, we know this resistor is 0.0001k or less.
Capacitor (left, Red line) is measuring... well, it doesn't want to. I'm less concerned about the capicitor, we can throw a larger cap there and be fine, it's most likely just for voltage smoothing.
I'm gonna have to just take these bits off the full size vtx so that I can properly measure them. We'll reinstall them afterwards.

I finally decided to splurge and spend $3 on some Mechanic Aac-14 tweezers from aliexpress. I'm glad I did. In addition to having a ridiculously fine tip, the inside of the tweezers are roughed up a bit- that should provide for some grip, and hopefully minimize smd components flying out of my tweezers never to be seen again.


Alright, off the board, that cap is... so damn small that I can't even see it, and I have no idea where it went. Crap. We might have just destroyed one board trying to fix another, folks.
I found what I thought might have been it, but, if it was, it disintegrated as soon as I rolled my finger over it.
Whelp. As I said before, I'm less concerened about the cap. Maybe we can now get a proper reading on the resistor with that cap off the board. Fingers crossed.
Alright, that resistor is testing at 0.09 ohm.

Let's slap a random cap back on this full size v1 vtx, and then get back to the 1s.

Alright.
1s pads tinned:


Missing cap and resistor installed:


Nobby, if applying power to this causes magic smoke to release because my gear wasn't accurate enough to properly measure that resistor and I ened up installing one with too low of a value that allowed too much voltage to go through it, I sincerly apologize.
Alright. Let's scrub this thing down with isopropyl alcohol, and give it power and see what happens...........
Fingers crossed.
Hrmm. No signs of life, but also no magic smoke.
Let's put a cam on this thing, I don't know if one is necessary for power up.
Hrmm. We're pulling amperage- and not so much that I think we have a short.
Chips are getting warm, which means they are getting voltage. That means our vreg and Dialog chip are definitely good.
I'm not really sure where to go from here on this one, I need to think about it for a bit.

Alright, let's take a look at the full size v1 vtx in the meantime.
Nobby wasn't sure what the status of this one is. It might work, it might not, but, he said he never trusted its range.
If this one has a problem, well, the first thing we'll look at is that cap I destroyed.


Visual inspection looks okay.
Let's scrub it down with iso and apply power to it and see what happens.
Okay! this bugger not only shows signs of life via the led, but it actually has video after binding to my vrx.
Regarding the range concerns- I'll have to test thoise in flight. If there are any range issues, I've got replacement SKY85743-21 rf amplifiers that I can toss into it. We're not going to do that unless we determine it is necessary.


The 1s vtx- at this point I'm thinking it is likely one of the main chips that is toast- if that is the case, it will be a parts/donor board, it is not worth repairing. We can try putting a different value resistor in place, or, given that it was a pretty low value, we could even try just bridging those pads- but that is just asking for smoke. We'll probably try it anyway, because we don't have anything to lose at this point.

top stuff lemony, cheers for trying. no worries re the disclaimer, you're all good...
that 1s did really take a hit, would not be surprised if bits flew off it. hard enough to destroy canopy, the cam got ripped off too, that would explain the mipi damage.
hope the v1 works out, i bought it second hand and always thought the range was shit, may have just been my expectations. after i crashed it yet again, fecked the antenna pigtail and the heatsinks fell off, i just went for a new v2 pro.

I love my hot air station... it's crazy how much fun it can be messing with old boards and doing surgery!