Many many boards have a specific order of which slots get filled. One stick can only be in one slot, two sticks in two specific slots, etc.
DIMM #2 is normally the first one to be populated, and the only one that will work if there is only a single stick. HP for all their other failings usually has pretty good documentation available for things like this. Look for a 'Service and Maintenance Guide' or something similar, instead of the typically useless Users Guide.
Cardinal Fang
Senior Member
Last Activity: 02-06-2026, 04:38 AM
Joined: 12-17-2025
Location: Meridian MS
-
Measuring resistors in-circuit, if the resistor is good you can still get a reading significantly lower simply because of other parts of the circuit; however, a good resistor cannot give a reading higher than actual because of the other parts of the circuit. This doesn't count if there are capacitors elsewhere in the circuit that are charging/discharging while measuring, but that will give you a constantly changing resistance number. If you really want to be sure you need to remove the resistor (or lift one side to break it from the circuit).
Are you sure you have the exact right...Leave a comment:
-
GTX1050ti Samsung to Hynix swap - BIOS changes?
GTX1050ti with some bad Samsung on it, found a cheap source for a donor board but they are Hynix. I have schematics giving the straps config for either so that's not a problem, but are there any differences in BIOS from memory brand to brand? Or just swap chips, change straps accordingly, and it just works? -
Finally, no more shorts.
The core is confirmed dead, measuring from 1v8/NVVDD/FBVDDQ pads to ground pads shows all 3 are shorted. No idea what the deal is/was with that memory chip.
p.s. I had to give it 5 AMPS for the thermal cam to show anything. The 1v8 rail would also take the full 5 amps but never did make a hotspot anywhere on the card....Leave a comment:
-
NVVDD short is gone!
However 1V8 is still shorted. And it also now has a dead short on every FBVDDQ coil, which it definitely did not have before. No vram got bumped while it was hot, no caps got knocked around. No I haven't wicked off the board yet but there are no bridges, I have looked at each and every pad up close and they are all clear....Leave a comment:
-
Don't see how that's possible since there are no shorted rails, all the rails come up as normal when run from the bench supply, it draws ~1.8a, and the thermal images look absolutely normal - no hot spots where it should be cool and no cool spots where it should be hot.
However about 1/4 of the PCIe lanes are missing when checked with the tester board thing.Leave a comment:
-
This is the same area but under a UV curing lamp (Mechanic L1 Pro). I haven't seen any specs listed anywhere but these curing lamps are probably 405nm? Not sure.
That shows absolutely nothing that would lead you to believe that spot needs closer attention....Leave a comment:
-
Weirdly, the top of the board shows more discoloring in visible light than the bottom, but somewhat less in UV.
So... maybe somebody needs to see what's going on under that core
Oh that's not good.
I just hope the PCB isn't completely roached, I have another identical 3070ti with a cracked PCB, I'd rather swap those good parts over to this board than deal with a crack repair....Leave a comment:
-
Yes, the UV is really good at making heat-affected zones pop. You can even look at working but well-used boards and spot where the factory skimped on cooling and whether it would be a good idea to maybe improve on things a bit. The area doesn't have to be charred underneath for it to show up, if it's been hotter than the stuff around it, you will be able to see it. And this is on things that look completely 100% normal to the naked eye & normal photos.Leave a comment:
-
HA! Still doesn't output any video, but I managed to poke the right buttons to finally get it into Windows...
Also managed to pull the BIOS out of it. Why am I wasting time on this when I already have five other repaired & working 1050 Ti's? Better yet, why am I bothering to fix yet another 1050 Ti when I have four repaired & working 3060 Ti's......Leave a comment:
-
Olight Arkfeld Pro, however do not buy one of those. They cheaped out and used a plastic lens instead of fused quartz, and the UV melts/bubbles the plastic lens if you aren't super careful about how long you let it run.
365nm is not blue or purple at all, it is much brighter than what your eyes can pick up.
I suppose eventually I will be building one from scratch which will as usual end up being hilariously overpowered for what it needs to do.
30 amps? I WANT ONE NO WAIT, TWO I WANT TWO...Leave a comment:
-
UV is magic
Can you spot any problems on this board?
How about now? See anything that stands out?
This is 365nm, other wavelengths will highlight different things. 365 is the most useful for the most things, though. It also cures all the UV solder masks & resins better than whatever other wavelengths the off-the-shelf UV lights are using.
(Just don't ever use 365nm to look at your own skin, or especially your face close-up in a mirror. I won't say any more about that but you have been warned.)... -
Speaking of chip brushes...
Go to your local home improvement store and buy all the cheapest 1" chip brushes they have on the shelf.
(two of those are new, the other two (shortest & longest) have been used every day on my bench for at least the last 5 years and still in great shape)
You will need the sharpest pair of scissors/shears on the planet to trim these anything close to evenly - whatever you have right now, they aren't sharp enough. Trim to various lengths for various uses. Short for heavy cleaning/scrubbing, long and fluffy for light...Leave a comment:
-
If you have to work on boards with corrosion, read this.
Yuk. Let's fix that. (or at least quickly get it in a condition where it's possible to actually fix it)
HOW??????////
Literally that took 5 minutes, with a stubby chip brush, some BKF, water, and then a rinse in alcohol. It only takes off material that was already compromised, and leaves you with something that solder will actually stick to.
BKF has oxalic acid but it is very mild, and after all exposure time is like 5min.... -
Trusted source for GPU-specific stencils?
Anybody have a go-to source for this, or just poke-and-hope on the ol' Aliexpress? -
Even the Pomona can have issues with staying put, the shape of the little grabber finger... things... (the plastic bits, not the contacts) isn't quite right to grab onto the underside edge of the chip package. I have always had to do a little fine detailing with a needle file, they're much better after that.Leave a comment:
-
Thermals... looks normal to me.
Possibly of note, the core & mem voltages do change right at the point the motherboard times out and switches over to IGP. NVVDD starts at 0.804v and then steps down to 0.774v. FBVDDQ starts at 1.506v & drops to 1.365v....Leave a comment:
-
The test rig will eventually post if set to use integrated graphics first, it just takes forever. The VGA fail LED turns off after about 30 seconds, it beeps, and boots.
This is what the card has on data out/data in at the BIOS
out:
in:
Anyone have a sample of what this is supposed to look like?...Leave a comment:
-
1050 Ti headscratcher
Got this EVGA 1050 Ti, bought as non-working. It had some damage on the back of the card that I fixed up, and all of that checks out OK now.
No shorts on anything. All PCIe lanes intact. All voltage rails OK. 1.8v present at the BIOS.
3v3: none, not populated
12v: 12.150
5v: 5.037
1v8: 1.799
nvvdd: 0.804
pex: 1.008
fbvddq: 1.506
Core warms up like normal, card draws 0.46A (no fans connected) on the bench supply in line with other working 1050 Ti's I have here. When fully assembled, the fans run at normal default idle speed...
No activity results to display
Leave a comment: