DynaxSC
Badcaps Veteran
Last Activity: Today, 05:03 PM
Joined: 04-13-2021
Location: Parcela/Mazowieckie
-
You will almost certainly not find any documentation for these caps. The only way is to get a second faulty core of same model and transplant the missing caps. -
Try to take off the sound chip, and see if the board will start w/o it. Burning chip means usually internal short, which can down some often shared power lane. Usually the respective vrm can withstand this, as they have overcurrent protections.Leave a comment:
-
Regarding the small ohms resistance difference should be no problem, such difference can come from contact resistance and meter tolerance. DMM's do not measure such low resistances reliably. The 121ohm resistor is on contrary critically important.
Regarding my card i found the problem, the pcb is not flat under the gpu in the area of 6'th and 7'th vram, somewhat crooked, so during reaseating the gpu after reballing some balls have been pressed so strong in the respective corner that they joined together making a short. Found at least one data line shorted to gnd after taking off the...Last edited by DynaxSC; 03-07-2026, 07:22 PM.Leave a comment:
-
Hi, in that case voltage injection may not help, because even at 2V and 2 ohms the heat generated in a device will be 2W, so not very much, and you can't go with the voltage very much higher, but you can try to use the thermal camera raising the voltage to 2V, maybe 2.2V. 2 ohms suggest that the short is in a silicon chip, as caps usually short to 0 ohms. So the only way is to find all chips connected to the 1.8V rail, and take them ono-by-one from the board, as you already have done. After taking off the chips you may try to raise the voltage beyound 2V, but you must be sure having taken off...Leave a comment:
-
If you can inject 1.2V @ 2.6A then there is no short, at least not a total one, otherwise it would not be possible to have 1.2V, but the voltage should be very close to zero. So something is wrong with these parameters, and no analysis can be done on that. Measure the resistance w/o voltage injection. Measure the voltage on the motherboard injection points, maybe the voltage drop is on the cables from psu to injection points.
Further if there is an absolute total short on the 1.8V rail, there will be almost no heat generated as there is no resistance on which the heat can be generated, so...Last edited by DynaxSC; 03-03-2026, 06:10 PM.Leave a comment:
-
Hi, I'm wondering if there is not yet another possibility for memory errors. I have now a very stubborn GTX1080 card with similar issue. It shows errors on two vram chips, nr 6 (C0) and 7 D1) counting against clock direction. But, I have already replaced the vrams 2 times, replaced the gpu core for a known good working one, and replaced the vrams a 3'rd time - to no avail. My only explanation is the pcb is somehow defective, I see no other explanation. Tried also different bios'es from Techpowerup to go sure bios is not corrupted. Either some data/address/clk tracks or vias are broken and lost...Leave a comment:
-
Never tried this, but diode test might make sense for memory data/address lines too, however if you find an issue you will know, but if you find nothing, mem controller still can have a failure. You can test only statically the output/input drivers, but not the much more complex and huge memory control logic behind. The only way to confirm memory controller failure, is transplanting the gpu on a good known working card, or replacing the gpu on the non working card. Unfortunatelly mem controller failure is quite common, making a card repair financially nonsense, as replacement gpu's cost almost...Last edited by DynaxSC; 02-19-2026, 06:22 PM.Leave a comment:
-
Hi, I order such chips on Aliexpress, a good shop is Shenzen Hong Ming Electronics, they are not very cheap, but are a reliable supplyier, just order chips called New, chips marked Test are refurbished, but still a good option. However there are usually few other suppliers, just look if the shop has good rating, near to 5, and significantly over average. Maybe you can find them also on Mouser or similar distributors, but I do not know if thay have such chips in the portfolio.
Mooving chip as far as I can assess will not work. Bios would have to be modified. nVidia Bios is in the core...Last edited by DynaxSC; 02-18-2026, 04:53 PM.Leave a comment:
-
Power - very unlikely, all the chips have one common power lane. Likely are GPU memory controller failure (RIP), your chips from ebay are crap, cold solder contact under gpu, or traces/vias on pcb demaged, maybe also the pads under the vram chip are oxidated, did not connect with the ball correctly, or you have a short between two balls under the vram chip. These pads are sometimes oxidated and so very difficult for the tin to adhere, especially if use standard solder paste, which is not aggressive enough to clean such oxidated pads - 99% of solder pastes which you can buy in hobby electronics...Last edited by DynaxSC; 02-17-2026, 05:26 PM.Leave a comment:
-
I have an audio amplifier I made myself ca 42 years ago, full of electrolytic caps, polish production during deep communist era, it stood ca 30 years unpowered, I switched it on, and it still works, except some switches and potentiometers which need to be serviced, because they make noise.
Such caps can only dry out and loose capacity raising internal resistance, they should not explode or so. This can eventually lead to unstable power supply and instability of the board, normally should not demage the board, except they will be completely out of capacity, which is unlikely. Dangerous...Last edited by DynaxSC; 02-17-2026, 04:56 PM.Leave a comment:
-
You need somehow to rule out what is causing the problem, without some replacement hardware, or second pc you will not be able to do this. There is no magic crystall ball making a diagosis, only excluding the working elements one by one or putting the ram, cpu, etc. into another pc will end up with final diagnosis where the issue is.Last edited by DynaxSC; 02-13-2026, 05:15 PM.Leave a comment:
-
Either BIOS glitch resulting in not supporting this RAM (especially Gigabyte mobos are prone to have buggy BIOS'es relating to RAM detection), RAM chip failure, non-compatible RAM (check support list), CPU socket bent pins, dirty or demaged CPU or RAM sockets, cut CPU-MEM tracks (deep scratches on mobo pcb), CPU memory controller failure, some RAM power voltages missing or wrong value, chipset-RAM SPI chip bus connectivity failure (1 data and one clock signal), ripped off 15 ohms resistor on the RAM pcb (here are twin resistors used, so they are not so easy to rip off like single ones), and thats...Last edited by DynaxSC; 02-11-2026, 05:39 PM.Leave a comment:
-
This coil is related to one of the power lines of the chipset, probably the 1,05V. Chipsets fail quite often, they are much lower durability compared to cpu's, so your chipset is 95% faulty. Replacement is very difficult and you need skills, expensive preheater and hot-air equipment, experience and a spare chipset, which is a quite expensive part, making the repair risky and economically at best a zero value game. Better sell the board for 10-15 bucks, and buy a working one.Last edited by DynaxSC; 02-11-2026, 05:15 PM.Leave a comment:
-
Recently I had similar issue, gpu was also reballed, both mem chips in one channel replaced. The problem was a short between two memory lines of one channel under the gpu - two balls have joined during reseating the gpu to pcb, what produced memory errors. Just an last idea before declaring mem controller as faulty. Now I use balls 0,05 lower than nominal in diameter for reballing, this solved the issue of ball joining.Last edited by DynaxSC; 02-11-2026, 04:57 PM.Leave a comment:
-
Then I would explore why the vrm starts working when heated. This indicates a solder ball contact issue under CPU or somewhere else, or a transversely broken smd resistor or track. I found once such transversely broken smd resistor, this is very difficult to see, you need an x15-x20 lens to see such a crack. Might be also a shorted smd cap, which ends the short when heated. Finally check the low ohm VCC serial resistor of the vrm chip if not decalibrated. They are ususally 2R2, but can become degraded to higher value with time and load, especially if VRM failed before and got somewhat shorted...Last edited by DynaxSC; 02-07-2026, 05:05 PM.Leave a comment:
-
Hi, the photos are bad quality, not very useful. Most probably your controller is not well soldered to the pads, some pin does not contact correctly. This happens with such ic cases quiet easy, the soldering must be done with preheater, sometimes going around the chip with a solder iron helps, the tip must be quiet sharp in order to access the corner between chip case and pcb.
Its good to put tin onto the pcb pads before resoldering the ic, using only tin paste not always gives success, solder paste behaves unpredictable, and does not go always where it should. Make sure there are no tin...Last edited by DynaxSC; 02-04-2026, 03:39 PM.Leave a comment:
-
I believe S_SLPS3# should be asserted after power on, but is not. This signal is usually generated by PCH as response to the Power On signal passed through the Super I/O from the Power On pinheader to the PCH. It triggers all power lanes for the CPU and RAM, so it seems CPU has no power supply, and cant start POST procedure. Can you confirm this - no VCORE ?
For some reason PCH does not assert S_SLPS3#. It can be lot of things, eg. one of the two PCH clock gens degraded and not working, some status signal monitored by the PCH has wrong status or is shorted, PCH partially broken, Super...Last edited by DynaxSC; 02-03-2026, 06:17 PM.Leave a comment:
-
Hi, caps are fine, measured them for capacity, shorts and isolation. Pcie tester shows all leds OK.Leave a comment:
-
First check usb sockets for physical demages.
Second check the 2 inner usb data lines for shorts. Close to the usb sockets there must be quite small ESD protection devices, either 6-pin or 10-pin dual inline casing (even smaller than 6-pin). They might be shorted or low impedance due to electrostatic discharge. Usually the resistance should be in Mohm range (board unpowered). If shorted, remove them and see if short is gone. If not than usb driver or chipset is shorted on these data lines, dependent on to which of them the usb port is connected. Repair done by replacement of driver...Last edited by DynaxSC; 01-27-2026, 04:40 PM.Leave a comment:
-
Asked ChatGPT, its most probably a tantalum 10uF 4V cap.Last edited by DynaxSC; 01-25-2026, 05:02 PM.Leave a comment:
No activity results to display
Leave a comment: