Author Topic: Gigabyte GTX 1070 8gb Founders edition missing to SMD ceramic capacitos (Read 4057 times)

Koletsa · « **on:** October 08, 2021, 02:51:33 pm »

Hello folks! Posting for the first time here, so if it is in the wrong place, I apologise in advance.
As I said in the title, I have a Giga 1070 card, which is missing 2 ceramic capacitors (red circles on the pic). Can some one of you assist me in finding their right capacity so I know what to look for as a replacement. I know I could desolder one of the adjacent caps and measure it, but I keep that as a last resort measure. Thank you in advance! Looking forward to hearing from you.

richard.cs · « **Reply #1 on:** October 08, 2021, 05:13:54 pm »

Are you certain there are supposed to be capacitors there? it is common to have pads for components which aren't fitted, e.g. to select different functionality on a single PCB layout. The way to check would be to compare with another known-good card (or photos of one).

Koletsa · « **Reply #2 on:** October 08, 2021, 06:54:23 pm »

Hello! Yes, I'm certain. I googled the exact same model and there are caps on those 2 places. Also on the one bellow the gpu - one of the soldering sockets is missing the solder pad. I'm not aware how and why those caps were removed but is it a fine way to start repairing the card.

TheHolyHorse · « **Reply #3 on:** October 08, 2021, 08:46:14 pm »

Does it work as it should? If it does why even bother.

Vovk_Z · « **Reply #4 on:** October 09, 2021, 04:54:17 pm »

Quote from: Koletsa on October 08, 2021, 02:51:33 pm

I know I could desolder one of the adjacent caps and measure it, but I keep that as a last resort measure.

1) I agree with the previous that there isn't a need to repair what is not broken. There may be redundant bypass caps on the PCB, and the manufacturer made homework to reduce excessive redundancy.
2) Adjacent caps may have totally another rated value.
3) You may only be sure desoldering these exact caps from the same card.
4) Are they caps, but not resistors (for example)? I can't see clearly.

Johnny10 · « **Reply #5 on:** October 09, 2021, 07:16:47 pm »

Unless I de-solder the chips on my board I cannot be sure because of measuring in circuit.

The cap on right is 10uF.
All the schematics I see only accurately show the value of the capacitor on the right.
The SMD on the left in pic? not shown in same position.

HowdyPartner · « **Reply #6 on:** October 09, 2021, 07:32:39 pm »

Has it been baked in an oven in an attempt to bring it back to life?

Johnny10 · « **Reply #7 on:** October 10, 2021, 12:04:54 am »

OK, I just couldn't let it go...

I am not really set up for SMD but I already had 2- EVGA GTX 1070 cards on my bench that I got back from a damaged delivery.
The layout is exactly like the Koletsa's card in this area.
So I removed that Capacitor on the left of the pic and measured it.

And then I preceded to shoot it across the room off into space. LOL

You have to treat working on these size chips as you would watch parts.
Build a side walled work area in white or green so parts can't go ricocheting off into the nether country and you can find them again.

Oh well, Its a 22uF Capacitor.

Maybe Santa will bring me JBC Nano tweezers for Christmas.
Wouldn't that be something?

Koletsa · « **Reply #8 on:** October 10, 2021, 01:05:32 pm »

Hello and thx for the assistance! I bought that card as non working at all - fans don't spin, rgb doesn't glow, the core and the ram chips, incl. the mosfets don't get warm. But the PC stats as if I don't have a vga in the slot. So I started exploring the pcb and I noticed those 2 caps missing.
Then I star measuring and found new problems. I'm missing the 5 v completely. 12 and 3.3 v are both fine. Based on a scheme that I found on Repair Wiki, explaining the 5 v rail on Paskal GPU I understood that the 3.3 v are entering into one buck converter (I located it and the 3.3 v are entering there) but the 12 v are not and because ot that I don't get the 5 v anywhere on the PBC. The scheme shows that 12 v from the Bus should reach the convertor. The same scheme shows that between the Bus and the convertor should lay 2 caps. One of which most brobably is faulty. I checked the one tha lays right above the Bus itself and it is OK. But I'm not sure if it is one the two and which one it is exactly. Do you have any idea how can I find which of all caps are those two. I'm sharing below the scheme from repair wiki. I'd like to state that the components numbers there are not present ot my card - I don't have c501, nor c252. Thx in advance!

Johnny10 · « **Reply #9 on:** October 10, 2021, 02:13:58 pm »

I don't believe the C501 and C252 are matching numbers on the 1070 card.
C501 on 1070 PCB is part of stereo buffer circuit and C252 doesn't exist on this card.

I'll take a quick look at the Wiki site and get back to you.

Johnny10 · « **Reply #10 on:** October 10, 2021, 03:09:40 pm »

The wiki repair page is pretty specific about the area of 5V rail circuitry.
Pictured in the upper left hand of the Front of the card

U2 is 8-Pin N-FET Linear Regulator Controller UCC3837
u37 IS 5 PIN LDO Low Voltage Dropout regulator

I would be checking all the caps and resistors in that area and following the wiki 5V short pics

Koletsa · « **Reply #11 on:** October 10, 2021, 04:49:47 pm »

Hello, Johnny10! Thx for the assistance again! I have 1 last favor to ask at that point - does any of your 1070 cards (I understood that you have 2 EVGA) have Samsung ram chips. I'd like to know the resistance of the Vmem inductor choke (R33). It says it should be between 30-60 oms, but Samsung chips were using lower voltage than microns, which meas that mine inductor should probably have higher resistance than the one for micron rams). Mine currently reads 0.2 oms. I have the same question about the 1.8v rail inductor - mine reads 1 om but it should be 800 oms+. Obviously those 2 inductors are damaged and I will replace them before I start fixing the voltage issues. And Yes, I read that the 1,8v goes straight to the Vcore and short on that specific choke most likely means dead core, but I'd like to excercise on it. Thank you in advance!

Johnny10 · « **Reply #12 on:** October 10, 2021, 05:34:21 pm »

Yes, I can check that out later.
I was thinking about your problems. I believe you should get all the voltages back to normal. Inducters don't go bad very often.
If you checked the voltage going into the U2 IC and going out. You will know if it is because of bad Regulator or caps in the area.

Do you have between 3 to10 volts on pin 2 of U2?
Do you have 5 volts coming out of pin 3 or 5?

Also one side of L1 should be at 5V

Koletsa · « **Reply #13 on:** October 10, 2021, 07:43:29 pm »

Hello!
1) Answers to your questions: on pin 2 (Vin) I have 0 volts; pin 3 (SW) = 0.192 v; pin 5 (boot)= 0.260 v. Also - pin 1 = 0.192 v; pin 6 = 3.302 v; pin 7 = 0; pin 8 = 0.030 v.
2) On L1 I have 0.192 v (on both legs). I have that exact voltage on a lot of the nearby capacitors;
2) I checked the pins also on continuity mode and I got beeps only at 4 (ground) and 2 (Vin). Does this mean I have a short on Vin and what would that mean?
3) Just to make sure, I'm attaching the pick with the u2 module and the numbers of the legs as I measured them;
4) In one of your previous posts you posted a scheme of the PBC in black and electric green with all the components and layers. What is that exactly and how/from where I can get such a scheme for my model?
Thank you in advance!

Johnny10 · « **Reply #14 on:** October 11, 2021, 01:26:52 pm »

Here is the location of the two PS 5V_12V In Signal caps.
listed on your image as C501 and C252.
C1100 22uF and the one to the right C1002 0.1uF

I am always searching for schematics and other info on these cards of late.
Have to get more organized as to where all this info originates.
I believe a couple of youTube videos are geared towards the 1070 repair also.

Koletsa · « **Reply #15 on:** October 11, 2021, 04:56:39 pm »

Hello again! You are a Saviour! I had my suspicions on the c1100 because there I have 0 volts on it but all the other caps of his size and shape nearby show some volts, so that made me wonder. Also, I'm not sure how visible is on the pic, but I think part of the ceramic layer is chopped off. I would never have gotten suspicious about the tiny ones to the right. Just to make sure - c1002 is the one that I marked (believing that c1001 is the one in the middle)? For the sake of the quest - I have 0 volts on both tiny caps next to the right from c1100.
Also another question - If I have to replace my inductor chokes on the Vmem (R33) and the on the 1.8 v (R2R): are R33 and R2R standards for the induction of the chokes?
P. S. I have watched several clips on faulty 1070 vga-a, but non of them had my problems. One had really low resistance in the R2R and it trourned out that the cause was the nvidia chip, which makes me a little bit worried

. I also found some schemes for the pascal gpu but at this point am still awful at reading them and it is all Greek to me.

Johnny10 · « **Reply #16 on:** October 11, 2021, 05:33:00 pm »

I have not set up a test board for this Graphics card as of yet.

Have been tracing around board and reading up on the circuits for 5V line.

For your questions... C1100 is connected to the first capacitor to the right C1102. The second SMD is a 49.9K resistor the third is another 0.1uF cap.

They all travel over to the other side of the board to the PEX circuits R1007, R987 and back to Pin 2 of U2.
Later today or early tomorrow I hope to power this board on and follow voltages.

The 12 volt line comes from the first two pins on the PCIe connector. There is a direct connection to the bottom of C1100 and to Pin 2 of U2. So you should be able to check that also.

Koletsa · « **Reply #17 on:** October 11, 2021, 06:23:53 pm »

It took me around a month to find good, yet affordable MB and PSU compatible with the spare CPU, RAM and cooling that I had for the set up of my board.
At this point I'm ordering the capacitors and I hope I can complete the solder/resolder procedure without any unnecessary damage. I hope that fixes the lack of 12 v to the convertor and bring back to live the 5 v rail. Once again - I remove c 1100 and the tiny one right next to it?
As soon as I have some results, I will write back. It may take some times, based on the availability of those components locally. Best regards!

P.S. Just saw your pic. Yes, my 12 v form the PCI are fine. I have them on C44 (face side of the PCB, above and to the left of the 1st pin) and on C1078 (the one above your probe). That is why I believe one of those 2 capacitor must be faulty. Based on the appearance, it should be C1100.

wraper · « **Reply #18 on:** October 11, 2021, 06:55:35 pm »

Changing capacitors won't fix anything and will not make 12V to appear. Even if those capacitors fail short, they would blow a fuse or something else in series with that 12 rail. Or if that rail was connected to 12V power input directly, just case the PSU to shut down due to overcurrent protection.
So first you should check if that 12V rail is shorted to GND. And if it's shorted, then find what's shorting it. And find an open fuse, ferrite bead in series with that rail.

Johnny10 · « **Reply #19 on:** October 11, 2021, 07:06:20 pm »

That's correct I have just laid out the connections to various components.
Which components are good or bad need more testing of the circuit.

Something is keeping proper voltage from going to the capacitor C1100 and to the input #2 pin of IC U2.

Koletsa · « **Reply #20 on:** October 11, 2021, 07:48:24 pm »

Hello and thank for tuning in! My 12 v are present both on the PCI and the 6+2 PSU. The problem at this point is that the 12 v from the PCI don't reach the buck converter. I have 0 volts on pin 2 (Vin) and thus - my 5 rail doesn't exist. And based on the only scheme that I have found about the 5 v rail - between the 12 v from the PCI bus and the convertor lay only 2 capacitors (the scheme is for Pascal GPU's and it is for Repair wiki, see the attached pic, pls). I also tested for short to ground the first 3 pins on the pci bus and the C44 cap (the one that is tested on every video for the 12 v rail). If you have any knowledge on the matter, I will be very thankful if you share it. Thank you in advance!

wraper · « **Reply #21 on:** October 11, 2021, 07:54:01 pm »

Capacitors are in parallel with the rail, they have nothing to do with supplying voltage to this rail. You need to measure if the rail is not shorted and from where the voltage comes from.

Koletsa · « **Reply #22 on:** October 11, 2021, 09:47:42 pm »

Hello, wraper! Thanks for the info. I read a little bit more and I got your point. Back to square 1 I guess. I obviously have to find a faulty resistor. Thx for the advice!

Johnny10 · « **Reply #23 on:** October 11, 2021, 10:56:23 pm »

I thought you knew you had a short on the line when you said you got a beep on Pin 2 and Pin 4 earlier in your posts?

It means one of the capacitors died short or the controller is bad or something connected to the controller is bad.

I found the schematic for the 5V line though the voltage regulator with accurate numbering for our card including the high side load switch.

Better than trying to follow the generic one you have been using.

Koletsa · « **Reply #24 on:** October 12, 2021, 06:21:00 am »

I knew about the short, but believed that the reason was one of the two capacitors we talked about. Then Wrapper said it can't be a cap, because they are parallel to the rail. Can you share the 5 v rail scheme so I can start looking for the culprit as well?


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: Gigabyte GTX 1070 8gb Founders edition missing to SMD ceramic capacitos (Read 4057 times)

Share me