Question BIOS error or hardware defect?

Jul 26, 2021
2
0
10
0
Hi everyone!

I recently got 3 videocards of same type (sapphire nitro+ 570 8GB) as defect. I guess they came somewhere from a mining rig...

As price was very good I assumed being able to make at least one or hopefully two of them running for myself.

Very inital question for myself is - how can I check whether I am having a BIOS issue or a hardware defect?

I tried running all of them in my PC with GPU-Z... results were quite different... from not even recognizing, to detection as the rx570 after second or third trial (maybe it needs to warm up a little?!?) or detection as a standard vga controller, but all of them finally with GPU clock of 0 and memory clock at 0...

Nitro+ has a manual bios switch... I used poth positions, but at the end best result was detection as a RX570 but again GPU and Memory clock at 0...

Can anyone support in at least this separation to BIOS or hardware defects?

c ya

Michael
 

Ralston18

Titan
Moderator
Update your post to include full system hardware specs and OS information?

PSU: make, model, wattage, age, condition?

Are you trying to run multiple GPUS: i.e., 2 or 3 of the RX570s' installed at the same time?

What BIOS error(s)?

Where did you purchase the 570's - source? What was that "good price"?

If your computer was working fine with whatever GPU you had installed and none of the RX570s worked then the RX570s are likely defective or damaged. May even be counterfeit.

Methodically try each RX570 in another known working computer.

If the problem(s) follow the GPU's then the GPU's are the issue.
 

vov4ik_il

Respectable
I assumed being able to make at least one or hopefully two of them running for myself.
Well, the diagnostic is the hard part... Software issues do not just happen - you would have to mess with vbios to brick it, hence 99.99999 hardware. Check power phases first - they are most likely the ones to go. Why would you buy something to repair and then start learning repairs? :)
 
Jul 26, 2021
2
0
10
0
thank you for your replies...

@Ralston18: basically I know that they are defect. They were sold as defect units. I paid 200EUR for all three cards. I am using a GTX970 as basic video card and connected the 570 into another free slot (one by one). You asked for what BIOS erros - that is my basic question - how do I know whether its BIOS error or not?

@vov4ik_il: okay... 99.9999 hardware makes it a little easier ;) But can I check and how whether the installed BIOS has just too high settings for RAM or GPU so that it could not start? I already measured the video cards... There is no shortage on the board... there are no visible damages or blown out parts... all fuses are still working properly... the GPU heats up when connected to the mainboard... and Mosfets are most probably still working in right way... so chances for mistakes gets less... honestly I dont know where to search :(
 

Ralston18

Titan
Moderator
I will broaden my question to include errors of any sort from any source:

BIOS, POST, Event Viewer, Reliability History, Update History, a popup window, etc ....

Before something can be fixed discovering or otherwise knowing what is broken is the key requirement.

= = = =

Per @vov4ik_il - diagnostics are needed.

Do you ( @justmicha ) have a multi-meter in order to check voltages, resistances, continuity etc.?

With three GPU's you would be able to compare measurements. You still need schematics and specs to identify failed or defective GPU components.

For example (not an endorsement or recommendation):

https://www.repairlap.com/threads/rx580-rx570-rx480-rx470-rx-ms-v341-schematic.3820/

You can google for other similar sites and documents.
 

ttower2020

Respectable
Nov 5, 2018
369
72
2,090
75
thank you for your replies...

@Ralston18: basically I know that they are defect. They were sold as defect units. I paid 200EUR for all three cards. I am using a GTX970 as basic video card and connected the 570 into another free slot (one by one).
I just want to make sure the you shut down the PC each time you swapped a GPU. PCIe slots are not "Hot-swappable", and need to be powered off.
 

vov4ik_il

Respectable
can I check and how whether the installed BIOS has just too high settings for RAM or GPU so that it could not start?
You can read the binary off the flash IC and bin-compare it to a known good reference. Higher CLocks/Voltages are only used per driver request so it should display and post anyway.

I already measured the video cards... There is no shortage on the board... there are no visible damages or blown out parts... all fuses are still working properly...
So what voltage outputs are present in the VRM circuitry? How did you conclude that there are no shortages or blown-out parts?

the GPU heats up when connected to the mainboard...
This does not mean much, OK it gets power so at least one phase goes up.

and Mosfets are most probably still working in right way... so chances for mistakes gets less... honestly I don't know where to search
Well, if you want to go into repairs, you would have to be methodical.
  • Start by visual inspection for obvious visible damage
  • Next, measure power and data interfaces for abnormals (this will usually yield most problems)
  • Next, if the above shows safe to plug power, examine power in all phases
And yes, you will need tools and be able to use them. At least a good multimeter and soldering iron for starters... Regulated bench power supply, smd/qfn (bga if memory needs replacement) rework station, and oscilloscope preferred.

For schematic requests, be significantly more specific by posting make/model/version/revision. Otherwise, it sounds like "-... I have a 4-door Japanese sedan"
 
Last edited:

ASK THE COMMUNITY

TRENDING THREADS