The Odyssey of an Ethereum Miner (problems solved)

[Pertinax]

Hi everyone,

I am currently facing a problem while setting up my first mining rig. I've already done a lot of research, without any success. Help would be very appreciated  

My GPUs (rx 570 4gb) are not detected when I use x1 --> x16 pci-e risers. However, they are when I connect them directly into the pci-e x16 slot (on MB).
My motherboard is an Asus z270-a and this is what I already tried :

1. Activated 4G in BIOS
2. Set pci-e settings to auto and then in GEN2
3. Tried with different risers and SATA to Molex cables (I have got 6, at least one of them should work)
4. Updated the MB

My power supply has 1000W, and I am almost certain that my risers are correctly powered.

I am running W10 and tweaked it with a tool (and already runned DDU as well)

[DevelopmentBank]

You seem to have done research well. But just possibly, are you sure you are plugging in the USB cable from the riser to the motherboard? I can only think of that as the reason or probably your GPUs are detected its just that they are not installed yet. Check device manager to see if you have uninstalled "Display Adapters."

[Pertinax]

Yeah of course ! I even tried to change that cable, didn't make any difference.
Plugging in a 1x PCI-e or a 16x PCI-e slot doesn't change anything either.

My GPUs don't appear on GPUz nor on device manager (except the integrated INTEL one)

[Mamasita22]

I have same issue on some cards. Realized that this cards don't want to work in x1 PCI-E mode like all usb risers. But it work plugged directly. I didn't understand, videocard is the problem or riser, just plugged to x16-x16 trail riser with additional power, and works perfectly

[Pertinax]

All my GPUs cannot have that same issue,... That's almost impossible.
I have seen rigs working with the same cards and MB

I guess it has something to do with the BIOS ?

[smoolae]

It's not rare to get faulty risers when buying them. I have seen examples where 50% of bought risers are not working.

[Geraldo]

From what you describe, what you have done and also a few suggestions above, many possible things may be caused it, but I suspected this is a power issue. Have you tried to powering your riser directly from PSU? Try it, as we know (RX) 400 series has a power issue, and may your 570 has the same.

[smoolae]

Hi everyone,

I am currently facing a problem while setting up my first mining rig. I've already done a lot of research, without any success. Help would be very appreciated  

My GPUs (rx 570 4gb) are not detected when I use x1 --> x16 pci-e risers. However, they are when I connect them directly into the pci-e x16 slot (on MB).
My motherboard is an Asus z270-a and this is what I already tried :

1. Activated 4G in BIOS
2. Set pci-e settings to auto and then in GEN2
3. Tried with different risers and SATA to Molex cables (I have got 6, at least one of them should work)
4. Updated the MB

My power supply has 1000W, and I am almost certain that my risers are correctly powered.

I am running W10 and tweaked it with a tool (and already runned DDU as well)

Whose PSU are you using and what efficiency does it have? Those NoName PSUs often cheat with their output power claims.

[gt_addict]

You are supplying enough power to the card aswell yes? If you have 8pin and 6pin sockets on the card make sure you plug BOTH in to start with.

[NiklasFalk]

It sounds like you have tested the risers one by one connected to the "primary" 16x slot?
The the "only" power issue can be the delivery to the riser, if it has a 6pin connector, use it to take away any converter crap.
If you still have issues, the solution is not to fiddle around with the PSU.
Also, if you only use one card connected to the primary 16x slot the bios could be at default setting since the card will present itself as it was connected to the 16x (this would only test the riser part of the setup).

If you have any mining friends nearby, try to get hold of some riser of another type (preferably one that have worked).

[edwardceng]

I knew someone with the same mobo and same problems with you only GPU was different. It was on Reddit post, may it can help you to figure it out your problem.

https://www.reddit.com/r/EtherMining/comments/7q4hlk/asus_z270a_not_recognizing_pcie_risers/

[remauto1187ma]

what version risers?  Do you have sata power to riser attached?  Work with one card only until you get it resolved then start adding and installing in device manager one at a time after reboot and device manager shows "working properly".

[Pertinax]

Hey, thanks for all your kind replies  

The problem turned ou to be the SATA to MOLEX cables that were a bit too bented, and therefore couldn't power the risers !
I put them in a better position and the cards were finally detected.

Now, I flashed the BIOS of the cards (with Polaris and ATIflash), but the fans don't spin anymore (only a bit when I turn on the MB, but when W10 boots, they go off and get really warm). Prior the flash they were always spinning !
Another fun story : once I flashed them, the memory type changes on GPUz (from Hynix to Micron, and then from Micron to Samsung...) I didn't take a random BIOS on the net, I just clicked "One click timing patch" on Polaris, saved the new .rom file, put it in the ATIflash folder, runned CMD and flashed one card at a time...  However, on MB the memory type is the right one. So I reinstalled GPUz, and the memory went back to normal (except the fans, that still do not spin)

Should I wait more, and hope that the fans finally decide to spin? I'm afraid that they never will, and that the card will destroy itself  

This rig is really driving me crazy  
Help and advice would be very appreciated  

[theodrim]

Hey, thanks for all your kind replies  

The problem turned ou to be the SATA to MOLEX cables that were a bit too bented, and therefore couldn't power the risers !
I put them in a better position and the cards were finally detected.

Now, I flashed the BIOS of the cards (with Polaris and ATIflash), but the fans don't spin anymore (only a bit when I turn on the MB, but when W10 boots, they go off and get really warm). Prior the flash they were always spinning !
Another fun story : once I flashed them, the memory type changes on GPUz (from Hynix to Micron, and then from Micron to Samsung...) I didn't take a random BIOS on the net, I just clicked "One click timing patch" on Polaris, saved the new .rom file, put it in the ATIflash folder, runned CMD and flashed one card at a time...  However, on MB the memory type is the right one. So I reinstalled GPUz, and the memory went back to normal (except the fans, that still do not spin)

Should I wait more, and hope that the fans finally decide to spin? I'm afraid that they never will, and that the card will destroy itself  

This rig is really driving me crazy  
Help and advice would be very appreciated  

Firstly nothing preventing you from making fan curve in software or settings static fan speed in software/miner.
Secondly, unless you checked only card-by-card, gpuz isn't ordering card normal.
Thirdly - How did you save your current bios from cards? Did all of your cards have exact p/n and similar s/n? Did you mod each one, or only one and flash it to all cards?

[Pertinax]

1. I will certainely do that,... I just wanted to know if a GPU flash can higher the temperature the GPU needs to have before it activates its fans.
2. I am aware that they have an order, and checked them one by one. I am 100% sure none of my GPUs were Micron/Samsung (only hynix and some Elspidia)
3. Yes I made a BIOS backup of every card (with GPUz), and will use them to see if the fans work back to normal. My cards are all different, and I modded each one of them and flashed them one by one.

[theodrim]

1. I will certainely do that,... I just wanted to know if a GPU flash can higher the temperature the GPU needs to have before it activates its fans.

Tighter timings may have some temperature impact on memory temperature and indirect impact on core temperature (because if you have cards with proper cooling for vram chips they share same heatsink as gpu core, so you'll have slightly higher core temp if memory get hotter, but it's not end of the world. Fan speed depends on target temperature that you have in settings or bios (settings in driver will override bios ones, if set) and also acoustic limit.

2. I am aware that they have an order, and checked them one by one. I am 100% sure none of my GPUs were Micron/Samsung (only hynix and some Elspidia)

You can check memory manufacturer withing PBE itself (and you can have different memory inside single card, eg hynix/samsung). For example left one is hynix/samung combo (that will show hynix in gpuz, cause that have majority of chips); right one - elpida (whose manufacturer are micron, so technically microne == elpida):
https://image.ibb.co/en47Gn/image.png

3. Yes I made a BIOS backup of every card (with GPUz), and will use them to see if the fans work back to normal. My cards are all different, and I modded each one of them and flashed them one by one.

Again, I'm sure they just have so relaxed fan curve and only spin when gpu reach high temperature (for prolonging fan life cycle and reducing noise when idle), so just change it in software and be done with it.

[Geraldo]

The problem turned ou to be the SATA to MOLEX cables that were a bit too bented, and therefore couldn't power the risers !
I put them in a better position and the cards were finally detected.

Now, I flashed the BIOS of the cards (with Polaris and ATIflash), but the fans don't spin anymore (only a bit when I turn on the MB, but when W10 boots,

Glad to hear you fixed the previous problem.

RX series mostly like that when idle, and will start spinning its fans when reaching specific temps. But in your case, you said "a bit warm" how you know it? Touching? How about temps on software view?

You can try to set your card fan to manual with a specific fan speed, check it Its fan spinning or not.

All about bios, it's clear enough above.

[Pertinax]

Hey, once again, solved the issue, thanks for your advice ! Actually, after the flash, my AMD drivers wouldn't recognize my cards and therefore the fans won't work.
I installed a patch and everything is OK again.

I now overclocked and undervolted the cards with OverdriveNtool, everythings is OK, I manage to push my cards up to 30 - 31 MH/s, I want to chech if my cards do not have any memory issues, hence I install HwINFO64.

But surprise, surprise...

Another problem... 

My computer freezes when I launch the tool... 

Any idea ? I tried to uninstall / reinstall. Runned it in sensors only mode, normal mode... It crashes when it scans ATI scan #93 (??)

[Geraldo]

Any idea ? I tried to uninstall / reinstall. Runned it in sensors only mode, normal mode... It crashes when it scans ATI scan #93 (??)

Are you running HWinfo while mining?
GPU tools such as HWinfo, GPUz, etc., would load and use card resource as shown in information. So, sometimes your rig would be crashed, especially when you open it while mining process on.

[Pertinax]

Nah didn't do that, but I installed the beta tool and runned without hardware check. It works OK now.

So, after soooo many problems I finally managed to make this rig work (I've got a few hashrate drops on some GPUs (30 to 25 very quickly) but it doesn't seem to affect the
hashrate at all and the rig looks stable, perhaps it's a bit too undervolted (90w)))!

Whoever wants to begin mining, I can only say one thing : Be prepared and never lose hope 

Thanks for the help !