Jump to content

How to stabilize PCIe slot power?

orangecat

So to make a very long story short I have an AMD RX 460 that does not have an external PCIe power connector so it gets all of it's power directly from the PCIe slot. The problem I'm having is that if I load down the GPU with a 3D load at all an invisible counter starts ticking and eventually the GPU will stop working. Now before someone claims that the card is faulty or defective I can say with absolute certainty that it's not. I've tested this GPU in 3 seperate computers and it works flawlessly in one of them. The issue I'm having is clearly related to PCIe power but I'm not sure how to go about alleviating it. I've tested this card in a system with a Z68 board and it had problems. I also tested this card in a Z77 board and again it had problems. Lastly I tested this GPU in an X570 board and it works flawlessly. I'm 99.9% sure the issue is related to PCie power not being up to snuff but I don't know how to find a solution. I tested the GPU in the X570 board and with the card going full tilt with an overclock HWinfo was reading the PCIe input voltage as low as 11v. the voltage drop under furmark load is very high but the card never fails in that system. In the other 2 Intel boards the card will work fine until it needs to up the clocks and run a load. By the way I should mention I'm not overclocking the card I just did in the X570 system to make a worst case scenario and it passed with flying colours. I also tested the GPU in Windows, Ubuntu and MacOS and the problem persists on all the Intel boards I've tested. I think the issue is that when the card goes from having an average power draw to suddenly having a moment of high power draw like when a stutter occurs the voltage on the PCIe slot drops or spikes in a way that causes the GPU to lock up or something. The actual GPU and all that works fine but power delivery isn't good on those boards. I've been thinking about trying to find something that can maybe help stabilize the voltage but I'm not sure what that would be. Maybe a 24 pin adapter with some extra capacitors on it or maybe thing similar for the PCIe slot. that being said I don't know if such a thing even exists. I'm also pretty sure it's not possible to hack on a PCIe power connector as this GPU in particular wasn't built with one.

Anyone got any ideas?

Also downclocking the card to avoid this issue isn't an option. If I have to do that I would rather just buy a new GPU as that isn't what I would consider a proper solution.

Link to comment
Share on other sites

Link to post
Share on other sites

How could you go through so much so much thought and still not suspect the PSU to be the culprit? It's the one powering the PCIe slot in the first place.

CPU: i7-2600K 4751MHz 1.44V (software) --> 1.47V at the back of the socket Motherboard: Asrock Z77 Extreme4 (BCLK: 103.3MHz) CPU Cooler: Noctua NH-D15 RAM: Adata XPG 2x8GB DDR3 (XMP: 2133MHz 10-11-11-30 CR2, custom: 2203MHz 10-11-10-26 CR1 tRFC:230 tREFI:14000) GPU: Asus GTX 1070 Dual (Super Jetstream vbios, +70(2025-2088MHz)/+400(8.8Gbps)) SSD: Samsung 840 Pro 256GB (main boot drive), Transcend SSD370 128GB PSU: Seasonic X-660 80+ Gold Case: Antec P110 Silent, 5 intakes 1 exhaust Monitor: AOC G2460PF 1080p 144Hz (150Hz max w/ DP, 121Hz max w/ HDMI) TN panel Keyboard: Logitech G610 Orion (Cherry MX Blue) with SteelSeries Apex M260 keycaps Mouse: BenQ Zowie FK1

 

Model: HP Omen 17 17-an110ca CPU: i7-8750H (0.125V core & cache, 50mV SA undervolt) GPU: GTX 1060 6GB Mobile (+80/+450, 1650MHz~1750MHz 0.78V~0.85V) RAM: 8+8GB DDR4-2400 18-17-17-39 2T Storage: HP EX920 1TB PCIe x4 M.2 SSD + Crucial MX500 1TB 2.5" SATA SSD, 128GB Toshiba PCIe x2 M.2 SSD (KBG30ZMV128G) gone cooking externally, 1TB Seagate 7200RPM 2.5" HDD (ST1000LM049-2GH172) left outside Monitor: 1080p 126Hz IPS G-sync

 

Desktop benching:

Cinebench R15 Single thread:168 Multi-thread: 833 

SuperPi (v1.5 from Techpowerup, PI value output) 16K: 0.100s 1M: 8.255s 32M: 7m 45.93s

Link to comment
Share on other sites

Link to post
Share on other sites

12 minutes ago, Jurrunio said:

How could you go through so much so much thought and still not suspect the PSU to be the culprit? It's the one powering the PCIe slot in the first place.

I've tested in multiple systems with multiple power supplies that work flawlessly with other GPUs. I fail to see how a power supply would have a hardware specific problem like this considering I've used more power hungrey cards in those systems that had problems. For example the Z68 system currently has an Nvidia GTX 1650 with external PCie power. If there was something wrong with the 12v rail surely it would be obvious with a card that puts more load on the rail. I've also tested that same GPU in the Z77 system and it works fine too. Both of those systems work perfectly fine and I know the PSU's work.

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, Kilobytez95 said:

Anyone got any ideas?

Also downclocking the card to avoid this issue isn't an option. If I have to do that I would rather just buy a new GPU as that isn't what I would consider a proper solution.

But have you tried undervolting/underclocking, just to see?

This might seem nuts, but you could try one of these Totally Not Suspect AliExpress cards...
https://www.aliexpress.com/item/4000770345682.html

Their alleged use is for noise filtering, but the way that's achieved is... a load of capacitors on the PCIe power lines. You can get DIMM module versions too. No clue if they work, or are even safe. Maybe convince @jakkuh_t to be a crash-test dummy?

Link to comment
Share on other sites

Link to post
Share on other sites

5 hours ago, TehDwonz said:

But have you tried undervolting/underclocking, just to see?

This might seem nuts, but you could try one of these Totally Not Suspect AliExpress cards...
https://www.aliexpress.com/item/4000770345682.html

Their alleged use is for noise filtering, but the way that's achieved is... a load of capacitors on the PCIe power lines. You can get DIMM module versions too. No clue if they work, or are even safe. Maybe convince @jakkuh_t to be a crash-test dummy?

Undervolting won't solve the issue and even if it did that's not a solution. I'm not sure if that card you linked is really going to do much as from what I can tell some motherboards have seperate on board voltage regulation per slot or are at least not direct;y getting power from the same place. On my X570 board if I load down my GTX 970 the PCIe voltage drops a bit but the voltage in on the RX 460 doesn't change so it leads me to think they are not directly connected. I'm also kinda skeptical if that card you linked would even do anything for it's advertised use case. Seems like one of those audiophile snake oil things for people with high end internal sound cards. Ideally what I would need is one of those PCIe mining card adapters that goes from 16x to 16x with external power input for the slot but sadly they all seem to be 1x only. Also I'm not sure if @jakkuh_t would waste his time with this lol. That being said I'll definitely keep that link in mind just incase I really can't find anything else.

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×