Aorus Xtreme TRX40 completely dead after shutting down safely

Joined
Oct 23, 2018
Messages
1,011
I was using my system and everything was working normally. I shut it down in order to do some work on it (GPU swaps), but then decided not to do that work after all and tried to turn the machine back on it. Now, it acts completely dead.

There are no LEDs on the board, the fans don’t twitch, no relays click, etc. I’ve tried using the Power button on the board with the same result. It’s like there isn’t power connected at all. I tried unplugging the ATX plug from the mobo and put the jumper plug on it for powering pumps and stuff when filling loops. When I do this, everything responds like normal and all accessories power up. When I plug back in to the board, everything is dead again.

I tried swapping PSUs just in case and get the same result.

Any ideas? What else should I try?
 
Oh, and Gigabyte says it takes 3-7 business days for each message cycle from their tech support. That is stunningly garbage service, especially on an $850 board.
 
I was using my system and everything was working normally. I shut it down in order to do some work on it (GPU swaps), but then decided not to do that work after all and tried to turn the machine back on it. Now, it acts completely dead.

There are no LEDs on the board, the fans don’t twitch, no relays click, etc. I’ve tried using the Power button on the board with the same result. It’s like there isn’t power connected at all. I tried unplugging the ATX plug from the mobo and put the jumper plug on it for powering pumps and stuff when filling loops. When I do this, everything responds like normal and all accessories power up. When I plug back in to the board, everything is dead again.

I tried swapping PSUs just in case and get the same result.

Any ideas? What else should I try?

Is there a clear CMOS button? You should try depressing that for about 8 seconds or so and then releasing. You can also try powering up the system with only one DIMM installed in the furthest slot from the CPU. In your case, the one towards the ATX power connector.
 
Pull the cmos battery, disconnect power, and hold the power button for a while. Reinstall battery and plug in power, should be back up and running.

Edit: you may be the first TR user to have a GB board which exhibits this issue (on the forum, at least).
 
Last edited:
Is there a clear CMOS button? You should try depressing that for about 8 seconds or so and then releasing. You can also try powering up the system with only one DIMM installed in the furthest slot from the CPU. In your case, the one towards the ATX power connector.
Tried both of these just now. No change :-\
Pull the cmos battery, disconnect power, and hold the power button for a while. Reinstall battery and plug in power, should be back up and running.

Edit: you may be the first TR user to have a GB board which exhibits this issue (on the forum, at least).
Will try this in a little bit. Looks like the motherboard has to come out of the system for this because there's a cover that has to be removed and it is held in with a screw from the back of the board.
 
Test the case power switch by connecting the reset switch to the power switch pins. Nevermind I missed the onboard power button test.
 
Pull the board out and run it on a table. See if that gets anywhere.
 
Oh, and Gigabyte says it takes 3-7 business days for each message cycle from their tech support. That is stunningly garbage service, especially on an $850 board.
And I thought that ASUS service sucked big time.
 
Pull the cmos battery, disconnect power, and hold the power button for a while. Reinstall battery and plug in power, should be back up and running.

Edit: you may be the first TR user to have a GB board which exhibits this issue (on the forum, at least).
Winner winner! Pulling the CMOS battery fixed it! Thank you!

It took a literal 23 screws in order to move all of the different covers and covers-on-covers-covering-covers in order to gain access to the battery. I wish I were making that up. And, true to Gigabyte's reputation for taking iffy designs and then manufacturing them poorly, only four of those screws were so much as finger tight (and likely only because those four are spring loaded heatsink screws).

Happy to have it working again, but wow. What a pile of garbage. The next time this happens, the board is coming out and I'm switching back to Intel. Looks like it's a recurring problem for those who have encountered it, too.

Still no response from Gigabyte support, of course. It's going to take so long to work it out with them that I might just keep the conversation alive so that there's an RMA ready for the next time this happens.
 
Oh, and Gigabyte says it takes 3-7 business days for each message cycle from their tech support. That is stunningly garbage service, especially on an $850 board.
It's the same service whether you buy the $1000 board or the $100. Just be glad it isn't Asus, they would probably claim you damaged it and offer to repair/replace for $999.
 
Still waiting for response from EVGA for issue with my SR-3 since december. I submitted a custom bios request for my Gigabyte board a week or so ago, wonder if they are on vacation lol.
 
Still waiting for response from EVGA for issue with my SR-3 since december. I submitted a custom bios request for my Gigabyte board a week or so ago, wonder if they are on vacation lol.
From my recent experience you are better off calling evga then dealing through their site/email if you haven't yet.
 
Side note for those still considering buying anything from Gigabyte: I still haven't heard back from them.
No experience with the gig. 10 years AMD/Radeon/Asus/AMD ram and SSD/G Skill Trident Z/Sabrent GEN4 NVMe/Watercool/EK Not one failure. Boots out of the box. Still works when it's so old that I junk it for newer TECH! Guys refuse to use a static mat and a static strap.
 
Winner winner! Pulling the CMOS battery fixed it! Thank you!

It took a literal 23 screws in order to move all of the different covers and covers-on-covers-covering-covers in order to gain access to the battery. I wish I were making that up. And, true to Gigabyte's reputation for taking iffy designs and then manufacturing them poorly, only four of those screws were so much as finger tight (and likely only because those four are spring loaded heatsink screws).

Happy to have it working again, but wow. What a pile of garbage. The next time this happens, the board is coming out and I'm switching back to Intel. Looks like it's a recurring problem for those who have encountered it, too.

Still no response from Gigabyte support, of course. It's going to take so long to work it out with them that I might just keep the conversation alive so that there's an RMA ready for the next time this happens.
How is a shitty Gigabyte board an AMD problem? C'mon mayne! What next, you'll blame win10 updates on Intel and switch to ARM?
 
Still waiting for response from EVGA for issue with my SR-3 since december. I submitted a custom bios request for my Gigabyte board a week or so ago, wonder if they are on vacation lol.
actually thats possible with chinese new year.. they take as much as a month off in asia and taiwan.
 
How is a shitty Gigabyte board an AMD problem? C'mon mayne! What next, you'll blame win10 updates on Intel and switch to ARM?
Motherboard makers work from the reference design that the chip maker gives them, and the chip makers sets the quality standard. I'm also running into some AMD-specific issues as well. For example, there are only two Threadripper motherboards on the market that can fit 4 GPUs, whereas there's a bunch of them on the Intel side. Plus, with Threadripper, those 4 GPUs require either case mods or a 9-slot case in order to fit because the PCI slots are all offset one lower than spec. This isn't a problem on the Intel side.

AMD rightfully dominates the budget gaming market, but this isn't a budget gaming build.
 
Motherboard makers work from the reference design that the chip maker gives them, and the chip makers sets the quality standard. I'm also running into some AMD-specific issues as well. For example, there are only two Threadripper motherboards on the market that can fit 4 GPUs, whereas there's a bunch of them on the Intel side. Plus, with Threadripper, those 4 GPUs require either case mods or a 9-slot case in order to fit because the PCI slots are all offset one lower than spec. This isn't a problem on the Intel side.

AMD rightfully dominates the budget gaming market, but this isn't a budget gaming build.
And some motherboard makers "fudge" that reference design or spec, for various reasons. Have to decide if that's on the reference design not being strict enough, or on the maker for "fudging" too far (for whatever reason). I've been plenty frustrated with AMD in the past for various things that are potentially their fault (not making VRM requirements strict enough for Ryzen, for instance, allowing low-end VRM designs out is arguably ~partially~ their fault - not saying it is or isn't, but you CAN argue that with a straight face), but at the same time, I've also had plenty of boards where it was clearly the vendor that got too "creative" interpreting the spec (Gigabyte NF3 boards put the ram slots too far out for the spec - which resulted in all sorts of fun instability except on certain ram, or if you got REALLY lucky and had a tolerant CPU).

Given Gigabyte's current issues with AMD Bios, but everyone else seeming to have no issues, I'd argue this is more on Gigabyte than on AMD... but that's a personal opinion (I also personally tend to avoid the Gigabyte + AMD combo, after 3 times running into them being super creative on how they interpret the specs - but I gleefully buy them for Intel builds all day long).
 
Motherboard makers work from the reference design that the chip maker gives them, and the chip makers sets the quality standard. I'm also running into some AMD-specific issues as well. For example, there are only two Threadripper motherboards on the market that can fit 4 GPUs, whereas there's a bunch of them on the Intel side. Plus, with Threadripper, those 4 GPUs require either case mods or a 9-slot case in order to fit because the PCI slots are all offset one lower than spec. This isn't a problem on the Intel side.

AMD rightfully dominates the budget gaming market, but this isn't a budget gaming build.

Well this seems alot irrelevant for the situation at hand. Talk about mobing the goal post
is that really how you think ?
"Intel has motherboard with better spacing for multi gpu so it is AMD faults my mobo went bad" , or can you see that it was irrelevant to bring up ?

also how did you exclude bad quality of mobo for it beeing AMD designs ?
seems li ke you are relly trying to take on some giant blinders to get to you specified pre-biased opinion
 
Well this seems alot irrelevant for the situation at hand. Talk about mobing the goal post
is that really how you think ?
"Intel has motherboard with better spacing for multi gpu so it is AMD faults my mobo went bad" , or can you see that it was irrelevant to bring up ?

also how did you exclude bad quality of mobo for it beeing AMD designs ?
seems li ke you are relly trying to take on some giant blinders to get to you specified pre-biased opinion

The simple fact is that AMD's design choices make it so that a 9-slot case is required for the same hardware that Intel's design choices allow with 8. The simple fact is that AMD's design choices led to only two motherboards having capacity for 4 GPUs whereas Intel's design choices led to a relative plethora of options. I'm sorry this hurts your feelings.

I'm not saying the Intel side is perfect either, so you can dry your eyes. Part of the impetus for switching to Threadripper is that my X299 Asus board had a known design fault wherein the 4th x16 slot can't actually be run above x4.
 
The simple fact is that AMD's design choices make it so that a 9-slot case is required for the same hardware that Intel's design choices allow with 8. The simple fact is that AMD's design choices led to only two motherboards having capacity for 4 GPUs whereas Intel's design choices led to a relative plethora of options. I'm sorry this hurts your feelings.

I'm not saying the Intel side is perfect either, so you can dry your eyes. Part of the impetus for switching to Threadripper is that my X299 Asus board had a known design fault wherein the 4th x16 slot can't actually be run above x4.

Again how is this relevant to you motherboard not working?
You are not hurting my feeling. You are hurting my brain with you illogical though patterns.
 
The simple fact is that AMD's design choices make it so that a 9-slot case is required for the same hardware that Intel's design choices allow with 8. The simple fact is that AMD's design choices led to only two motherboards having capacity for 4 GPUs whereas Intel's design choices led to a relative plethora of options. I'm sorry this hurts your feelings.

I'm not saying the Intel side is perfect either, so you can dry your eyes. Part of the impetus for switching to Threadripper is that my X299 Asus board had a known design fault wherein the 4th x16 slot can't actually be run above x4.
This has nothing to do with amd, you looked at intel boards lately? lol Neither amd or intel give 2 shits about what form factor their partners make the motherboards or how many slots they have.
 
Suspect that’s due to the size of the physical CPU?
If it were an intentional choice rather than an oversight, the reason would have been to protect their higher margin Epyc line. The WRX80 boards turn the CPU 90deg to resolve this (it also matches the Epyc orientation). Those boards, of course, are priced the same as what Supermicro sells their Epyc boards for with CPUs already in them.
Again how is this relevant to you motherboard not working?
You are not hurting my feeling. You are hurting my brain with you illogical though patterns.
It's logical if you're smart enough to think about more than one thing at a time. There are a number of issues with the current setup, and having a motherboard that might require a full day of downtime every quarter just to turn it on is the final straw. If When the board dies again, I'm not going to keep making the same mistake. I'm going to switch platforms so that I can make new ones.
 
Most WRX boards are EATX or SSI-EEB though I think... been a bit since I looked
 
If it were an intentional choice rather than an oversight, the reason would have been to protect their higher margin Epyc line. The WRX80 boards turn the CPU 90deg to resolve this (it also matches the Epyc orientation). Those boards, of course, are priced the same as what Supermicro sells their Epyc boards for with CPUs already in them.

It's logical if you're smart enough to think about more than one thing at a time. There are a number of issues with the current setup, and having a motherboard that might require a full day of downtime every quarter just to turn it on is the final straw. If When the board dies again, I'm not going to keep making the same mistake. I'm going to switch platforms so that I can make new ones.

You have yet do show link between one motherboard have room to 4 gpu has anything to do with you motherboard not starting.
Please let me know how a motherbaord with room for more GPU would have changed the outcome if your motherboard not starting AND would results in it beeigns AMD fault you motherboard would not start.
 
Rather than crapping over someones personal decision because you don't like it, how about proposing an actual solution?
 
Rather than crapping over someones personal decision because you don't like it, how about proposing an actual solution?
If they didn't reseat the cpu they could try that and memory too.
 
If they didn't reseat the cpu they could try that and memory too.
If that didn't help the guys in the other thread, I doubt it'll help here. Even if it did, the issue can take months to years before it recurs, or it could happen tomorrow. There appears to be no rhyme or reason, other than overclocks and fiddling with any bios setting can make it happen more often...maybe.
Edit: you may be the first TR user to have a GB board which exhibits this issue (on the forum, at least).
 
I mean, it's not a bad idea, but his system had been stable until he shut it down. If it were the cpu or memory, I would expect some OS errors at least– especially since whatever it is caused the system to act completely dead.

If you haven't read that other thread, you probably should. They've tried just about everything trying to figure this thing out.
 
Small update:
Unrelated to the OP, the mobo went up in smoke all by itself. Time to see if they will RMA it. I guess it's true that Gigabyte puts the "hot" into "hot garbage."
 
  • Like
Reactions: Nobu
like this
Sorry, I didn't see this before I created a post.

Hello all,

I'm new to this site, so I'm sorry if this isn't the right place to post this. I left my system on all night (I would normally power down) and when I came into my office to start work, I noticed my RGB lights on, but my screen saver not working. I pressed the spacebar to wake my computer up and nothing. I did a hard power reset and I could hear my power supply click. I have another power supply that I tried with the same results. I started unplugging cards to troubleshoot. The only way I could get my system to boot was to leave one of the 8pin CPU power connections on the motherboard unplugged. My CMOS battery is hidden and I would have to drain my loop and take my system most of the away apart to access it. I've opened a case with Gigabyte, but I'm sure it will be a dead end. Can some one please point me in right direction to fixing this issue? Again I have my system up and running now with only one 8pin power connection. My system has been running fine for the past 11 months with no issues until today.

System Specs
Gigabyte TRX40 AORUS XTREME Rev 1.0
Threadripper 3960x
Bios F4p
128GB Memory
RTX 5000 GPU
EVGA 1300 G2 PS (using my Corsair AX1500i now)
Windows 10 64Bit
Running since 5/20 just fine

Thanks
 
For me, removing the battery was the only thing that brought it back to life (before it went up in smoke). Even with all of my QDCs allowing me to do this without draining and loops, it was still a real pain to have to completely remove the board from my system in order to do this.
 
For me, removing the battery was the only thing that brought it back to life (before it went up in smoke). Even with all of my QDCs allowing me to do this without draining and loops, it was still a real pain to have to completely remove the board from my system in order to do this.
Yeah, I can see that being an issue. I don't want to RMA my board as god only knows how long it will take to get another. Still waiting on Gigabyte to reply to my ticket.
 
I think my recommendation would be to buy another motherboard from a different company and switch your machine over to that. RMA the broken Gigabyte board and then sell the replacement when it arrives. Any replacement they send you is just going to have the same issue (it's a known bug which hits multiple models) and, if you're especially lucky, it will die a smoky death just a few months after you receive it anyway.
 
Back
Top