ASRock z370 Taichi Power Issue

Message Topic Search Topic Options Post Reply Create New Topic Printable Version Translate Topic

   Board: ASRock z370 Taichi @ UEFI 1.20
CPU: i7 8700K
RAM: Corsair 2x 8GB 3000MHz (supported on the QVL)
Graphics: EVGA GTX 1070
PSU: bequiet! Pure Power10 700W

Hey, it seems like my new z370 Taichi has a power issue. At first I had problems with my 3000MHz Corsair RAM which is on the QVL list. The system would freeze as soon as the RAM was under load when using the XMP1 profile. I fixed this by turning the XMP switch on the board to "On". RAM is working fine now. I can run 10 loops in MemTest64 without any problems. I can also run Small FFTs in Prime for hours without any problems. 
The problem I have now is when I run Large FFTs in Prime which consumes a lot of power the system freezes/crashes. I also tried overclocking my gtx 1070. When I allow the card to use up to 1.09V at 112% power target the system crashes. The UEFI settings are default except for the RAM XMP1 profile

Im not sure if its the board or the PSU. 

Edited by ParadoxX - 26 Nov 2017 at 3:51am

Author	Message Topic Search Topic Options Post Reply Create New Topic Printable Version Translate Topic
ParadoxX Members Profile Send Private Message Find Members Posts Add to Buddy List Newbie Joined: 26 Nov 2017 Status: Offline Points: 6	Post Options Post Reply Quote ParadoxX Report Post Thanks(0) Quote Reply Topic: ASRock z370 Taichi Power Issue Posted: 26 Nov 2017 at 3:50am
	Board: ASRock z370 Taichi @ UEFI 1.20 CPU: i7 8700K RAM: Corsair 2x 8GB 3000MHz (supported on the QVL) Graphics: EVGA GTX 1070 PSU: bequiet! Pure Power10 700W Hey, it seems like my new z370 Taichi has a power issue. At first I had problems with my 3000MHz Corsair RAM which is on the QVL list. The system would freeze as soon as the RAM was under load when using the XMP1 profile. I fixed this by turning the XMP switch on the board to "On". RAM is working fine now. I can run 10 loops in MemTest64 without any problems. I can also run Small FFTs in Prime for hours without any problems. The problem I have now is when I run Large FFTs in Prime which consumes a lot of power the system freezes/crashes. I also tried overclocking my gtx 1070. When I allow the card to use up to 1.09V at 112% power target the system crashes. The UEFI settings are default except for the RAM XMP1 profile Im not sure if its the board or the PSU. Edited by ParadoxX - 26 Nov 2017 at 3:51am

Xaltar Members Profile Send Private Message Find Members Posts Add to Buddy List Moderator Group Joined: 16 May 2015 Location: Europe Status: Offline Points: 22793	Post Options Post Reply Quote Xaltar Report Post Thanks(0) Quote Reply Posted: 26 Nov 2017 at 7:08am
	If you can, try with another PSU. That is a multi rail PSU, it may not be sufficient for your setup. Multi rail PSUs have their watt rating split between their rails so if it has 2 12v rails each rated for 350w then the PSU is rated as 700w. The problem is, if you are drawing more than 350w from either rail you are overloading it. Your unit has one 30a 12v rail and one 36a 12v rail If all your components are using the weaker 30a rail then you are in essence only using a ~350w PSU. Try connecting your GPU to the more powerful 36a rail and your CPU power (8pin) to the 30a rail. The details should be in the unit's manual. If that does not work then you probably need a better PSU. BeQuiet states that it's dual rail PSUs are for office use and entry level gaming (GTX 1050ti for example). Dual rail PSUs are no good for mid to high end gaming systems. This is something that really needs to be made obvious in the marketing material for PSUs...... For most people, 700w sounds like a lot but when it's a dual rail PSU it really isn't. Edited by Xaltar - 26 Nov 2017 at 4:10pm


ParadoxX Members Profile Send Private Message Find Members Posts Add to Buddy List Newbie Joined: 26 Nov 2017 Status: Offline Points: 6	Post Options Post Reply Quote ParadoxX Report Post Thanks(0) Quote Reply Posted: 26 Nov 2017 at 6:08pm
	I just tried the other 2x 6(8) cable which is labeled as "VGA 2" but unfortunately it didnt help much. First I tested Prime Large FFTs and EVGAs KBoost for the GTX 1070 @ 2020 MHz and 1.093V: The system crashed after ~15 minutes. Then I tested Prime Large FFTs without any GPU OC: The system crashed after ~10 minutes. I ran Small FTTs and Blend for 1 hour without any issues. I forgot to mention that im using a Corsair H115i AiO, 1x SDD, 2x HDD, 1x additional USB 3.0 Card which draw additional power. Could this, even without the GPU, be too much for the 12V 30A rail ? I found this on the side of my PSU: https://www.dropbox.com/s/2nr8z4l7cybt774/20171126_121647.jpg?dl=0 The PSU manual states that when the system draws too much power the PSU will turn off. This doesnt happen in my case though. In my case the system freezes. I can still see the screen and all fans are still running. Edit: I just ran Prime Small FFTs and at the same time MemTest64, as soon as the memory test reached "Bitpattern Test" the system froze for 5 seconds but recovered. MemTest64 only still works fine. Edited by ParadoxX - 26 Nov 2017 at 8:47pm

Xaltar Members Profile Send Private Message Find Members Posts Add to Buddy List Moderator Group Joined: 16 May 2015 Location: Europe Status: Offline Points: 22793	Post Options Post Reply Quote Xaltar Report Post Thanks(0) Quote Reply Posted: 26 Nov 2017 at 9:57pm
	Honestly, if I were in your position I would return that PSU and pick up a decent single rail 600w+ model. You are not the first on the forums using that PSU range to report issues. I am not at all impressed with BeQuiet's marketing on that product line, it should be made very clear on the product page and even on the box itself that these are not meant for systems using mid - high end GPUs. It may not be responsible for your issues (it certainly could be) but it wouldn't hurt to replace it with something better suited to your system. If you ever plan to upgrade your GPU or even add a second one, you would most certainly need to replace it, even if it is not the cause of your current issues. For now, until you can try another PSU or replace this one, try to balance the load across the 2 rails. If your GPU uses 2 8pin power connectors then try using one from VGA 1 and one from VGA 2, connect the 36a rail to your motherboard (8pin power) and all other devices to the 30a rail. The motherboard power also provides the 75w power to the PCIe slot your GPU is connected to so you will probably want the better rail going to the board. I had hoped that simply allocating the entire 30a rail to the system would be enough but it does not appear to have helped. A few other questions, on the off chance the PSU is not behind your issues: Are your RAM modules installed in A2 B2? Have you tried clearing CMOS and testing the system at default values (no XMP etc)? Is this a fresh OS install?


Xaltar Members Profile Send Private Message Find Members Posts Add to Buddy List Moderator Group Joined: 16 May 2015 Location: Europe Status: Offline Points: 22793	Post Options Post Reply Quote Xaltar Report Post Thanks(0) Quote Reply Posted: 26 Nov 2017 at 10:00pm
	Also, if the PSU is operating at or near it's threshold (prior to safety shutdown) it can be unstable, providing the system with fluctuating current (ripple). This is one of the most common causes of system instability with regards to power.


ParadoxX Members Profile Send Private Message Find Members Posts Add to Buddy List Newbie Joined: 26 Nov 2017 Status: Offline Points: 6	Post Options Post Reply Quote ParadoxX Report Post Thanks(0) Quote Reply Posted: 26 Nov 2017 at 11:26pm
	Im not 100% sure yet but it seems like the so called "Smart Switch - XMP Switch" is causing this issue. At first I changed the RAM profile from XMP1 to Auto and tested again. The system froze after a few minutes. Then I set the XMP Switch to "OFF" and so far the system is stable. If this is really causing the issue then its quite a problem for me. I dont really like paying for 3000MHz RAM that only works at 2400MHz with different timings and everything. RAM is installed in A2 B2 yes. I also tried a fresh Win10 install without success Edit: I just stopped Prime Large FFTs after 2 hours. The system is stable now but as I said im not happy with this solution. It looks like I have to change my board for a maximus x hero which supports the XMP settings of my RAM. Edited by ParadoxX - 27 Nov 2017 at 12:40am

Xaltar Members Profile Send Private Message Find Members Posts Add to Buddy List Moderator Group Joined: 16 May 2015 Location: Europe Status: Offline Points: 22793	Post Options Post Reply Quote Xaltar Report Post Thanks(0) Quote Reply Posted: 27 Nov 2017 at 5:06am
	The switch being off does not disable XMP, it being on enables it automatically without having to set it in the UEFI. Set the switch to off then enable XMP from within the UEFI. See if that works for you. If not, check to make sure XMP is setting your voltages and sub timings correctly for your kit. Sometimes they are not, especially on new model boards (early Z170, Z270 etc). Usually these issues get resolved with BIOS updates pretty quickly. The fact that you are stable at 2400 is a good sign, it means the kit isn't outright incompatible. I wouldn't swap boards for a memory issue. You have absolutely no way of knowing if it is a problem with the board or the kit. More often than not the issue is either a problem with compatibility (corrected by a BIOS update) in which case waiting a little while for an update will save you a hassle or, worse, the RAM itself is defective in which case you wasted a lot of time and effort for nothing. I would say this regardless of the board you have. I am guessing your board will work fine with XMP enabled in the UEFI but set to off via the switch. If not, try contacting Tech Support directly before going as far as returning the board, they will probably be able to help you with a BIOS update or particular UEFI settings to get things stable at 3000mhz. Good luck. Edited by Xaltar - 27 Nov 2017 at 5:09am


ParadoxX Members Profile Send Private Message Find Members Posts Add to Buddy List Newbie Joined: 26 Nov 2017 Status: Offline Points: 6	Post Options Post Reply Quote ParadoxX Report Post Thanks(0) Quote Reply Posted: 27 Nov 2017 at 7:40am
	I mentioned in my very first post that the RAM with xmp profile wasnt stable without setting the switch to "ON". I couldnt even run 2 loops of MemTest64. UEFI settings are exactly the ones for my RAM so thats not the issue either. I used this RAM kit in a different system for 1 year now and never had any problems using the xmp profile. So I highly doubt that the RAM is faulty. A UEFI update for my RAM could take a while and im not really sure whether they would do it at all. And if they dont/cant fix it im stuck with this board. Right now I can just return it and get my money back but I wont be able to do that in a few weeks/months/years. Im not a fan of gambling. Edited by ParadoxX - 27 Nov 2017 at 7:49am

Xaltar Members Profile Send Private Message Find Members Posts Add to Buddy List Moderator Group Joined: 16 May 2015 Location: Europe Status: Offline Points: 22793	Post Options Post Reply Quote Xaltar Report Post Thanks(0) Quote Reply Posted: 27 Nov 2017 at 2:18pm
	I see, so the RAM was from a previous system. You can try contact Tech Support, make sure you select your region by clicking the crosshatched globe at the top right of the (ASRock) page. This will ensure your ticket goes to the nearest tech department to you. If you do contact them, include your RAM model number, board number and a brief description of the issue, you can also link them to this thread. Best of luck to you, sorry I couldn't be of more assistance.