New build with two distinct motherboards, fails to boot after resetting CMOS unless...
New build with two distinct motherboards, fails to boot after resetting CMOS unless...
I'll simplify this for you. Here are the main issues:
- Motherboard won't power on after sleep.
- Hard resets (like disabling Hyper-Threading or XMP) don’t work.
- Power button clicks but needs long press (1.5–2 sec) to boot.
- Best success was disconnecting the GoXLR Mini – then it boots with a long press.
- Even after removal, it still fails to boot from sleep or BIOS changes.
- Full setup: specific hardware and peripherals.
- Tried updates, different ports, resets, voltage changes, but nothing helped.
- Peripherals were fine before, now they cause problems.
- CMOS cleared, single tap works sometimes, but Windows 11 won’t launch.
- No TPM access in BIOS settings, which is confusing.
- Case airflow isn’t working—needs cleaning every few weeks.
- Support replies were unhelpful or misleading.
- Still unsure if hardware fault or just setup quirks.
The only possibilities that come to mind are VCCIN and the PLL voltages. They’re mainly needed for memory overclocks above 7000 and for fixing bugs while using LN2. They shouldn’t interfere with normal booting at room temperature. If you’re concerned, raising VCCIN can improve low-temperature booting but it also limits maximum memory speeds. I’d prefer to keep it under 2V. The PLL voltages have an optimal range that changes by chip, so you’d need to test each one until it works. Ideally, I’d aim for values below ~1.1V, though most boards can’t reach that and the Strix board likely won’t allow it either (my Unify-X tops out at 1.135V for all models). The main consideration would be which BIOS version to try first—new releases don’t always guarantee improvement. I recall reading about Maximus threads where a Z690 Apex had several problematic BIOS versions, so revisiting an older release might help. Fast Boot has also shown odd behavior where the system refuses to restart, though that usually doesn’t stop it from starting at all. Another point worth checking is the CMOS battery voltage; some boards are very strict about maintaining a specific level and can behave strangely if it drops (like Luumi’s SR-3 Dark).
Remove the board from the case, leaving only the CPU, RAM, boot drive, and power connected. Begin by shorting the power pins using a screwdriver with jumper wires. Observe its behavior. Occasionally, the case or standoffs can interfere with communication. This process also removes any peripheral problems. If it starts working, gradually reconnect components one by one until the issue reappears.
Regarding BIOS updates, I started with the version that came with the board (7xx), then upgraded to 8xx. After that, another update was released while I was working on the first board "904," which is currently the latest. All models are showing identical issues. I also attempted to turn off fast boot in both BIOS and Windows power plan settings. Concerning the CMOS, it seems the battery might be hidden inside the heatsinks—it's not out of the question to access, but I prefer preserving the motherboard if I decide to report it as defective or return it. It’s interesting you mentioned the Apex Z690. I bought one last year, and thankfully while I waited for the 12900KS, I noticed my Apex batch matched the one everyone was experiencing problems with. I returned it right away, canceled other orders, and chose to wait for the platform to stabilize a bit.
I purchased it from the same batch. The seller claimed it was a solid model, but he didn’t tell the truth. It performed worse than the Z790 Aorus Elite for memory overclocking, even with the 13th generation controller. It wasn’t the worst among similar units; I managed to reach 7600MT/s, but stability dropped below 6600 with M die and 6600 with A die—while the Unify-X handled 7000 with M die and 7800 with A die (the latest BIOS seems promising). The BIOS offered numerous voltage options across four submenus, which is a recurring issue with other Maximus boards since Z170. This one felt particularly chaotic, with hidden settings scattered everywhere. Overall, it was a frustrating build. The only positive was the Turbo VCore utility, which gave slightly better stability than Dragon Power on the Unify. You didn’t lose much, but it was definitely a hassle. If you don’t want to risk the CMOS, consider returning it and buying another board. The Z790 Edge is well-reviewed and more affordable.
I was lucky to avoid another hit with that apex and the overall setup. Still, I ended up taking another one! It’s interesting how, after returning the first Z790I Strix, I felt really upset with Asus’s support and their forums. While I compared other ITX board options from various makers, I noticed the same edge and my mind decided to give Asus another shot. To be honest, I’m not too sure the board was the main problem. As I mentioned, how did I manage to get two boards with manufacturing dates separated by four months that all had the same issue? I didn’t encounter any problems like mine, but some customer reviews mentioned a DOA—though that term can mean many things. Could you explain more about PLL voltages? I’ve dealt with bios tweaks before, but PLLs are something I never worked with or adjusted.
There are five distinct PLL configurations, and according to what I remember, ASUS boards allow you to configure boot and runtime versions for each (though this might apply only to the Maximus series). One setting applies to the entire CPU, another to the Ring, one to the System Agent, another to the E cores, and a third to the memory controller. In very low temperatures, raising these values can aid system startup (this is typically what LN2 mode does), whereas at normal temperatures each of these helps maintain stable memory performance above 7000MT/s. At regular temperature, my 13700K prefers 1.035V on the CPU, Ring, and SA, while the MC runs at 0.990V. I left the E core setting in auto mode since it offers little benefit unless you're pushing those cores to their limits, which I don’t want to do. These settings really started gaining importance around the 12th generation, assisting Ring, CPU, and memory overclocking by slightly lowering the required core voltage, but they make a noticeable difference on the 13th gen as it can determine whether DDR5 6800 or 7600 becomes functional. I’m not sure what you need to configure, and I haven’t used a 13th-gen chip in LN2 without checking the recommended settings for each model—every board is unique, so experimentation might be necessary. Remember to keep voltages under about 1.1V, and remember these values are set in 0.015V steps according to Intel standards; don’t try to adjust them in tiny increments like my Apex does.
Just a note on this discussion: After initial testing, the technician confirmed the power supply was functioning properly. Another technician raised concerns about the 5V rail, suggesting it could handle up to 1000W without failing—but noted that the current supply might be too high or too low, leading to boot issues. He continued testing and discovered some compatibility problems between the PSU and the board or peripherals. Eventually, he requested further checks without charge.
I clarified that I’d allow him to test it more if Amazon refused a return. Amazon’s support had already advised me to return the motherboard for a full refund, but I explained to them the situation. They suggested an exception for my Thermaltake power supply, which I received plus a $50 credit—something very kind.
On the other hand, the first technician who checked with the old motherboard ended up costing the seller the full value of that board. Since it was marked as damaged, they couldn’t ship it back and now have an open used board. Now I own an Asus Tough 1000W Atx3.0 that’s fully operational—no boot failures, peripherals stay connected, and sleep transitions are smooth. Thank you, RONOTHAN##, for your help!