Z170 ProM4s build help

Excuse me bios is version 2.60 sorry

I'm giving up on this board. I will start an RMA with ASRock. Thanks for all the input from all. Wish me luck on the next try

invasmani wrote: I'd also advice checking system agent voltage if that's set too low for a given CPU clock frequency/BCLK frequency not sure which is the cause it can fail to post or cause post failure messages to do with the boot drive being inaccessible. Generally speaking I've found SA needs to be 1.05v minimum or higher in typically up to around 1.15v to avoid such issues. I have no idea in the case of M.2 drives though that might require even higher SA in conjunction with other SATA drives is my initial suspicion based on what I've seen at least. I'd try raising SA to 1.25v and seeing if it helps and from there lowering it to the point where the boot drive inaccessible errors begin to occur and raise it up anywhere from 0.05v to 0.15v above that.

System Agent voltage (VCCSA) being to low causing POST failures and drive recognition issues?

I'm glad you brought this up, since it might be a factor in this issue, but possibly not directly for the reason you explained. Your idea is definitely worth testing IMO.

I wonder what the default/Auto setting in the Z170M Pro4S' UEFI for VCCSA voltage provides.

I've never had the problem with my drives that the others in this thread have, with the Auto VCCSA setting in the UEFI resulting in 1.072V for VCCSA. I had three M.2 SSDs in my Z170 Extreme7+ board with that VCCSA voltage.

VCCSA is related to memory stability in Skylake systems. We have seen some strange issues that were fixed by changing to memory with better compatibility. Is this another example of that?

Increasing VCCSA is done when over clocking memory, and the XMP profiles of high speed DDR4 memory (3000+) set VCCSA and VCCIO voltages to values significantly higher than the defaults. My G.SKILL DDR4 3200 memory's XMP profile set VCCSA to 1.250V, and VCCIO to 1.150V. I found that those rather high voltages could be reduced significantly, without loss of memory stability or boot problems, with the memory at its highest rated speed. But that depends upon the model of memory being used, and some DDR4 memory designed for X99 systems does not always work well with Skylake systems.

If anyone tries to increase the VCCSA voltage to the value suggested above, you might find the voltage value suddenly changing to a red colored font. That is normally done when a voltage setting value is above the range suggested by Intel. The voltages in red are not automatically dangerous, but there are limits to any voltage, beyond which damage can occur.

That is what I meant by increasing VCCSA not directly being a fix for this issue. That is, increasing VCCSA does not directly benefit the M.2 and SATA IO system itself (part of the Z170 PCH), but fixing any memory stability problems on these boards has also cured other strange problems like this one.

Another hint about this is, I reduced the PCH +1.0 voltage from the Auto setting, which was 1.016V by default. I wanted to see if that would affect the PCH temperature, and it did. The PCH voltage is the main voltage for the Z170 PCH chipset, which provides all the M.2 and Intel SATA ports. The default PCH +1.0 voltage would be another worth checking on the Z170M Pro4S.

Skylake is a rather different system from Intel compared to their recent past. The voltage regulators (FIVR) on the CPU die itself, found in Haswell and Broadwell processors, is no longer used in Skylake. Skylake also is the first non-HEDT platform to use DDR4 memory, and is the first Intel system to use PCIe 3.0 lanes in the PCH/Chipset, for use with the PCIe (M.2) and SATA storage interfaces. AHCI and NVMe use system memory to store and communicate IO requests to disk drives. Put this all together, and we can see the potential for new issues like this one.

The one time I don't suggest trying another PSU......