ASUS Maximus Hero cold boot

[Leoplate25]

Hi, i have a Maximus Hero XI WiFi. When i have the PC unplugged and then i plug it and, obviouly, i do a cold boot it powers up, then shuts down and then powers up again. I read in some forums with no luck till today. Setting the XMP profile do that, but if i change the CPU VCCIO voltage and the CPU System Agent voltage it starts normally, even in cold boots. The question is: what values should i enter for these 2 voltages? Is 1.1v ok? They has really high values, near 1.250v. I'm using a 9600k and i'm not OCing it for the moment. Thanks!!!

This is my setup:

Intel i5-9600k at stock (not OCing for the moment)
ASUS Maximus Hero XI WiFi
Corsair Vengeance LED White 2x8gb 3000mhz (XMP enabled but VCCIO and SA too high)
Corsair H100i v2
EVGA GTX 1070 TI SC
BitFenix Whisper M 750w

[Leoplate25]

mdzcpa wrote:
You should double check that setting your SA/IO voltages to 1.1v cured your double boot issue when power is lost to the motherboard. This would be the first report of this. Power loss require the M11 series of boards to cycle ONE time. If memory is not tuned, it may require a second boot cycle. So setting the right memory settings and SA/IO can reduce the number of boot cycles needed. I have never heard, however, that this will eliminate the reboot cycle completely.

Yes, if you set your ram to 2666, that is what it will run at. It overrides the XMP setting.

For example, I have 4200mhz. But I like an extra margin of stability so I set XMP, then adjust the timing to 4000mhz, and I also manually tune SA/IO to 1.215/1.1v respectively.

I doubled check that. I turn off the pc, unplug from the wall, wait a couple of minutes and voilá, issue solved. I'm running vccio and sa at 1.1v, ram at 3000mhz and i5 at stock settings, 2 or 3 days passed and the problem disappeared. In fact, i'm gonna turn off the pc now, unplug it from the wall, wait 5 minutes and turn it on again and see what happens. Do you think 1.1v is enough for my settings?

[Leoplate25]

mdzcpa wrote:
You should double check that setting your SA/IO voltages to 1.1v cured your double boot issue when power is lost to the motherboard. This would be the first report of this. Power loss require the M11 series of boards to cycle ONE time. If memory is not tuned, it may require a second boot cycle. So setting the right memory settings and SA/IO can reduce the number of boot cycles needed. I have never heard, however, that this will eliminate the reboot cycle completely.

Yes, if you set your ram to 2666, that is what it will run at. It overrides the XMP setting.

For example, I have 4200mhz. But I like an extra margin of stability so I set XMP, then adjust the timing to 4000mhz, and I also manually tune SA/IO to 1.215/1.1v respectively.

I did what i said and it cold booted fine. Don't know how or why, but it booted fine...

[Leoplate25]

leerock360 wrote:
I have the maximus extreme and a 9900k. I left both voltages on auto and my pc would double boot *each time from cold boot. I lowered both values to 1.15v for vccio and the system agent voltage and now my pc no longer double boots from a cold boot. So happy. I have a normal pc again. When i left on auto both were on 1.42 volts *

When is setted on auto, vccio reaches 1.312v and is coloured purple (danger maybe?). I think it is something to do with that and no with memory training at all. If it's memory training, how is it possible that lowering vccio and sa voltages solve the "issue"? I'm glad my question helped you.

[Silent_Scone]

It's more important to check the memory is stable than be concerned with AC cycling...

Some folks seem to have strange priorities :D.

[Leoplate25]

Silent Scone@ASUS wrote:
It's more important to check the memory is stable than be concerned with AC cycling...

Some folks seem to have strange priorities :D.

So far memory is stable. I don't know what is the problem of wanting to have only one ac cycle. You are not being nice at all.

[Silent_Scone]

Leoplate25 wrote:
So far memory is stable. I don't know what is the problem of wanting to have only one ac cycle. You are not being nice at all.

How have you tested the memory?

There isn't a problem, which is ironically the point - it should make no odds to the user if the board requires one AC cycle, as it takes a matter of seconds. Arbitrarily lowering signal voltage rails simply only to avoid these things is very peculiar indeed, but it's a free world 😉

[Kelutrel]

Silent Scone@ASUS wrote:
How have you tested the memory?

There isn't a problem, which is ironically the point - it should make no odds to the user if the board requires one AC cycle, as it takes a matter of seconds. Arbitrarily lowering signal voltage rails simply only to avoid these things is very peculiar indeed, but it's a free world 😉

Sir, the Intel specifications state that the default voltages are 0,95V for VCCIO and 1,05V for VCCSA on 8th and 9th generations processors. Also, having those 2-3 AC cycles each time the PC is booted after switching it off feels pretty weird somehow. I, for one, considered that an issue, like a lot of other people in this forum if you check the previous threads.

Most of us are overclockers, so it is not a big deal to change a BIOS setting if it makes our PC boot up faster and it is also stable. I verified my RAM stability in plenty of ways, from MemTest to gaming and benchmarks, and I can state that it is pretty stable, probably more than with the higher voltages that the BIOS automatically used before.

It feels like you don't like that this issue has a solution.

[Silent_Scone]

Kelutrel wrote:
Sir, the Intel specifications state that the default voltages are 0,95V for VCCIO and 1,05V for VCCSA on 8th and 9th generations processors.

This is for stock operation. The auto rules scale with frequency. I'm not sure you can quantify the behavior as "weird somehow". It's there because it is needed. It may have some impact on memory training by lowering these values, but it's fruitless if the system isn't stable.

It makes no sense to do so simply to avoid ac cycling if not testing for instability. Higher frequencies require more VCCSA voltage, so advising users to lower these values for these reasons is ,simply put, bad advice.