high voltage still: ryzen 3600

[CrazieChipmunk]

i have put ryzen balanced power plan still high voltage , i have updated drivers for cpu and ryzen master. I have a feeling its the beta bios for msi b450 tomahawk causing this, my temps are around 45oc now which is an improvement i guess but still very high voltage for idle 

[Jurrunio]

it's AGESA update from AMD's fault. Seems like B450 and X470 boards from other brands have the same problem, likely B350 and X370 as well.

[Sorenson]

I think this is the way ryzen 3000 is. I tried manually undercoating a 3700x through ryzen master and it wasn’t stable at the higher frequencies. It could just be early firmware, but it seems like it’s on pretty much every motherboard. 

Edited August 1, 2019 by Sorenson

[CrazieChipmunk]

7 hours ago, Jurrunio said:

it's AGESA update from AMD's fault. Seems like B450 and X470 boards from other brands have the same problem, likely B350 and X370 as well.

 

7 hours ago, Sorenson said:

I think this is the way ryzen 3000 is. I tried manually undercoating a 3700x through ryzen master and it wasn’t stable at the higher frequencies. It could just be early firmware, but it’s on pretty much every motherboard. 

Ahh i see, so i guess we wait for amd to give the motherboard vendors the new update?

[Jurrunio]

1 minute ago, CrazieChipmunk said:

 

Ahh i see, so i guess we wait for amd to give the motherboard vendors the new update?

just roll back to older bios for the mean time

[Sorenson]

7 minutes ago, Jurrunio said:

just roll back to older bios for the mean time

My 3700x ran at high voltages on with or without the AGESA update. It was an gigabyte X570 board. 

[Jurrunio]

11 minutes ago, Sorenson said:

My 3700x ran at high voltages on with or without the AGESA update. It was an gigabyte X570 board. 

What makes you think older boards use different AGESA versions compared to newer boards?

[Sorenson]

5 minutes ago, Jurrunio said:

What makes you think older boards use different AGESA versions compared to newer boards?

I never said they did. All I said was I had high voltage with or without the update, and when I manually reduced the voltage through ryzen master the CPU wasn’t stable. I think the higher voltages are just a downside to otherwise fantastic CPU’s. I’m basing this opinion off of Hardware Unboxed and Optimum tech’s discussion of the high ryzen 3000 voltages. 

[Jurrunio]

4 minutes ago, Sorenson said:

I never said they did. All I said was I had high voltage with or without the update, and when I manually reduced the voltage through ryzen master the CPU wasn’t stable. I think the higher voltages are just a downside to otherwise fantastic CPU’s. I’m basing this opinion off of Hardware Unboxed and Optimum tech’s discussion of the high ryzen 3000 voltages. 

then why quote me in the first place?

[Sorenson]

8 minutes ago, Jurrunio said:

then why quote me in the first place?

Because from my understanding you were saying the high voltage was caused by the AGESA and that was not my experience. I didn’t think the information you were giving was accurate. 

[paulmohr]

I have the same cpu and the same board. I installed the 1.A0 bios and the new amd chipset drivers. I am using the AMD performance powerplan. I run mine with the default bios settings, as in I select "default settings" or "optimum settings" I forget whats it called. The only thing I changed was turned on my memory xmp profile and changed my fan curves. Save and exit bios.

 

I just checked cpuz while typing this and my voltage jumps between .3 and .7. And that is with firefox running with 5 tabs open.

 

I can check it with pbo on, and then check it with a quick 4.2 overclock if you really want me to.

 

*** I just checked ryzen master and it shows me at 1.375 volts steady. I don't trust it. I would suggest using cpuz or something and see what it says your voltages are. All these monitors seem to have an issue somewhere in them. Ryzen master shows my voltages being higher, but it shows really low clock speeds. Like 300 to 450 or it shows them all being asleep. Cpuz shows really low voltages but core speed doesn't go below 3500. Something else I just noticed. After I turned Ryzen master on to check what it says, and closed it, cpuz shows my voltages are above 1 volt now and my core speeds are bouncing between 3600 and 4200. I am going to restart my computer and see if I duplicate that. If it happens again I am uninstalling ryzen master. It looks like it is effecting the load while running and leaving something running after you close the program.****

Edited August 1, 2019 by paulmohr
Added information

[paulmohr]

Yep, it was repeatable. I rebooted my computer and opened cpuz. Voltage at idle was around .3-.4 volts, core speed bounced between 3300 and 3600. I opened firefox and core speed stayed at 3600 and voltages jumped between .3 and .8 volts.

 

I re booted and opened cpuz again, same readings as above with just cpuz running. I opened ryzen master and cpuz showed my voltage pegged at 1.375 volts and core speed was pegged at 4200. I closed ryzen master and my voltages were between .8 and 1.1 volts. Core speeds were bouncing between 3600 and 4100. I opened firefox again and voltages are now reading over 1 volt. Like 1.0-1.2 volts. Core speeds are between 3600 and 4200.

 

My conclusion, ryzen master isn't that great for monitoring and I wouldn't use it unless I had to because my bios wouldn't let me overclock something and I will be uninstalling from my computer.

 

Like I have said before, none of these software based tools are that accurate and they all have issues to some degree. The only way to really do it accurately and know for sure is to use the proper equipment to measure voltages, current and heat at different parts of the board. Which somebody has probably done, or will do in the very near future. You will just have to search around to find the results.

[Mathieu9836]

It's an issue with monitoring software and how Zen2 behave, monitoring software tend to poke the CPU to ask what is it's current state and voltage, so the CPU think it's workload and get out off idle OR your monitoring software show you voltage and clockspeed BEFORE the cpu went into idle.

[CrazieChipmunk]

@Mathieu9836 do you think it will ever be fixed?

[CrazieChipmunk]

im honestly worried about my cpu overheating while 3d modelling/rendering as well as gaming

[Mathieu9836]

Here is good read about this ''issue'' my friends:

Hope it will help you understand what is going on with your CPUs and monitoring software

[CrazieChipmunk]

18 minutes ago, Mathieu9836 said:

Here is good read about this ''issue'' my friends:

Hope it will help you understand what is going on with your CPUs and monitoring software

i have read this before but its still not fixed, my fan is being weird it will randomly just go brummmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm and then silent and then 10mins- 30 mins later brummmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm

[paulmohr]

Could something be running in the background on your computer causing it to load up at certain times? Have you tweaked and tuned windows and checked your start up menu?

[Mathieu9836]

10 minutes ago, CrazieChipmunk said:

i have read this before but its still not fixed, my fan is being weird it will randomly just go brummmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm and then silent and then 10mins- 30 mins later brummmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm

For how many seconds does your fans goes crazy?

[CrazieChipmunk]

@Mathieu9836 for like a few seconds

[Dubba]

There's a few things connected. First, the spikes we see in voltages at idle are mostly for short single core boosts and apparently Ryzen also reports temperatures from the hottest core. Updating to the latest chipset driver helped idle behaviour somewhat on my X470 platform but I still see some spikes to +1.4V if I have anything running, even when not using, like a web browser.

 

Because the manufacturing process is so small, and the transistors are so tightly packed together we have these really intense spikes in heat output. At least our older motherboards might not understand this and just panic with the fan curves. I was able to stabilize my CPU fan craziness by just adding a few seconds of delay to the reaction time in BIOS. If your CPU temp spikes to a max of 84°C like mine does on all-core synthetic tests, your chip is totally fine. While gaming it stays a lot cooler.

 

Btw, I'm happy you got your board to finally boot : )

 

 

[CrazieChipmunk]

2 hours ago, Dubba said:

There's a few things connected. First, the spikes we see in voltages at idle is mostly for short single core boosts and apparently Ryzen also reports temperatures from the hottest core. Updating to the latest chipset driver helped idle behaviour somewhat on my X470 platform but I still see some spikes to +1.4V if I have anything running, even when not using, like a web browser.

 

Because the manufacturing process is so small, and the transistors are so tightly packed together we have these really intense spikes in heat output. At least our older motherboards might not understand this and just panic with the fan curves. I was able to stabilize my CPU fan craziness by just adding a few seconds of delay to the reaction time in BIOS. If your CPU temp spikes to a max of 84°C like mine does on all-core synthetic tests, you're chip is totally fine. While gaming it stays a lot cooler.

 

Btw, I'm happy you got your board to finally boot : )

 

 

Thank you dubba, really appreciate the help from everything to the board not booting etc, you helped alot and ty everyone else 

[Mathieu9836]

Buildzoid tested the 3700X why an oscilloscope

 

 

[paulmohr]

Damn you, I haven't looked at his channel this week. Now I am going to fall down a 2 hour or so rabbit hole catching up on his videos lol. This man has a wealth of information if your brain doesn't melt trying to take it in. He also does videos for Gamers Nexus. This is what I am talking about when I tell people stuff needs to be tested the proper way with good equipment and not trust on board sensors and software.

 

After watching the video I can see they are actually what I kind of thought they were doing. They are controling current. When we are looking at our freebie software we tend to look at voltage, heat and maybe wattage. We are only seeing part of the equation by doing this. Even though the voltages might get to what we consider high, the current doesn't. They are keeping it at a safe level so it doesn't destroy itself.

 

And oddly enough my cpu doesn't get hotter on single core workloads, its the opposite. I don't get over 50c when running R15 on single core. I use HWmonitor though, not Hardware info like he is. Maybe my software is lieing to me. I might try that software instead. I kind of like the layout better anyway after seeing it.

 

And it looks like your fan going nuts is actually a thing. Its odd, mine doesn't do that. Or its doing it and I just don't notice it. I find that hard to believe though since I have 10 fans in my system and some of them have LED's. It gets pretty loud under a load when they ramp up. And the LEDs get brighter when they spin up. If it was ramping up all the time I think I would hear it and the LEDs would be pulsing at me. I guess you could try what he said and change the delay times for your fans if its an option in your bios. Mine has it, but I didn't mess with them. I think they are at 1s or something. Like I have mentioned before though, my windows config is pretty stripped down and tweaked. If a service or something doesn't need to be running it isn't. Heck I got mad because Mystic light is pulling 20mb of ram lol. Which is stupid considering I have 32gb. I will probably paint the leds on my motherboard or something though so I can uninstall it. I also use the amd performance power plan and tweaked that too. Whether any of that actually makes a difference I have no idea honestly. Its just something I have always done when setting up a computer.

 

I am not shocked about any of this either. AMD did say they made these hyper responsive on purpose. So I guess that is the trade off. I think their might be a setting where you can slow this down if its a real issue.

[Mathieu9836]

I wonder what this guy do for living, he knows way more stuff then 99% of the reviewers. Thats the kind of videos i like to watch, very very in depth information and tests.