[SOLVED] How to stop CPU from overvolting

asot

Distinguished
Apr 7, 2012
49
0
18,530
0
Hi guys,

I just got a new cpu today (i5 9600k, using it with asus prime b360m-a) and for "some reason" it's running hotter than expected (can get to 78C without being in full load). At first I was thinking that I maybe screwed up mounting the cooler but then I noticed that instead of having 1.2 Vcore, somehow it gets to 1,450 value most of the time when doing smth.

I have not tried OC (not sure if its even possible with my mobo and using default bios settings.

Any suggestions on what raises the voltage so drastically and how I can stop it?

Thanks!
 

Karadjgne

Titan
Herald
Paste shouldn't be an issue, a little movement is usually no worries, it's the remove/replace or up/down that is worrisome, especially with the old paste, as it tends to leave air pockets/bubbles. Which aren't healthy for a cpu or its temps.

You have removable HDD trays, so if using just the bottom blue led fan, remove the tray and set any drives in the top tray. They do nothing but hinder airflow and can make any front fans next to useless for cooling, especially the gpu area.

Power = Volts x Amps. So with the cpu in Power Limit throttling, and pushing upto 70A or so, it's basically saying it could do more, but has reached a point at 107w where it's uncomfortable going any further. At 107w output, that's @ 1.5v on the cpu at 70A. That's normal under a stress test where the cpu can be expected to be pushing limits. It's a I5-9600k, that's an unlocked cpu, that's capable of more than the motherboard (B360m-A) is capable of, being locked, so the VRM's and other voltage regulatory circuitry are not designed/built to handle the kinds of power/stress regulation a Z board is.

So yes, reaching Power Limit throttling is entirely possible/plausible with that motherboard and brutal tests such as IBT or ITU.

1.5v is high, I agree with that, but without the ability to step vcore down manually, which is odd, every bios I've ever seen has the ability to 'Lower', only OC mobo's have the ability to 'Raise' past stock values. You just have to take vcore off 'Auto'.

You have to take posts/reports/videos with a grain of salt. There's going to be a high degree of variation in results. Unless you see someone using your exact cpu, on your exact motherboard, running the exact same game, with the exact same settings, same cooler, same ram, same drives etc then their temp is going to be somewhat different. My i7-3770K at somewhere around 180w currently, games at @ 55°C. Stress test at 70°C. That's at 4.6GHz. At 4.9GHz, over 200w, and my old X61 aio, I got the same results. Huge difference in capacity between your 150w rring12 and my 300w+ AIO.

So yes, it's possible ppl are gaming at 60°C, with that cooler, and even with that cpu/cooler combo, but other factors such as which mobo, vcore, LLC, OC, applied bios settings etc can and will change things, even somewhat drastically.
 
Last edited:
Reactions: asot
Hi guys,

I just got a new cpu today (i5 9600k, using it with asus prime b360m-a) and for "some reason" it's running hotter than expected (can get to 78C without being in full load). At first I was thinking that I maybe screwed up mounting the cooler but then I noticed that instead of having 1.2 Vcore, somehow it gets to 1,450 value most of the time when doing smth.

I have not tried OC (not sure if its even possible with my mobo and using default bios settings.

Any suggestions on what raises the voltage so drastically and how I can stop it?

Thanks!
All modern CPUs have a turbo boost frequency as opposed to their BASE frequency. TDP numbers and “stock” voltage are based on BASE frequency. Once the CPU goes into turbo boost, it is basically a factory overclock. The voltage increases to increase the speed, which causes more heat. Many people think that they don’t need a decent CPU cooler or case airflow because they “will never OC” when in fact, their CPU will OC everyday if stressed.

If you have a Z370 or Z390 motherboard you can set a standard frequency and voltage so that it does not fluctuate. You basically want to find a speed you are happy with and then find the lowest voltage needed to keep this speed STABLE.
 
Reactions: asot and RodroX

Karadjgne

Titan
Herald
Your cpu is unique. There's no other cpu exactly like it. Some in the same model are extremely close, but others are somewhat different. That's the nature of cpus. Individuality. So being unique, it has its own exact requirements. Another cpu might have higher expectations, another lower. Intel isn't about to test every single cpu to determine what the exact needs are, so they choose a higher than needed voltage, to cover every cpu, no matter how different. Intel guarantees that your pc at stock settings will be stable. They don't guarantee it'll run cool. Being as stock voltages are set way high, this leaves room to lower voltages until you do become unstable, then bump up 1 or 2 notches. That's the basic form of overclocking, finding best speeds at best temps with lowest voltages to remain stable. It's not uncommon for older cpus to do a basic multiplier bump up 400-500MHz and not change any other setting, simply because stock voltages are high enough to accommodate the speeds. And be stable.

Mostly though, stock voltage settings are not entirely your friend, and are generally somewhat higher than necessary, doing nothing but adding to heat output, raising temps.
 
Reactions: asot

asot

Distinguished
Apr 7, 2012
49
0
18,530
0
Hi guys - thanks a lot for the replies.

I did a short miss in my post, it's actually an i5-9600kf, not sure if the lack of onboard gpu support makes things any different.

I did read a bit yesterday and found several posts which mentioned that having close to 1.5 voltages are detrimental to the CPU and was worried because the voltage seems high during windows operation and also sometimes the temperature seems to be fluctuating like mad...

@bmockeg I don't have a Z motherboard, I'm using an ASUS Prime B360m-a

@RodroX Couldn't find an MCE option, all I could find was a "CPU power enhancement" one and disabled it. Doesn't seem to be doing much though :/

@Karadjgne My original purpose was to find out if the crazy temperatures spikes/voltage are coming somehow from the motherboard (initially after I installed the CPU, turbo boost wasn't working and it was stuck @3,7 Ghz, but then I did a BIOS default reset and started working) or due to my incompetence of mounting the cooler.

I'm using a Thermaltake Riing Silent 12 cooler and people reported running this in under 60C for games. So I kinda panicked when I saw it reaching up to 78C, even if it would drop to 60-70 later.

So what you're saying is that I could have possibly gotten a CPU from the lower end of the batch that runs hotter.

But what do I do next? Can I undervolt my CPU with my current mobo ? (there's a manual option in BIOS but doesn't really let me choose a value, only for DRAM). Is there any point to re-mounting the cooler ? (it was painful as it was so not really looking forward to that TBH, but if that's the cause I will)

Here's a sample with 5 mins of windows operation having only a game launcher open, in a room with 20C temperature:

https://gifyu.com/image/v5aN

Edit: also did an Intel Extreme Tuning Benchmark here:
https://gifyu.com/image/v5YJ

Not sure if that power limit throttling is suppose to be "yes" tho :)

Thanks!
 
Last edited:

asot

Distinguished
Apr 7, 2012
49
0
18,530
0
Using a Zalman Z11 Neo. Left side of the case is open.

Pea sized to the middle. Although I admit, I may have shifted the radiator a bit while I was trying to screw it on one of the sides of the front plate.
 

boju

Champion
Have you tried closing the case to create pressure and exhaust stagnant air?

Another fan on that cooler can help a bit, dragging more air through the fins at a quicker rate.

I believe thermal paste is fine. It wasn't set long enough so moving heatsink whilst screwing it down wouldn't have done anything.
 
Reactions: asot

Karadjgne

Titan
Herald
Paste shouldn't be an issue, a little movement is usually no worries, it's the remove/replace or up/down that is worrisome, especially with the old paste, as it tends to leave air pockets/bubbles. Which aren't healthy for a cpu or its temps.

You have removable HDD trays, so if using just the bottom blue led fan, remove the tray and set any drives in the top tray. They do nothing but hinder airflow and can make any front fans next to useless for cooling, especially the gpu area.

Power = Volts x Amps. So with the cpu in Power Limit throttling, and pushing upto 70A or so, it's basically saying it could do more, but has reached a point at 107w where it's uncomfortable going any further. At 107w output, that's @ 1.5v on the cpu at 70A. That's normal under a stress test where the cpu can be expected to be pushing limits. It's a I5-9600k, that's an unlocked cpu, that's capable of more than the motherboard (B360m-A) is capable of, being locked, so the VRM's and other voltage regulatory circuitry are not designed/built to handle the kinds of power/stress regulation a Z board is.

So yes, reaching Power Limit throttling is entirely possible/plausible with that motherboard and brutal tests such as IBT or ITU.

1.5v is high, I agree with that, but without the ability to step vcore down manually, which is odd, every bios I've ever seen has the ability to 'Lower', only OC mobo's have the ability to 'Raise' past stock values. You just have to take vcore off 'Auto'.

You have to take posts/reports/videos with a grain of salt. There's going to be a high degree of variation in results. Unless you see someone using your exact cpu, on your exact motherboard, running the exact same game, with the exact same settings, same cooler, same ram, same drives etc then their temp is going to be somewhat different. My i7-3770K at somewhere around 180w currently, games at @ 55°C. Stress test at 70°C. That's at 4.6GHz. At 4.9GHz, over 200w, and my old X61 aio, I got the same results. Huge difference in capacity between your 150w rring12 and my 300w+ AIO.

So yes, it's possible ppl are gaming at 60°C, with that cooler, and even with that cpu/cooler combo, but other factors such as which mobo, vcore, LLC, OC, applied bios settings etc can and will change things, even somewhat drastically.
 
Last edited:
Reactions: asot

asot

Distinguished
Apr 7, 2012
49
0
18,530
0
@boju Closing the case would probably create more heat tbh - doubt my fans are that strong. Another fan on the cooler... sounds a bit hardcore :)

@Karadjgne Thanks a lot for the detailed reply! You are awesome for taking the time to explain in such detail :)

I checked again in BIOS the voltage is auto set to either 1.184 or 1.200 it seems. It gets to 1.3/ 1.4 /1.5 during Windows operation once the boost to 4.3ghz kicks in. I CAN set it to manual and input the value after all (my bad) but after setting it to 1.150 windows would no longer boot. There's also an offset option, if I'm not mistaken.

About the power throttling, this seems to be occurring even when I download games on Steam at over 70MBs. That shouldn't be considered "brutal" in the least... Not sure how much that will affect my performance overall, but combined with the potential temperatures it's rather concerning.

I wanted to avoid the hassle to change MB (and reinstall everything) and get the best CPU I could get for it - the 9600kf was the best(i7s too expensive for me tbh), even without OC, according to my conclusions at least.

But based on your explanation it would seem that the main culprit is the MB (PSU is 650W gold+ so should be sufficient?) and my best bets are:

1 - #dealwithit
2 - Get a nice/decent Z motherboard
3 - Return CPU, throw MB off the window and go AMD =)

Oh well. It makes me appreciate consoles so much more... you turn them on and they simply work. Not get some new (compatible) parts and spend hours on the internet looking for answers and bothering people. FML :/

Many thanks again for all your help, ppl!
 

dstln

Distinguished
Jun 8, 2007
283
15
18,815
12
They're going to briefly spin up the voltage as needed, it's fine as long as they're not constantly running at those numbers. Your CPU fan and cooling seems adequate. Honestly things look okay to me. The CPU starts working, fan speeds up to cool things and it stabilizes at reasonable temperatures.
 

Karadjgne

Titan
Herald
1.184 dropped to 1.15 might have been a little much. There's no instant fix, definite number. So I'd suggest much smaller drops, such as 0.08v at a time. And test for stability each run.

The offset is used in combination with VID. VID is what the cpu tells the motherboard to supply. VCORE is what the cpu actually uses. Little bit different. If the cpu sets vid at 1.30v and then tries to use 1.40v, then you'd need a +1.0v offset to make up the difference. If cpu sets vid at 1.40v and uses 1.10v, you could apply upto a -0.3v offset. Ideally you'll want to keep vid about 0.05v higher than vcore, so the cpu uses all it demands, without overtaxing VRM's or creating instability.
 
Last edited:
Reactions: asot

asot

Distinguished
Apr 7, 2012
49
0
18,530
0
@dstln Considering the CPU has zero OC and sits in a rather cool room (now it's like 19C) I wouldn't want to imagine how it runs in the summer. If it was summer now and it had a moderate OC as well, then yeah, would have totally agreed with you that it's fine as it is.

@Karadjgne I am once again honored to bathe in the rays of enlightenment provided by your knowledge! I'll try with smaller increments like you mentioned. Reading over the inet I saw multiple mentions of people successfully undervolting with like more than -100 so I thought that -0.50 would be by far in the more reasonable limit. I'll try to tinker a bit later on that and see if I can get any benefit.

@bmockeg Would a MSI Z390-A PRO do the job? Any other suggestions ? :) My current MB is mATX, but I guess ATX should work as well in my Z11 Neo.
 
Last edited:
@dstln Considering the CPU has zero OC and sits in a rather cool room (now it's like 19C) I wouldn't want to imagine how it runs in the summer. If it was summer now and it had a moderate OC as well, then yeah, would have totally agreed with you that it's fine as it is.

@Karadjgne I am once again honored to bathe in the rays of enlightenment provided by your knowledge! I'll try with smaller increments like you mentioned. Reading over the inet I saw multiple mentions of people successfully undervolting with like more than -100 so I thought that -0.50 would be by far in the more reasonable limit. I'll try to tinker a bit later on that and see if I can get any benefit.

@bmockeg Would a MSI Z390-A PRO do the job? Any other suggestions ? :)
You can’t undervolt AND expect the CPU to boost up to its top speeds and maintain the voltage. When you undervolt you also have to set the CPU at a set frequency.

Gigabyte Z390 Gaming SLI is a solid 12 phase cheap board.
 
Reactions: asot

ASK THE COMMUNITY

TRENDING THREADS