[SOLVED] Upgraded CPU worried about VRM and voltage

OnlyAlex

Commendable
Sep 16, 2021
303
22
1,685
Hi guys,

I recently upgraded my CPU from a Ryzen 7 1700 to a Ryzen 7 3700x in a B350 Tomahawk Plus mobo. Since doing this whenever I am gaming the fan is rather loud. Looking at Ryzen Master the temp of the CPU at idle is around 42 which is a bit hot and then gaming 65 which I understand is ok?

In terms of voltage it's at 1.43 which I understand is high, so I have set it to max 1.35 volts in manual settings.

Basically I think the PC is safe to be using as it's not going above 80 degrees temp, but I am not sure about voltage. It's very noisy when gaming, never used to be but I understand that's probs due to VRM and lack of cooling on the board.

Do you think I need to get like a B550 or is the PC perfectly fine gaming, just deal with the louder fan - like it's not going to break on me?

I barely game to be honest and just need it to be safe if that makes sense. Cheers guys for any advice.
 
Solution
I did a very short stability test on the CPU and I believe the CPU temp and mobo temp is fine in this photo - CPU temp 66 at full load? AMD Ryzen Master max temp for CPU shot up to 95 degrees though. Am I missing something here? https://ibb.co/Jq6cdxS
Suggest to get HWInfo64 for more reliable temp monitoring of Ryzen CPU's.

But I'd be a bit concerned about temp of a 3700X going up to 95C in RyzenMaster. What CPU cooling are you using? the stock cooler is barely adequate. A good tower air cooler or 240mm AIO can have pretty much the same effect as overclocking from the way Ryzen performance boosting works.
Take a look at: Motherboard VRM Tier list
A tier list is a compilation of ranking of a certain component (in the case above, a motherboard) based on their testing. It conveniently cuts down on research time on the part of the consumer to see which model is better than the other.

There is no calculation/formula to quantifiably determine if a CPU x is paired good with motherboard Y in terms of VRMs or pre determine how good it will perform on a certain motherboard. It can only be revealed through actual testing.

AMD Chipset Naming Scheme:

A = Budget
B = Mainstream
X = Enthusiast
TRX – High-End Desktop, Threadripper (HEDT)
X – High-End Desktop

Note: Trying to run a Ryzen 9 (R9) CPU with an AMD A chipset motherboard will obviously be detrimental as the motherboard cannot handle it even if they are technically compatible, not to mention you have to consider how well the VRM of the motherboard can handle the CPU.

Personally I would definitely not be using a high end power CPU on an A-series AMD motherboard.

There are some cases however, that you can "cheap out" or "get away" using a Ryzen 9 on a B series motherboard because that B series board has good VRMs.

View: https://www.youtube.com/watch?v=CTM0EeRikFg


In this video the tester uses an Asrock B550 Velocita
It can be seen in the tier list I linked earlier that it's a Tier A motherboard in terms of VRMs:

MsJNswV.png



There are however some guidelines:

-The more powerful the CPU is, the more you want to look at a higher tiered motherboard that has good VRMs (and this is probably you're main concern).
-Generally, an X series AMD motherboard will have better VRMs than a B series motherboard, unless if it has been revealed through actual testing that these VRMs suck
(There have been such cases in the past). If you are willing to go for a B series motherboard, then research further to see if it can handle your CPU.

-Any fan that runs at a max speed over prolonged periods of time is bound to fail sooner than a fan not running at max speed, especially if its build quality is substandard (or cheap generic fan). Fans however are easily replaceable as long as its not the stock CPU cooler you're using. The bearing of the fan will usually be the common point of failure in this case.

-The higher the CPU's voltage is running at constant load, the better the VRMs need to be on the motherboard to be able to handle that power delivery at a consistent rate.

-Voltage is the push of electricity, so if you guess it, the CPU when it is at a higher voltage, is drawing more power in order to execute its work load efficiently.
-If the VRM can't handle the "higher push" of electricity well enough, it may cause hardware issues such as overheating and potentially electrical damage.

-Modern CPUs can run up to 1.5V, but that doesn't mean all CPUs should safely be able to.
Part of the problem is that, the manufacturer (AMD) doesn't list safe operating voltage in the CPU spec sheet, so its left for us consumers to guess "how high is too high" of a voltage. A rule of thumb I would use is don't go 10% beyond your idle operating voltage, which you can see with software like HWInfo > Sensor settings and under the parameter "CPU Core Voltage". Read the SVI2 reading, not the VID

Zq3GAZ9.png

For example, this is what my voltage looks like on my Ryzen 9 5900x on HWInfo. It's currently running at 1.456V and drawing 16.996A on the CPU Core. Temp is 39-42C idle. This is normal according to some observations.

SVI2 is the actual voltage that the CPU is getting, VID is what's being requested, not actually being used.

SoC Voltage definition:

vmgGLaI.png

From (Glad I found this): https://www.corsair.com/corsairmedia/sys_master/productcontent/Ryzen3000_MemoryOverclockingGuide.pdf

In the case of "Undervolting CPUs" which you did:
Reducing the amount of voltage does reduce temperatures and is perfectly safe, but if the motherboard's VRMs are inherently of substandard quality or not up to your CPU, it is only a stop gap solution.


VRM Explanation:
You can think of VRMs as "mini power supplies". They further step-down voltage down to 1V or 0.5V depending on what the component needs. The VRM is not a single component like a PSU but is actually a collection of multiple components

LsUjaOr.png


1.) MOSFET - Your CPU will tell this mosfet it needs 1.5V, then the Mosfet will deliver it. This is the gate switch that allows or restricts voltage to control the amount of voltage.

2.) Chokes (Block structures) - usually next to the MOSFET
3.) Capacitors - Temporarily stores electricity and protects against sudden voltage spikes

The quality of your VRMs will dictate the longevity of your PC. Cheap VRMs can lead to instability and crashes. Very poor quality VRMs can mess up components and damage them. When overclocking, VRMs are very important because they need the right voltage and consistent delivery.

-You can count the number of chokes in the motherboard. The motherboard splits up power delivery to the CPU into a number of phases. The more phases = less load on each phase = greater stability

Each choke on your motherboard corresponds to 1 power phase. Usually 1-2 of them are reserved for your CPU Cores

Entry level motherboards usually have 3-4 phase for CPU power
Higher quality boards will have 6-8 or even more

Note: The chokes are usually near the CPU socket slot

It is also good to pick up a board with solid capacitors. Cheaper capacitors may contain a conducting liquid which may cause problems if they are not made correctly. Even if they are made correctly, they can bulge or rupture (sign of of capacitor failure). Solid capacitors usually have wider temperatures and a longer lifespan.

Coincidentally, someone had asked the same question with the same specs you have.

Edit & TLDR: After reading many articles, Ryzen 3000 CPUs generally seem to run at 1.4V. There's nothing much you can do about it, but if you feel you don't trust the VRMs on your board, you can go for a higher tiered one on the list.

I have never tried a Ryzen 3000 series CPU, but it appears that people found issues when monitoring the voltage at idle, giving inaccurate readings because the monitoring software itself is pushing Ryzen to do some load. If you want a serious attempt to determine your idle voltage reading, see the link below:

Use CPU-Z to monitor Idle Voltage on 3rd Gen Ryzen CPU
(Follow the how-to in this reddit to take an accurate measurement)

Im starting to think this is only going on with certain bios versions on certain motherboards. I've seen people posting about high temps at "idle" a lot with MSI. I don't have one on hand to test unfortunately and it may be why I have not seen this issue across multiple other motherboards. This appears to possibly be a bug where the CPU is actually just sitting at 1.4v regardless of load and not shifting dynamically like it is supposed to. If you read the AMD post carefully they seem to be aware of this and are probably working on a fix. A sub par cooling solution would be severely taxed in this scenario.


Key take aways:

The CPU has a power and temperature budget. If the load, current, and temperature conditions are correct it will use that power and temp headroom to boost the CPU as much as possible. Whether it dumps a bunch of voltage and frequency at one core or less voltage and frequency at a bunch of cores it will use it's temp and power budget. The power plans that come with the chipset drivers also allow the CPU to boost on a hair trigger. I'd say you don't really need to worry about anything but heavy (high current) load temps. In a modern OS the CPU is almost never truly idle so light tasks will trigger boost and higher "idle temps".

In terms of voltage it's at 1.43 which I understand is high, so I have set it to max 1.35 volts in manual settings.

Generally, when a CPU is running at a high voltage, it must also be drawing a lot of current (inference), else why would it need to run at a high voltage in the first place? The issue when Ryzen 3000 series came out is that people sort of "panicked" to see they were running at 1.4V on light loads or even at "idle". It came out at this state. The metric I would be more concerned is the temperature of your CPU under load, which is correlated to how much amps your CPU is drawing (more current draw = higher temperature). Hence yes, I wouldn't worry too much about it running at 1.4V, but you can under volt your CPU if you desire in an attempt to improve temperatures.

Since doing this whenever I am gaming the fan is rather loud. Looking at Ryzen Master the temp of the CPU at idle is around 42 which is a bit hot and then gaming 65 which I understand is ok?

This is ok temps

Do you think I need to get like a B550 or is the PC perfectly fine gaming, just deal with the louder fan - like it's not going to break on me?
What model is your CPU cooler anyways? Did you happen to max out your RPM on your BIOS' fan curve by accident? (i.e. set the RPM to 100% when temp reaches 60C, since your gaming temp is 65C then your fan spins at 100% RPM)
 
Last edited:
  • Like
Reactions: OnlyAlex
Hi guys,

I recently upgraded my CPU from a Ryzen 7 1700 to a Ryzen 7 3700x in a B350 Tomahawk Plus mobo. Since doing this whenever I am gaming the fan is rather loud. Looking at Ryzen Master the temp of the CPU at idle is around 42 which is a bit hot and then gaming 65 which I understand is ok?

In terms of voltage it's at 1.43 which I understand is high, so I have set it to max 1.35 volts in manual settings.

Basically I think the PC is safe to be using as it's not going above 80 degrees temp, but I am not sure about voltage. It's very noisy when gaming, never used to be but I understand that's probs due to VRM and lack of cooling on the board.

Do you think I need to get like a B550 or is the PC perfectly fine gaming, just deal with the louder fan - like it's not going to break on me?

I barely game to be honest and just need it to be safe if that makes sense. Cheers guys for any advice.
Yah, putting a 3700X on a B350 board is a bit sketch since they have notoriously poor VRM's. But so long as you keep the CPU in stock settings you'll probably be OK, that means no overclocking not even using PBO.

The good thing about the Tomahawk is it's one of the better B350's and has decent sensor outputs so you can read the VRM temperatures with the CPU under heavy loads. Gaming is NOT a heavy CPU processing load. Rendering, e.g. Cinebench 20 in multi-thread mode, is a good heavy and realistic processing load.

Set up the CPU in full stock (clocks and voltage in AUTO, no PBO) and run the rendering for 20 min's at least while monitoring VRM temperature (HWInfo will graph temps to make that easy). VRM max spec'd temps are probably around 115C-125C (typical for modern MosFET's). But I'd want to see 95C-105C max. If it's staying under or around that in CB20 I'd say any gaming you do will be just fine and it's still got the headroom to handle any heavy processing that might come along.

Leave CPU Vcore voltage in AUTO so the boosting algorithm can lower voltage as often as possible in routine use and that will lower the stress on VRM. If you MUST adjust voltage only do it using offsets so the algorithm can still adjust core voltage to help both CPU and VRM.

Ryzen 3000 works way different from a 1000 CPU's. When left in AUTO, Vcore will vary considerably, raising up to 1.5V when it boosts a single core to max clocks. That's by design and to be expected so do not be concerned. Temps also run a bit hotter...95C is Tjmax for the CPU but it's desireable to be operating in the mid-70's to keep the CPU boosting well in heavy workloads. It only boosts to max clocks in light, bursty work loads though, but that's typical of gaming and what you want.

When it boosts a core it will spike temp along with it, that's what makes fans pulse and run loud if you don't set a custom profile to ignore the spikes.

Yes, moving up to a B550 board would be a good move, for several reasons. But be sure to get one with a decent VRM as there are low-spec ones there too.
 
Last edited:
  • Like
Reactions: OnlyAlex
Take a look at: Motherboard VRM Tier list
A tier list is a compilation of ranking of a certain component (in the case above, a motherboard) based on their testing. It conveniently cuts down on research time on the part of the consumer to see which model is better than the other.

There is no calculation/formula to quantifiably determine if a CPU x is paired good with motherboard Y in terms of VRMs or pre determine how good it will perform on a certain motherboard. It can only be revealed through actual testing.

AMD Chipset Naming Scheme:



Personally I would definitely not be using a high end power CPU on an A-series AMD motherboard.

There are some cases however, that you can "cheap out" or "get away" using a Ryzen 9 on a B series motherboard because that B series board has good VRMs.

View: https://www.youtube.com/watch?v=CTM0EeRikFg


In this video the tester uses an Asrock B550 Velocita
It can be seen in the tier list I linked earlier that it's a Tier A motherboard in terms of VRMs:

MsJNswV.png



There are however some guidelines:

-The more powerful the CPU is, the more you want to look at a higher tiered motherboard that has good VRMs (and this is probably you're main concern).
-Generally, an X series AMD motherboard will have better VRMs than a B series motherboard, unless if it has been revealed through actual testing that these VRMs suck
(There have been such cases in the past). If you are willing to go for a B series motherboard, then research further to see if it can handle your CPU.

-Any fan that runs at a max speed over prolonged periods of time is bound to fail sooner than a fan not running at max speed, especially if its build quality is substandard (or cheap generic fan). Fans however are easily replaceable as long as its not the stock CPU cooler you're using. The bearing of the fan will usually be the common point of failure in this case.

-The higher the CPU's voltage is running at constant load, the better the VRMs need to be on the motherboard to be able to handle that power delivery at a consistent rate.

-Voltage is the push of electricity, so if you guess it, the CPU when it is at a higher voltage, is drawing more power in order to execute its work load efficiently.
-If the VRM can't handle the "higher push" of electricity well enough, it may cause hardware issues such as overheating and potentially electrical damage.

-Modern CPUs can run up to 1.5V, but that doesn't mean all CPUs should safely be able to.
Part of the problem is that, the manufacturer (AMD) doesn't list safe operating voltage in the CPU spec sheet, so its left for us consumers to guess "how high is too high" of a voltage. A rule of thumb I would use is don't go 10% beyond your idle operating voltage, which you can see with software like HWInfo > Sensor settings and under the parameter "CPU Core Voltage". Read the SVI2 reading, not the VID

Zq3GAZ9.png

For example, this is what my voltage looks like on my Ryzen 9 5900x on HWInfo. It's currently running at 1.456V and drawing 16.996A on the CPU Core. Temp is 39-42C idle. This is normal according to some observations.



SoC Voltage definition:

vmgGLaI.png

From (Glad I found this): https://www.corsair.com/corsairmedia/sys_master/productcontent/Ryzen3000_MemoryOverclockingGuide.pdf

In the case of "Undervolting CPUs" which you did:
Reducing the amount of voltage does reduce temperatures and is perfectly safe, but if the motherboard's VRMs are inherently of substandard quality or not up to your CPU, it is only a stop gap solution.


VRM Explanation:
You can think of VRMs as "mini power supplies". They further step-down voltage down to 1V or 0.5V depending on what the component needs. The VRM is not a single component like a PSU but is actually a collection of multiple components

LsUjaOr.png


1.) MOSFET - Your CPU will tell this mosfet it needs 1.5V, then the Mosfet will deliver it. This is the gate switch that allows or restricts voltage to control the amount of voltage.

2.) Chokes (Block structures) - usually next to the MOSFET
3.) Capacitors - Temporarily stores electricity and protects against sudden voltage spikes

The quality of your VRMs will dictate the longevity of your PC. Cheap VRMs can lead to instability and crashes. Very poor quality VRMs can mess up components and damage them. When overclocking, VRMs are very important because they need the right voltage and consistent delivery.

-You can count the number of chokes in the motherboard. The motherboard splits up power delivery to the CPU into a number of phases. The more phases = less load on each phase = greater stability

Each choke on your motherboard corresponds to 1 power phase. Usually 1-2 of them are reserved for your CPU Cores

Entry level motherboards usually have 3-4 phase for CPU power
Higher quality boards will have 6-8 or even more

Note: The chokes are usually near the CPU socket slot

It is also good to pick up a board with solid capacitors. Cheaper capacitors may contain a conducting liquid which may cause problems if they are not made correctly. Even if they are made correctly, they can bulge or rupture (sign of of capacitor failure). Solid capacitors usually have wider temperatures and a longer lifespan.

Coincidentally, someone had asked the same question with the same specs you have.

Edit & TLDR: After reading many articles, Ryzen 3000 CPUs generally seem to run at 1.4V. There's nothing much you can do about it, but if you feel you don't trust the VRMs on your board, you can go for a higher tiered one on the list.

I have never tried a Ryzen 3000 series CPU, but it appears that people found issues when monitoring the voltage at idle, giving inaccurate readings because the monitoring software itself is pushing Ryzen to do some load. If you want a serious attempt to determine your idle voltage reading, see the link below:

Use CPU-Z to monitor Idle Voltage on 3rd Gen Ryzen CPU
(Follow the how-to in this reddit to take an accurate measurement)




Key take aways:





Generally, when a CPU is running at a high voltage, it must also be drawing a lot of current (inference), else why would it need to run at a high voltage in the first place? The issue when Ryzen 3000 series came out is that people sort of "panicked" to see they were running at 1.4V on light loads or even at "idle". It came out at this state. The metric I would be more concerned is the temperature of your CPU under load, which is correlated to how much amps your CPU is drawing (more current draw = higher temperature). Hence yes, I wouldn't worry too much about it running at 1.4V, but you can under volt your CPU if you desire in an attempt to improve temperatures.



This is ok temps


What model is your CPU cooler anyways? Did you happen to max out your RPM on your BIOS' fan curve by accident? (i.e. set the RPM to 100% when temp reaches 60C, since your gaming temp is 65C then your fan spins at 100% RPM)

Thank you so much for the detailed reply, I really appreciate it! I don't understand a lot of it because I am a noob haha, but I got the overall gist! I have the stock cooler Prism for the AMD Ryzen 7 3700x. I did under volt by accident so I set everything to default and then went and adjust the fan curve. I've now got the temp idle at 35 degrees and gaming 50 / 60 degrees and the fan is now a lot quieter, I set the curve to quiet instead of default mode. Does that all sound ok? Thanks for your help again. I won't be messing around with the voltage ever again unless I know what I'm doing lol!
 
Yah, putting a 3700X on a B350 board is a bit sketch since they have notoriously poor VRM's. But so long as you keep the CPU in stock settings you'll probably be OK, that means no overclocking not even using PBO.

The good thing about the Tomahawk is it's one of the better B350's and has decent sensor outputs so you can read the VRM temperatures with the CPU under heavy loads. Gaming is NOT a heavy CPU processing load. Rendering, e.g. Cinebench 20 in multi-thread mode, is a good heavy and realistic processing load.

Set up the CPU in full stock (clocks and voltage in AUTO, no PBO) and run the rendering for 20 min's at least while monitoring VRM temperature (HWInfo will graph temps to make that easy). VRM max spec'd temps are probably around 115C-125C (typical for modern MosFET's). But I'd want to see 95C-105C max. If it's staying under or around that in CB20 I'd say any gaming you do will be just fine and it's still got the headroom to handle any heavy processing that might come along.

Leave CPU Vcore voltage in AUTO so the boosting algorithm can lower voltage as often as possible in routine use and that will lower the stress on VRM. If you MUST adjust voltage only do it using offsets so the algorithm can still adjust core voltage to help both CPU and VRM.

Ryzen 3000 works way different from a 1000 CPU's. When left in AUTO, Vcore will vary considerably, raising up to 1.5V when it boosts a single core to max clocks. That's by design and to be expected so do not be concerned. Temps also run a bit hotter...95C is Tjmax for the CPU but it's desireable to be operating in the mid-70's to keep the CPU boosting well in heavy workloads. It only boosts to max clocks in light, bursty work loads though, but that's typical of gaming and what you want.

When it boosts a core it will spike temp along with it, that's what makes fans pulse and run loud if you don't set a custom profile to ignore the spikes.

Yes, moving up to a B550 board would be a good move, for several reasons. But be sure to get one with a decent VRM as there are low-spec ones there too.

Thank you so much for your reply and for putting it in plain English for a noob like me lol!

I only play The Sims 4 if I'm honest on it and I want to save the money for a new rig. I have adjusted the fan curve and I now have idle temp of 35 / 40 and gaming 50 / 60 degrees and the fan is quieter now I've set the fan curve better. I have left the volts and CPU OC to default. I know getting a B550 board would be better but can I get away with this? I think it's safe as you said and I'm not going to play any other games. 😁
 
I did under volt by accident so I set everything to default and then went and adjust the fan curve. I've now got the temp idle at 35 degrees and gaming 50 / 60 degrees and the fan is now a lot quieter, I set the curve to quiet instead of default mode. Does that all sound ok?
Yes that's better. Culprit was the fan curve all along.

I know getting a B550 board would be better but can I get away with this?
Probably, if you insist. You can be the first data point if you're able to prove it after X years have passed :)
Just leave everything as is and don't OC or turn on PBO boost as suggested.
 
Yes that's better. Culprit was the fan curve all along.


Probably, if you insist. You can be the first data point if you're able to prove it after X years have passed :)
Just leave everything as is and don't OC or turn on PBO boost as suggested.
Yes that's better. Culprit was the fan curve all along.


Probably, if you insist. You can be the first data point if you're able to prove it after X years have passed :)
Just leave everything as is and don't OC or turn on PBO boost as suggested.
Haha thank you. Well those figures seem safe so as long as it doesn't jump etc. and with my light gaming I hope it will be ok for the next few years. Fingers crossed. It does say on the AMD website that the board does support the CPU. 😁
 
I did a very short stability test on the CPU and I believe the CPU temp and mobo temp is fine in this photo - CPU temp 66 at full load? AMD Ryzen Master max temp for CPU shot up to 95 degrees though. Am I missing something here? https://ibb.co/Jq6cdxS
Suggest to get HWInfo64 for more reliable temp monitoring of Ryzen CPU's.

But I'd be a bit concerned about temp of a 3700X going up to 95C in RyzenMaster. What CPU cooling are you using? the stock cooler is barely adequate. A good tower air cooler or 240mm AIO can have pretty much the same effect as overclocking from the way Ryzen performance boosting works.
 
Solution
Suggest to get HWInfo64 for more reliable temp monitoring of Ryzen CPU's.

But I'd be a bit concerned about temp of a 3700X going up to 95C in RyzenMaster. What CPU cooling are you using? the stock cooler is barely adequate. A good tower air cooler or 240mm AIO can have pretty much the same effect as overclocking from the way Ryzen performance boosting works.
Thank you, got it all sorted now. :)