VRM cooling question

[frank30fps]

Yeah 4.0@1,35or1,40v

 

[trickson]

Just stopped working while I was watching a video so I sent it for RMA and the new one should arrive in about 4 days

Okay wait I am confused now.
Did it just die?
Well then it's over.
Reset bios? Is it gone for good?

 

[frank30fps]

Okay wait I am confused now.
Did it just die?
Well then it's over.

What's over ?

 

[EarthDog]

What's over ?

Not the confusion...lol

Trickson. It died.. not within the course of this thread, but he killed it and returned it. Currently an apu is in use.

 

[trickson]

What's over ?

If your computer is dead then there is no fixing it with anything other than an RMA. or new MB. so yeah it's over for this thread and MB. or am I missing something?

 

[frank30fps]

If your computer is dead then there is no fixing it with anything other than an RMA. or new MB. so yeah it's over for this thread and MB. or am I missing something?

I had a ryzen 2700 die on me 3 weeks ago, while I wait for the replacement I am using a ryzen 3200g

 

[EarthDog]

If your computer is dead then there is no fixing it with anything other than an RMA. or new MB. so yeah it's over for this thread and MB. or am I missing something?

You're missing something. I've explained and the op did for second time (with some detail) above.

 

[trickson]

Hi

I have a motherboard without any VRM cooling, will it have any effects on temps and oc stability if I put a fan directly over it without a heatsink ?

First off this kind of a MB should never be used to OC anything (JMHO) NO VRM heat-Sink= DO NOT OVER-CLOCK ,DO NOT USE I/O plate get fan blowing directly on the VRM area and if possible get a heat sink for the NEXT MB you get and put it on.

Hope you can get it up and running again. be careful next time.

 

[eidairaman1]

Watch those to understand

 

[JackCarver]

Each phase should be capable of 69A, as the low side fets are onsemi 4c06b.

https://www.onsemi.com/pub/Collateral/NTMFS4C06N-D.PDF

The most load goes via the low side fets, so their current capability is important here and it has 2 low side fets per phase, so that they can split the load.

If cpu draws for example 80A, your output voltage is 1.4V and switching freq is at about 300kHz, the conductive losses of the low side fets should be at about 0.71W.
Switching losses in high side fets should be at about 1.08W.
Total losses a little higher, so approx 0.85W per low side fet and 1.2W per high side.
As it has 8 low sides then 6.8W low side total and as it has 4 high side then 4.8W high side total. So 12-14W total power loss in the mosfets at a load of 80A.
The data sheet of the low side fets says heat dissipation junction to ambient at about 49 degrees per W. With an ambient temp of 25 degrees they would be at about 25+49=74 degrees, as they are a little less than 1W power loss maybe even less than 74 degrees.
The high sides have 50.3 degrees per W. With 1.2W power loss they could be a little higher than 50.3+25=75.3 degrees.

Anyway in the case of 80A total load the vrm should be in safe temps even without direct air cooling or heatsinks.

 

[trickson]

Anyway in the case of 80A total load the vrm should be in safe temps even without direct air cooling or heatsinks.

Key word here is "Should"
I personally have never purchased a MB would out proper heat-sinks even on the north and south bridges. No WAY!
It is a cheep way to make and sell a MB, In other words you get what you PAY for.
This could never be more clearer than in the FX DDR3 MB days. They all were burning out at a massive rate, And when you look at the way they did the VRM's for the FX chips the cheep MB used NO heat- sink that burned out the moment you over clocked the FX CPU past 4.0 Ghz! The ones that had LIKE the high end MSI MB had VRM heat-sinks that were massive! Mine still gets hot I took the I/O plate off so the CPU fans could cool it and cool it better it has. Dropped the temp's 10*f!
So when you go cheep or NO VRM heat-sink you should know there is a limit to just how hard you can push them.

 

[eidairaman1]

Key word here is "Should"
I personally have never purchased a MB would out proper heat-sinks even on the north and south bridges. No WAY!
It is a cheep way to make and sell a MB, In other words you get what you PAY for.
This could never be more clearer than in the FX DDR3 MB days. They all were burning out at a massive rate, And when you look at the way they did the VRM's for the FX chips the cheep MB used NO heat- sink that burned out the moment you over clocked the FX CPU past 4.0 Ghz! The ones that had LIKE the high end MSI MB had VRM heat-sinks that were massive! Mine still gets hot I took the I/O plate off so the CPU fans could cool it and cool it better it has. Dropped the temp's 10*f!
So when you go cheep or NO VRM heat-sink you should know there is a limit to just how hard you can push them.

My favorite board, see sys specs.

(If Asus had focused on a Non ROG RGB board under TUF for AM4/TR40 that is same quality or better I would get 1)

R2.0

Version 1

 

[Countryside]

My favorite board, see sys specs.

(If Asus had focused on a Non ROG RGB board under TUF for AM4/TR40 that is same quality or better I would get 1)

R2.0
View attachment 144710

Version 1
View attachment 144711

Yeah they dont make them like they used to i owned three am3+ motherboards GA-990FXA-UD3, M5A99FX PRO R2.0, 990FXA-GD80 they all had solid vrms and were one of the best am3+ motherboards.

Asus hasn't even managed to release the newest bios on x300 series boards while many others have released them long ago.

 

[JackCarver]

Key word here is "Should"

Sure you can‘t definitely say that a mosfet reaches that temp, it‘s only an approximation. They only assure junction to case thermal resistance, the junction to ambient depends on how the fets are mounted on the board, how many layer the pcb has, the ambient temperature or if there are heatsinks or fans used. So it depends on the Mainboard manufacturer and the user.

The junction to ambient resistance of 49 degrees of the low side fets are for example measured with the fet „Surface−mounted on FR4 board using 1sq−inpad,1ozCu.“ like the data sheet says.

But I think this should be approximately the situation on a Mainboard plus/minus a few degrees.

Plus/minus a few degrees at about 70-80 degrees isn‘t that problem for a mosfet with junction max temp of 150 degrees.

And I agree, I also wouldn‘t buy a board completely without any heatsink on the mosfets. But even my MSI Z370 Gaming Pro Carbon with heatsink on vrm throttles in OCCT benchmark, using AVX512 with small dataset, cause of reaching too high vrm temps. It has only 4 real phases and a vrm with more than 4 real phases is definitely better if your goal is really high oc.
Or if the vrm is using power blocks (high side and low side fets in one package) or power stages (high side/low side and gate driver in one package) or as in the case of many b450/x470 boards high side/low side fets and driver as separate chips. The power stages, especially that from IR are one of the best here in terms of efficiency and power loss. Another thing is if they fake the phase count by doublers or they use twice components for every phase. Here you have also to look which one is used. In terms of conduction losses the use of doublers is better than the use of twice components per phase.
So the heatsink isn‘t always the guarantee to reach higher oc as a board with 8 real phases and power stage mosfets without heatsink will maybe better overclock than a board with 4 real phases and cheap fets with heatsink.

Edit:
Here is a link to the Gigabyte Ultra Durable technology:

-- GIGABYTE --Geeks Column of the Week - Ultra Durable 3 Technology

As you can see they use several 2 oz copper layers in their PCB design, so the mosfet junction to ambient resistance should be at least at the same level as the OnSemis testing scenario or rather better.