The effect of replacing a VRM heatsink on a budget board using the example of Gigabyte b450m s2h

The success of this or that budget motherboard on the market is determined by what consumer qualities and functionality it can offer for the same price or even less than that of competitors. At the same time, to reduce the price, it is not necessary to save only on the quality of components: you can simply order a gigantic volume of a certain product from a supplier, increasing the payback of his equipment and request a reduced price for yourself.

The latest Gigabyte b550m s2h uses the same heatsink as the Gigabyte GA-Z77MX-D3H from 2012 …

The very presence of a heatsink on the VRM of a budget motherboard already puts it in an advantageous position over its competitors, however, you can often hear from someone that another motherboard for a price of two or three thousand rubles more has a more powerful VRM and that you need to choose it. At the same time, it is overlooked that, given the low retail price, these same 2-3 thousand lead to an increase in spending on a motherboard by one and a half times, which is not acceptable for everyone, and the circle of applicants for a purchase in this price range will be completely different.

In the same price range as the Asus TUF b450m PRO Gaming, its rival will be the full-size MSI X470 Gaming Plus MAX with improved mosfets.

At the same time, if we compare in terms of VRM power the budget board Asus b450m-k and the one and a half times more expensive Asus TUF b450m-plus gaming, it turns out that all the differences are only in the use of a massive heatsink in the latter. And the Asus TUF b450m PRO Gaming, which is considered to be quite successful in terms of the component base, does not differ much – the same mosfets, but with a doubling in the low-current upper arms, as well as paired chokes in the phases of powering the CPU cores and a more rational arrangement of these phases under the radiators. Considering that according to the specification, the maximum permissible current in the most loaded mosfets 4C06B of the lower arm of the VRM power phases from the nominal 69A for 25 degrees drops to 52A for 80 degrees, then the use of decent radiators is absolutely justified here.

AD0-E116 Exam Dumps

NSE6_FWF-6.4 Exam Dumps

Cisco 350-801 Practice Questions

SAP C_SAC_2102 Practice Questions

Salesforce Sales-Cloud-Consultant Questions

NSE6_FWB-6.1 Practice Test Questions

HPE6-A72 Test Questions

810-01 Test Dumps

GCP-GC-REP Practice Test Questions

VMware 5V0-32.21 Questions

JN0-1332 Practice Questions

Atlassian 150 Practice Questions

Scrum SSM Dumps

Scrum SAFe-Agilist Dumps

PMI DASSM Exam Dumps

But back to our “budget”, which in some cases have to take the rap for their older relatives, especially after processors with 6-8 cores became affordable. I previously tested Ryzen 2700 @ 3.9GHz on a Gigabyte b450m S2H motherboard
Then not the most ventilated case and the location in the niche of the computer desk played their part, warming up the VRM in a stress test in 10 minutes to 85 degrees.

After that, I perforated the front panel of the system unit and replaced the blank computer niche with a shelf:

The temperature situation improved, 79 degrees in 20 minutes:

However, I was haunted by the fact that on the same Asus TUF b450m PRO Gaming on similar OnSemi mosfets, the VRM temperatures are much lower. Therefore, I decided to experimentally check how the VRM temperatures of this budget motherboard will change with a more powerful radiator on the mosfets of four one and a half phases. By analogy with the Asus TUF Gaming b450m-PRO S, it was decided to use all the available space between the socket and the rear panel of the interfaces for the radiator, adjusted for the shape of the profiles of commercially available aluminum rolled products. After weighing the price offers and the height available under the Thermalright Macho 120 Rev. A processor cooler, I chose the domestic AVM-053 profile made of AD31 alloy:

In addition, a compact ribbed heatsink with a thickness of 5.5 mm was ordered on a well-known Chinese marketplace for use as a backplate-heatsink, or an improved version of the reinforcing plate on the back of the Asus TUF b450m PRO Gaming board. Comparative dimensions of the used radiators:

Naturally, it was not without the purchase of a thermal pad, in my case 1mm thick:
I had to remove one rib from the workpiece so that the heatsink would fit between the 8-pin CPU power connector and the 4-pin fan connector, and also remove some metal in the area of ​​the audio output rack. In addition, in order to preserve the useful cooling area as much as possible, it was decided only to make circular recesses at the base of the radiator for three high and one low capacitors. The structural features of this section of the motherboard can be seen in the photo of the Gigabyte b450m H, devoid of a VRM radiator:

GIGABYTE B450M H motherboard

The last photo of the board with the stock VRM heatsink:

After installing a new heatsink, the board looked like this:

There is enough distance between the VRM heatsink and the bottom aluminum plate of the CPU cooler:

It is worth pointing out that at first I fixed the radiator without a backplate and with springs on the screws, as a result of which it was not possible to ensure uniform pressing of the radiator to the mosfets, and moreover – it seemed to me that it stood a little with a blockage to the side opposite to the area of ​​concentration of the mosfets, and the textolite slightly bent over in this place. This was later confirmed by a high-quality imprint on the thermal pad of only a couple of mosfets closest to the screw. As a result, the first temperature measurement in the stress test did not reveal any positive changes.

Then I cut out the backplate heatsink, removing also a little metal from the side of the edge, which would have touched the case when installing the motherboard back on the brass racks to fix it:

This time I refused to use springs for mounting the radiator, and by tightening the M3 screws I tried to ensure the maximum uniform pressure in the area of ​​concentration of the mosfets. That everything turned out well, I realized literally in the first minutes of starting the stress test, when the VRM temperatures began to rise unusually slowly. As a result, 59 degrees in 10 minutes:


Even taking into account the correction for the difference in ambient temperature by a couple of degrees relative to the previous result, the resulting gain of ~ 20 degrees speaks for itself. Thus, we can conclude that the potential for lowering the VRM temperature with one and a half phases on the 4C06N / 4C10N mosfets widely used by Gigabyte due to the use of high-quality cooling is quite good. It could be used, for example, in Gigabyte b450m Gaming