News MSI Z790 motherboards reportedly failing with cracked PCH chipset — a manufacturing error may have affected a few hundred units

Admin

Administrator
Staff member
Apr 2, 2024
1
2
10
Oh dang. I just finished a new build and had the MSI z790 tomahawk in my cart for weeks. The day I went to order it I noticed the Asus Z790 was on sale for a reasonable chunk of money less, so ordered that instead. I don't have that kind of luck often, but I'll take it.

Hopefully MSI is quick to rectify the issue and looks after those affected by it
 
Apr 3, 2024
2
3
15
Oh dang. I just finished a new build and had the MSI z790 tomahawk in my cart for weeks. The day I went to order it I noticed the Asus Z790 was on sale for a reasonable chunk of money less, so ordered that instead. I don't have that kind of luck often, but I'll take it.

Hopefully MSI is quick to rectify the issue and looks after those affected by it
I picked up a MSI z790 pro a few days ago as part of a bundle. I haven't yet opened any of the boxes. What to do what to do?
 
  • Like
Reactions: artk2219

Max The Dragon

Honorable
Jul 24, 2018
14
9
10,515
I have high doubts about this post and "defects", the post on MSI forum is related to bend CPU socket pins, but its not related to PCH Defects.

YouTube video (reported a similar failed BOOT process)... well its can be anything, single CPU debug led shows that PCH is working fine, because CPU debug LEDs are hardwired to PCH and CPU signals:
JxQ11mm.png

And if PCH is damaged, there will be no PGOOD signals at all and Debug LED will not lit up or all will be lit up.
So maybe guy just damaged something in progress of building his rig.

About Joshi Repair vid. The damage on PCH crystal looks like physical damage (something impacted the PCH radiator and applied force cracked the PCH). And i assume its just a typical DOA, but Joshi just added some hype to it.

TLDR
So. IF you allready have that MSI Z790 Tomahawk MB, you have nothing to fear, because chipsets are does not blow by itself. And i guees there is some hype point, because i have high doubt about qty of defective mb.
BUT, YOU need to always check ANY PC part for DOA. Because any Manufacture can have DOA.
 
  • Like
Reactions: artk2219 and KyaraM

Max The Dragon

Honorable
Jul 24, 2018
14
9
10,515
I picked up a MSI z790 pro a few days ago as part of a bundle. I haven't yet opened any of the boxes. What to do what to do?
Test any new hardware for DOA before assembly, just put it together outside the pc case (put mobo on cardboard box, install ram and cpu, put cpu cooler and connect the psu) and do a test boot. If it works - there is nothing to worry about. Just carefully put it inside your case and test again, and it should be in working condition.
 
  • Like
Reactions: d8mach and artk2219

Max The Dragon

Honorable
Jul 24, 2018
14
9
10,515
BTW, i guess i found way how they are destroying PCH. Maybe some users being uncareful and just jam GPU into PCI-E slot with force.
On this and on other mobos with big rad on PCH, you can apply extensive force to one side of PCH radiator very simply, when you jam GPU into slot you can press GPU into PCH radiator. And since mounting holes are located near the PCH and fastening screws are spring-loaded, you can just crush the PCH when installing GPU uncarefuly.
So, always be careful when assembly PC`s.

I marked area of possible force concentration zone, especially if the GPU is long:
x5rASsr.png

And mounting holes are doesn't overlap with the force application area, so Radiator may sag into force application direction and crush the chip (and maybe there is problems with spring loaded bolts, that can't prevent oversag):
7Cp7gMy.png

And, well, you can see how damage to PCH is being given here and the damage on PCH is more concentrated on GPU SLOT side:
JR4L9wy.png

And i assume that damage is given when someone inadvertently pressed too hard on the edge of the radiator, so its more Mechanical damage, than MSI fault.
And well, i guess someone over-hyped here.
 
Last edited:
  • Like
Reactions: artk2219

Max The Dragon

Honorable
Jul 24, 2018
14
9
10,515
nothing is perfect, msi mobo's are still good lol - do you not "have a few cracks" somewhere too?
my msi B660 mainboard under a review still has the lowest temperatures compared to the rest,
some mainboards (gigabyte, asus) were hitting 100 celsius..
View: https://youtu.be/PEa0vcJ0lRY?t=999
As electrical engineer, i can tell that ~100 degrees C, is kinda OK temperature for N-channel MOSFETS. Since their rated working temperatures are up to 125-150C (depends of structure).
And only way here to lower the temperature is:
Lower the switching frequency (modern PWM IC are smart enough to read thermals from MOS and adjust frequency on the fly, as you can see here:
tnuMthN.png
zFtT0Dj.png
)
Or add more metallization (bigger ground plane, more PCB layers or thicker layers themself), radiators on top of MOS are not effective, since the most of thermal conductivity are between solder joints, because there is metal to metal connection (not metal from rad to thermal pad to plastic case of MOS).

qfn32.jpg

TLDR and so, thermals is not a big deal if we speak about "stupid" stuff like FETs in VRM. When we speak about "smart" stuff likeCPU, APU, GPU, etc - thermals are big deal here.
And tests like hardware unboxed did, are not so representative, since they are doesn't measure the VRM switching frequency and so on. Plus VRM components can vary between different revisions of same motherboard.
 
Last edited: