after CPU VRM burned, CPU still good?

[ChaosLegionnaire]

Re: after CPU VRM burned, CPU still good?

Remember, there were some weird ASROCK boards that dropped phases as CPU activity decreased. Those didn't work too well and could be unstable, since they were 4-phase designs that actually single-phased at times. As you can imagine (look at Fig.4 diagram for reference), they tended to be less stable at part-load, especially when those Evercons aged- more ripple for a given duty cycle, as N decreases. Supposedly, they were compensated so the loop didn't oscillate, but switching-freq ripple still made it to the CPU. Silly ASROCK!

well, asrock is a company meant for engineers to try silly stuff and silly ideas on so no surprises there. going by murphy's law, 80% of the ideas they come up with will turn out to be bird brained while the remaining 20% of ideas will turn out to be truly revolutionary and will shape future mobo design to come...

[Stefan Payne]

Re: after CPU VRM burned, CPU still good?

Update from my side:
I've found some M4 screws, so I put my Xeon E5 6Core Low Power Thingy in the socket.
Put a heatsink on and used the M4 screws. Coudn't use much force because I can't reach the screwheads with my screwdriver, so my hand had to be enough...

Installed it in a case, put a harddrive in it and used a PSU (a Hiper 680W CWT PSH thingy with that 140mm Martech blue LED fan).

But haven't booted the rig up yet :ugly:

[kaboom]

Re: after CPU VRM burned, CPU still good?

well, asrock is a company meant for engineers to try silly stuff and silly ideas on so no surprises there. going by murphy's law, 80% of the ideas they come up with will turn out to be bird brained while the remaining 20% of ideas will turn out to be truly revolutionary and will shape future mobo design to come...

no longer true- it's a myth that persists from when the "original team" left asus to start ass-crock

even then, I've got doubts. Far as I can tell, they may as well be the same- are the "backwards" capacitor silkscreens any hint?

that's a stupid fanboy excuse- "throw shit at the wall and see what sticks- you never know, it'll be the next big thing in computahhzz."

don't buy that story- don't even rent it like a cheep prostitute

[kaboom]

Re: after CPU VRM burned, CPU still good?

Update from my side:
I've found some M4 screws, so I put my Xeon E5 6Core Low Power Thingy in the socket.
Put a heatsink on and used the M4 screws. Coudn't use much force because I can't reach the screwheads with my screwdriver, so my hand had to be enough...

Installed it in a case, put a harddrive in it and used a PSU (a Hiper 680W CWT PSH thingy with that 140mm Martech blue LED fan).

But haven't booted the rig up yet :ugly:

I don't know that I'd have it in a case yet. If there's any possibility of the heatsink not making proper contact, you can always test the MB on several pieces of cardboard (to raise it up so the tails of exp slots are clear) on a table or something.

This is so you can always put gentle pressure on that heatsink to keep it in place- don't want to overtighten it and pull a "Socket 478!"

Now I confess to "testing" intel's THERMTRIP by using a wee heatsink from a monitor on a 2.8G Prescott... THERMTRIP... tripped, and the CPU was undamaged.

Boot it up already- you know you want smoke and flames!

[momaka]

Re: after CPU VRM burned, CPU still good?

I removed the defective MOSFET and the Board kinda works again.

Do you think it's safe to use or not?

Depends which MOSFET it is. If it's a "lower" MOSFET and the motherboard has 3 MOSFETS per phase (1 "upper" and 2 "lower"), then it is probably safe.
But if that is not the case or if you have removed an upper MOSFET, you're probably better off disabling the entire phase vs. leaving some parts of it still connected. It all depends on what MOSFET drivers the motherboard uses, though. If the drivers don't have much protection to determine that a MOSFET is gone (O/C), I think that can create problems as the other working MOSFETs are switched ON/OFF.

I don't know that I'd have it in a case yet.

Agreed.

I test plenty of motherboards without a case... usually on a plastic box about the size of the motherboard - that way, I can add cards in the slots too, if needed.

As for the heatsink - you don't need any pressure at all on it. Just lay the board flat and the heatsink weight will be enough. Just make sure not to knock it down by mistake when the board is running or to have a wire push it off the CPU. I don't even add new thermal compound on my "dedicated" test heatsink - it has a thick cruddy layer on there already, which I never change or remove. Good enough for testing, though. I haven't had any CPUs overheat so far. Most CPUs will run only 5 to 10C higher that with a properly installed heatsink with fresh thermal compound.

don't want to overtighten it and pull a "Socket 478!"

Or socket 775/1156/939/AM2/AM3 for that matter. I've seen warped boards with all of these. For AMD, it's the bracket behind the board that always warps with the board. For Intel, it's those stupid plastic push-pins... which also have the tendency to break after a few installations. I wish they brought s462/370 mounting mechanisms back - now those were a lot more legit.

Now I confess to "testing" intel's THERMTRIP by using a wee heatsink from a monitor on a 2.8G Prescott... THERMTRIP... tripped, and the CPU was undamaged.

Well, I forgot to put a heatsink whatsoever on a 3 GHz Prescott HT (but I did have the CPU fan attached), and the PC displayed the boot screen only for a second before locking and and emitting a continuous beep. Found my mistake, corrected it, and off it went working again.

left asus to start ass-crock

I like that name.

Far as I can tell, they may as well be the same- are the "backwards" capacitor silkscreens any hint?

Yes, they are the same (or were at some point). Ass-Crock (there, using the new name you gave them ) is/was ASUS's "experimental" division. And indeed some of their motherboards were quite crap while some others were rather interesting. I think my 939Dual-Sata2 could fall in the latter... if it wasn't such a finicky piece of turd, though.

[kaboom]

Re: after CPU VRM burned, CPU still good?

As for the heatsink - you don't need any pressure at all on it. Just lay the board flat and the heatsink weight will be enough. Just make sure not to knock it down by mistake when the board is running or to have a wire push it off the CPU. I don't even add new thermal compound on my "dedicated" test heatsink - it has a thick cruddy layer on there already, which I never change or remove. Good enough for testing, though. I haven't had any CPUs overheat so far. Most CPUs will run only 5 to 10C higher that with a properly installed heatsink with fresh thermal compound.

Unless it's a 100W+ Prescott device!

Sometimes, the heatsink moves as the compound heats and flows, especially with little heatsinks and lots o' heat.

I wouldn't push my luck with "naked" S370s/462s though..

Or socket 775/1156/939/AM2/AM3 for that matter. I've seen warped boards with all of these. For AMD, it's the bracket behind the board that always warps with the board. For Intel, it's those stupid plastic push-pins... which also have the tendency to break after a few installations. I wish they brought s462/370 mounting mechanisms back - now those were a lot more legit.

Along those lines, I know it's no longer possible with high-density, high-bandwidth stuff today, but I miss thru hole CPU sockets. Especially now with RoHS BGA ones.

Well, I forgot to put a heatsink whatsoever on a 3 GHz Prescott HT (but I did have the CPU fan attached), and the PC displayed the boot screen only for a second before locking and and emitting a continuous beep. Found my mistake, corrected it, and off it went working again.

So you didn't "netbust" the netburst?

I like that name.

Well, either name is shorter than "what's the lousiest junk we can make, that at least some will buy?"

Yes, they are the same (or were at some point). Ass-Crock (there, using the new name you gave them ) is/was ASUS's "experimental" division. And indeed some of their motherboards were quite crap while some others were rather interesting. I think my 939Dual-Sata2 could fall in the latter... if it wasn't such a finicky piece of turd, though.

Wait, it's not a classic like my A7V? Does your 939 have the "crock" of phase dropping too?

[momaka]

Re: after CPU VRM burned, CPU still good?

Sometimes, the heatsink moves as the compound heats and flows, especially with little heatsinks and lots o' heat.

With new compound, yes. But old silver compound (particularly Arctic Silver) tends to go brown and sticky rather than dry out. I wouldn't use anything with baked dry white compound, though.

I wouldn't push my luck with "naked" S370s/462s though..

Me either. Though I've seen plenty of Xbox 360's survive accidental "heatsink-less" tests.

Also, there this old video:
https://www.youtube.com/watch?v=06MYYB9bl70
(song name is Project X by Allister Brimble, in case anyone is interested in these geeky tunes )

Along those lines, I know it's no longer possible with high-density, high-bandwidth stuff today, but I miss thru hole CPU sockets. Especially now with RoHS BGA ones.

Nothing like a thru-hole socket with shiny leaded solder anymore, indeed. Which is why I still keep all of my Pentium II/3 and AMD stuff.

Wait, it's not a classic like my A7V? Does your 939 have the "crock" of phase dropping too?

Nah, none of that nonsense. Southbridge BGA is going bad, I believe. Either Ali didn't disclose the full heat dissipation of their chips or AsRock didn't install a big enough heatsink.
Anyways, the VRM on that motherboard is the only thing that is solid - classic 3-phase design with 3 MOSFETs per phase and toroid-core inductors with medium-to-large inductance (hence, probably much lower switching frequency than modern stuff and many spots for large capacitors).

[kaboom]

Re: after CPU VRM burned, CPU still good?

Me either. Though I've seen plenty of Xbox 360's survive accidental "heatsink-less" tests.

"I didn't leave gaming. Gaming left me."

Also, there this old video:
https://www.youtube.com/watch?v=06MYYB9bl70

Improper use of IR temp gun. Shiny surfaces (heatspreaders and cores) have low emissivity- reading will be lower than actual temp.

What were we supposed to expect with that "test?" Even back in the day, who wouldn't know what would happen? Hey, my username resembles it!

Nothing like a thru-hole socket with shiny leaded solder anymore, indeed. Which is why I still keep all of my Pentium II/3 and AMD stuff.

That.... and win98!

Nah, none of that nonsense. Southbridge BGA is going bad, I believe. Either Ali didn't disclose the full heat dissipation of their chips or AsRock didn't install a big enough heatsink.

Do you know the TDP of the SB? I'm curious now and will probably look this up. I take nothing for granted; whatever it is, I'll double it. I wonder if those SBs commonly fail..

Anyways, the VRM on that motherboard is the only thing that is solid - classic 3-phase design with 3 MOSFETs per phase and toroid-core inductors with medium-to-large inductance (hence, probably much lower switching frequency than modern stuff and many spots for large capacitors).

1-hi, 2-lo, per phase? Sounds like something "borrowed" from a more sensible design...

[kc8adu]

Re: after CPU VRM burned, CPU still good?

my understanding was that asrock is asus economy line.

[momaka]

Re: after CPU VRM burned, CPU still good?

Improper use of IR temp gun. Shiny surfaces (heatspreaders and cores) have low emissivity- reading will be lower than actual temp.

That's true indeed. Thermocouple would have shown a much more accurate temperature. Most likely, those Pentium 3 and 4 CPUs probably throttled down somewhere around the 60-80C mark.

What were we supposed to expect with that "test?" Even back in the day, who wouldn't know what would happen? Hey, my username resembles it!

I think I heard, that apparently it was acceptable in the older days (socket 7 and possibly earlier) to try a quick motherboard POST without installing the CPU heatsink, just to see if the system works at all (like, for example, to see if the motherboard works in the store before completing a sale). Those older CPUs, of course, didn't exactly have a very high TDP, so I think that's why it was possible to do this.

This is also why I think some motherboard manuals from the Pentium 3 and AMD Athlon era were so specific to never test the motherboard without a CPU heatsink, even for brief periods of time.

Do you know the TDP of the SB? I'm curious now and will probably look this up.

No idea, but it is an ULi M1567. NB runs very hot too, and it is a ULi M1695.

1-hi, 2-lo, per phase? Sounds like something "borrowed" from a more sensible design...

Yes.
It's very likely an (older) ASUS design. If there is one thing ASUS does fine, it it CPU VRMs. For example, the P5GC-MX has only 6 MOSFETs total, and those motherboards are actually designed for Pentium 4 and Pentium D CPUs, including the 130+ Watt TDP Pentium D mini-heaters. Go figure.

my understanding was that asrock is asus economy line.

I think that applies only to the newer stuff nowadays. The older AsRock boards really did have some interesting features (like my 939Dual-Sata2 having AGP 8x and PCI-E 16x, both being full-speed slots and also a "future_CPU_port" slots for a special add-in card to support socket AM2 CPUs and DDR2 RAM).

[Stefan Payne]

Re: after CPU VRM burned, CPU still good?

Was able to test Windows today.
And everything seems to be working perfectly.

All 6 Cores of my E5-2630L are working, though it's only a 60W TDP CPU...

And no, you can not overclock this CPU
Tried it on two boards (my other X79 Board is a MSI X79A-GD45 PLUS with a i7-3820)...

[kaboom]

Re: after CPU VRM burned, CPU still good?

^- Nice to hear!

Good luck with it.