K7D kinda burned

[spldart]

After 1.75 years of nearly constant uptime I all of a sudden developed an instability the day before yesterday. This is my dually workstation with a pair of Barton mobile chips modded and overclocked in it. All the 5V lines burned away :eek: The supplmental 4 pin 12v plug was fine. I ended up hardwiring the 5v lines to the powersupply and carved out the burned areas of the atx plug. Also had to replace one cap in the powersupply that failed.
Mobo MSI K7D dual socket A, powersupply Antec True 430. I'm posting from it now ;)

http://67.10.210.195:91/burn1.jpg

http://67.10.210.195:91/burn2.jpg

[dustybyrd]

what speeds/voltages were you running those mobiles?

i had an antec tp550 and dual mobiles at 2.4ghz and one of the caps started leaking inside the psu and the 5v line dropped to 4.6-4.8...

these cpus probably still suck hard on the 5v line, even if there is a 12v 4pin connector...

[spldart]

what speeds/voltages were you running those mobiles?

i had an antec tp550 and dual mobiles at 2.4ghz and one of the caps started leaking inside the psu and the 5v line dropped to 4.6-4.8...

these cpus probably still suck hard on the 5v line, even if there is a 12v 4pin connector...

They are at 1.5Vcore and 2304mhz apiece. The cap that failed in the power supply was a 2200uf at 16V.

I had this happen once before in a maxed out old Abit with socket 7.

[fdoggiedog]

Funny I had this same exact thing happen to my K7D about two weeks ago. Only it really burned the last two in the atx connector. I also was using an Antec power supply but mine is a TP550. My board has no other damage other than the atx connector.
I was running two 1700+ XP's with the L5 mods at 2.3 with 1.75V.
When did you buy your board?<edit> I guess I should have read your post a little better. <edit> I got mine in Feb of this year.

[spldart]

Bought mine 1.75 years ago with 2 1700 tbred b's. Ran those for a year at 2ghz then bought the barton mobiles and ran those at 2.3+. Now this happened. I'm still stable so I assume the board has no other problems.

[dustybyrd]

i guess these boards were never meant to deliver that much power to these overclocked and overvolted mobiles...

my old asus a7m266-d board also had an additional auxillary power connector and the 4 pin 12v connector...i used all three power connections to the mobo and never had a problem at 2x 2.415ghz and 1.775v for 6-8 months and before that for 6 months at 2.1-2.3ghz with 2100's and 2600's and 1.75-1.85v...and actually before that i think the original owner had it for a year or two at 2.1ghz and 1.65v

[Mr B]

i guess these boards were never meant to deliver that much power to these overclocked and overvolted mobiles...



That makes no sense though.

Old Palimino based MP chips have a much higher default vcore than the 1.5v he was putting to the Bartons he has. I believe they're the same as the 1600+ Pally's I modded for SMP, which is 1.75v default.

If what you say is true, I would have fried things long ago. So would most everyone else (using Pally based chips).

[dustybyrd]

That makes no sense though.

Old Palimino based MP chips have a much higher default vcore than the 1.5v he was putting to the Bartons he has. I believe they're the same as the 1600+ Pally's I modded for SMP, which is 1.75v default.

If what you say is true, I would have fried things long ago. So would most everyone else (using Pally based chips).


no...it's not the voltage, per say, that causes the problem...it's the increase in voltage on chips that weren't designed to run at that voltage...and more importantly, the speeds that can be attained with the thoroughbreds and bartons are much higher than the palominos...

those 1600+ pallys used 65 watts max a piece at that speed and voltage...but 2.415ghz bartons use about 105 watts max a piece at 1.775v...

that's 80 more watts the board has to supply to those chips through the same wires in the ATX connector...

[spldart]

But what do Mobile Bartons pull at 1.5vcore?

[dustybyrd]

But what do Mobile Bartons pull at 1.5vcore?


if you ran them at 2.3ghz and 1.5v then they suck 70 watts a piece max...

is that what you did? if so, those were some good chips

[spldart]

is that what you did? if so, those were some good chips

They are at 1.5Vcore and 2304mhz apiece.

;) :p

[spldart]

http://67.10.210.195:91/tempvolt.jpg

[dustybyrd]

did you ever try to really push those?

could you get 2.5ghz out of them with more voltage?

[spldart]

Nah. Never really tried to push them hard.

[dustybyrd]

and i guess now might not be a good time to try ;)

i don't know if i could stop myself, though :D

[spldart]

It would be a fine time if I were up to it. The repair is still doing great and it's solid. I'm just happy with 2304 apiece and frankly don't want to take chances with the setup. If I did blow something up it costs money, I might not get it as stable as it is now, I might not get it back at all.... The risks aren't worth an extra 2 to 350 hund mhz.

[Albigger]

Glad to hear the repair is holding up so far.

I just want to point out that this seems like a more common occurrence, and not just on the K7D. My good friend has an intel rig and this happened to the extra power connector on his 9800 AIW, except he caught it when it was black and charred but had not yet melted away any of the connectors.

So, it seems that the connectors currently in use may not be quite rated for the amperage going through them? Is there anything we could do (short of soldering wires directly to the motherboard) to help overcome this issue?

[spldart]

I was thinking some dielectric grease would be a good idea.