2 Phase to 3 Phase VRM

[Per Hansson]

Re: 2 Phase to 3 Phase VRM

Alright update time... I downloaded the Distributed.net client and ran it for 10 minutes...

My tempsensor now displays "HHH" this means the temp is over 100°C, even more amazing is that it shows temps 10°C too low (10°C in a room with over 20°C temp...) so it is infact over 110°C hot there now...

[dood]

Re: 2 Phase to 3 Phase VRM

You need to get yourself one of these bad boys... http://www2.towerhobbies.com/cgi-bin...&I=LXEMG5&P=ML

I used to use one for checking the temps of my RC car engines. They'd be great for what you're trying to do... and much more accurate.

[Per Hansson]

Re: 2 Phase to 3 Phase VRM

dood; the tempsensor I've got now is really accurate. It's this one.

After calming down a little I brought the system back down to 1.6v and 2.3ghz... Now the MOSFETS are running about 73°C...

I was thinking that perhaps the reason they get so hot is because the caps have started to fail, they are not visibly bulging but I still think something is wrong for them to get this hot... If they indeed have failed or degraded that would cause the MOSFETS to have to do more work and thus become hotter...

[Rainbow]

Re: 2 Phase to 3 Phase VRM

Yes, bad caps can result in hot mosfets. Even so hot that they burn and desolder from the board or even worse (I saw a picture of GigaByte board somewhere with mosfet fried through! the PCB)

[gonzo0815]

Re: 2 Phase to 3 Phase VRM

you should consider the substantiall higher ripple current in dual phase vrm too. So, if you drain mutch more power as it is desingt for, the specified ripple current for the inputcapacitor can easily exceeded. Even good caps won`t witstand them for long time. 3 phase vrm desing need s mutch fewer caps with lower ripple current rating and they will therefore last longer.
But with newest improvements in capazitors, like KCZ and comparable series, the dual phase circuit will be sufficient for all socket A CPU, if properly designt.
If i ncompare my losy K7s8xe+ vrm with only 4 x 3300 caps, for output, and 4x 1500uF input the you could feel comfortable with 8x3300uf. Especially the inputcapacitors are realy close to the ripple current specs. Therefore i will replace them, even if they are good ucc ones.

[999999999]

Re: 2 Phase to 3 Phase VRM

In many overclocking experiments, I had added a jumper wire on the back of the board when it was a 5V rail supplying the VRM subcircuit, between the 5V socket pins and the VRM input inductor leg. Typically that jumper was 12-14 gauge, standard solid core household wire I'd epoxied to the board for durabilty / permanence. It tends to result in at least 0.2V higher supply to the VRM at full load, unless the ATX connector was moved very close to it by the board designer.

Even then, if you have one of the last generations of CPu running from a VRM off 5v rail, you will see large current and pseudo-accurate software/bios measurements of voltage. All you can do then (reasonably) is measure the PSu voltage at the connector to determine if it's the PSU or motherboard falling short of the goal.

[Super Nade]

Re: 2 Phase to 3 Phase VRM

Hannson, put the temp probe on the VRM inductors. You may be in for a surprise