Need expert advice - powering a Q9200 CPU

I have put a Q9200 into my HP 8530W, which is only the dual core CPU version. There is a quad version laptop/mobo available, at a price!

It works fine*, but I have found that the proper quad core mobo actually has a 3rd PSU circuit, whereas mine only has 2, it still has all the pads for the missing components. I have had some issues with BSOD on XP boot, probably due to the mobo not being able to supply enough power (* details here). So...

I have pinmodded the CPU to drop the default voltages by around 10%. By my calculations, that should give me the headroom I am missing. But I would really like a second opinion...

The intel datasheet for the mobile quad says the PSU should be designed for 64Amps, the so called "ICC for Processors Recommended Design Target" .
The datasheet for the supported dual cores lists this value as 60Amps.
So given that the Dual core's 60A is nearly 94% of the quad's 64A requirement, I should be good to go with the 10% undervolted quad, no?

 

Well, for basic math, yes it might be, after all if you really achieved 10% then that's a good one. But voltages and stuff have these habits of not always being straightforward, so you might have to research further into the topic.. (too tired to go googling now, lol)

 

I would

Just because it would be the most fun thing to do.

 

I may just have to do the scavenge thing out of necessity too...

The undervolted quad is much more stable than it was, but I still get the BSOD on boot when the laptop is stone cold, second and subsequent boots are fine, even boots when it is very very hot are good now. But...

I was thinking about it a bit more, and it struck me that the core2 quad can put a massive transient load on the PSU compared to the dual core2. eg. if they are idling, with only one core switched on, then suddenly they get the signal to enable all cores, the quad then enables 3 times the load that the dual does (3 more cores as opposed to 1).

Looking at the datasheets some more, this seems to be confirmed by the Icc_core_inst rating. Which seems to be for transient power requirements. This is listed as 85A for 35us for the quad, but only 57A for 35us for the dual.

So it looks like I'm gonna have to get hold of a SMD rework station and have some fun!

 

Most cpu run at stock speed even when undervolted by a good margin.

If you have it running stable w/ undervolt now I wouldnt worry about it and enjoy the power/heat saving.

 

Yeah, I've had good undervolts in the past on other dual C2D's, but only using RMClock, this is the first time I've VID-pin modded one. You are very limited in what you can do when pinmodding, since each mod affects all other VID levels.

In this case the VID mod I wanted didn't work, since IDA mode became too low (@1.1V) to be stable. The next available option was IDA @1.1500V, HFM @1.1125V, LFM @0.9500V. Once into windows I can set them how I want, but for booting, this is the lowest I could go and still be stable with the pinmod options I had. It's the HFM that is still a tad higher than it could be with this setup.

 

In case anyone finds this post, I should add some new info.

I tried it on a different mobo, still a dual core mobo, and it is working fine, so there must be something iffy about that first mobo I tried.