Precision 7550 & 7750 Owners' Thread

They just don't care. "Within spec" or whatever they say. 113C is ridiculously high for the chipset.

Actually, my 7540 had the fix for the pch temp. They put a fat thermal pad (or cube) to bridge heat to the palmrest. It seemed to work well (either that or I never plugged in any TB3 devices so I wouldn't really know).

Sent from my iPhone using Tapatalk

 

I've always wondered why on Earth notebook manufacturers leave the PCH naked. Is it that hard to route a tiny heatpipe and heatsink over it?

The worst problem is that using Dell's fan algorithm, the PCH temperature also contributes to fan speed, so even if the CPU and GPU aren't doing anything much, but the I/O is being hammered, the PCH temperature jumps, and the fans ramp up, too.

It's also not like the fans help with this, anyway.

 

Have you seen the MacBook Air from the last 2 years? Fan not connected to tiny heatsink. 100C all day


Sent from my iPhone using Tapatalk

 

Meh, Apple has never been any great shakes at cooling anything, and they'd sooner let their notebooks thermal-throttle than have any semblance of fan noise. I'd much rather a Precision that I can at least open up and fix myself (still a shame that I have to, but oh, well) than a MacBook with proprietary screws and a glued-down battery.

 

So I've left PCIe Link State Power Management on maximum (default) and just let it be. The PCH runs hot, whatever. Nothing I can really do about it.

I did however make an interesting discovery. Arctic MX-4 outperforms Kryonaut!



I made sure to test both on a flat surface and on my stand. It makes practically zero difference because the temperature is now dependent on how well the heatsink is mated to the silicon, not how much airflow there is (there is plenty).

MX-4: 85C @75W
KPx: 88C @75W
Kryonaut: 87C @75W (pitted die)
Gelid: 89C @75W (scratched and pitted die)

Two of the pastes did scratch up my CPU and GPU. Not too happy about that.

This machine has good mounting pressure. Apparently less viscous is better in this scenario. We'll see how it holds up but longevity isn't an issue for me at all since I sit at my desk all day and a repaste takes +/- 5 minutes now. The new chassis design is actually growing on me due to how easy it has become to open up the back and change the paste. I like messing with things

Some good news for 7750 owners concerned about CPU performance:
Over at the 51NB forums, it was revealed that the 7750 would have a PL1 limit of 105W compared to the 7550's 75W (in Ultra Perf. mode).

 

Interesting / disappointing. I guess I haven't had any systems that support modern standby. I don't remember seeing anything about it in the Precision 7510 or 7530 BIOS.

I've been using MX-4 when I need to repaste in my M6700. I've seen benchmarks that claim that Kryonaut is better but apparently they have poor QC, with some good batches and some bad batches, or something?

 

Wow... That's exciting. Do share your first impression and thoughts with us. Btw, there is not a single review on YouTube about 7750...

 

Kryonaut is an enthusiast grade paste. They really aren't formulated well, and by well, I mean practical and reliable.

This is the third tube I've purchased from three different sources, and all three have abrasives! The price to amount you get is really awful and it's nothing more than a gimmick. If 1-2C better (as tested in those desktop test bench setups) is really all that important, it means the heatsink and fan setup isn't good enough to begin with.

Most laptops are designed to never be taken apart. I've taken this 7550 at least 10-15 times already but I would imagine that most others would have had a few broken clips or damaged plastics at this point. Thankfully, mine has no issues since I try to be extra gentle with disassembly and reassembly.

KPx, Kryonaut, Gelid etc are all a PITA to apply. Some are too gummy, some are too sticky, and others have the solids separate from the silicone oil. I'm clearly a little frustrated at these pastes

When I applied MX-4 I was really surprised how easy it was to work with. It's more "runny" but that's what I was looking for after I realized that the less paste there is in between the top of the CPU and the heatsink, the better my temps were.

I think the whole pump-out issue is overblown. The issue really boils down to using crappy expensive paste that is designed to last long enough to post top scores by PC hardware reviewers and that's it.

EDIT: Here's a photo of the CPU die. Applied is MX-4. You can see the scratch and pitting damage done by Kryonaut in the previous application. So much for that...

 

In that review, the owner didn't take any performance review yet. The comment claiming 105w PL1 was not confirmed by any test nor referred to any validated source.

 

That is true, but the owner quoted the correct PL2 limit of 135W. I figured that the only way to find that out is to have HWiNFO installed or some other app that is able to read the set power limits. I guess the only thing to do is to wait until someone here gets their hands on a 7750.