*** XPS 15 7590 Owners Lounge ***

I just walked through Costco a few moments ago and played around with the x1 Carbon keyboard. I really liked that over the Dell XPS 13.... makes me want to stray from Dell...

Well, it looks like the X1 Extreme gen 2 is up for order. Interesting that they dropped their OLED display and their 8core cpu options:
https://www.lenovo.com/us/en/laptops/thinkpad/thinkpad-x/X1-Extreme-Gen-2/p/22TP2TXX1E2

Looks like it says coming soon at the top, and you still aren’t able to order one at present.

 

I can't see why Dell didn't include the upgraded heatsink for everyone. We will have to see how the crippled the i9 performs vs the crippled i7.

The notebookcheck article shows us how Dell upgrades cooling of i9 CPU vrm, which looks quite basic. We can DIY the same solution.

The pictures are fuzzy and I only have vrm schematics of the earlier XPS models but can speculate . It looks like the i9 has additional metal strips and fluffy cheap thermal pads above the 4 CPU power mosfets & maybe 2 power caps. I think only the mosfets are generating significant heat so would ignore the power caps but check with a scope or other more "basic" heat sensors.

For DIY, a simple metal strip shaped like an "L" with fluffy thermal pads linked to the heatpipe could get one close to Dell's engineered solution. Just make sure the metal strip is well adhered and does not move and has no chance of shorting out the board or any components during transport lol.

 

It is surely possible to DIY sth like this, it might work also in older XPS models. Asuuming that the CPU and the GPU don't get throttled more.
I'm curious if the two versions of the heatsink also differ in the heatpipes and fins. I recall Chinese were selling enhanced (double heatpipes, copper fins) aftermarket heatsinks for the 9570, but I don't remember any proper assessment. Maybe there's a business niche for aftermarket exhaust for laptops too

 

That's an interesting editorial to write for notebookcheck if you are ever interested

 

Does anyone have one of these in their hands with an OLED display? It says mine will be delivered by 8/30.

 

got mines earlier this month. i'm not sure if this is oled by design but on black background there seems to be dark areas like the backlight isnt working and only on dark backgrounds if i bring up like a browser with white background or any other color all is normal.

 

That's perhaps because an OLED has no backlight at all. That's the main advantage of OLEDs, each pixel is an LED in itself. So they can turn off when displaying black, giving far better blacks.

 

First Findings | Dell XPS 15 7590 OLED/i7-9750H/16GB/GTX 1650 in review
Douglas Black, 07/24/2019, notebookcheck.net
https://www.notebookcheck.net/First...7-9750H-16GB-GTX-1650-in-review.427940.0.html

 

The CPU performace regression is rather indicative.

@Eason I loved your review.

I believe the CPU performance regression and the GPU downclocking to 300Mhz are rather scary findings.

People (rightfully) complained in the past that VRM modules were getting too hot. However, Dell's solution to **simply** extend the current heatsink to cover the MOSFETs will do more harm than good.
The cooling solution now has to transfer and dissipate more heat (CPU, GPU & MOSFETs) but the radiators' size and the fans are the same.

 

I'm probably going to have to go with LM for this laptop.

 

@clayton006 LM won't do much because the "thermal bottleneck" are the fans/radiators not the thermal conductivity between CPU die and heatsink.

I have repasted my 9570 several times with Kryonaut and I noticed only a couple degree improvement over the stock paste.
Moreover, the heatsink warps easily and it's very hard to get a good/even contact between CPU die and heatsink. This is particularly important for LM applications.

 

I've had mixed luck with Kryonaut. I will probably try phobia nanogrease extreme first to help with contact and see how that does.

 

If someone cooks up a modified heatsink with extra fins/radiators like the ones people were buying from China, the VRM modification Dell made might be more usable.

 

I noticed that with my HS as well. The CPU side is naturally curving a bit away from the board, and this goes all the way to the radiator on the CPU side. The heatpipes were also a bit dented, before I even touched it. Hopefully, that doesn't affect the functionality of the heatpipes.

After ~4 applications, I found my best results (w.r.t. temps) was when I did:

• apply heavy amount of thermal paste (a tiny bit less than pea-sized, which is a lot for these little dies)
  • This is more than I usually apply on a desktop, but this isn't a desktop heatsink
• not apply any pressure at all to the heatsink, specifically the parts over the ICs/dies. The only pressure came from the screws/mount

I did this because I found my previous applications didn't even seem to have any thermal paste residue, and the laptop was throttling, even with a ~100 mV undervolt (at the time, it was applied through XTU; I later switched back to ThrottleStop). The lack of thermal paste residue is not normally alarming. However, that really is only true for computers with good heatsink mounts and large, flat heatsink contact/interface surface. This XPS 15 heatsink was bending before my eyes, and the paste didn't even seem to make it to the entire die.

Now, with the same 100mV undervolt (100.6mV in ThrottleStop), the laptop doesn't bounce off the 100C thermal limit under CPU-only loads anymore. It's around a stable 95Cish in CPU-only loads. I've also applied a max ratio limit (effective frequency 3.9GHz 1-2 cores, 3.6GHz 4 cores) and a long term power limit of 45W in ThrottleStop (this doesn't seem to stick or work, according to HWINFO and XTU). With those settings, the severe throttling (in gaming) seems to be gone. It still reports throttling, and BH PROCHOT is still triggered, but it's not the 15fps lagfest it was before.

Again, I know the quad core die is probably the oldest and least refined of the Coffee Lake configurations. But if even a quad core takes this much effort to tame, I really have to tip my hat to those who are wrangling the 6 and 8 cores. Maybe they are just that much more efficient than the quad core. Who knows?

 

My guess is that by using a temp alarm in Throttlestop of around 20 DTS (CPU) and 73C(GPU) set to trigger a lower-clocked profile will work to control temps and keep the GPU from downclocking to 300 mhz only. Maybe set the max turbo limits to 3.2 Ghz or so on the 2nd profile you have the alarm trigger.

 

That is a good choice. I got good results (also more durable ones) with IC Diamond.

 

Kroyonaut breaks down above 80C. Excellent choice for water cooled loops that never go above 60C, but in laptop bouncing of 100C it will deteriorate in matter of weeks. Or maybe even days, not sure, it took me three weeks to look at the manufacturer website and forums after I noticed 20C higher idle temps as compared to immediately after repaste.

IC seems to last forever, and it is dense, kind of helps holding the fragile,bending laptop heatplate in place.

 

It seems to last a long time IME. On my 2007 MacBook Pro, when temps started to rise, I knew it was time to change. I repaste every say 2 years, but all 8600GT graphics cards are defective so run extra hot, pushing the thermal paste to the limits.

IC can scratch so I don't use it on newer equipment.

 

Do you have a source for this? I've ran Kryonaut on many laptops/desktops and never had such a thing happen...

 

From their own product page:

https://www.thermal-grizzly.com/en/products/16-kryonaut-en
"Kryonaut uses a special structure, which halts the drying out process at temperatures of up to 80° Celsius."

Sounds like something that Apple marketing would come up with.

 

Really that doesn't back up your statement, had you said "Kroyonaut can dry out at temps above 80C" then that sounds like a reasonable assumption.

I've never had kryonaut appear dry, even after many months of use/abuse. They also state their temp range as "-250 °C / +350 °C."

Just for fun I checked my wife's laptop and even though it's had kryonaut on it for about 18 months, Prime95 temps are roughly the same as when I first tested it (actually 1C cooler, but that's within the margin of error).

 

Whatever you call it, dry out, break up, something bad is happening above 80C, otherwise they wouldn't put this in writing.

Mine went to crap after 3 weeks, but I pushed it daily above 90C for couple of hours, averaging probably 95. And when I opened it again the paste look normal, not dried out. I probably wouldn't notice this on my old 9550, I don't remember seeing more than 80C on it, but got myself 6 core GS65, removed all the power brakes and let it run at thermal limit - and it quickly killed the kryonaut. The GPU was fine, exactly the same temps and benchmarks, but that rarely went above 70C.

EDIT - one additional observation, the spread between cores become very big, initially it was about 5C difference between them, after 3 weeks it was like 15. I can't really say whether that was due to molecular changes in the TIM due to high operating temperature, or flimsy MSI construction. Or combination of both. But that was at high load, at idle all cores were close together around 50C, while before it was 30C.

EDIT2 - I did notice that the die and heatsink were scratched after I removed the kryonaut, didn't really give much thought to it, but just now found this: Kryonaut abrasive?? Bad batch, not grinded good enough, company has no way to track and recall individual ones so just keeps them all on the market. I'll send them an email and ask about this and the rapid TIM degradation at high temps. Word of advice - don't buy Kryonaut right now.

 

@custom90gt I think the problem is that Kroyonaut has a low viscosity and is more prone to the "pump-out" effect in uneven contact scenarios.

 

Personally, I have never had a problem with Gelid GC extreme. Prefer it to Kryonaut.

 

I've heard a couple of users talk about the "pump-out" effect on their over powered laptops, but again I never saw it on my OP15 either. I do always take time to ensure the heat sink is as flat as I can make it though. Also there have been bad batches, so that may be apart of it, I may have just been lucky with the last three 11g tubes that I've purchased. Even with 100's of hours of logged Prime95 and CPU temps >85C, it's never dried out or pumped out for me.

 

Hello,
I bought the XPS 7590 Oled i7. The only problem I got is that bluetooth dissappeared after some time (it worked fine at the beginning). Device manager indicates that bluetooth is not connected (error 45) and BIOS even show that there is no bluetooth. I am guessing that it is related to bios or driver update. Does anybody get the same thing?
Thanks

 

I don't get the same thing, but the Killer BT drivers are just the Intel AX200 BT drivers. Just update them separately from the Killer drivers. In the device manager, uninstall the drivers and check the box that says something like remove software, too. Then install the up to date Killer or Intel drivers (both will work, Killer doesn't do anything to/against the BT drivers).

 

Just saw this

 

Firmware update out for the Toshiba XG5.
https://www.dell.com/support/home/u...M&oscode=WT64A&productcode=xps-15-7590-laptop

Claims to have improved performance. Only posting this since my Service Tag didn't show this as being available, despite having one of the listed SSDs in my system.

 

i have the g60 1tb so i guess it doesn't apply to me.

nope but i ran the dell assist support and updated all firmware,drivers, softwares.

 

Had to give it a 3/5. https://www.ultrabookreview.com/29145-dell-xps-15-7590-oled-review/

Dell won't be happy, but they need to wake up.

 

Sorry sir I can’t give you more than NBRs forums max reps. +5. We need much more honest reviewers. As long the others give it ok reviews nothing will change. Thanks!

 

Thank you! I really feel like whoever has been in charge of product testing and QC just stopped caring a long time ago and needs to be replaced. Just my uninformed guessing opinion.

 

What Azor the CEO for Dell’s Gaming brand throw out in Twitter messages was just a huge lie. (Not so long after he went to AMD). Only to sell more crippled garbage and scam new/old buyer/customers! Just sad. I wonder when people will discover this. Maybe when they are on the hook on no way back? See... Outside return window.

 

That is a pretty good rating considering Dell simply has been recycling the same chassis for 4 generations.

 

Someone send me one and I will review every little issue of it it with all the hate and anger I have left.

 

Thanks for the review! That's a bummer, I was really looking forward to this iteration. Gave my 9550 to my brother in law, GS65 to my son (he's 5 years old, but at least I can buy another 15 inch without my wife going bananas). If I wasn't going for vacation I'd order Precision 5540 (to have 8 cores and 2.5 inch drive). Some questions:

1. Did you try a clean install, without any Dell software? That GPU throttling is insane, would you be able to check if that's hardcoded in VBIOS or Dell Power Command/some Dell driver is causing it?
2. OLED screens all seem to have that banding effect, so I wouldn't hold my breath that you will get one without it. This plus image burn in after a while. And it doesn't even have to be a persistent image for a long time (like with plasmas or LCD), just frequently in the same spot. I'm the FHD guy, don't like trouble.
3. Just my 2 cents about the amount of thermal paste. It looks like too much is not a bad thing, but bad quality is, most likely you saw improvements after repaste because of better TIM quality, not that Dell flooded it. Too little is definitely bad, so it is actually better to err on the side of 'too much' especially in case of flexible laptop chassis with three screws on heatplate as opposed to lapped desktop heatsink.
4. Your CB15 results, the repaste+undervolt is awfully similar to only repaste, like within measurement error. I guess it is the power limit causing it, would you be able to repeat this with power limits removed? I honestly don't know if and how you can do it on a Dell. If it is not possible - what about the temperatures? They should be significantly different.

 

#4: The first CB run didn't change because it hasn't warmed up enough to saturate the cooling solution by that time in both scenarios. However, the undervolted + repasted results are much improved in the 10-run CB loop vs the repasted or stock results only I think. I'll see if I can figure out what kind of throttling it engages in

 

Best non-conductive thermal paste is 1 to 2 orders of magnitude worse thermal conductor than the metal of the chip and the heatsink. If the gap is wide (poor alignement, bent heatsink...), thick paste is better than air, but it isn't good. If the gap is reasonably thin and even, the problem with lots of "thick" paste (printed in rectangles on the heatsink) is that it doesn't get squeezed out of the gap under normal screw pressure. The mess around the chips may look ugly but is not an issue.

 

https://www.notebookcheck.net/Dell-...590-to-fix-GPU-throttling-issue.428456.0.html

well well well!

 

just found this:
https://www-techradar-com.cdn.amppr...date-messed-with-your-pcs-audio-heres-the-fix

maybe it will solve the latency problems with dell's laptops?

 

Everyone who owns a 9570, please open a ticket at Dell about this. This cant be allowed, they "fix" this on the new model, and leave the 9570 in the trash. The 9570 throttles to 300MHz too which is unplayable at 75°C. They could do the same, throttling to around 1150MHz which would still give decent performance in most games.

 

Are you leaving the score the same on your review?

 

I opened a ticket today as I am having issues and I am only playing a video online, emails and surfing the net the latency is really high video and audio will cut out. Dell sent an email to install Intel Dynamic Platform and Thermal Framework, restart, run the diagnostics and it should fix the issue but it made it worse the fans are on most of the time and the latency is worse that it was before.

I emailed them back with some samples of the latency and am waiting for a reply back as to what they will do.

 

Dell XPS 15 7590 System BIOS 1.2.3

Download = https://downloads.dell.com/FOLDER05...-438f-8845-5813e7302380&fn=XPS_7590_1.2.3.exe

Fixes & Enhancements
- Enhances the performance of the system.

 

No, but I will make a comment that the same issue is present on the 9570.

"On this topic, however, I have been told by a number of 9570 owners that a similar throttling issue as originally present on the 7590 still exists on the 9570. That is, the XPS 15 9570 with GTX 1050Ti exhibits the same throttling to 300 MHz at 75C that the 7590 initially did. I find it strange that the XPS 15 9570 has been out for a year or so and had this problem for months without any resolution, and that makes me more than a little bit wary of Dell's commitment to supporting these machines."

 

Does it have a space for a 2.5 inch SATA SSD?

 

Versions with the 56Whr battery should, but options for it haven't shown up in the Dell configurator. Those options have shown up for the Precision 5540, a closely related variant of the 7590..

 

I'm curious if the new bios actually fixes the GPU throttling during gaming. If I didn't pick up this Aero15x, I would consider the 7590 if it actually works...

 

I use a 87w usb c charger. This forces the dell laptop not to reach the full gpu potential but at the same time, it will never reach the 75c throttling point. I will get a stable 1200mhz core during my lumion load.