Dell XPS 15 (9550) list of hardware and software problems

If a CPU core has something to do, the clocks should be at 100%. This is where a CPU is most efficient in terms of performance per watt. A fast CPU core can get work done quickly and this allows the core to spend more time in the low power C7 C State. If all of the cores have nothing to do, the entire CPU package can enter a lower power state and save more power. Here is an interesting paper that questions running a CPU slowly.

Power Optimization – a Reality Check
http://people.eecs.berkeley.edu/~krioukov/realityCheck.pdf

A CPU that is slowly plodding along at a task spends longer in the C0 state, consuming power. Much better for the core to spend as little time as possible in the C0 state and as much time as possible in the C7 state where it is disconnected from the voltage rail and consuming next to nothing. In the deeper C States, the reported MHz becomes meaningless. The core is not getting any signal from the clock generator so it is running internally at 0 MHz. Software does not accurately report idle MHz when the C States are being used because it would have to wake the core up to measure its speed. One advantage of ThrottleStop over XTU is that ThrottleStop allows cores to spend more time in C7. This reduces heat and power consumption.

Also, don't use power consumption data calculated by ThrottleStop or any software. This data is not measured power consumption. It is an approximation calculated by the CPU to control the Intel Turbo Boost feature and that is all. It is not accurate when a CPU is mostly idle in C7.

 

So you're saying the CPU staying at the "maximum state" instead of dropping and rising consumes less energy and gets less hot? Is that why you guys recommend us to put min processor state at 100% in advanced battery options? The problem with that option is I'd like my laptop to run at a lower voltage if possible when plugged in as well because it's not like I'm gaming all the time. And lower voltage means the fans don't run as much (or at least that's what it seems like since that's what happens when my dell is not on battery). But if I keep the CPU at max, the voltage stays at about .900-1.00v and the fans run longer

At one point, the voltage for my battery setting was around .600v, but when the funky issue happened, it wouldn't go beneath .8 and no matter what I tried, it wouldn't go back to the .6. I'm not sure what happened.

 

Just applied some coollaboratory ultra on my xps 15. It definitely helped my GPU out but my CPU is staying about the same.

 

Just applied some coollaboratory ultra on my xps 15. It definitely helped my GPU out but my CPU is staying about the same.

 

Just applied some coollaboratory ultra on my xps 15. It definitely helped my GPU out but my CPU is staying about the same.

 

What I am saying is that people spend too much time worrying about MHz and voltage when they should be worrying about C State residency time. Here's an example.



When idle, my CPU cores are all spending almost 99% of their time in C7. In this C State, the core clock is stopped so it is moving at 0 MHz and the core is disconnected from the voltage rail so they are getting 0 volts. When a core is in C7, whatever MHz and voltage monitoring software is showing you is meaningless because it is going to be wrong 99% of the time. You need to eliminate background apps that are interfering with your CPU spending its idle time in C7. Completely exiting XTU is a good place to start.

 

I erased XTU and am using only throttlestop. How do I achieve what you're saying? Right now my cores spend 87+% in C7. And is it good to enable it in battery mode as well as performance/gaming mode? Will it affect performance?

 

I think that the relatively low values for the XPS 15 might be some sort of driver problem. What I am getting in idle is around 95% for the cores (which is already a bit low), but the package itself does not budge much. (C6 2-3%). I found that this is likely to relate to the optimus function of the laptop.

 

I have an xps 9350 and my core c states are similar to yours, but my package states are in the low single digits. Is this an issue in terms of battery life? The only way i can boost the package c states is by disabling the graphics driver..

Thank you for your info and ofc. TS!

 

hey all, so just got my refurbished 9550 from the recent $500 off deal dell was running. got a model with the i7-6700hq/256 samsung pm951 nvme ssd/8gb ram and the fhd screen for $780. i have read through a bunch of the threads here, and i actually did start off by clean installing with gonzo's iso, which created a lot of blue screens until i installed the samsung nvme controller and that helped. all is pretty good now except for the following:

1. fans run extremely loud/constant when playing games. i've been playing fallout 4 and they run at full speed nearly 100% of the time and it's pretty aggravating. this is with a -150 mv offset in throttle stop on cache and core as well, but my guess is that the gpu is causing more of the heat possibly. is it worth it to repaste/put thermal pads on the cpu/gpu or is there some sort of external solution like a cooling pad that would help more overall?

2. the samsung ssd has this odd habit of little micro stutters/hangs from time to time. i think i've read a couple reports of that being an issue with the pm951. is there any solution to this or just a quirk i have to put up with until i can upgrade to the samsung 950 possibly?

3. during fallout 4, i've noticed a lot of times when i'm trying to move and look left (right doesn't seem to be an issue most of the time), that it won't move with the touchpad which is really frustrating. is this a game only issue or is there something up with the touchpad (or mine in particular)? the thing is, during normal usage in windows there isn't any hangs when moving right to left.

anyway, thanks for the help guys.

 

That is a screaming deal.

I did an informal review of 9550 posts on this site for claimed temperature improvements for undervolting & repaste/thermal pads (for both i5 and i7 models). I don't have my notes here but average improvements were typically:

~3-5*C improvements with both repaste & vram repad.
~8-12*C improvement via undervolting.

This was a tiny sample and people used different stress tests and measurements so YMMV. I saw over 15*C improvement doing both on my i5.

Remember the bulk of the effort is around the nVidia GPU (thermal paste & VRAM pads) so I think that should have a better than indicated impact... Not too difficult either

 

yeah, i read through the thread and i did read about some improvements with the fan noise as a result. i have done cleaning/repasting before but never the thermal pads, and i know that you can get some cooling pads for around the same price/cheaper than it'd take to buy thermal paste plus thermal pads. my main issue is the fan noise... just would like a way to deal with that. i don't really think a cooling pad would be all that silent but i suppose it depends on the size of the fans.

 

People have identified 2 major thermal defects that are easy to improve on the 9550 (thermal paste and VRAM pads). I think they are worth fixing as a first step as that is getting directly to the core of some (but not all) thermal issues and have show good results by some knowledgeable people.

I suspect if one VRAM chip is uncooled and running hot, nVidia will throttle all graphics. So the logical step is fixing that pad (or two pads).

You can do a search for a few peoples' reviews of a cooling pad with the 9550. I recall it being marginally effective and maybe worth 1*-2*C after all the above is completed.

 

okay, well that sounds pretty useless then, haha. repaste and pads it is. how difficult is it to shape the pads to fit everything correctly for a first timer?

 

AS you have probably seen, this thread will get you started:

http://forum.notebookreview.com/thr...rature-observations-undervolt-repaste.785963/

Great set of pictures for VRAM issue:

http://forum.notebookreview.com/thr...undervolt-repaste.785963/page-2#post-10168639

Disassembly pics:

http://www.laptopmain.com/dell-xps-15-9550-disassembly/

While changing the thermal paste, I quickly changed the 4 vram thermal pads with Fujipoly 17 w/mk (2x1.0mm and 2x1.5mm for left side by the heat pipes). I think I cut 14mm x 15mm squares but someone measured the actual dimensions and they are posted in the above link. Thermal pad replacement was simple: it entailed removing the old pads, cleaning the area above and below the pads, placing the new pads properly and took just a few minutes. There are some good pictures on this site of the area to get you started.

 

i know that you used 1.0mm and 1.5mm, but given that i'll probably never have another use for thermal pads again, would i be fine buying the 100 x 15 x 1.0mm 17 w/mk sheet and using that for all the vram?

 

Hello,

I have the XPS 15 9550 and have had multiple times the issue that after closing the lid, the laptop doesn't go correctly to sleep, instead becomes blazingly hot. I assume the issue is well known since there is already a long reddit page about the issue ( https://www.reddit.com/r/Dell/comments/3w2t0k/dell_xps_15_9550_overheated_while_in_sleep_mode/). I have installed all BIOS, windows 10, and driver updates.

My questions:
- has the root cause of this issue been found in the meantime?
- Does Dell finally acknowledge this issue, or do they still pretend they never heard about it, every time a user complains to them?
- Is there a known fix for the issue? I am getting really scared of this thing...

Thanks & best,
Matt

 

Hello,

I have the XPS 15 9550 and have had multiple times the issue that after closing the lid, the laptop doesn't go correctly to sleep, instead becomes blazingly hot. I assume the issue is well known since there is already a long reddit page about the issue ( https://www.reddit.com/r/Dell/comments/3w2t0k/dell_xps_15_9550_overheated_while_in_sleep_mode/). I have installed all BIOS, windows 10, and driver updates.

My questions:
- has the root cause of this issue been found in the meantime?
- Does Dell finally acknowledge this issue, or do they still pretend they never heard about it, every time a user complains to them?
- Is there a known fix for the issue? I am getting really scared of this thing...

Thanks & best,
Matt

 

The laptop wakes up whilst in sleep normally due to the network card allowing wake on lan.

 

so i installed another game, mass effect 3, and i noticed the same behavior with the touchpad as in fallout 4. trying to look left while walking/running with the wasd keys is very delayed. given that this issue is only occurring in games, does anyone have any clue what i'm dealing with? do i have a faulty touchpad or is this some other issue?

 

both the 9530 and 9550 do not play well with a trackpad, use a mouse instead.

 

so this is a pretty common thing then? just wondering whether it's worth possibly using my warranty to get a new touchpad installed.

 

No, it is not a faulty touchpad, if it was then all 10 I have had pass through have been the same...

 

haha, got it. did you experience the same issue with the looking left while moving?

also editing this to ask, is there something i have to change to be able to right click and left click at the same time? in mass effect 3, you have to right click to aim but i can't shoot while holding it down for some reason.

 

The trackpad plain and simply is **** for gaming, use a mouse!

I couldn't WASD for more than a second while trying to pan with the trackpad.

 

haha well i'm not a serious gamer so i might just buy a ps4 gamepad to play with. i plan to tackle gta 5 next (yeah, i know, i'm catching up on old games that i couldn't play because my previous laptop sucked) and that probably would play better with a gamepad anyway.

 

Hi,
no, this is not the reason. A) I checked in BIOS, and "wake on LAN" is disabled (and so are all other BIOS wakeup options), and B) this problem is known to be a bug, was described several times here in this thread and also in the reddit thread I quoted.

Does anybody else who knows about this problem, know the answer to one or more of my questions (see below)? Thanks in advance!

 

I found it helped to disable "wake on magic packet" and "patterm match" in the network adapter advanced settings.

 

OK I will give it a try... Thanks.
I'll report back if it made a difference.

Matt

 

If you look at the 9550s thermal thread, the VRAM defect is that the heatsinks do not touch 1 or 2 VRAMs, so larger pads are needed there. Custom90GT has taken apart ~8 9550s and found this problem in every single unit. That, combined with the observations of some other (clearly) knowledgeable people here, makes it clear this is an expected manufacturing defect on every 9550, but just for 1 or 2 VRAMS. So thicker pads for those defective areas are the logical fix path:

http://forum.notebookreview.com/thr...ndervolt-repaste.785963/page-30#post-10262419

There are not a lot of data points here but I saw experienced people using one of two combinations: [1.0mm+1.5mm] or [0.5mm+1.0mm]; I tried both and had virtually the same results.

If you want to save money, try going with slightly lower rated foam in 2 thicknesses. But a laptop's heatsink clamping force is limited and putting the same size dense foam across the board is not an elegant solution to an easy fix; it could have unintended consequences.

I believe that the 17 w/mk pads were quite dense (like modeling clay) so much less compressible than the fluffy stock foamy pads. Some lower rated (less expensive) pads might be more fluffy so more tolerant to depth variation. One guy here complained it was really tough to crush his dense pads from 2mm to 1.5mm.

A Fujipoly spec sheet shows compressability vs thermal conductivity (along with other specs) here:

www.fujipoly.com/usa/assets/files/Fujipoly 2015 Sarcon Catalog for web.pdf#page=13

Also given the poor thermal conductivity of even the best pads, perhaps using the thinnest possible for each VRAM is the best option.

 

The BIOS options are for powered off state, standby is different.

 

Indeed- I don't think you should replace all the vram pads; only the ones that don't touch. The more soft the pad the better, imo, because it allows the pad to compress/expand as needed to fill the gap.

 

Hello,

I've recently purchased the Dell XPS 15 9550 (Core i7-6700 2.60GHZ, 16GB RAM, GTX 960m 2GB RAM, 512GB SSD). I've noticed that after gaming for a while, I get drops in frame rate. I'm guessing its throttling of some kind. Sometimes it lasts for a few minutes, sometimes it persists much longer. So I downloaded Intel XTU to monitor my XPS while gaming and I see my Core Frequency dropping down to 0.80 GHz and Power Limit Throttling at 100% at the time frame that matches when I experience my frame rate drops when gaming. I've noticed that it happens when the Package Temperature reaches 79C, Is this normal?



Thank you.

 

See this post and following comments by none other than the Throttle Stop creator!

 

Thank you.

 

would the fujipoly 11 w/mk with the 1.5 and 1.0mm be a decent option?

 

Just the other day I was thinking that would be the easiest option at similar cost for you...

A few post above Eason just recommended replacing the left one or two with thicker pads and leaving the factory pads on. I think a few of my factory pads ripped when taking off the heatsink, making them junk so you will need to be careful if you follow that route.

 

in some need of some help, guys. yesterday i tried dell update for some reason, it installed 4 things, intel graphics drivers, realtek usb audio, tb3 update and asmedia usb update. after restart... blue screens. no system restore so tried booting into safe mode, nothing. i have a full backup plus incremental backup images made with macrium reflect and the rescue media on a usb stick, so i booted into the rescue environment and restored the latest incremental (not the initial full backup) and still the same.

do i have any other options on how to fix things? maybe roll back the drivers somehow? or am i going about the restore process in reflect wrong?

 

can you get into safe mode?

 

i tried safe mode with command prompt and safe mode with networking and both just rebooted me.

 

ok, I didn't know when you said nothing meant you couldn't get into safe mode at all but that sucks, can you disable the thunderbolt and usb in BIOS and see if that gets you back in?

 

uh, wow, well i tried that, didn't work, took me right back to a blue screen with code 0xc0000185 and then when i restarted it, it was a screen full of pixels and all it does is show pixels every time i start it up.

 

hold D down whilst the laptops fully off then tap power to go into screen diagnostics, if it still does it them i'm afraid it's time to call Dell and explain it's new motherboard time!

 

So after sending my laptop in because it would no longer charge the battery and waiting 7 days, I've finally received it back.
Thankfully the dell tech helped me with the following:
1. Stretched out the plastic piece that holds the service tag door down so that it no longer aligns properly
2. Saved me some weight by not including a bunch of screws (including the one for the SSD)
3. Broke a standoff to a daughterboard
4. Stripped almost every phillips screw head in the laptop
5. Didn't send me my AC adapter back

Spent an hour talking to dell, they really don't care. They offered to send a tech to my house to fix it all, sadly trust levels are low. I asked them to just send me a new refurbished laptop and was denied. I'm sending it back to the depot as I'm heading out on my honeymoon come friday. Maybe when I get back I'll have a working laptop, or maybe I'll waste some more of my life phoning dell again...

 

that's working, it's flashing the white, black, blue, green and red. but after it finishes i still get the screen full of pixels. maybe a bios reset would help?

 

okay so i pressed f2 even though i figured it wouldn't work but now i essentially see there's one visible strip but the rest is all pixels. i can't read enough to actually see anything to change settings though.

 

Just demand a new one. That is just ridiculous.

 

FFS, unfortunately you are not the 1st and won't be the last with a story like this, call back and ask for resolutions team, at worst you get a full refund at best you get it replaced but obviously you have lost all confidence in the service team so they are no longer an option, give the phone support another go as the resolutions team have to be avoided and they can make decisions above the humble line managers.

They got me a 32gb ram, 1tb SSD upgrade to shut me up!

Having worked for a company offering national service to most of the big boys, Toshiba laptops, Samsung TV's & printers, Epson's projectors and printers to name a few I know what a piss poor range of tools you get given to do the jobs, I bought my own high end tools to save being embarrassed on site and charged special kit to the company who tried to scare you out of doing it to save money, Dell employ 3rd party companies who in the UK use self employed paid per job guys. Last one I had was made to use my screwdriver set as is resembled a poundshop special, he was after his own Wera kraftform micro screwdriver set at the end of the job!

Yes it has fixed it before, at least we know the parts are working in the diagnostics. Fingers crossed!

 

is there a key i can hold while powering up to a force a bios reset? because as it is now i can't see anything barely haha.

 

also assuming that none of this works and i'm stuck with this dead, rainbow pixel laptop, what number should i call for best results? the xps support line? or home consumer support? my laptop is refurbished and honestly i'm hoping i can either goad them into sending me a new one and/or something upgraded rather than be serviced, but i'm guessing they'll push service before anything.