XPS 15 9560 owners thread.

Somethings changed as speedfan used to allow you to control the fans, now it shows nothing under fan controllers.

 

Go to Configure > Options > Enable DELL support and then close and relaunch Speedfan. The two Fan controllers should now be accessible. PWM1 is left fan, PWM2 is right fan.

However you will notice that any attempt to control the fans will result in a sudden speedup and then quickly return to the original speed (most likely 0). This is the BIOS or OS regaining control of the fans. On most computers you can prevent this by setting the fans to a fixed speed in BIOS, or in SpeedFan go to Configure > Advanced > Chip (select DELL) > PWM Mode > Manual.

But you will notice the DELL chip on the 9560 does not expose a setting for PWM Mode. So the only way to use SpeedFan with the 9560 is to fix the fan speed through another method (possibly EFI as you have mentioned). Considering that Dell provides an application (Power Manager) that already provides access to hidden BIOS settings (i.e. thermal management mode) I suspect there may also be a hidden fan speed mode setting as well. I will reach out to Dell about this.

 

I doubt dell will help even if they could, biggest bunch of idiots I have come across. I had forgotten about the stake Dell option, it's been a couple of years since I installed it.

Sent from my SM-G920F using Tapatalk

 

Hey, which is it? Please give more info about

 

doesn't work, it enables options for hardware locked features.

Sent from my SM-G920F using Tapatalk

 

Thanks @GoNz0 for the reply

so isn't a good idea install it?

 

Do whatever you want, I already told you they don't work

 

Anandtech's review is up!

http://www.anandtech.com/show/11670/the-dell-xps-15-9560-review-infinity-edge-part-two

 

But they haven't been particularly thorough, haven't spotted power limit throttling.
They did find display gamut problems reported by some (failure to calibrate properly), probably this has sth to do with Dell's regular video driver woes.

 

Wrong. New XTU efi hacks are working as it should be. Voltages stick even after uninstalling XTU. and no xtu services running, and even after 10-15 reboots. Its like permanent undervolt at bios level. LEGENDARY.

 

Well I have the efi settings, it will not change multipliers, voltage control is the same a before the hacks, I can change voltage and it will stick until it decided to reset.

Sent from my SM-G920F using Tapatalk

 

I am not sure if having something that throttle stop can do everytime I restart my computer is "legendary"
If some how you could edit GPU voltage I would be excited.

 

Well if something can undervolt without running any app or service is what i like. I may try finding gpu undervolting but it may not be in bios. Other app maybe

 

GPU voltage is controlled by the bios, someone may have found a work around with MSI afterburner (I believe that is the tool).

 

Hi. Just picked up a XPS 15 9650 7770HQ / GTX1050 / 512 M2 / 4K / 92W batt. I have owned a bevy of laptops over the years and this will be complimenting my Lenovo Thinkpad S1 Yoga 12.5 that I also have. So far I'm very pleased and don't have a wobble, don't have screen flicker, don't appear to have dead pixels, don't have a strange space bar--everything seems great and I honestly can't complain.

I also got it for a steal @ microcenter for $1359 as an open box (display was 1599) and Dell's site right now is $2200 lol. I cherry picked it out of three XPS 15 open boxes that I looked at this one wasn't even removed from the plastic seal so brand new really. Can't go wrong there.

I can live with a bit of ghosting as I'm used to IPS and I don't ask the world of this laptop and I'm just impressed it can even PLAY games like Doom and BF4 which I'll occasionally do a bit of. However, my main reason for getting it was for web development & photography/graphic work. For that purpose it'll work just fine.

I still need to get the colors exactly where I want them but I'm close. I'm very picky and I really ought to purchase a calibration tool. Anyway I'm in this thread for the following input if anyone has any:

Looking for a top-load tight fitting non-velcro (in other words magnets for quiet operation) laptop sleeve for this. I plan on using a sleeve in a backpack. I've been searching but haven't found one just right yet.

EDIT: Correction... I blame being tired but I did want to mention two complaints. I am not happy with the wifi card and have a 8260 on order to replace it until the 9260 is out.

Also I'm starting to get used to it but I really do not like the lack of a dedicated home/end key. Would it really have been that hard to get them in there? We have lots of room on these. I knew this going into buying it but figured I'd give it a try. We'll see how it goes.

 

Can confirm that my undervolts are holding here after two weeks, booting mostly to Linux. I believe the XTU interface is not fully implemented which is why it isn't reporting or setting everything accurately, such as the GTU voltage offset and the errors regarding applying offsets when it has in fact applied. As a result, setting/checking using the EFI variables is a more reliable option.
As for the multipliers, the CPU is locked so setting in XTU or EFI won't work even though the AMI BIOS contains it.

Regarding the "Pmic Vcc IO Level, Variable: 0x2B9" I believe there is a small additional benefit to be had if anyone is interested. 0.85v seems to be stable so far in conjunction with the other undervolts.

Please don't make any changes unless you are comfortable with potentially bricking your laptop!

 

Just got my 4K / 16 GB / 512 GB XPS 15 9560 and I am very happy -- there is only one problem: I hear faint static (white noise/hiss) through the headphone jack when using small Sennheiser earbuds. Furthermore, when no audio is playing, the static disappears completely (I'm assuming Windows disables the audio output to save battery). Once audio starts playing again, the audio is extremely loud for a split second before falling back down to the level I had it set is. Sounds like a bomb going off in my head. I have not tested with any other type of headset and there is no static through the internal speakers. My drivers are up to date.

Have anyone else encountered this problem? I can live with the faint hiss, so if there is any way to force windows not to disable the audio output when audio is not playing, that would be an acceptable solution.

This thread is fantastic, by the way. Awesome dedication in here.

 

I think I have the exact same problem. I haven't found or even looked for a solution. You pause a music or video, it goes way louder for a split second upon resuming. It's so annoying, especially when you forget and you're using headphones. #bomb

Sent from my G8141 using Tapatalk

 

Maybe disabling Maxxaudio will help. Optionally you can try this modded driver http://forum.notebookreview.com/thr...drivers-for-dell-xps-15-9560-and-more.807060/

 

try reinstalling the audio driver.

 

So loving my laptop so far. Great battery life and no issues to report. Undervolt is working great but wish I could go more. Lol

Is it just me but it seems the power draw of the Kaby Lake chips are terrible...take a look at benchmarks of the Apple fusion 10 or 10x...its comparable to the i5-7300hq and beating it in some benchmarks...how come? Apple has a like a 6w cpu that beats the 45w Kaby lake chip...kinda disappointing for Intel...is this right or are benchmarks deceiving? Could the new iPad run excel and 20 tabs of chrome open smoothly like the i5 can assuming it had enough memory to do this?

 

You can't compare Intel's and Apple SoC. Then why would anyone buy the 10 cores or more when A10 can surpass them in benchmark. Unless you can run Mac OS on ARM and Intel that share majority of the code then the benchmark is comparable.
Pair an i5 w/ 8G of RAM and an SSD, you can achieve what iPad did.

 

I spent a few more hours playing with cooling system in my 9560 last night. VRM always has heatsinks, other parts optional. I tried creating "air channels" with thermal pads and with sticky 2mm foam strips - options with only foam, partially foam and thermal pads, sealing fan intake with foam (strip in a circle around the fan intake). 5 or 6 different configurations.
There is very little difference in performance.
In the end I removed almost everything, but kept the VRM heatsinks since they definitely keep VRM cooler when only CPU is loaded (Prime95). It makes the bottom cover hot, but only in the back area and only at full load.

The only noticeable difference is when I add thermal pad between GPU/CPU heatsink and bottom cover. Obviously, it takes some heat away from GPU and high load benchmarks can run a bit longer - maybe 10min to throttling instead of 3-4min. But after 10 min the whole bottom cover becomes hot (not just thermal pad contact area), i.e. something like thermal saturation happens and cooling performance drops. Also, VRM gets hotter if GPU is thermally connected to it through the bottom cover. Not really worth it, I think.
None of my configurations of "air channel to VRM" make any significant difference. I gave up.

Quite often after high load test and VRM throttling, laptop goes in some strange state when it starts detecting power limit (PL1 I think) even when all temperatures decrease to 40-45 degrees.
Any small load on CPU immediately causes PL limit detection and CPU clock drops to 800MHz. All the temps are still below 50.
The only way to get out of this state is to reboot. Happened several times. It might be DPTF bug or hardware.

I still like the laptop just wish it had separate PgUp, PgDwn, Home, End keys. On my other 7 year old laptop I replaced keyboard 3 times due to arrow keys failing after extensive use. I imagine it will be worse with combined keys.

 

Tnx for reporting. Mostly matches the experience of others. I think iunlock must have tested simpler mods, like the air channel without cutting the copper foil, before coming up with his final mod. He also used Conductonaut.

 

Undervolt the CPU using ThrottleStop. Also disable clamp option in TPL. Use same value of undervolt for CPU, Cache, System Agent and Analog IO.

 

Undervolting and repasting was done on the first day

 

Did you enable SST, Intel Power balance (favour CPU than iGPU)?

 

Disabling Maxxaudio didn't work. I tried the modded driver, but no luck there either, the issue persists. I also tried the updated modded driver, but that reported an error and would not install. The newest drivers directly from Realtek did not solve the problem either. Uninstalling altogether and reinstalling the stock Dell drivers also failed to fix the issue. Running out of options here.

Edit: Several other threads describe this problem. According to mudito on this thread, it appears to be a driver issue. http://en.community.dell.com/support-forums/laptop/f/3517/t/20008728.

Edit 2: It seems there are registry entries for Realtek Audio power settings which should force the audio driver to not enter sleep mode. See this thread: https://www.reddit.com/r/Dell/comments/5wd1uk/xps_15_9560_loud_popping_sound_when_playing_music/. Problem is, all registry keys are reset after a reboot, and I have not been able to change permissions in such a way that this does not happen. I'm just about ready to give up on this.

 

SST yes, but not power balance. I still don't know if it is OK to reduce Intel GPU power. I know in old systems with switchable graphics, video output was connected to iGPU. I guess it is not the case with xps 15 and the output is directly connected to nVidia GPU.

 

I did the same thing, but with the aim of using this. I put the laptop directly on top of the fan. It has big rubber feet which hold it very well. Of course, it ruins it for typing, but it's good for external monitor/keyboard work. It keeps the bottom much cooler, though the fans still run the same and the overall temps are pretty much the same. I haven't done any lengthy temp test with it though.

 

If you don't have optimus don't fiddle with it. I prefer dGPU anyways.

 

You're wrong, everything is output via the iGPU. the nvidia card can be considered a co-processor.

 

The XPS' case bottom is a poor heatsink.

So external laptop fans don't seem to help too much. Furthermore, I suspect maybe the turbulent air from the external laptop fans reduces the performance of the onboard fans.

Simply raising the back of the laptop a bit with a long bar at the back rubber feet (without blocking the air intake) can help increase air intake and cooling a bit. If you don't use a solid bar, you increase the amount of hot exhaust being sucked in at the intakes. So using a small pencil rubber/eraser to prop up the computer would be suboptimal.

The 9550 Temperature thread has a lot of good data regarding laptop cooling fans...

 

Yeah, it's not great as a heat spreader. Too thin. It might be ok for longer sustained loads, though. Also the contact area of the CPU/GPU is only the copper shielding and the covered vents. Turbulent air has nothing to do with it, since it doesn't blow into the fans. What will work is getting another and placing the swept area over the intakes for the CPU and GPU. This will work pretty well I suspect.

 

Not even that. The intake slots are too small. Oh well, it works really well as an anti beer tower.

 

This is completely accurate, the iGPU has the framebuffer that does everything. It's basically like a toll road the data from the 1050 has to go through as the IGPU is what's physically connected to the display ports. That said, you can still basically neuter the HD clocks/voltage as it just needs to be physically enabled to allow the Framebuffer to be available for the 1050.

Also, you were correct about the XTU settings, they are available and able to be tweaked but fail to stick after reboot.. Which I find extremely odd. Still trying to do a NVRAM/CMOS reset as all of the reset methods don't seem to do anything for me. I just purchased some thermal pads to do the VRM's and some aluminum ducting tape to insulate the fans. Finally, I'm going to remove that copper shielding over the bottom vents. Maybe over the next week or so, that will aloow me to pull the CMOS battery while I'm in there.

On a side note, has anyone else experienced extremely high battery degradation on the 97Wh battery? I've had the thing less than a month and I'm around 15% decay... Which is unbelievably poor, that's with disabling the BIOS longevity/reserve features...

Would anyone also be interested if I purchased a FLIR E8 thermal camera? I've always wanted to buy one and I think tackling this thermal issue at a professional level with actual images and data would be a pretty awesome way to kick it off.

 

They are about 50% of the area, I think mesh is about 30%
What we need is someone with a cnc to widen the slots a little, removing them is asking for trouble and I have to say only a complete idiot would do that unless the laptop never moved due to the risk of ingesting something to stop the fan spinning.

Yipes, big post.

XTU is a pain in the arse, I have seen no change since flipping the bits to unlock the menu options. undervolting is the same as before bit flipping, works as expected but has a fallback if it thinks the system crashed, I tried undervolting and disabling the service but that just rebooted to stock voltage.

Why do you want to pull the cmos?

All mine come with various levels of battery wear level, this arrived at 9% then after a couple of discharge cycles is settled at 4% so it may be worth running it flat, then go into BIOS and let it sit until it shuts off and refuses to boot. After they it is suggested to leave it off and fully charge it but I skipped that part and still wound up at 4%.
Dell are thick ****s when it comes to the wear level out the box and quoted 10% is expected but I threw it back that it should be what you pay for, a 97wh not 85wh etc, in the end (due to my superbly high returns rate) they just said they would refund me. I decided to keep it as I have taken the piss enough getting a full refund after 2 years 11 months into my 3 year warranty, 2 free upgrades and at least 7 other system swaps.

15% is enough to get it replaced under warranty as that is really bad but give it a discharge and see before you call them.

 

That's what I figured as well, my degradation has gotten worse overtime rather than better...
http://imgur.com/A3FcdxX

What a pain in the ass to have to go through a replacement. Do they just get the battery? Or a whole new machine?

 

That's bad, mines 92wh.

Depending on the warranty level, an onsite would be a battery replacement onsite, maybe you could talk them into shipping the part to DIY if you prefer but it's a 10 minute job to let them do it.

 

After doing the cost comparisons...I think I'm going to get a new xps 15. and keep my inspiron 13 5000. the xps is 1500 dollars cheaper than a MacBook pro 15, has 2 times the ram, touchscreen, and you don't have the starbucks factor either. sweet deal imo.

 

Good to know! I will try changing power balance then.

 

There are some thermal photos of the 9550 (Precision well 5520 equivalent) which give some clues. The CPU & GPU are obviously different and the heatsink has been tweaked slightly for the 9560.

You can find the thermal photos in one of the following threads (can't remember which). Read through the threads as they will give you some idea of key thermal issues and what does not work:

http://forum.notebookreview.com/thr...ures-benchmarks-xps-15-9560-kaby-lake.802345/

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

A few quick thoughts, many of which you already know but for others:

-The 7700HQ & 1050 can overwhelm the small chassis in more aggressive use. I think the fundamental issues are:

1) the radiators are undersized
2) the vrm is both underspecced and essentially uncooled

- Dell's thermal paste is terrible so a repaste of the CPU & GPU is probably a good first step. Also note the lousy 3 screw heatsink over a rectangle in the 9560 so you need to make sure fitment is flat.

- Undervolting helps also, although not as much as on the Skylake units

- Another easy and effective tweak is simply raising the back of the computer a bit with a long solid bar which mirrors the rubber foot (and does not block the intake)

==> Cutting out the grills under the intake fans and replacing with mesh will increase fan performance. I think you will need the back of the computer raised to realize the full benefit here. Not sure how the undersized radiators and 90* exhaust bend @ the screen will perform with more airflow. Somebody documented benefits with in those threads (can't remember if he "overclocked" his fans or he ran without a bottom cover (danger!)).

- I found enabling SpeedShift to be a helpful performance tweak that should help thermals slightly on a relative basis (there is a dedicated thread here). EPP=0 max performance, EPP~78 (you need to test as 78 might not be optimal for your processor) will allow both max processor performance and some power savings, EPP=255 max power savings.

- The vrm heat is coming from the mosfets, which are slightly cooled by some stagnant air above and the board below. There is a hand-drawn preliminary schematic of the 9550's vrm in the thread linked above. The 9560 vrm is a similar design with different components

1) A few of us tried cooling the vrm by connecting the chokes or mosfets to the case bottom with thermal pads. But that case bottom & keyboard just got really hot fast and delayed throttling just a few minutes. The hot case bottom also superheated the fan intakes, making the CPU/GPU radiators less effective. Simply adding alu heatsinks to the mosfets with stagnant air is a tough proposition

2) A few tried increasing the airflow over the vrm but that is not easy to do. One mod to increase airflow to the radiators by patching leaks with electrical tape likely will decrease airflow around the vrm. You could divert some air from the fan towards the vrm but need to make sure there is a place for that air to exhaust (behind the vrm is a foam pad that GoNz0 said prevents hot exhaust from being sucked into the laptop).

- note that the 9550 heatsink did not make good contact with 1 or 2 vram chips. The fix was to replace factory thermal pads with low-quality fluffy thermal pads. I have seen no cooling issues with 9560 vram.

- User iUnlock did a few creative extreme mods which are worth checking out.

- You can get around some power limit throttling disabling Intel Dynamic Platform & Thermal Framework. Another alternative is to boost the power settings in the flakey XTU which you will need to check every few days as it tends to drop settings. Either is a good path to melting down the laptop as you know. . .

 

Geez, that's nearly an entire hour of normal runtime between our batteries.. I just pulled up my warranty and I do apparently have On-Site for another year luckily. I think I'll take them up on that and get it replaced. I gave it a completely unstartable discharge last night, and kept it off until this morning while it was charging. I went up to 83Wh, so it did gain a bit but still way too low.

@pressing That's an excellent summary, holy smokes! I already did a repasting, and undervolt around ~120mV system wide. Temps are pretty solid, but combined load still gets up there. I was playing a game and didn't realize my shorts were blocking one of the vents. After some stutter I pulled up HWiNFO CPU @ 93*C.

 

The battery degradation might be due to BIOS settings. I forget the exact naming and don't want to restart, but in BIOS there are Dell Power Management settings that claim to increase battery life and allow you to set the charge mode for the battery ranging from adaptive to a settings where it only charges during off-peak times. Mine came set to adaptive, which progressively limited the battery full charge point similarly to what you described. Turing this off reset my battery capacity back to the rated value after a few restarts and charge/discharge cycles. Might be worth a try before you initiate repair

 

If anyone is thinking about modifying the bottom panel, it might be a good idea to grab a spare off of Ebay if you are concerned about your warranty. Modify the spare. Keep the original intact.

 

I should have noted that the 9550 seems to throttle when temps hit about 78*C (the vrm sensor if I recall correctly).

And I think Dell set the same artificial threshold for the 9560. There are plenty of posts about that for more info. But you can easily see it for yourself with HWiNFO and a light benchmarking program that draws both CPU and GPU (ROG RealBench stress test is not very rigorous but quickly shows the weaknesses of the XPS design).

My best guess from back of the envelope calculations of CPU & GPU draw and vrm component shelf specs was that Dell's engineers pushed the mosfets to their wattage limits. The mosfets get really inefficient as heat and wattage rise, so stop producing enough power and quickly spiral out of control thermally. Hence Dell's 78*c cap.

The two easy ways to get around the 78*C ceiling (via XTU or disabling Intel's thermal software) are noted above. That may provide a bit more headroom but not for "extreme use" as the vrm loses efficiency quickly.

Thus, I think the real solution here is to find a way to keep the mosfets cool so they can provide max wattage "all the time."

 

I see them on eBay for like $15 dollars all the time, this was my thought exactly. At that price it's most likely less than cost lol.
So I took your guys's advice and disabled all the BIOS battery trash, well that really screwed up windows hahaha.

Windows has no idea what's going on with the battery now, needs a major calibration. I was still playing Bioshock 2 for an additional 14 minutes after I took this screenshot haha.
http://i.imgur.com/CCt6avG.png

What do you guys think of this battery pack, a bit pricey, but this should even under load get me about into the full day battery life range. Couple days on web browsing.
https://www.amazon.com/MAXOAK-50000mAh-Portable-External-Notebook-Most/dp/B00YP823NA

 

I have set CPU to 30 and GPU to 9, so every time CPU is preferred more than a GPU. Do this only if have a dGPU like AMD or Nvidia. If you have Intel GPU set CPU to 30 and GPU to 31

 

I calibrated my battery about 5 months ago and have been using a battery meter. After calibration, my battery wear was about 3.9%. A few months later it went to 6.9%. Now it's showing 12.9%. I am not having very good battery life either.

Should I be contacting Dell and requesting a replacement battery (if this warrants it) or would you do something else first?

 

Yes definitely contact Dell.