XPS 15 9550 temperature observations (undervolt + repaste)

I think the tubing is generally a specialized sealed environment so a puncture or kink could cause a leak, crippling the thermal conductivity.

Do some research around the internet beforehand. Some things to consider:
- you want very clean surfaces
- make sure not to scratch the delicate chips or the soft copper mating surfaces
- do not overapply the thermal paste

I have seen people complain about lint problems with cotton towels, paper towels, cotton, etc. Some people use coffee filters but those seem abrasive. Some people use microfiber cloth but they get dirty quickly so a small one is not enough (maybe a few decent quality ones will do). Someone here (I think) recommended using photo lens cleaning tissue (but never reusing a side); that seemed to work well for me.

Previously I used high purity isopropyl alcohol (e.g. above 99% to prevent leaving impurities on components). I recently purchased the Arctic Silver 2-part cleaner and thought that worked well. Ventilate well and use caution as these chemicals can react negatively with plastics, rubbers, adhesives, your skin, etc...

 

So went on Amazon and bought thermal grizzly and artic cleaning stuff. Not sure if I can see a difference in temperatures. I am still getting "power limit throttling" on Intel XTU. I think this is different than actual thermal throttling. I haven't undervolted and don't want to go through the hassle of doing it. I hear it slows down boot times and is not a good idea when playing games etc.

 

Power limit is the CPU package staying under 45w so undervolt could help prevent it as lower voltage = lower wattage, of course stress testing is needed to stop it crashing in game.

 

Undervolting has no effect on performance and or boot time. If you didn't see a temp difference between stock paste and a repaste than you may have put too much on. Like GonZo said, power limit throttling is related to power output during turbo, an undervolt would help alleviate that.

 

Pkorgt,

I had "power limit throttling" and you can search what worked for me (and others in that thread). A few things I remember:

0. Several users noted Intel XTU had some bugs with the Dell 9550. One of those was causing "power limit throttling" during my benchmarks at ~79*C and apparently low wattage situations. Delete the program

1. Undervolting had the biggest thermal impact on my 9550 (~11*C improvement average over several benchmarks)

2. New thermal paste and new thermal pads had significant thermal impact (~6*C improvement average)

3. Note that 1 or 2 VRAM pads are too thin; I think an over heated VRAM chip will trigger throttling although I don't know how it might be classified (could be misclassified). That is a defect that needs to be fixed by the user or Dell and could be the bottleneck during gaming

4. ThrottleStop does not launch at boot so does not impact boot times. I manually launch it every morning. It is a simple program and the guide and warnings by Eason will take you 5 minutes to read. A quick review at NBR forums of 9550 i7 users, show some people reporting 140-160mv undervolts on the CPU. If I were in your shoes and did not want to bother stress testing, I might just plug in some conservative figures into the FIVR control screen, and go with that (e.g. CPU core -140mv, CPU Cache -140mv and Intel GPU -100mv)

http://www.ultrabookreview.com/10167-laptop-undervolting-overcloking/

5. According to the maker of ThrottleStop, you can not have XTU and ThrottleStop installed on your computer at the same time as they conflict with each other

 

I simply applied a thin coat across both CPU and GPU. The method of application really doesn't matter unless you put too little or too much.

 

Application can make a difference. If you created air bubbles while spreading it won't transfer as well. If you spread it too thin and the heatsink doesn't make great contact then you'll have issues... Best way to ensure your method works well is to do a test fit. I would honestly be surprised if you didn't see a drop when going from the stock paste to ANYTHING else.

 

Hello there,

So I started using Throttlestop and re pasted my laptop (FHD/256Gb/i7) yesterday.

I have set -150mV CPU / -105mV iGPU profile for the time being and I wanted to go further once I have everything else set.
The problem is that I really don't get much difference after repasting, I did one yesterday and the results on Prime95 (small ftt) weren't very convincing, 90°/+ C (with throttle from 3,1 to 2,6 GHz).

Given the results I saw here I figured that my repasting wasn't good enough/well applied. So I did a new one today, but even though I get some better results (Around 85°C) it still seems quite high (comparing to, the standard 70/80°C, that I've read about here.).
On top of that the CPU also throttles at 85°C after a few minutes ( Also 3,1 -> 2,6 ).

EDIT1 : On the other hand my idle/low use/browsing temps seem good, in between 30° and 35°C.

So I was wondering, is it my paste that isn't good enough ? (Gelid GC - 2) Or perhaps I did a bad job applying the paste ?
I am also thinking that maybe, my throttlestop profile isn't set properly ?

I am new to this stuff so I try to get to learn the most I can, but here I have no clue of what can be going on !
(Hope the answer is not anywhere on the thread and that I missed it.)
Thank you very much for this thread it has been very useful !

 

How did the paste look on the heatsink after removing it the first time? Spread evenly and not too far over the sides? What are you using to test temps? 90*C with and undervolt and repaste in prime95 only would probably point to a bad heatsink or something.

 

The first pasting I did, I spread it with the little plastic tool and after reading more carefully I saw that spreading before could led to air bubble, which was the case, and I think that there wasn't enough paste. So yeah, not very even, flat and "bubbly".
I use Prime95, to stress the PC, and HWmonitor to monitor the temps.
How could I tell if my heatsink is bad ? Is there something that I could see ?

 

Smoath - sorry you are having these issues. Thanks for the detailed comments. My preliminary thoughts follow - people here are incredibly helpful so keep posting. Apologies if some of the questions sound insulting - they aren't!

Undervolt software
To be clear, in ThrottleStop under the FIVR window, did you adjust the "OFFSET" for these THREE items NEGATIVE: CPU Core -150, CPU Cache -150, Intel GPU -105? Did you click the on button (the bottom right button should read "turn off")?

You can also read Eason's guide to ThrottleStop to make sure you are using it properly (find one of his posts - it is linked at the bottom in his signature)

Also, do you have Intel XTU installed? If so, that causes some conflicts in 9550 forcing the computer to throttle. It also does not play nice with Throttlestop. Uninstall that XTU software.

Sample thermal improvements in 9550 - tests, thermal paste & undervolt
Our resident professional laptop engineer GoNz0 recommended adding other diverse tests; I also used RealBench (benchmark and stress test). He said Furmark is too stressful and can kill a laptop so stay away from that.

For reference, I have an i5 6300HQ and was able to reduce temps on Prime95 small FFTs (15 minutes, room temp about 22*C) from 75*C factory to 62*C with ThrottleStop to 56*C with ThrottleStop & repaste/repad VRAMS ( -19*C total). So most of the change was from undervolting. Other benchmarks have similar results but improvements were not as high.

Scanning the NBR site, I think the i7 has less potential for thermal improvement in the 9550 as it is pushing the tiny thermals a lot more. My buddy builds supercomputers and he thought the same. Regardless, every chip has different scope for improvement. I have read the HQ chips are low binned so variance is quite high; not sure if that is true (really LQ chips lol). You can scan the NBR posts but they are a bit difficult to interpret in my quick and incomplete table below (e.g. is improvement just repaste or repaste+undervolt, what was the test). You might search those names to see if they were running the same Prime95 test so you know where an i7 should be on that test (or message them if not detailed):

mrpeaches (77.7-72*C) -5*C
Bullhonkie (88-79*C) ( -140mv) -9*C
superdavemaine ( -170mv) -5*C
splitframe (85-75*C) -10*C
R#lph (-150mv) -10*C

Thermal paste application - please research and study
I have used Gelid thermal paste and it ranks well for low pressure clamping systems in a few internet shootouts in what appear to be reasonable tests. If you have an issue it might be application. My supercomputer buddy had a full-time TEAM of engineering students just repasting GPUs every afternoon; he said they had the experience and training to do the spread method without many bubbles (and the stats to back it up) but we don't. You should research the internet for videos and articles to learn about best practices and popular methods like the grain of rice or dot or x method. With exposed CPU and GPU in the 9550, you want paste to cover the entire surface (of BOTH chips), just barely. It should not be spitting out the sides of the chip. Thermal paste is a terrible heat conductor and is only to fill microscopic scratches and imperfections better than air so the less paste, the better.

Even GoNz0 does a test paste, clamps down the heatsink, then unclamps it to see if paste was properly applied, cleans up the paste a second time then puts laptop back together. Tedious but that is how the pros do it.

FYI this is too much paste:
http://forum.notebookreview.com/thr...ndervolt-repaste.785963/page-17#post-10222921

Also please research excellent cleaning methods; think clean room. Artic Silver makes a popular 2 part cleaning kit. You can also use very high purity iso alcohol made for electronics (think higher than 99.9% pure); anything else will leave lots of residue. Be careful not to get the cleaning fluids on other plastics, rubbers, resins, etc. Also use a zero lint, non scratching cleaning "rag."

One good trick I saw was to heat up the paste before using it. That makes it easier to squeeze out and apply. I put the syringe into a ziplock bag, remove the air, then insert that bag into a cup of pretty hot water for 15 minutes (replacing the hot water once).

Next steps
Research the thermal paste cleaning and application procedures. Try again and see how your benchmarks compare to those of others. My sense given the data above, my i5 results on Prime95, an absolute ceiling of Intel chip generally ~100*C, is that you should be 70-80*C running Prime95 with a good repaste and undervolt. I'm not sure thermal throttling at 85*C is an issue; that might be spec protecting the chip or perhaps some other sensor is getting too hot (which you can see in a temp monitoring software). Intel XTU had a bug that gave me a massive power throttle at 79*C.

As you asked and custom90GT noted, you could have a bent / defective heatsink. I have read at NBR about one or two people who have had that issue (I think one guy had his replaced and another rebent his, which is risky). My view is that all 9550 heatsinks are defective and that is why that VRAM thermal pad(s) does not contact the heatsink properly. Regardless, I think the probability of a defective heatsink is low but still possible.

Good luck

 

First things first, thanks for the detailed answer !

So I got, what was wrong, I didn't see the "Turn On" tab on throttle stop... (I can't believe I didn't see that ...)
Well, at least that sorted out a bit of the issue.
Nonetheless, I don't use XTU, so it won't interfere !
I am going to run other benchmarks to have a better overview.

Now my prime95 test, I didn't run it more than 10 minutes, indicates lower temps, towards 80°C which seems more accurate.
(I still got some throttling along the tests..)

Besides I played a 30 min Cities Skyline session and the temps were also much more decent. CPU at 75°C.

However, something bugs me, is it normal that one of the cores is always 5°C higher than the others ? (Idle and load.)
I don't know much, but couldn't that indicate poor pasting ?
(e.g: when I stress test with Prime95, all the cores are at mid 70's °C, whereas the "hot" one is bursting in between 80 and 85°C.)

I think, I'll try a third time, I use 70% (+) alcohol and Q tips to clean the paste, as I saw often on YT. The only issue I have with that is the chips on the GPU that "stick out" and that I am afraid to damage.. (Still some factory paste on these things.)
Is it really worth it to use higher % alcohol ? Since it is easily obtainable might as well get some, oh and iso alcohol made for electronics ? Should I use that instead ?
(Tried to find the artic silver cleaning kit without any luck ...)

I'll try to get my method right this time.

I'll check my heat sink properly when I open it up.
And I intend to get some pads quickly so that gaming isn't (as much) turning my laptop into a burning potato.

Thanks







Sent from my iPad using Tapatalk

 

Hi together,
when I got my XP 15 9550 with i7/512GB I untervolted using XTU and got thermal throttling which I described here. After repasting and undervolting with ThrottleStop, my problems seemed to disappear. In the meantime I did not game on my XPS, but last week I played a bit of Dota2 on it and the FPS dropped considerably. Starting from ~90-80 FPS they dropped to 25-30 and from their alternated between 25 as the lowest and 60 at the highest.

Investigating this using HWInfo I found that the CPU went to 80-ish °C and dropped then to 55 and started alternating between 55 and 65°C. The GPU Temps did not show up in HWInfo, but they were also safe. Obviously the problem was a VRAM chip, which reached over 100°C resulting the system to throttle, as shown in the plots as 'Ambient Sensor'.

http://imgur.com/a/JFhkc

I wrote Dell about the throttling and temps and they will replace my mainboard and the cooler. My question to you is, why do you fix Dell mistakes yourself when you could get Dell to fix them? Is the Dell service bad in your country / location? I am from Germany and here the technician doing the repair comes to my place.

Maybe I could have fixed the problem myself with some pads for the VRAM, but considering all the reports of bad mainboards and/or coolers I though getting Dell to do this is a good idea.

 

They generally use used parts for replacements so odds of getting more troubles than you had before are quite high in my opinion.
In my case I followed the guide, repasted and added some thermal pads and I don't experience any throttling right now.

 

To add to what was already said, most of the time, it does not fix anything.
I have had 3/4 motherboard replacements, and 5 LCDs...

 

After a lot of research and testing, here's what I've discovered.

Hopefully this explanation helps anyone experiencing this issue, both to better understand what's going on and also with some solutions/workarounds noted below.

The main culprit for this throttling is the PL1 power limit which is slowly being reduced as the cooling system becomes saturated. If you run either CPU /or/ GPU this does not happen because you have an excess of heat-removal capability but if you load both CPU /and/ GPU at the same time (aka gaming) then, depending heavily on the ambient temperature, PL1 slowly decreases until it reaches its lowest point, 7W. This correlates with the 800MhZ (actually 8x100MHz) speeds being reported and is the lowest speed possible (each CPU running at the bus clock frequency directly with no multiplier).

That said, there is another gremlin at play here as mentioned elsewhere: the VRAM temperature. In poking around the hardware this sensor is actually located at the very back of the board more by the voltage/current regulators than by the VRAM, but colloquially that is what it's being called. This placement has zero forced air movement and is expected to be cooled by convection out the rear of the case. If you want to see how little airflow there really is for yourself just open your favourite monitoring software (I use HWiNOFO64), find the second listed "Ambient" sensor readout, and then blow gently just below the Dell logo at the bottom of the screen. You'll see the temp drop for this sensor very quickly, especially if it's already pushing 100C. Minimizing the difference between this sensor temp and your CPU/GPU sensors is key to limiting your throttling.


So with the "why" out of the way, what can we do to minimize the practical impact? We can't get around the fact that the cooling system in the XPS 15 is designed at the limits, especially if you're running the i7-6700HQ CPU. So here are a couple of "solutions" (which are more like workarounds) you can do to minimize the throttling effects:

1) Clean your heatsinks: Just pop off the bottom of the case, remove any dust/lint/hair/whatnot and then give the cooling fins a good blast from a can of compressed air. I would recommend doing this every few months, more if you have pets or are in a dirty environment.

2) Control the ambient temperature as much as possible: In my testing, a fully loaded system with a room temp of 22C (72F) kept PL1 above 22W. Increase the room temp to 27C (80F) and PL1 dropped to 9W. This roughly corresponded to clock speeds of 2.1GHz and 900MHz respectively - a HUGE difference.

3) Use the laptop on a hard, flat surface: After testing with an elevated laptop (front/back/both) and forced air laptop coolers, the best overall result was to not do anything special and just use the unit directly on a table. This actually makes sense as this would have been the test condition the engineers used during development. Cool air is drawn in from the sides (because of the rubber feet) and hot air is directed out and upwards in front of the screen. Elevating the laptop allows exhaust air to be sucked back in the fan intakes. Note: using a forced air cooler kept CPU and GPU temps lower overall but I still saw increased throttling due to increased temps reported near the VRAM/voltage regulators.*

4) Undervolting: I haven't gone down this road personally because 1) I like my system stability and 2) the software to set this does not reliably start after every reboot. That said, lowering your CPU voltage by even 50-100mV should significantly reduce the amount of heat being generated.

5) Repasting: Yes, Dell is notorious for using too much thermal paste and my GPU actually had a large air bubble covering about 20% of my GPU die! (GPU temps weren't my primary issue so I didn't actually see any ill effects from this thankfully.) If you're comfortable repasting then it's worth doing, even if it will only gain you at max a few degrees. If you don't like using paste directly you can always get it in pad form (AIT Cool Silver Pad or similar) - just make sure you're not using a regular thermal pad and that you clean all of the surfaces well before re-application!


Hopefully this sheds some light on the issues/constraints we're dealing with. The XPS 15 is a great unit overall, has a lot of diverse uses, and works well in a lot of situations, but it isn't a gaming laptop first and foremost which shows as the cooling system is quickly saturated during heavy loads on both CPU/GPU. This is compounded by (relatively) poor thermal engineering around the regulators but hopefully the workarounds above will help you get the most you can out of your system!

Cheers!




*Logs show that with an elevated laptop (even with a forced air cooler) the VRAM/Regulator temps climb quickly to 118C (!) at which point the CPU throttling kicks in heavily until the VRAM/Regulator temp drops to 100C. PL1 (and thus your clock speed) is then varied to keep the VRAM/Regulator temp right at 100C - resulting in very slow clock speeds and a very bad user experience. When using the laptop flat on a hard surface the difference between CPU/GPU/VRAM/Regulator temps is less drastic and thus less CPU throttling occurs as PL1 can remain higher. As noted above, minimizing this difference is key to getting the best performance you can.

 

Thanks for the lengthy post. It makes me wonder if perhaps creating a makeshift heatsink for the sensor (good pics can be found on page 27 of this thread) could fix all of the throttling people are experiencing. Wish I had some time to look at it.

 

Hahaha!

 

I take it you or someone else has tried to make one?

 

DIY gives you the opportunity to really clean the chips well and carefully apply high quality paste.

You will also have the opportunity to look at the VRAM thermal pads and correct; Dell will not look at that defect so it is either the owner or leave the defect...

 

Hi custom90gt,

I think the idea is great. Actually this summer, I was thinking about ideas to improve the cooling around the voltage regulators but just overriding or fooling the thermal indicator is cleaver workaround. Of course if there is a thermal meltdown, that would be problematic as there is virtually no cooling or air circulation in that part of the laptop.

I did some preliminary research on the voltage regulator area, and responding to my query regarding voltage regulator cooling, GonZ0 posted a detailed response as to why he thought cooling that region was "challenging".

 

I recall your post about the VRM cooling with a fan. I was thinking something much simpler and a lot less elegant. Perhaps a sheet of copper that sits on top of the taller R22-604-111s (from the pic on page 27) with thin thermal pad between them and a thicker thermal pad to contact the VRM sensor. All of that followed up with a thicker thermal pad to sink it to the case. I dunno, I'm only pondering things in between studying for my cardiovascular test on tuesday.

I don't think silicon thermal pads would pose a problem with electrical conductivity, but perhaps someone else can chime in on that. Also they may not transfer enough heat and may actually compound the problem.

 

GonZ0 loved that fan idea!

Thermal pad manufacturers typically provide electrical data in their sheets. For example, Fujipoly published some data on pages 15-16 below:
http://www.fujipoly.com/usa/assets/files/Fujipoly 2015 Sarcon Catalog for web.pdf#page=8

In the draft picture, I jotted down a quick and dirty schematic guessing the source of the key components; maybe that will be useful for someone some day. Some of those parts are special order-not easily identified. The picture should be considered a starting point for your own research.

- I would guess the 3 TI87351 MOSFETs (PQ1000-PQ1002) and the 3 Alpha Omega MOSFETS ZS019QI (PU1200-PU1202) produce quite a bit of heat as my audio equipment has massive heatsinks attached to the MOSFETs.

- You will also see several nearly-microscopic tantalum SMD caps in that area (PC1003...). Those can be sensitive to heat (had to replace a bunch of them in my car's instrument cluster last year due to heat - difficult to diagnosis and tough to solider anything at that size).

 

A heatsink really won't help because there isn't any airflow apart from that generated by convection.
What is needed is air movement over this area.
I just ran a test where I placed a small fan under the Dell logo blowing air on the regulators and it kept the temps in check to a point that the GPU was the main cause of throttling:
- VRM ~75C
- CPU ~77C
- GPU ~89C
This was while playing Overwatch on high settings at above 1080 res capped at 60FPS.
Without the fan the VRM hovered at 90C with PL1 at 15W.
With the fan the VRM was around 75C and PL1 at 20W.
Room temperature was 25C (77F).
My new main logic board (replaced recently due to the audio problem that is common among these machines) seems to have less difference between VRM and the other components as my first one, and the temps leveled at 90C instead of 100C - but maybe that's because of the new (v13) BIOS? Too many variables to tell with any degree of confidence. I'm happy that the VRM temp is roughly tracking with the GPU temp though! Less throttling overall.

 

Sat on my phone so I can't be arsed to go back and start quoting but whoever it was that suggested cooling the vrm sensor needs to slap themselves!
It is an ambient sensor so it monitors the area. I found it when Aida64 reported the vrm as memory so the developer could correct it. Cooling the vrm's would be great if the chassis could dissipate the heat. Who knows, I may give it a go one day to see how well it works.

Sent from my SM-G920F

 

lol, you need to get off your high horse. You're the one that used a q-tip to cool the sensor down to locate it. You can either cool the sensor to trick it or go ahead and cool the rest of the VRMs down. The copper plate may allow you to cool everything down and sink it to the bottom. You also assumed that the VRM wasn't causing issues, when it seems that during gaming people are having throttling due to the VRM temps.

 

Not on a high horse but why would you want to cool the sensor and mask the problem until you break the mainboard when the vrm's crack?

Should my none contact temp probe ever arrive from China I will see if I can get a temp from the vrm's as I expect they are running a lot higher.

Sent from my SM-G920F

 

I think half the problem is we don't know how hot the VRMs are actually getting. I've read threads where someone uninstalled the intel thermal platform drivers and their VRM sensor readings were lower. I don't know how well setup the VRM sensor is to compensate for the spacing from the actual VRMs.

If I still worked as a mechanical engineer I could throw some thermocouples on the VRMs and sensor and see what was happening, but I don't have access to any of that anymore.

 

Brian - Super interesting solution! Could you elaborate on what you are saying here? Do you mean the Dell logo just below the screen? Do you mean you put a small fan just above the f7/f8 keys effectively laying horizontally on top of those keys? How big was the fan? Which fan was it?

Which audio problem (lol)? Do you mean the crackling sound when adjusting brightness via f11/f12 keys?

 

I know that Dell uses used parts as replacements. Beside the throttling my XPS also suffers from quite a bit of coil whine. Chances are that I end up with a better board than before In the case Dell does not fix the problem after the second attempt I can return the notebook and get a refund (according to german law, I do not know how this is handled in other countries):

Back to topic. Yesterday I repasted again and one VRAM chip definitely does not make contact with the heat sink. I put a thicker pad on this vram (had not enough pads for all four), but temp still go above 90°C, which still leads to throttling (1.8 - 2.3 GHz).

 

Ive just had an Overwatch gaming session (low-medium settings) with hwinfo in the background...
The lag kicked in perhaps 15 minutes in, but it was not as severe as usual.. after checking the numbers it is true that PL1 Power limit reached a minimum of 7W, and the cpu clock minimum was 800MHz.
Maximum values for temparatures reported in the block containing "ambient" meters are:
CPU: 82
Ambient: 92
Ambient: 77
Temp3: 51
Ambient: 70
Ambient: 74
Ambient: 54
DIMM: 65
CPU-fan ~5000RPM
GPU-fan ~5000RPM

I dont know what these specific ambient readings represent, but the 2nd in the list has a relatively low value of 77C, this should not be in the problematic range?
Granted, this time the lag did not manifest in a completely wrecked framerate with cracking audio, but still in the unplayable range.
Weirdly enough lately I am having issues with ping spikes (1000+) and rubber banding and im not sure if these are all related, or if the bluetooth/wifi is crapping up on me (awaiting on card replacement).

 

Gerty,

You can see in HWINFO64 some extreme benchtesting on an i7 in R#ph's post quoted below.

As you might guess, running Prime95, his repaste drops CPU temps but does not really drop Ambient sensor #1 temps (pictures 1 &2)

Surprisingly, when he runs very extreme benchmarks, CPU temps fall with the repaste but Ambient sensor #1 skyrocketed to 114*C (pictures 3&4)! Caution as those temps can shorten lifespan of many electronic bits. Part of that temp spike could be a sensor that is not too accurate at extremes, or because his repaste work included using the bottom cover as a cooler via thermal pads to the heat pipes (and the cover got really hot causing as gonZ0 called a something like a thermal feedback loop where hot air was being pulled in by the fans). Or the system just going out of control at extreme limits.

For your reference, today I ran my i5 6300HQ with RealBench 2.44 stress test for 15 minutes in a (very) hot room (30*C), HWiNFO64 shows the following max values (FYI CPU and GPU were repasted, VRAM thermal pads were replaced, no undervolt)

Room (est) 30
CPU: 83
Ambient: 78
Ambient: 76
Temp3: 48
Ambient: 61
Ambient: 63
Ambient: 56
DIMM: 66
NVIDIA 88
PL1 min 15.375
CPU min 1.1GHz
CPU-fan ~5000RPM
GPU-fan ~5000RPM

In a cool room (~22*C) with undervolt, the results are much better as you can see in a prior post. Undervolt makes no real difference today in this hot room

I think you could download trial Aida64 to cross reference exactly what these ambient sensors are; I think #1 is the H56 located in the voltage regulator area.

 

Yes, I just grabbed a small USB powered fan and sat it between the F7/F8 keys and the Dell logo pointed down.
Here is a graph of the temp difference:

Unlike my old board, the VRM temp roughly tracks my GPU temp even without the forced cooling which is acceptable to me, but I wanted to post this to illustrate the difference.

The audio problem is one that either (randomly) causes no sound or a BSOD. After a lot of reading I think it was caused by my lappy going thermonuclear when Windoze Update decided to wake it from sleep in my backpack. When I found it two hours later it was so hot I couldn't touch it - 95% of the battery discharged in a backpack in a notebook sleeve. Shortly after that I started having the audio issues described here. The only fix, after many many tests and confirmed by others, was to have Dell replace the hardware.

 

Thanks for the detailed response Brian. Glad that USB fan idea helped thermals. Hope I don't end up with that cooked motherboard as a refurb from Dell "service"!

Some quick thoughts:

- A few guys here noticed thermal improvements just lifting the laptop into the air (like 5*C)

- GonZ0 also noted the long rubber feet in the back and front are cleverly designed barriers so the fans intake air from the sides, rather than suck hot exhaust from the back

So a few simple ideas I had:

- Passive cooling - Could improve general thermals just by building 2 long solid feet (25cm x 1cm x 0.2cm) say with concave tops to accommodate the 9550s rubber feet. That would allow more airflow into the fans and maintain some of the barrier system. Maybe connect the solid feet with a collapsible X structure for strength and portability

- Active cooling - I like something like the cooler master U2 plus because you can shift the fan position easily. I don't like the bulky feet much, however. But it might work better adding a DIY barrier system

- With the active cooling pad above, it could make sense to buy a used 9550 bottom cover and poke some holes for fans to directly cool the voltage regulators. I think (but am not sure) the voltage regulators are just below the middle grill section that Dell covered with some mylar/copper which would be easy to puncture. As a used bottom cover is $20, that is not an expensive experiment. It is a pretty "brute" solution as the air into the voltage regulators has no real place to easily go but a rough exhaust pipe could be built I suppose for some balance.

 

So just for S&Gs I decided to throw thermal pads on the R22-604-111s and sink them to the case.

I ran 2 (yes only two) loops of 3dmark time spy before the thermal pads and had a max VRM temp of 79*C.

After that I tried with thermal pads with two more loops of 3dmark time spy and my max VRM temp was 67*C.

When I have more time to mess with it I will update this thread.

 

Only issue I found when using the case as a heatsink is eventually the case heats up in turn heating the incoming air causing an increase in temps until the entire laptop heats up then it can't cool and well you can probably guess the rest

 

Interesting option!

I saw several people here earlier this year used the bottom case as a heatsink by connecting the CPU/GPU heatpipes via thermal pads. I think everyone removed them. GonZ0 noted the intake air got heated and caused a bad thermal loop. Some also complained the bottom / top got too hot to touch.

But the VRM produces little heat and needs little cooling so maybe this is a good option. What thickness pads did you use? Which pads did you use? Unfortunately, those pads get quite useless as they get thicker...

 

Right, I wasn't a fan of cooling the heatpipes from the cpu/gpu due to that issue, however I don't know how much energy is being dissipated by the R22s and the case in comparison. I would have to spend a lot more time than I have to figure if the juice is worth the squeeze, but I'm trying to throw out ideas for the people having throttling issues.

 

Let it run daily stuff, surfing the net and very little load, the sort that spins the fans on low now and again. That is all I did with the pads from heatpipe to case and after an hour the palmrest heated up and the bottom was about 40 degrees causing the thermal loop and fans spinning up to 3000 until I gave up.

Would be great if the VRM's (is that the R22's as I was browsing on my phone the past week on holiday so I haven't really kept up to date due to alcohol abuse!) didn't cause a thermal loop!

 

Yeah I'm trying to run loops of 3dmark while I study on the desktop to see if it gets heatsoaked...

I used 1mm thick just cheap silicon thermal pads. They aren't the greatest, but if they can knock 10*C or so off it may be worth it.

 

Yeah I'll leave them on for a week while at school and what not to see how it does. Any increase in fan noise or palmrest temps will be pretty easy to notice since I typically use the laptop 8+ hours a day.

 

Typically, how would you expect the heat of the CHOKES (6xMagicTec R22) compare to the heat of the MOSFETS (3xTI 873510 & 3xAO ZS019QI)?

 

I will let you know if my none contact probe arrives!

 

That probe is lost in the post!

Sorry for repeating detailed elementary questions. I ask as I thought MOSFETS were generally heat generators and very sensitive to heat, whereas chokes are typically just wound wire which don't generate much heat and are not too sensitive to heat...

 

So sadly I don't have thermocouples or even a cheap temp gun to check so here is why I decided to sink the chokes.
1. With the case open and the system on, they were the hottest things to touch within the first couple minutes of powering it on.
2. They are tall enough to easily be sinked to the bottom of the case.
3. I don't have time to make copper shims for the MOSFETs

After running 8 loops of Time Spy the max VRM temp was 68*C - down from 78*C after only 2 runs of Time Spy.

The bottom of the case was not significantly warmer and thus far I have not experienced any changes in fan duty cycles.

I do welcome anyone here to try sinking their MOSFETs or coming up with other ideas. Sadly I don't think I have thick enough thermal pads to sink them (also that's quite the distance, the thermal conductivity would likely be poor anyway).

 

@custom90gt could you post a picture of where and how you placed the thermal pads on the sensor? I am aware of the drawing posted earlier here, however I would feel safer if I could see an actual image of the pad application.

Regarding the actual VRAM temps: I replaced the stock pads with 'Alphacool Eisschicht' pads, they claim to have 14 w/mK. Eisschicht is a german word and means icesheet, or layer of ice. My first tests (playing Dota2 for at least 30 minutes, that is also how I recorded the temps posted earlier) show only a minor improvement when comparing 'Ambient' temps from before to the new pads, around 1 to 4 °C. One could argue, that this makes no difference when considering room temperature and margin of error. However, I will investigate this further.

@Brian Anderson when I was playing Dota2 on my XPS yesterday ans had HWInfo64 open on a second screen. Flowing gently down behind F7/F8 reduced the 'Ambient' temps also for me by ~10 °C accodring to the temps reported by HWInfo. Just to confirm your findings Good Job!

 

I only put thermal pads on the larger R22 chokes. I can take a picture later today after I get out of school, but if you look at the picture on page 27 you can see the large grey R22s.

 

Thanks for the page number! This makes it clear for me where to put the pads, I guess. Just to be safe, you mean the the ray square things I marked green? I hope GoNzO does not mid that I alterd his image ^^

http://imgur.com/U4qgbLp

 

Yeah that's where I put them because it was the easiest thing to sink. Nothing weird has happened yet, they are constantly cooler for me and I don't have any real change in the temp of the bottom of the laptop.