[Alienware 17R4 / 15R3] - Disassembly + Repaste Guide + Results

Regarding the details on the pads, I'm not entirely sure as I sent my device to @iunlock for his repasting service. I tagged him so hopefully he can chime in about that.

Sent from my VS995 using Tapatalk

 

You can use 0.1mm thermal tape and yes the 8w m/k pads will work great as they have a higher elongnation percentage.



::iunlock::

 

Hey iunlock I have a question. So I re-pasted with Kryonaut and it did lower the temperatures of my 15 r3 significantly but I am considering using Conductonaut because I'd still like to lower the temps even further. My Core temp differential is pretty good and always within 5C of each other. Would you still recommend switching the thermal pads out?

 

Absolutely. Thermal pads do make a difference. Also, I'd stick with Kyronaut on the GPU and Conductonaut on the CPU. Great combo and it lets both shine at their finest as the GPU temps seem to be better with Kyronaut over Conductonaut due to the solidity of the coverage.

 

Ah ok. I'll give that a shot then. Thanks for your help!

 

If your temps are great already with your repaste, why even open it up again to get that extra 2-3 degrees at this point unless you just like to tinker with it. Chasing the extra degree or two isn't worth the possibility of damage or a damage screw.

 

Well I never said the temps were "great" they were significantly lower It went from max temps of 95C to about 85-86C. Also if I thought all I would get is an extra 2-3 degrees lower I probably wouldn't do it. I have looked at many people's results and they have had much better temps going from Kryo to Conductonaut have had double digit drops in temperature after swapping thermal pads and using Conductonaut compared to Kryonaut. If I could keep the temps in the 70's then that's what I'm going to shoot for .

 

Hi everyone. I thought to replace the GTX 1070 heatsink to 1080 and found a couple of options:
1. ( Link) ~85$ (0K2PKV, AT1QB001ZC0)
2. ( Link) 177$ (0WM7X0, AT1QB001ZC0)
But, I'm not sure about originality. What do you think about this?

 

I'm on Conductonaut and intense gaming session for many hours with a 4GHz OC stays below 80C with the back raised no problem

Sent from my SM-G935F using Tapatalk

 

Any thoughts on the Fujipoly Ultra Extreme 17 w/mk thermal pads instead of the popular 11 w/mk version?

 

I'd avoid as they are so much more difficult to work with being so brittle for very little (if any) gain. Unless you're an absolute pro like the Team LHz guys you won't be able to get the best out of them. I used the 14w/mK ones and just about got it right as these were quite brittle also

Sent from my SM-G935F using Tapatalk

 

Has there been any updates on the PCH heatsink mod? Usually I keep up with this thread but it disappeared from my alerts and I forgot to check it haha

 

I finally remove the 0,5mm pad, and the core differential is still important but the whole temps are much better (max 73°c for the hottest core after 1 hour playing bf at stock clock and 83°C OCCT 4ghz).

 

I'm going to do this early next week. Using a 0.5mm pad with a 1.0mm copper shim then a 1.5mm thermal pad on top

Sent from my SM-G935F using Tapatalk

 

I'd love to know if it affects max GPU/CPU temps in a good way. I still haven't gotten around to checking out my LM to add more or use a copper shim I bought, I'd like to think the temps can be a bit better

 

After repaste with Grizzly Conductonaut and Fujipoly thermal pads the standby temperature is 35° with OC 4.1. The stress test with OCCT after 1 hours, the maximum temperature per core is 66°/68°. Thks @iunlock...

Room temperature: 24°
6820hk + gtx 1070 + bios 1.0.9

 

did u undevolt for 4.1GHz overclock?

 

Nop... What is it?

 

was your system tuned by team LHz? if it is, it's most likely undervolted too. u can check that in intel XTU

 

Today change the motherboard and heatsinks that Dell sent me. In addition, replace the original thermal pads by fujipoly and apply Grizzly Conductonaut in the cpu and gpu. (I followed the @iunlock guide) I did not make any more changes.

I have no idea what you mean by undervolted... can you explain me?

look

 

Try lowering your core voltage offset a little to perhaps 0.100V.

Great temps btw...good job.

 

wow. really good temps even without undervolt. well done

undervolt is basically lowering your core voltage offset so u feed less voltage to the CPU which can lead to lower temperatures, but it will lead to instabilities and blue screens as well. so u need to find the right balance.

 

I just read about it. I just change OC to x41 and pl1 pl2 to 100.000.

only core voltage offset? Or do I have to change another parameter? When set -0.100 Windows crash...

I can only leave it at -0.070. But, the Score on XTU benchmark down to 1280 (before 1380)

 

@iunlock, when run OCCT, after 49s say this stopped. Error detected in core # 3. With or without core voltage offset tuning.

He does not tell me what the erroris. What can be? Temperatures do not exceed 63 in each core... OC to x41 and PL1/PL2 100.000. Does not happen in OC x40 PL1 / PL2 100.000

 

i dont think 6820HK, generally speaking, can sustain 4.1GHz without some kind of overvolt. Even with default voltage it has trouble getting 4.1GHz stable. My own cpu can get 4GHz stable on a -50mV undervolt. anything more than -75mV i'll be asking for trouble.

-100mV is near impossible to achieve on a 4.1GHz OC so i wont even be bothered with that unless you have a darn good silicon.

 

On 4.1GHz OC and -40mv undervolt is sable for gaming, play videos and use apps like Photo Shop, Chrome, makemkv or handbreake to compress videos... Windows never crash. Only fail when run OCCT test... but windows dont crash, only stop the OCCT test.

En 4.0GHz OC + -40mv never fail or crash, even on OCCT test.

Anyway, before repaste and changing mother+cpu+gpu+heatsink, could only take it to 3.8mhz and had temperatures of over 95 °. Now, I am very happy with my max temp of 66°.

 

I was testing with bios 1.0.12 with OC 4.1Ghz PL1/2 100.000 without undervol/overvolt. OCCT did not fail again after 1h. Then, I put HandBreake to work by converting a uncompressed REMUX video file of 35GB 1080p23 H.264 to 720p30 H.265 (compressing audio too) While running other applications and surfing the internet with Chrome watching Netflix whit Edge. The maximum temperature was 69° (avg 43°) after 1h 20m of hard work. And everything worked perfectly. 6820hk OC 4.1Ghz definitely is stable with bios 1.0.12. Windows responded perfect, never hung up.

 

iunlock ,
1mm thermal pad does not fit the two elements marked on the picture after the replacement (need 2.5mm maybe) ...
I incorrectly chose the thickness of the thermal pad?

 

Changing the parameters of the integrated video card (intel hd 530) improves the CPU performance or temperature?

 

Nope, leave it it's a waste of time

Sent from my SM-G935F using Tapatalk

 

Is it possible that when using Graphics Amplifier the CPU temperature drops more?

 

Good job

Sent from my SM-G935F using Tapatalk

 

Graphics are old with bios 1.0.9. With the bios 1.0.12 OC 4.1Ghz PL1/2 100,000 without undervolt the temperature never exceeds 69 °.
But, I use GA ... maybe that influences because the iGPU is disabled.

 

Yes because it's a shared heatsink

Sent from my SM-G935F using Tapatalk

 

That's normal don't worry about it, are you sure you've properly unlocked the power limits? If you haven't then it will seem like it's running at 4.1, temps will also be good but the CPU will just underperform. Ask @Papusan or pathan

Sent from my SM-G935F using Tapatalk

 

Yes, adjust accordingly as the size needed will differ depending on the heat sink as they are not all created equal.

::iunlock::

 

I use PL1/2 100.000, at that limit you mean? Also try 240.000 and 300.000. The result does not change much by 2 or 4 degrees of difference per core.

 

Use what is needed. A limit is just a limit. Too low will cause throttling.

 

100.000 work fine. Cpu Packard power up to 84.xxx. The performance is very stable. Windows never crash even when I run really heavy tasks like creating .mkv h.265 that take 1 or 2 hours with high priority. How do i realize if the cpu runs underperform?

 

Not sure what you mean. But everything below the maximum turbo boost in all types load 24/7/365 is underperforming.

 

I mean what this guy says here...

 

What @Papusan and @Pete Light are saying, is by not setting power limits correctly, your cpu will hit the OC'ed value sometimes and not consistent.

 

I get it. But, I only know two ways to set pl1 and pl2 ... I do it from the bios or from XTU. in both cases. I use 100,000

 

Or ThrottleStop

 

Using ThrottleStop to overclock or overvolt your CPU or to disable a laptop manufacturer's throttling scheme may damage your computer and is at your own risk.

Very risky for me.

 

See above. You talked about XTU All too high Core Voltage / Heat. Yes, then we agree. But not otherwise.

 

XTU is an Intel official. I think it's more reliable, or not?

 

NOPE. XTU eating more of your processor cycles. Not wanted. Pretty graphs - more clock cycles does not improve performance, right?
@unclewebb

 

I'm going to try it.

I tell them something. Dell sent me a new display for my 17r4, when install it ... surprise! It was defective. A gray strip in the middle. Is incredible. Now I have to wait another 30 days to send me another one.

 

What with Next Business Day? Dell always offer Premium Support. First year is included