*** MSI 16L13 (Eurocom Tornado F5)/EVOC 16L-G-1080 15.6" Owner's Lounge ***

Here are the results for El Cucuy...

 

Now I am going to follow what Brother @Papusan did with the Bitspower IHS lapping and see if that helps the CPU in this little micro-beast run cooler.

Spoiler: Why was this necessary? Here's why...

Most stock IHS are also not flat.

Just getting started...

About half-way there...

Almost done...

More to follow... stay tuned for results.

 

But, but Yoo will lose the nice silver shining bro Fox Design before functionality. Yoo know what I mean? Hint... First letter starts with J, T or B

 

So while my 1070 is much smoother and the clocks don't fluctuate as much with the new bios Talon had linked earlier, my card will still no boost above 900 mV. I am at a loss here, does anyone have any ideas? GPU Z estimates its only using 93W so I have much more thermal headroom left.

 

It will not boost above 900mv because boost 2 is disabled. With boost 2, when it goes above 900mv, you hit power limit/vrel right away anyways.

 

Is there a fix for this by chance? Should have plenty of room to go above if the TDP was set a 120W correct?

 

Default TDP is 115. 8A vbios does hit 115+. I hit it when doing Heaven Extreme settings.

 

What clocks and voltage do you hit when running the heaven benchmark?

 

0.881v and mine hit 117W and Power Limit throttle it back. Mine is overclocked 200+/+300 (core/memory). Clock at 1700-1835. When Heaven does not run in extreme mode, constant 1835mhz.

Edit: 0.818v

Sent from my iPhone using Tapatalk

 

Btw, looks like nvidia knew about this? I am reading the nvidia control panel supposedly has a debug mode that downclocks the gpu to nvidia default spec? I guess they expected factory overclocks of boost 3.0 to cause a bit of problem.

 

Gotcha, thanks for the replies. Wish they would bump the voltage up a bit for us because i feel like the cooling could handle a bit more than the 115W.

 

If 8A gave us voltage control, we could in theory hit higher clock speeds but be limited by the power limit.

 

Yea i dont need voltage control, but control on the power limit. Unlocking the power limit would let us a run a bit more voltage through the cards and reach a bit higher clocks.

 

Yup, encrypted vbios sucks.


Sent from my iPhone using Tapatalk

 

By the way. This little tweak lowered by PCH temps by about 15-20°C and I know Brother @Donald@HIDevolution would agree with my use of copper here. The shim I added a while back didn't hurt, but it really wasn't enough to have a meaningful impact on the PCH temps. However, it made the perfect foundation for adding these three finned copper heat sinks. Now I'm seeing PCH temps in the low 50's, where before it was mid to upper 60's and sometimes in the 70's.





And, for those that may be thinking about buying pads, here are the size of what I have on my MSI 1080. Reducing all by 0.5mm should also be doable. That's what I will do next time. Below are the sizes of the pads as it ships from the factory.

 

FWIW, I bought a 230W PSU along with a spare 330W when I ordered my EVOC from HID Evolution. I keep the 330W at my two working locations (I have a desk in my office and a rolling cart in the living room for evenings). But I travel with only the 230W, and I watch my power usage more diligently when I'm mobile. No problems so far. I don't do heavy benchmarks or GPU-melting games when I'm traveling. But the 230W is plenty for programming, productivity tasks, and *light* gaming.

 

Sweet! Copper approved with results like that!

 

Wow, big difference in temperatures! I wasn't expecting that given the virtually no differences you saw on your other laptop - do you have any thoughts why the change is so much larger on this one?

 

Maybe the Unified Gpu HS design(the big VC Grid)?

 

The unified heatsink design - you're referring to the Tornado aren't you rather than his other laptop. VC Grid, what's that? Why do you think unified heatsink enables better temperatures with a different paste in contrast to a heatsink design where it's not unified across CPU & GPU?

 

Yoo look after why it make a difference. I tell you what the differences is between the two machines. If yoo made equal good application, correct amount of thermal paste on both machines and have equal ambient temp... What have you left?

FYI. The Grid in DM3 is Unified HS. Aka Shared between both graphics.

 

well i just checked with higher boast and i can hold them fine during benchmarks but games will drop them to ~1600. also im hiting 0.8930V but only 93watt in hwinfo.
also i dont get any perf cap.
Edit: i can not exceed 98watt even with 300+ on core, i dont get any perf fap in gpu z and voltage remains 0.8930.

 

Ha, Papusan, you're talking in riddles! First you tell me it could be due to a unified heatsink, and then you tell me both of his laptops have unified heat sinks - what sense does that make, or can you tell me where I'm missing something? I'm not really understanding your last 2 posts.

 

Sorry if you don't understand my <very good> English. But both use same thermal paste, same amount of paste, tested in about same ambient temp, same good paste application. And the job is performed by a experienced user.

But the same thermal paste used on both will give different results. One better and one not as good. What have you left... Only 2 alternatives remain. The heatsink and how the thermal paste works with the different setups. You have ruled out everything but two options is left. I haven't the sources here now, but I have read somewhere the whole setup can make a difference although you use same thermal paste.

 

I'm not thick, I do understand logic and the scientific method, you confused me by saying initially in one post that the difference could be due to a unified heatsink, but then in your second post you say both laptops have a unified heatsink - this is where the logic fails, which is why I asked you if I was missing something.

 

I don't meant you are thick in the box
I should probably be more clear. The first post I posted actually tells the whole... Read it one more time. The Heatsink design. Or shall we say different cooling setup. Can give different results.

 

Ah Papusan, that's not any clearer, but it is more general & vague! I do appreciate that different cooling setups can give different results, that's a very general & vague statement though. What are the specific differences between the two systems that you think explain the better success of Phobya over Kryonaut? (Be aware I'm not aware of the exact cooling layout of his other laptop, but I'm familiar with the Tornado from reviews). Doesn't matter if you're not sure, I'm just curious, and I suppose Mr Fox can always provide some thoughts on why Phobya performs so much better on the Tornado in contrast to his other laptop. I'm just curious about it.

 

It can be as easy as how the paste spreads + paste thickness on HS for the two different setups. The two paste are somewhat different in viscosity as you have seen. The spread can be good on both but the results can be somewhat different. Many tests two different thermal paste brands and get very good results with one specific brand. And other can use same two brand paste but will come to different results although both users are experienced aka good with applying thermal paste. We don't talk about very big differences here. As I said in previous post. Have read that the cooling setup can and will make a difference for what thermal paste you should go for. Nothing is put on stone. Damn, some don't even getting the best results with Liquid metal vs. good thermal paste although heatsink is good enough for use with it.

 

Ok, thanks.

 

Just finished to pay my order by @woodzstack for this "little" beast!

Here my config:

EUROCOM Tornado F5 MS16L13 (No Branding/Barebones kit)
15.6-inch (39.6cm); 120Hz FHD 1920x1080 AUO Panel
VGA / GPU Graphics Option : GTX 1080N
Delidding of CPU + Liquid Thermal paste
Processor / Intel Core i7-6700K
RAM Corsair CMSX16GX4M2A3000C16 2 x 8GB 3000MHZ CL16
Keyboard (Language) Backlit; English; USA/Canada; QWERTY
WiFi INTEL 7265
CMOS wire modification
Modified bottom Panel by HIDevolution/EVOC
Standard Battery (internal) 8-cells Smart Li-lon Battery
330W PSU
Standard Power Cord - included USA / Canada
1 Year Return to Factory Depot with 1 Year of Tech Support

Hope it will arrive soon

Edit:

added more 2 years of warranty for a total of 3!

 

Congrats - a wise config!

 

As yoo can see below, not the big changes from Intel. They continue in the same way as before. But also AMD is making their lid equal as Intel now.


Amd's change from what they used before.. Compare with the pict above.


Edit. See also this post

 

OK, here is the "more to follow" delivery part to show how the before and after results compare.

Inside of the spoilers you will find test results on the Tornado F5 when I first got it from @Eurocom Support and delidded the CPU and added CLU. Compare the temps in the spoilers to new results produced today. Bear in mind, the results in the spoiler were already a major improvement over stock. Considering the examples in the spoilers were a huge improvement over stock (15°C to 20°C temperature reduction) back in January, adding the most current results to the picture and the change is nothing less than phenomenal.

Spoiler: Reported 1/23/2017 - Stock 7700K after Delid and CLU

New results after lapping Bitspower IHS (compare to above spoiler) - CPU clocked at 45x4, Conductonaut on die and on top of IHS...

Spoiler: Reported 1/23/2017 - CPU 47x4 after Delid and CLU

New results (today) after lapping Bitspower IHS (compare to above spoiler) - CPU clocked at 47x4, Conductonaut on die and on top of IHS...


@Papusan @bloodhawk @Prema @Johnksss@iBUYPOWER @Donald@HIDevolution @Ted@HIDevolution @thattechgirl_viv @Zoltan@HIDevolution @Phoenix @Eurocom Support @Meaker@Sager @TBoneSan @j95 @dspboys @iunlock @Mobius 1 @GTVEVO @Cass-Olé @Khenglish @KKthebeast @hmscott
(if I didn't tag you it's because you're already visiting this thread frequently - no offense intended if your name was omitted)

 

Nice one, low temps! I noticed though, your stock CPU at stock voltage is drawing about 13 more Watts than your overclocked CPU - you lowered the voltage & overclocked - so the 8 degC temperature drop you're seeing between your stock run and the 4.7Ghz lapped Bitspower run are coloured by the fact that you're running 13Watts more through the stock setup. This kind of now looks like the temperature drop is not that great with the lapped Bitspower & liquid metal on heatsink. Have you worked out what the temperature difference is if you run the same Watts through the CPU?

 

Those are some seriously impressive results!!!
I really need to lap my Bitspower IHS as well. Thank you for doing this. Do you remember what your ambient temps were like?

 

Stock is stock and I did that to show what people might expect if they let Intel make decisions for them. Obviously, that's never a good idea.

Look at the 4.7GHz results to see a direct comparison.

One also has to remember that cooler temps allow for lower voltage. The hotter a CPU gets, the more voltage it takes to be stable, so using less voltage may not even be an option if the CPU is not running nice and cool. It becomes a vicious circle. That may also explain why Intel's stock voltage is so insanely high, as they have no control of the wide variety of temperatures that might be encountered "in the wild" so to speak.

Side note: I ran the memory at 2400 for today's tests because that's all I was able to run the memory at when I did the original testing in January. Memory overclocking was not possible then because MSI stock BIOS is a joke.

You're welcome. I need to do it on the DM3 now.

My temps are always the same year round in my house. Thermostat is set to 70°F.

 

I still think the only way to truly understand the value of your bitspower lapping & liquid metal mod is to run the same Wattage through the CPU, otherwise you're not comparing like for like in terms of the cooling ability of your newly modified setup. For people to jump on the Bitspower lapping wagon they'd best understand what kind of temperature reductions they're gonna see when running the same Watts through the CPU.
EDIT: or just run the same CPU frequency & voltage and then compare the two (that's comparing apples to apples!), that's more accurate than just running the same Wattage through the CPU. Yeah, so you gotta run the same CPU frequency & voltage to see the effect of your Bitspower lapping mod. (In bold because this is better than my earlier recommendation to run the same Watts through the CPU - takes into account your temperature/voltage/wattage viscious cycle you mentioned)

 

If you don't like that stock example, just ignore it and look at the 4.7GHz result only. You will see the load voltage is essentially the same. The 4.7GHz applied values are the same, but the real-time values will vary by CPU temperature and load. It's not possible to lock them at a fixed precise value, as KBL and SKL exercise some degree of autonomy in spite of applied settings.

The overclocked results are more important anyway. I don't run my CPUs stock.

I did run the CPU frequency and voltage at the same values for 4.7GHz. That's the one that is most important.

The 45x4 example is an overclock, not a direct comparison to the BIOS defaults example. But, it does show it running cooler with an overclock than stock. I don't have a 45x4 example from before because I just never saw any reason to run the CPU at such a low clock speed.

 

I edited my previous post while you were writing yours, have a look at the bit in bold I added. If you don't do that then people can't really see how valid the cooling effect of the Bitspower Lapping Mod really is, because you're not isolating that variable, you'd have too many variables going on otherwise - not comparing apples to apples.

 

I replied above. People do have a perfect example to look at. The 4.7GHz example is using the same CPU voltage settings. 1.100V static voltage in both examples. It never works out to that precise voltage amount with SKL or KBL, but that is what the voltage setting was.

 

Ah, thanks for explaining, I missed that one included there under that particular Spoiler banner in your original post! That means you're running 8 degC cooler with the Bitspower lapping mod under the same CPU settings. That's great, my mistake for not noticing that one, apologies!

 

No worries, bro. Glad we're on the same page now.

Isn't it sad that BIOS defaults using Intel reference values runs hotter than using "right" settings at 4.7GHz? Just stupid, isn't it?

 

Yeah, it's mad how much voltage they wanted to pump through a stock 7700K!

 

@Mr. Fox awesome man. What method and grits did you use or find best for lapping the IHS?

I have a bitspower IHS sitting around that I've been meaning to put on. You're also still using the Skylake one on your Kabylake correct?

 

thanks for the posting results, improvements are huge!.
also the bold part explains why i can keep offset to -150mv fine as long as AC is on these days but im getting BSODs without AC. i never got these crashes before when temps since temps werent as high as its now here(high 30s).

 

Sorry Bro Fox, that I pushed you to take this step sandpaper ain't cheap And your conclusion? Probably same as mine... Intel's concave lids will never work properly for cooling the chips!!

Edit. The lid on Kaby is probably worse than the one Intel used on Skylake. As you know.. . A lot have experienced higher temp with 7700K vs. 6700K if you pushed same Cpu Package Power. They can most likely get better results if they take the extra step and time lapping the lid.

Exactly as expected

 

I used 400 wet sand to cut it down, then polished it up with 1200. Use a piece of glass to keep the sanding surface flat. If you don't have one, a really cheap picture frame works well. I bought an 8x10 picture frame from Walmart for $1.99 to use for this. It also makes a nice surface for cutting thermal pads with a razor blade. Yes, I am using the taller Skylake Bitspower IHS on both of my laptops.

My conclusion is that it is worth the extra effort.

 

For someone who doesn't have the extra time or cash would you recommend the Skylake lid with Grizzly Conductonaut on the cpu and phobya on the gpu?

 

What did you do to contain or remove the metal particulate and prevent it from contaminating the motherboard and causing a short? I'm sure if anyone could find a way, it's you, but this sounds like a tough problem to solve.

 

The Bitpower IHS is a separate piece. I am sure he sanded/rinsed it away from the mobo and other computer components.