CCD MI-6: Performance in a 6.7L MIcrotower

I have gone back & looked at the post(s) again; and I stand corrected; I was reading the 1060 as 1080…

I definitely need bifocals at this point in my life…!

But I will not concede the point on the cable measurements assisting with extrapolation on cable lengths needed for similar locations…

If a MB cable is listed as (purely hypothetical here) 18cm, and another MB has that same connector about (eyeballed) 3cm up; then I could safely assume (extrapolate) that a 21cm or 22cm length cable would be needed…

This the last I will discuss on the matter; to defer to the (implied) wishes of the Thread Creator / Owner…

Back to MI-6 related meanderings…

So…! I checked Superbiiz… They currently list mITX-sized 1060s from EVGA & Zotec, and the 1070 from Gigabyte, but none are showing as in stock…

But if we go by the links over in the ITX Pacsal GPUs thread, I am gonna guess you are going all in for the Gigabyte 1070…?

I REALLY wish they would have rethought the output choices on that card though; virtually every other Pascal card out there has the 'standard' DisplayPort x3, HDMI x1, & DVI x1 setup; Gigabyte gives us DisplayPort x1, HDMI x1, & DVI x2… WTF…?!?

For me personally, the 1060 (already making sense due to the lower pricing) will allow me to run triple HD monitors via the x3 DisplayPort output…

I will be starting with one HD monitor, adding the other two as budgeting allows; the Gigabyte GPU would make me want one of those curved screen 34" Acer Predators…

But I can totally understand the need to grab one for testing in your chassis, as it is probably the 'hottest' GPU available in mITX, so it would be best for testing temps…

Load the MI-6 with the top power usage / heat output components for testing (within reason, the X99 platform does not come cheap), any lesser wattage components could only result in overall lower temps…?

Looking forward to the temp testing with the i7 6700K & the mITX 1070 in the chassis…!
 
Firewolfy said:
Thx for the pics of the mb's. Good for quick reference.

Looks like most of the setups will have the 24 pin and the 8 pin conns at somewhat similar distances from the psu, so that is good.
Click to expand...

Well, I'd say the 8/4 pins (some Z170 have 4 pin connectors) is 70/30 being top (close to I/O) or front (close to ram slots). Still, since positioning for said components is up to the manufacturers I wouldn't bother too much measuring cable length. Keep in mind that not every psu has the cables in exactly the same location. Example (Silverstone SFX 450W Gold and Corsair SF450 both with the fan facing down)

sf450-600_03.png


12-Front.jpg




It all simply ends up adding up. Lets not forget, also, on the cables used by each psu.

hs_SFX450w_042.jpg


ECDB6A309CBA4B089C319C17CB5FA0B1.ashx


Corsair cables are very flexible and easy to hide because they are flat. On the other hand, the Silverstone ones are very stiff as they aren't flat and have a sleeve on top of them. In your pics with the Corsair psu you show how you doubled-down the cables to save space. It would have been much, much harder to do so on the Silverstone cables because they aren't flat. I know I suffered in the NCASE M1 with those cables...


All in all there is so much variance involved in the whole thing (psu connector placement + psu cable type + motherboard connector placement) that every user should just acknowledge that they might have to trim every cable to the specific length required unless they find a way to tuck the cables away. Even on an M1 cables are an issue... on cables half the size, 4 times the problem.
 
Ok, I performed some preliminary tests. Here are the results:

System: Mi-6 prototype case black textured powdercoat (perforated inserts for ventilation), i7-6700k at stock speed (4.2GHz), Asrock Z170 Fatality ITX MB, 2x8G memory Corsair LPX 3200, CPU Cooler Skythe Big Shiriken 2 revB (58mm), Corsair SF450 PSU with stock cables, Samsung 850EVO M.2 500G SSD, WD Blue HDD 500G, LiHeat riser.

Test equipment: Multimeter thermocouple temp monitor, IR temp sensor for ambient room temp, KillaWatt plug power meter.
Software: Prime95 v26.6, AIDA64, Realtemp monitor, CPUID HWMonitor,.
Logging methodology: Max measurements taken from monitor software, except max CPU temp defined as average of max temps of individual cores.

NOTE: No GPU installed during this testing, so this is just prelim and will be repeated with 1070 GPU.

Now, some results:
Prime95 v26.6 (using small FFT setting) for 10 minutes, 100% fans: No cover: 70C (26C amb, SSD 39C). With cover: 70C (25C amb, SSD 37C).
Prime95 v26.6 (using 8k-only setting ) for 10 minutes, 100% fans: No cover: 68C (26C amb, SSD 38C).

SSD Thermal testing:
AIDA64 Disk benchmark (linear read, looping, auto) for 40 minutes, 100% fans: No cover: 42C (26C amb), max read 531, min read 450. No drops or throttling.

EDIT: I started out using AIDA64 stability test, but found a lot of articles suggesting it was a weak test, so I switched to prime64. Used v26.6 based on suggestion of articles and posts. I can go into more detail on that if anyone is interested.

So, bottom line is that temps are looking good so far for a 95W TDP processor.

Another EDIT: I added ssd temps in the prime test results above.
 
Last edited:
How were the noise levels with fans @ 100%…?

Cover on, of course…
 
Firewolfy said:
The cooler fan isn't bad. That noctua 92mm slim on the bottom is pretty loud, which was surprising.
Click to expand...

I think any fan is going to seem loud if running at 100%…?

Maybe things will quiet down once you set some fan curves & stuff; not just running everything full-tilt…?

I am so used to the constant 'idling turbine' that is my (almost) seven year old MacBook, and I usually have a 20" box fan running on low or medium in the room; so some fan noise would be alright, plus wearing headphones while gaming (I play WoW with my son & the headset is better than yelling back and forth from room to room) really negates the perceived fan noises…

Gotta have some sort of white noise to sleep as well, I remember one of the first nights I was out in the boonies in Maine,TOO QUIET, couldn't sleep for hours…!
 
Firewolfy said:
Ok, I performed some preliminary tests. Here are the results:

System: Mi-6 prototype case black textured powdercoat (perforated inserts for ventilation), i7-6700k at stock speed (4.2GHz), Asrock Z170 Fatality ITX MB, 2x8G memory Corsair LPX 3200, CPU Cooler Skythe Big Shiriken 2 revB (58mm), Corsair SF450 PSU with stock cables, Samsung 850EVO M.2 500G SSD, WD Blue HDD 500G, LiHeat riser.

Test equipment: Multimeter thermocouple temp monitor, IR temp sensor for ambient room temp, KillaWatt plug power meter.
Software: Prime95 v26.6, AIDA64, Realtemp monitor, CPUID HWMonitor,.
Logging methodology: Max measurements taken from monitor software, except max CPU temp defined as average of max temps of individual cores.

NOTE: No GPU installed during this testing, so this is just prelim and will be repeated with 1070 GPU.

Now, some results:
Prime95 v26.6 (using small FFT setting) for 10 minutes, 100% fans: No cover: 70C (26C amb, SSD 39C). With cover: 70C (25C amb, SSD 37C).
Prime95 v26.6 (using 8k-only setting ) for 10 minutes, 100% fans: No cover: 68C (26C amb, SSD 38C).

SSD Thermal testing:
AIDA64 Disk benchmark (linear read, looping, auto) for 40 minutes, 100% fans: No cover: 42C (26C amb), max read 531, min read 450. No drops or throttling.

EDIT: I started out using AIDA64 stability test, but found a lot of articles suggesting it was a weak test, so I switched to prime64. Used v26.6 based on suggestion of articles and posts. I can go into more detail on that if anyone is interested.

So, bottom line is that temps are looking good so far for a 95W TDP processor.

Another EDIT: I added ssd temps in the prime test results above.
Click to expand...

10 minutes isn't enough to saturate the case with heat (which is the interesting point to look for). Remember that you are testing the case, not the coolers inside of it. This is why you should look for sustained testing, IE what temperature will I have if I let my computer rendering for 4 hours? As it is the SSD temperatures climb 12ºC ambient in only 10 minutes. And only using the cpu.
 
prava said:
10 minutes isn't enough to saturate the case with heat (which is the interesting point to look for). Remember that you are testing the case, not the coolers inside of it. This is why you should look for sustained testing, IE what temperature will I have if I let my computer rendering for 4 hours? As it is the SSD temperatures climb 12ºC ambient in only 10 minutes. And only using the cpu.
Click to expand...

As i mentioned, this is prelim testing. It is the first testing done. It is not meant to be conclusive. It is first cut testing to make sure nothing crazy is going on

I shared it to give a quick update on what is going on. It is not conclusive testing and conclusions should not be drawn.

I have mentioned in posts previously that I will be doing testing to address case heat soak. You are preaching to the choir, as this is in the testing planned that i already discussed in posts.

I will continue to post interesting updates even if not conclusive or exhaustive results. I will spell out specifically when i think i can make a conclusion and will supply rationale to back it up .

Thx
 
Boil said:
I think any fan is going to seem loud if running at 100%…?

Maybe things will quiet down once you set some fan curves & stuff; not just running everything full-tilt…?

I am so used to the constant 'idling turbine' that is my (almost) seven year old MacBook, and I usually have a 20" box fan running on low or medium in the room; so some fan noise would be alright, plus wearing headphones while gaming (I play WoW with my son & the headset is better than yelling back and forth from room to room) really negates the perceived fan noises…

Gotta have some sort of white noise to sleep as well, I remember one of the first nights I was out in the boonies in Maine,TOO QUIET, couldn't sleep for hours…!
Click to expand...

Lol, yes it should be better when not 100% wide open throttle!

Although its good to know you'll be able to get the white noise if you need it...
 
Good day today, got the GPU!

Initial installation pics--------

1070%20box.jpg
GPU%20installed%20w%20cables.jpg


top1.jpg
gpu%20top%20clearance.jpg


Back%20panel.jpg
cpu%20side.jpg


The cabling is kind of routed, but as I mentioned earlier, there is too much of the USB and audio cables. The PSU cables take some work but I think I got some clearance above the case fan.
I started doing some prelim testing with the GPU this evening. I'm trying Furmark for the gpu, and some AIDA64 Disk benchmark testing at the same time for the SSD (all with the cover on) CPU isn't doing much, so if anyone has some severe test software that taxes the cpu, gpu and ssd, let me know.

I'll try to get some data up tomorrow.

Finally, it's more than cad modelling and daydreaming!
 
Just popping in to say I love this case. Keep up the good work!

My Dan A4 won't even ship for a few months, yet here I am already ogling this smaller case. Granted, I'm sure it will be a great deal longer before this one is available.
 
Thx for the feedback. As far as space, I want to leave as much space for working and airflow as possible. But it is getting to be a little like a console case or that steam machine case a year or so ago, where everything fits in like sardines in a can.

EDIT -- One note regarding the riser for that gpu. The li heat 300mm that is in there showed up as gen 3.0 so that is good, and i ran a bunch of games and benchmarks last night and had no pcie connectivity or emi problems!

Woohoo
 
Nice clearance on the GPU side, the 1070 looks comfortable in there.

I really wish they'd use a bigger fan on it though. Given how much height they gave it, the fan kinda looks off-centered...
 
Hahutzy said:
Nice clearance on the GPU side, the 1070 looks comfortable in there.

I really wish they'd use a bigger fan on it though. Given how much height they gave it, the fan kinda looks off-centered...
Click to expand...

Yeah, i had to give it some room there. The gpu pcb sticks up esp in the back where that case press-in nut is, so i had to add clearance to be able to lift it up out of the pcie conn. More room the better i guess.
That gpu has fins that run fron to back. It pretty much blows out of the back and the front. Very little radial flow at all out the top or bottom, so it is good that it is less than 7" long so there is room for redirecting the flow up out the vents.
 
Reading up over in the Hutzy XS chassis thread…

Got clued into the fan that the Project Orthrus chassis is using on the NH-L9i heat sink, namely the Scythe Kazu Jyu Slim…

This is a 100mm slim fan with options for either 100mm mounting holes or 92mm mounting holes…

Wonder if this could fit in the bottom of the MI-6…?

Only 0.5 cfm less airflow than the NF-A9x14, 200rpm slower & supposed to be quieter…

Discuss…!

EDIT TO ADD:

Could also work as a replacement for the NF-A9x14 on the NH-L9x65 heat sink; a little extra size, a little less noise, and no ugly 'coffee & cream' color palette…!?!
 
Great progress!

I noticed that cpu cooler is orientated front to back. Is it possible for you to rotate it 90 degrees and see if that helps with cpu cooling?

I think once noctua releases the 120mm slim fan we will have a very potent cpu cooler combo of scythe big shuriken + noctua.
 
I noticed the back of the graphics card is oriented towards the front panel of the chassis,

since hot air vents out the back of the graphics card towards the front panel it might cause slightly hotter temps.

unless that hot air is being exhausted by the 92mm fan but it couldn't hurt to make sure imo.

still I've looked at the temps and its pretty good.
 
prava said:
Huge temperature disparity between ssd and hdd. Where do you have located the SSD? I see the HDD by the fan but can't see the SSD.
Click to expand...

The ssd is an m.2 sitting behind the motherboard. There's very little airflow getting behind the motherboard which is the norm for most cases. I could see a 950 pro overheating though.
 
Yup, the SSD is M.2 on back on MB. The OS and everything is pretty much on the SSD, so it is working during the test, while the HDD is pretty much idle, and getting a lot of airflow.

The SSD temp is below any throttling limits. I would like to get the temp down. Tying it to a heat sink would help, but I have a big cutout behind the MB so there is nothing nearby for the SSD to contact. Well, the test was a synthetic worst-case, so a case could be made that we are fine, as real-world usage won't see that kind of stress....

However the SSD has a SATA M.2 protocol, so if someone sticks a PCIE M.2 back there the temps will likely be high in that synthetic test.

Edit -- M.2 ssd's behind any mb in any case won't get much ventilation, but it still stinks.
 
Last edited:
Just eyeballed the build pics; the Scythe Kaze Jyu Slim 100mm fan will NOT fit in the bottom of the chassis… The Noctua NF-A9x14 is almost (if not actually) touching the PSU…

Hey, in regards to the temps issue(s), what if you increased the height of the chassis ever so much (say, 14mm) to allow for a pair of slim 92mm fans up top exhausting the chassis…

That would be bumping the chassis closer to the 7 liter mark though…
 
A couple thoughts:
  1. the Puget Sound testing on a high-end M.2 pcie ssd here: https://www.pugetsystems.com/labs/articles/Samsung-950-Pro-M-2-Throttling-Analysis-776/ , tested the 950 Pro (pcie). It throttled in any location under torture testing, just quickest under the MB.
  2. That info, coupled with the fact that my 850EVO didn't throttle at 54C peak, and all the SSD's I just checked had operating limits of 70C: (Samsung 750 EVO (sata), Samsung 840 EVO (sata), Samsung 850 EVO, PRO (sata), Samsung 950 PRO (pcie), Crucial MX200 (sata), Crucial M500 (sata), OCZ RD400 (pcie)), tells me that the 54C peak with Furmark and Prime64 running may not be too big of a deal after all.
So the temp isn't a show stopper. I'll just do some noodling to see if there is an simple way to get more airflow to the back of the MB or a heatsink for the M.2 SSD.

Boil -- yes, we could add some fans to the top as a last resort.
 
Can we just make the entire chassis out of fans…?!?

I kid, i kid…!
 
alex1577 said:
The ssd is an m.2 sitting behind the motherboard. There's very little airflow getting behind the motherboard which is the norm for most cases. I could see a 950 pro overheating though.
Click to expand...

Ah, ok, now I understand why I couldn't see it anywhere :sneaky:

Firewolfy said:
Yup, the SSD is M.2 on back on MB. The OS and everything is pretty much on the SSD, so it is working during the test, while the HDD is pretty much idle, and getting a lot of airflow.

The SSD temp is below any throttling limits. I would like to get the temp down. Tying it to a heat sink would help, but I have a big cutout behind the MB so there is nothing nearby for the SSD to contact. Well, the test was a synthetic worst-case, so a case could be made that we are fine, as real-world usage won't see that kind of stress....

However the SSD has a SATA M.2 protocol, so if someone sticks a PCIE M.2 back there the temps will likely be high in that synthetic test.

Edit -- M.2 ssd's behind any mb in any case won't get much ventilation, but it still stinks.
Click to expand...

Don't worry about the SSD. The only solution that might exist to cool it down would be to have some sort of thermal adhesive and have it make contact with one of the panels of the case. This should work because the case (as per your test) maxed at 4x ºC whereas the SSD was 10ºC hotter. But that would mean that the cut in the panel would have to go and I'm not sure that is a good trade-off for most.

In any case, 5x ºC temperatures for a component being in the worst possible place (cooling-wise) isn't anything to be scared of. Not that you can do anything to fix it in any case so I wouldn't think about it.
 
Alright, some more test results. I did 3 more tests (G2, G3, G4) after the test I posted the other night (G1).

I wanted to see how bad temps can get if the case fan is not used. This defines whether a case fan is needed depending on the CPU and GPU you use.

Also, I had a feeling that the top vent area was a bit low, so I checked behavior with the top vent panels off. Here is a pic with the vent panels off so you know what I mean.
20160821_214433r1.jpg


Here are the results:
G1-G4%20results.jpg


With the case fan off, the HDD temp rises to the max allowable (50-55), with or without the top vent panels. So that settles the question: If you go high-end and use a HDD, you need a case fan.

Now, opening up the top vent area by removing the panels helped a bit (2-3 deg). I expect 1-2 change to be in the noise, so this is just believable as a relation.

And a bonus--- I took a couple pics of the built up case with a bit better lighting--
20160821_190909r.jpg

20160821_190730.jpg
 
Firewolfy said:
Alright, some more test results. I did 3 more tests (G2, G3, G4) after the test I posted the other night (G1).

I wanted to see how bad temps can get if the case fan is not used. This defines whether a case fan is needed depending on the CPU and GPU you use.

Also, I had a feeling that the top vent area was a bit low, so I checked behavior with the top vent panels off. Here is a pic with the vent panels off so you know what I mean.
20160821_214433r1.jpg


Here are the results:
G1-G4%20results.jpg


With the case fan off, the HDD temp rises to the max allowable (50-55), with or without the top vent panels. So that settles the question: If you go high-end and use a HDD, you need a case fan.

Now, opening up the top vent area by removing the panels helped a bit (2-3 deg). I expect 1-2 change to be in the noise, so this is just believable as a relation.

And a bonus--- I took a couple pics of the built up case with a bit better lighting--
20160821_190909r.jpg

20160821_190730.jpg
Click to expand...

Duration of the "no fan" is 20 min vs 60 min for standard testing? In any case, the drive results are to be expected, to be honest since they will get cooled by the temperature differential of the air surrounding them... and if there is no forced cooling around, they will get cooked by the hot air inside the case.

Keep up the testing.
 
Firewolfy said:
I took a couple pics of the built up case with a bit better lighting--
20160821_190909r.jpg

20160821_190730.jpg
Click to expand...

Good Deal on the continued testing…!

The case looks AWESOME…! Only thin bugging me (as always) is that damned Noctua color scheme…! I REALLY wish they made a version of the NF-A9x14 in black, or even the black/grey of their Redux line…

I fully realize that no one will really SEE the coffee & cream colored fan when the chassis is all buttoned up; but I will know, I WILL KNOW…!

Keep up the good work, Firewolfy…!!!
 
Prava,
I cut the testing short when i hit above 50C on the one test, then just continued that duration for comparison purposes.

Boil,
Yes that fan is "different" looking, lol.

I finally got that IS-60 in the mail, so i will be trying it out. Should be fun to get some cooler comparisons!!
 
  • Like
Reactions: Boil
like this
Looking closer at the newest pics…

Is that the MB 24-pin you have bound up like a Japanese schoolgirl in the pic of the CPU side…? If so, impressive work getting that big hunk of cable into such a neat bundle…

You should, at some point, pull the PSU cables (the ones bundled & bound) and take a few multi-angle shots (maybe lay a scale/rule in the shot as well; you can pretend you are taking pics for CSI…) so those of us with Corsair SF-series SFX PSUs might take advantage of your advance origami work…!

Towards the SATA, you could always take off the end connector & trim the cable back to a single connector (replacing the one at the first spot with the one removed from the end)…

This is a chassis where I would ask a custom cable manufacturer (IceModz or Ensourced, maybe even CableMod) to make custom cables without individual sleeving, just make each needed cable assembly into a bundle of black wires in a single larger sleeve…? I dunno, might save space & make for less disruption/blocking of the airflow…?

I would be interested in results of both the stock fan on the IS-60, and if it were replaced with something like the Fractal Design Venturi 140mm (regular 25mm thick) high static pressure fan… This fan IS 140mm, but with the ability to change to a 120mm bolt-up pattern…

But there is that thing I have yet to really read up on, something about having a fan too close to the vent surface (which the combo of the IS-60 and an 25mm thick fan would do), something something laminar effect…? I dunno…

You really need to increase the size of this chassis some, so we can include a mini-optical drive; gotta have my built in cup holder…!

Would be funny to see an actual mini ODD that fit those tiny CDs that one could never use if one had a slot-loading ODD…! ;^p

Not that I have actually had ANY need for an ODD in, what, the last decade or so…?!?
 
Took me two days to stop laughing over your japanese comment. So politically incorrect!
 
  • Like
Reactions: Boil
like this
Yes, I will pull the cables at some point and get some pics and measurements.

My next work will be running the system with the GPU hanging out of the case and moving that riser all around and see if I get any crashes.

Then I will swap in that IS-60 cooler and rerun some of the tests.

One interesting result: I have left the wattmeter connected to the power plug. The computer pulls about 40w when I am just web surfing right now with a few windows in the background. When I am running world of tanks at max possible settings, it pulls about 200w. That's a i7-6770k stock clock and Giga 1070 itx. All fans on standard profile. I would have thought it would be higher. But hey, I'm not complaining and have a lot of headroom for the 450w psu.
 
  • Like
Reactions: Boil
like this
Quick Status:

My next work will be running the system with the GPU hanging out of the case and moving that riser all around and see if I get any crashes. -And if I do I'll try gen 2.0..

Then I will swap in that IS-60 cooler and rerun some of the tests.


wha? double post?
 
2 quick tests completed--

1) Disconnected the end of the riser from divider panel, then connected 1070 to it hanging outside the case. Booted the computer and ran same Furmark I've used for other tests (1600x900 fullscreen 0AA), and monitored with GPU-Z to check PCIE bus interface. GPU-Z showed 16x 3.0, then ran test for a few minutes, waving the GPU around, flexing the riser at both ends near the connectors to give maximum stress. No change in bus interface and no crashing or freezing at all. No indication that the riser movement had any effect. So that is good.

A bit of info on the riser pedigree: I purchased it from Li-Heat directly in the spring. It is 300mm, B-type (the GPU end has a 90 deg mounting pcb), and has the black plastic strain relief bars over the solder joints. I used it as a bend and fit check model in my workhorse chassis to determine locations of bends and such, so I could generate a model and confirm that 300mm was the right length. I bent it varying radii and ended up with a radius of about 6.5mm (13mm for a180deg bend). It is a stiff cable, but does hold bends. As with most cables it tends to bend right at the strain relief edge, so I tried to force the bend out a bit, maybe starting 3-4mm away.

I have been using it with the 1070 in this version 2 prototype case with no issues. The only issue I had was before I got the 1070, -I was using a gpu from an old Dell computer where I got a hang at boot a couple of times. The system booted on second tries and I did not pin down the cause, but the gpu required Dell drivers so I have a suspicion it didn't play well with the mb.

2) Installed the ID-Cooling IS60 heatsink-fan unit. Here is the website link: http://www.idcooling.com/Product/detail/id/19/name/IS-60
http://www.idcooling.com/Product/detail/id/19/name/IS-60
Well it looked good on paper, 6 heat pipes, copper base. Execution was another story, however. When I received the cooler it was cushioned top and bottom, but no side cushioning, so it rattled around during transport and showed up with some dented fins at the corners. Not a big deal I thought as far a function. I can always straighten them out to look pretty later.

Looked easy to install, no back plate, just nuts. Then I took the plastic off the copper base and found a surface that actually could be used as a file or a rasp. We might call it a brushed surface, except that the grooves appeared way too deep. When I ran my finger across it I was concerned that it would cut me. I'm not kidding. I tried to clean it with an alcohol wipe and there was all sorts of bits of the paper wipes left behind in the grooves. I have to believe the production operators missed the final surfacing steps.

I am not the brightest tool in the shed, as you know already, so instead of sticking it back in the box and returning it, I thought "I'll just polish that surface myself, woo hoo, how tough can it be, and get some test results." I ended up spending time with 1000 grit sandpaper trying to get it smooth. Halfway through I got to thinking about how I had no way to keep that surface flat using a piece of sandpaper and my thumb. Even with a sanding block. The only way to get a flat surface would be to take it to a machine shop, have them indicate off of some datum surface to know what axis to use (maybe the 2 surfaces where the brackets attach), and have them finish mill and hone it. Heck of a lot more than the $45 price tag. But I digress. I of course continued, stopping when I thought the finish looked good enough, and maybe the flatness wasn't too messed up, and crossing my fingers.

To make a long story longer, I cleaned it up, used the same thermal paste as on the Big Shiriken 2 and installed it and tested it. Low and behold the results kinda suck:

G1-G5%20results.jpg


Test was G5 if you hadn't guessed, and can be compared to test G4 with the Big S2.

Conclusion: Don't DIY convert a rasp surface to a heat sink surface. I do not know if correctly manufactured IS-60s would behave better or worse than the Big S2. I have no cheap way to get a good surface on that HS and no interest in buying another, and my warranty is void. Another important lesson relearned.

I will pull it off and take a pic to show you what I tested.
 
Here are a couple pics of the IS-60 after I pulled it back off:
is-60.jpg
is-60%20closeup.jpg


Oh and I retested after I put the Big S2 back and temps went back to normal as shown in test G6 below (ambient temp dropped 2C since test G4 so that affected the temps):

G1-G6%20results.jpg


I changed the thermal paste for kicks on G6, but no difference seen.
 
Dude, I would put in for a RMA on that anyway, just deny doing anything to it yourself (beyond unboxing and realizing you just got raped by a giant gorilla named ID-Cooling)…
 
Well I waited 3 weeks for it to arrive on a slow boat from China. I don't want to deal with the hassle and wait another 3 weeks. You want it?
 
Thanks, but I am gonna have to call a "Hard Pass" on that textured CPU cooler…!

Real bummer that you have to eat the cost on the unit though…
 
You must log in or register to reply here.
Back
Top