You have to do some homework ;-) PCI bus powermanagment (ASPM) to auto
Insert in your OS opnSense the C3 state. If it wont drop to 30 degrees there is a physical issue with your cooling.
Topton Jasper Lake Quad i225V Mini PC Report
- Thread starter ReturnedSword
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Insert in your OS opnSense the C3 state. If it wont drop to 30 degrees there is a physical issue with your cooling.
I did literally zero thermal modification to my n5105 (in the 'basic' case). opnsense right now is about 66C. Sometimes it's down to 55C. Sometimes it's up to 80C briefly. My rule of thumb is I don't start to care until 90C
yeah I suspect I have an overly thick glob of thermal paste between my cpu and the copper.
But I would also probably make it worse if I tried to sand or shim anything.
Thanks, that's similar to as what I'm getting. I believe ours is limited by the smaller heatsink which just gets saturated faster, but as I said in my previous post, this doesn't seem to be limiting the CPU in any significant way especially when just running Opnsense, so I agree with you on the temps not being a big deal.
Do you also have the cheaper smaller case? There are like 3 or 4 ones, and this one is relatively new, and it's significantly cheaper with otherwise identical hardware. I applied those tunables but temps didn't really lower that much.
OK this is pretty funny, I think I found the issue.
Brilliant! that's the best example I have ever seen of the problem.
I know some people filed a bit the standoffs and put a copper shim, but then the pressure is applied by the stiffness of the motherboard basically, so the more pressure, the more the mobo flexes. Not sure I wanna go that way, besides it wouldn't apply too much pressure. I got some ideas, will share if it works.
Over a week of uptime on the CWWK N6005 V5 build with no issues so far
Over 16 days uptime no issues so far with opnsense on proxmox
Just be mindful that there are some clear plastic strips around the CPU so too much pressure won't make it work any better, you just need a firm contact to make the CPU heat aware that there is something cooler it can expand/balance into. That's the best way I can explain it.
Yes, it lacks 4 standoffs/screws all around the cpu to level it and make the correct pressure to the heatsink.
Still the fixes described in the forum work to make a very good contact and reduce heat to normal levels.
From what I read we all got a very good results, only doing the tweaks described in the forum, so that should be the way to go, but I imagine you could put something not conductive and a bit flexible (oposite side of top the cpu die), that could make the same pressure your finger is doing, when you close the lid, taking care not to crush any component.
In my case (I already had good temps.), I didn't need to sand the standoffs, but needed to level the board on one side a tiny bit, because was flexing too much.
looking at the design we have, if we make pressure down to the hdmi or usb3 inserting cables, that is also pulling the MB/CPU away from the heatsink.
Last edited:
Yeah something like that is what I was thinking. What kind of temp difference did you get before and after?
Problem people report is a gap between the cpu die and the heatsink, factory will fill that with thermal paste and that's really bad.
The solution is to have the cpu die parallel and in perfect contact with the heatsink and good thermal paste.
In my case the cpu was already making very good contact, so I already had good temps and didn't took note what temps were before.
I only test the worst case scenario and let the heatsink saturate, waiting for the temp to stabilize.
I use full load of all cores (N5105=4x2.8GHz always) with prime95 or AIDA FPU test and BIOS P1=P2=20W: https://forums.servethehome.com/ind...225v-mini-pc-report.36699/page-93#post-356099
There are also different boxs/heatsink sizes, so we need to watch that temp is not too high while doing stress tests.
Last edited:
Some Intel N100 M9 mini pc benchmarks with power wattage and performance numbers (take with some internet salt)
Appears to draw similar idle watts with n5105 and n6005, same Max watts as N6005 but N5105 much less power draw at max watts.
Another video test here
Looks impressive the N100 with its gpu performance and even AES, Chacha20 performance over N5015/6005. Its a shame its using more watts similar to N6005 to do it though but possible with some tweaking could be improved.
Appears to draw similar idle watts with n5105 and n6005, same Max watts as N6005 but N5105 much less power draw at max watts.
Another video test here
Looks impressive the N100 with its gpu performance and even AES, Chacha20 performance over N5015/6005. Its a shame its using more watts similar to N6005 to do it though but possible with some tweaking could be improved.
Last edited:
that is the most servethehome-specific tiktok video I can possibly imagine, and it's awesome.
I guess it's mainly mounting pressure.
if cwwk is listening, maybe they redo their case so that there's a few screws closer to the cpu that attach the back half of the motherboard to the top half of the heatsink, to get mounting pressure.
Dell has been given crap on desktops for having the CPU heastink mount directly to the case... but for THESE guys it would actually make sense.
Nope model A I read in the reviews, under the product, that model C (yours) has less cooling surface and remains hotter than model A or B. Which make me decide to go for A.
Which make sense, the later model (mine is from December) have an extra copper plate in between from the factory. Therefore you need less cooling paste which improves the heat conduction.
Here I'm upgrading the cooling paste to a better one. Coolingpaste is in between the copper and on the cpu side.
Not anymore, the factory applied the copper plate in between, sees my previous post.