Interesting.
So, from your experience do you see failure rate jump up significantly with temperature or you just wear out TBW before the 5 years of warranty run out?
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.
So, from your experience do you see failure rate jump up significantly with temperature or you just wear out TBW before the 5 years of warranty run out?
Although I have 8 NVMe drives (totalling 14TB) in production, I don't think that quite amounts to the level of data. It's still well within anecdote territory. I think I just push these drives a bit harder than the average user. Of the 8 drives I have had 3 x 2TB fail within the first 2 year and well under their TBW counts.
Got all the goodies. Unit looks like it'll be here this week according to tracking. Just need to figure out the best way to get this fan to fit the tiny fan port on the board
Well, I replaced the RAM and my CWWK N6005 V4 unit has been stable for longer than it's ever been. I didn't redo any TIM or add any fan, and the only UEFI settings tweak I made was reversing the ridiculously aggressive power settings they put in on the latest update, and temps have been good. So I guess now I have some hardware being extremely under-utilized as just a router, lol.
First post here after finding Patrick's video on STH youtube. Thanks for all the work everyone's put into finding more about these units as it's help me with the setup of mine.
I ordered a barebones 5105 V4 barebones unit (https://www.aliexpress.com/item/1005004697822666.html) from Ali on Sep 6 and received it on Sep 28 to Canada. I dropped in 16GB of SK Hynix 3200 DDR4 RAM and a 256GB SK hynix NVME.
Seeing all the issues people previously had with temps, I first replaced the thermal paste with some Artic MX4 I had. This didn't make too much of a difference as the temps would still hit around 80-90 degrees under constant load for both CPU and NVME (yikes!).
I then took it apart again and removed the copper heatsink for the CPU and sanded off the paint between it and the main case.
I then reapplied the thermal paste between the case and copper heatsink before reattaching. I also took off the heatsink for the NICs so I can check for a CPU gap and lo and behold there was one even though I have a V4 unit! It wasn't huge but there was definitely an air gap which I think they were hoping to fix by applying a lot of thermal paste.
I ended up taking a file to the motherboard standoffs and shaved off about 0.15mm from each post (measured by a digital caliper). Also on the CPU were two transparent strips next to the actual core (?) which I assume was to prevent it from getting damaged. This also prevented it from making proper contact with the copper heatsink from what I can see as it was slightly higher. After removing these two strips. I was able to get very good contact on the copper heatsink and the CPU. I also bought a generic NVME heatsink and a Noctua 40x10 5V fan which I wired up from the USB header power:
Running benchmarks for Prime95 and Crystal Disk Mark saw temps max out at 68 degree for the CPU and 48 degrees on the NVME. Definitely a big improvement and one that I was more comfortable with since I'll be putting this to work as a bare metal OPNSense firewall/router.
Also not sure if anyone else noticed but seems like the NVME speeds are severely capped as I've only been able to get around 850 mb/s read and 700 mb/s write in benchmarks (not that it will matter much for OPNSense).
Overall I'm very happy with this box and the cost to build. Going to run some more stress testing this week before putting it into actual use for my home network.
I ordered a barebones 5105 V4 barebones unit (https://www.aliexpress.com/item/1005004697822666.html) from Ali on Sep 6 and received it on Sep 28 to Canada. I dropped in 16GB of SK Hynix 3200 DDR4 RAM and a 256GB SK hynix NVME.
Seeing all the issues people previously had with temps, I first replaced the thermal paste with some Artic MX4 I had. This didn't make too much of a difference as the temps would still hit around 80-90 degrees under constant load for both CPU and NVME (yikes!).
I then took it apart again and removed the copper heatsink for the CPU and sanded off the paint between it and the main case.
I then reapplied the thermal paste between the case and copper heatsink before reattaching. I also took off the heatsink for the NICs so I can check for a CPU gap and lo and behold there was one even though I have a V4 unit! It wasn't huge but there was definitely an air gap which I think they were hoping to fix by applying a lot of thermal paste.
I ended up taking a file to the motherboard standoffs and shaved off about 0.15mm from each post (measured by a digital caliper). Also on the CPU were two transparent strips next to the actual core (?) which I assume was to prevent it from getting damaged. This also prevented it from making proper contact with the copper heatsink from what I can see as it was slightly higher. After removing these two strips. I was able to get very good contact on the copper heatsink and the CPU. I also bought a generic NVME heatsink and a Noctua 40x10 5V fan which I wired up from the USB header power:
Running benchmarks for Prime95 and Crystal Disk Mark saw temps max out at 68 degree for the CPU and 48 degrees on the NVME. Definitely a big improvement and one that I was more comfortable with since I'll be putting this to work as a bare metal OPNSense firewall/router.
Also not sure if anyone else noticed but seems like the NVME speeds are severely capped as I've only been able to get around 850 mb/s read and 700 mb/s write in benchmarks (not that it will matter much for OPNSense).
Overall I'm very happy with this box and the cost to build. Going to run some more stress testing this week before putting it into actual use for my home network.
Each of the M.2 ports are connected via a 1 x PCIE 3.0 lane so throughput will max out at just under 1GBps after overhead
I have that fan and my unit arrives this week. What cable did you use between the fan and the USB header? I'm guessing you weren't able to find a way to use the tiny fan header on the board?
Ah that makes sense. Thanks for confirming!
I used a standard 4 pin female header (not sure if this is the right term) that I had in my electronics parts bin. I connected only 2 of the pins based on the USB header pinout pins 1 and 4 for VCC and GND. That's soldered to the 2 pin female fan connector that will come in the box with your Noctua fan. If you don't want to solder, it also comes with these 4 crimp connectors. I was actually surprised at how fast and how much air this small little fan moves at only 5V. A bit noisy but still very good.
Also forgot to note in my previous post I tried using an older Targus laptop adapter I had from many years back. It's a 19V, 65W adapter that's about 25% the size (very thin) of the 12V adapter that's supplied.
Damn, I was kind of hoping the fan would be quiet since you can't control it's speed
From what I know the 5V has less options as Noctua doesn't include their lower noise fan adapter in the kit which the 12V version does. If noise is a concern you might be better off getting a 120mm fan and putting that on the exterior which should be much quieter. With the 40mm 5V, it puts out enough noise that I wouldn't want to have that sitting on my desktop for general use.
Wonder if it'd be worth experimenting with a resistor to slow the fan down a bit? Was honestly hoping it didn't even need a fan.
The problem i think is the case.
Patrick stated in his last Topton Unit Test, that the case can make a big difference concerning temperature. And you get all this cases in the test also with a N5105
Patrick stated in his last Topton Unit Test, that the case can make a big difference concerning temperature. And you get all this cases in the test also with a N5105
Have 12V version of the same Nocuta fan installed on mine, and noise is barely audible.
Given, installing a 40x40x10mm fan on this mini box leaves no room for second NVME. I will be experimenting installing two 20x20x10mm fans on the side grids instead. Would be interesting to see how much noise they make running at quoted speed of 9,000RPM.
Given, installing a 40x40x10mm fan on this mini box leaves no room for second NVME. I will be experimenting installing two 20x20x10mm fans on the side grids instead. Would be interesting to see how much noise they make running at quoted speed of 9,000RPM.
Borrowed 12V supply from SATA power cable that came with the box.
I tried a similar fan setup (i.e. 5v Noctua fan wired to the SATA header). It was very noisy. I could hear it in the adjacent room. Ended up unplugging it for now and instead temporarily placed another small fan on top of the box to cool the "fins".
As a long term solution, I'm waiting for this cable that someone suggested earlier. I have no idea if it'll fit but I'll give it a shot.
LINK
Another option is buying an official cooling fan from Topton. They ask you to pay $9 via a special link and ship you one. I ordered one this way also a few days ago.
As a long term solution, I'm waiting for this cable that someone suggested earlier. I have no idea if it'll fit but I'll give it a shot.
LINK
Another option is buying an official cooling fan from Topton. They ask you to pay $9 via a special link and ship you one. I ordered one this way also a few days ago.
Noctua 40mm fans be it 5V or 12V are rated <20dB at full-speed of 5,000RPM.
20dB fall towards lowest level of typical noise spectrum. Not sure how can it be that audible. Could it be that 5V - 12V wires on SATA header was accidently swapped? Fan on overdrive?
There is more to it than just a fan... Maybe it's 20db in open air. But in this small enclosure, with little clearance on the bottom, it was generating quite a bit of noise (almost like the whistling you get when you close your home HVAC registers 75% of the way). The noise was better if I lifted the enclosure but I eventually gave up on the whole thing. I'd be willing to put up with it for short periods of time during high load... But not all the time.
I had the air blowing towards the processor. Maybe I should try reversing it...
Good point. I wonder if dust shield air flow obstruction might be creating the noise. Mine did not come with dust shield installed, just grill holes on the aluminum body.
Last edited:
Its not just about the case, but the thermal solution inside of it...
idk why @Patrick never shows the thermal solution inside these, pretty sure that the better case that he mentions is the one with the new design with heat pipes and etc, in other words, this one:
Pretty sure that the other 2 cases that he tested and said that they are bad, don't have these heat pipes and the copper block is smaller.