Newbie poster so sorry on advance. I've just ordered an i3-N305 from CWWK and wondered if anyone has one and has tried 32gb ram in one ? If so what was the outcome ?
CWWK/Topton/... Nxxx quad NIC router
- Thread starter Becks0815
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
There's 3 pages of posts mentioning 32gb in this thread. That search box up top lets you restrict the search to the current thread with a checkbox.
How to did you request one?I asked HUNSN for a fan kit via Amazon, shipped to UK in 1-2 weeks.
The pitch is definately bigger than 1.0mm difficult to tell if its 1.25mm but I'd guess so => Molex PicoBlade
Height is about 4.0mm => Molex PicoBlade
Has 2 bumps/lock => Molex PicoBlade
I think it's a Molex PicoBlade 4 pin, the trouble is probably finding somewhere that will sell one!
My bare metal OPNsense jumps around ~50-57C idle, old NVMe Samsung 950 pro 59-62C
I've received a fan kit to install sometime, I did not see a gap between case and CPU when I checked but will probably repaste the CPU anyway.
I think the NVMe probably contributes to the heat it's a quite an early model that I had lying around and is notorious for running quite hot and thermal throttling, if the fan kit doesn't help, I'll probably replace it with a WD Blue SN570
I think the fan kit - or at least the cable should be included with all of these products, the SSD/Memory side has barely any airflow and just heats up.
edit: also I hear virtualised in proxmox runs cooler but I prefered to run bare metal
I'm seeing 40-41 for my idle temps and running a Samsung but not sure what model
On the Amazon order details page there is a "Product question? Ask Seller", I just sent them a message as briefly as possible explaining I have high temperatures and could they supply a fan kit, they confirmed address and all in it took 2 hours and I had a tracking code. Great service!
I expect it's similar via AliExpress, the only reason I ordered via Amazon was I assumed it would ship from Europe avoiding customs risks but they shipped it via DHL from China and it arrived to the UK in 6 days.
My temperatures where much better when I tested it without the NVMe drive, left it idling for a while and ran memtest.
So I think my issue is mostly just the SSD heating under the CPU up.
I mentioned the WD SN570 because a few STH videos mention it as a good low power option, I'm sure any modern ones would be an improvement over mine which I think was Samsungs first NVMe SSD from 2016
Greetings, I bought a HUNSN RJ44 bareones device (see https://www.amazon.com/gp/product/B0C8J529FV) from Amazon which shipped out of HK and took about a week to Texas. I went with them for the amazon protection and the faster shipping and reports of decent customer service. I've been in contact with them even before the purchase asking them questions and it's been good so far. I have attached pictures of the hardware
I installed SAMSUNG 970 EVO Plus 500GB NVMe m.2 SSD and Crucial RAM 32GB DDR5 4800MHz CL40 Memory CT32G48C40S5 also from Amazon. At first I could not get it to boot BIOS until I reseated the memory, now it boots. I have run memtest86+ for 2 passes and installed proxmox-ve_8.0-2 with no virtual machines or load on it, just letting it sit on my desk idle. I also asked for the fan kit like someone else on the forum did and received that and installed it as well.
In BIOS screen the CPU is about 32C and it draws 11.7W (measure with kill-a-watt) with no fan. With fan in BIOS screen it uses about 13.7W. Booted up into proxmox it was using about 8W with the fan.
I will note that the memtest86+ (about two passes) showed the RAM has no errors at all although it did get very hot memory reaching over 80C on some tests, but would drop to 60C on the sleep test. Chassis also got up to 52.6C during these tests, but idle the chassis temp is about 37C. With proxmox booted, the CPU idle temps average about 35C and the SSD averages about 40C.
The only serious issue I have noticed so far is that it seems to reboot after it's been on for a while and I touch it or plug a video cable or USB into it. It's not easy to reproduce and seems to happen after its been on for hours and I move something. I have tried jiggling the HDMI cable, putting my USB hub (w/ keyboard and mouse plugged in since I am switching it between this and my laptop). I have now plugged it directly into the wall instead of a power strip. It's sitting at my work desk and I am monitoring it all day for stability and performance. I am not sure if the unit is very sensitive to power or something else is going on with it. It was packaged with a DAJING ADP-36C2 power supply.
Just now, the unit ran 11.5 hours, but when I removed the USB hub to my laptop, it rebooted again. It really seems like it reboots when I touch it, but I cannot reproduce it easily by jiggling things. I asked HUNSN about it and they said to install esxi as there is a driver issue with PVE, and they can't find the reason and make it work smoothly. So maybe there is some weird BIOS issue? I am perplexed since PVE is based on Debian.
I do see an ACPI bios error from proxmox as first message on bootup. I have just now added "acpi=off" to the grub bootloader startup options. I will see if that makes a difference but so far it looks like it lowered my console resolution and the sensors command can no longer read the CPU temperature, so that doesn't sound like a great long term solution if that is the issue.
If anyone has any things I can try, please let me know. You can it's the same Changwang motherboard that is used in all these Topton, CWK devices for the N100 CPU. I want to make this thing work but I want it to run proxmox and OPNsense and be able to use open vswitch to build a lab that has a device that can do network monitoring, support a DMZ subnet, VPN and run a few VMs. I'm just doing gigabit speeds and haven't tried testing 2.5 Gbps but I do have one other device to help test that with iperf. If this thing is going to reboot randomly it's kind of a non-starter for me since it will be a 24/7 perimeter device. It does boot quite fast after these restarts and I can find nothing in the logs that indicates a reason except some messages that look like this that pop up randomly here and there:
[sda] Media removed, stopped polling
[sda] Attached SCSI removable disk
These occur multiple times but the SSD is actual device name /dev/nvme0n1 (with BIOS boot, EFI system part and Linux LVM partitions) so I'm not sure what /dev/sda even refers to since fdisk can't access it. I have another Samsung 970 Plus 1TB m.2 NVMe that I just got today since they were running sale but I kind of doubt it's the disk so I was going to leave it sealed until I get this stable to swap it out with the 500GB SSD.
I installed SAMSUNG 970 EVO Plus 500GB NVMe m.2 SSD and Crucial RAM 32GB DDR5 4800MHz CL40 Memory CT32G48C40S5 also from Amazon. At first I could not get it to boot BIOS until I reseated the memory, now it boots. I have run memtest86+ for 2 passes and installed proxmox-ve_8.0-2 with no virtual machines or load on it, just letting it sit on my desk idle. I also asked for the fan kit like someone else on the forum did and received that and installed it as well.
In BIOS screen the CPU is about 32C and it draws 11.7W (measure with kill-a-watt) with no fan. With fan in BIOS screen it uses about 13.7W. Booted up into proxmox it was using about 8W with the fan.
I will note that the memtest86+ (about two passes) showed the RAM has no errors at all although it did get very hot memory reaching over 80C on some tests, but would drop to 60C on the sleep test. Chassis also got up to 52.6C during these tests, but idle the chassis temp is about 37C. With proxmox booted, the CPU idle temps average about 35C and the SSD averages about 40C.
The only serious issue I have noticed so far is that it seems to reboot after it's been on for a while and I touch it or plug a video cable or USB into it. It's not easy to reproduce and seems to happen after its been on for hours and I move something. I have tried jiggling the HDMI cable, putting my USB hub (w/ keyboard and mouse plugged in since I am switching it between this and my laptop). I have now plugged it directly into the wall instead of a power strip. It's sitting at my work desk and I am monitoring it all day for stability and performance. I am not sure if the unit is very sensitive to power or something else is going on with it. It was packaged with a DAJING ADP-36C2 power supply.
Just now, the unit ran 11.5 hours, but when I removed the USB hub to my laptop, it rebooted again. It really seems like it reboots when I touch it, but I cannot reproduce it easily by jiggling things. I asked HUNSN about it and they said to install esxi as there is a driver issue with PVE, and they can't find the reason and make it work smoothly. So maybe there is some weird BIOS issue? I am perplexed since PVE is based on Debian.
I do see an ACPI bios error from proxmox as first message on bootup. I have just now added "acpi=off" to the grub bootloader startup options. I will see if that makes a difference but so far it looks like it lowered my console resolution and the sensors command can no longer read the CPU temperature, so that doesn't sound like a great long term solution if that is the issue.
If anyone has any things I can try, please let me know. You can it's the same Changwang motherboard that is used in all these Topton, CWK devices for the N100 CPU. I want to make this thing work but I want it to run proxmox and OPNsense and be able to use open vswitch to build a lab that has a device that can do network monitoring, support a DMZ subnet, VPN and run a few VMs. I'm just doing gigabit speeds and haven't tried testing 2.5 Gbps but I do have one other device to help test that with iperf. If this thing is going to reboot randomly it's kind of a non-starter for me since it will be a 24/7 perimeter device. It does boot quite fast after these restarts and I can find nothing in the logs that indicates a reason except some messages that look like this that pop up randomly here and there:
[sda] Media removed, stopped polling
[sda] Attached SCSI removable disk
These occur multiple times but the SSD is actual device name /dev/nvme0n1 (with BIOS boot, EFI system part and Linux LVM partitions) so I'm not sure what /dev/sda even refers to since fdisk can't access it. I have another Samsung 970 Plus 1TB m.2 NVMe that I just got today since they were running sale but I kind of doubt it's the disk so I was going to leave it sealed until I get this stable to swap it out with the 500GB SSD.
Last edited:
@spiffinite - perhaps the power supply is the issue. It was reported in this thread and post 43 in this thread
Based in the model its a 36W max 3amp power supply. Is this enough power for it? It has rebooted idle or under load. See https://www.amazon.com/DAJING-12V3A-ADP-36C2-Certified-Converter-Motherboard/dp/B085TQZ78W. It says it is a 5.5x2.1mm plug.
Can some recommend a power supply off amazon prime if this power supply is crap?
My similar n100 came with a 12v 5a brick. Though a quite bad one. Anyway, 36W should be enough unless you do power hungry ssd/usb stuff. Currently i'm running 12v 3a from a gan charger, no problem whatsoever. It's not the power that's not enough, but probably the quality of the power supply is the reason for your issue.
Any recommendations on a new one for this kit? I would like to order something and rule this out as the issue ASAP. Is the key thing to get 12v with at least 3 amp supply and match up the plug? Is a gan charger the way to go?
As i'm, not in the US but in the eu, all i can say is that amazon is full of fake power supplies. Someone from us will tell you where can you source good quality power supply. Going the gan usb-pd to 12v conversion way is 3A max, costs a lot, and you have to buy a special cable from china probably, all this for ~95% psu efficiency instead of ~85%.
Latest answer from HUNSN: "There is a possibility that hot-swapping a device may cause a reboot due to a sudden change in current, and may you try using a 12V5A power supply, which will improve the situation. " I may need a little more juice and better quality as well.
Same here and you really don’t. My OPNSense is virtualised and the box is not even using 10% CPU.
Regarding having some redundancy, you don’t need to do anything if you have planned ahead
If internet is important for you, or your faimily then people who are going so far to use one of these boxes and ‘roll their own’ I would suggest they have some plan in case of issues.
Depending on where you are: Meanwell or Leicke. Or buy a good PD charger and one of those 12V decoys. All of them are in the range of 86-88% efficiency.
Pick one with 60W or more, solder a terminal to it, so you can use the whole thing as power supply for more than one machine, and with 2 pieces of hardware attached to it, you are always above the 10W threshold and in the range with the highest efficiency of the power supply.
I'm using a small 60w DC UPS. A liitle more expensive vut it will power a number of devices.
Your opnsense is running bsd. I don't think cake qos is available for bsd kernels. So we are not running the same workload. My box is rougly loaded at 40-50% when i max bandwidth in one direction. Opnsense != openwrt. Also other features can drastically can increase cpu load as well.Same here and you really don’t. My OPNSense is virtualised and the box is not even using 10% CPU.
Regarding having some redundancy, you don’t need to do anything if you have planned ahead
If internet is important for you, or your faimily then people who are going so far to use one of these boxes and ‘roll their own’ I would suggest they have some plan in case of issues.
Yep. My CPU goes from 1W to 12W when doing CAKE on a 1Gb download.
Saw the posts about Mean Well. Was going to pull the trigger on this one
MEAN WELL GST60A12-P1J 12V 5 Amp 60W Regulated Switching Table Top Power https://a.co/d/0lSd0uM
But then HUNSN said it has a 5.5x2.5mm barrel so I may need the p1m. Need to verify the size so I order the correct one. Disabling acpi seems to have corrected the issue. I re-eneabled and turned off features such as wake on lan, s3 suspend and other power features in the bios. Also been doing stress tests with proxmox/debian stress command and have not been able to reproduce the issue thus far even if repugging usb devices and jiggling cables.
MEAN WELL GST60A12-P1J 12V 5 Amp 60W Regulated Switching Table Top Power https://a.co/d/0lSd0uM
But then HUNSN said it has a 5.5x2.5mm barrel so I may need the p1m. Need to verify the size so I order the correct one. Disabling acpi seems to have corrected the issue. I re-eneabled and turned off features such as wake on lan, s3 suspend and other power features in the bios. Also been doing stress tests with proxmox/debian stress command and have not been able to reproduce the issue thus far even if repugging usb devices and jiggling cables.
Last edited:
The mean well p1m isn't as readily available as the p1j. Get a globtek. They're high quality and easy to get.
It came with the cheap Danjing adapter I mentioned above.
Thanks for the suggestion on this. I believe they are correct, it is a 5.5x2.5. I used a drill bit to measure which I looked up online for this trick.
5/64 inch = 1.9844 mm
3/32 inch = 2.3813 mm
7/64 inch = 2.7781 mm
a 5/64 bit will fit the 2.1mm barrel but not a 3/32
a 3/32 bit will fit the 2.5mm barrel but not a 7/64
In my case a 3/32 bit fit, but the 7/64 bit did not.