CWWK/Topton/... Nxxx quad NIC router

[slybunda]

my ones been running rock solid non stop now for many months. really like these cwwk units. as long as you pair them up with a good quality power adapter and genuine ram they do the job.

 

[pigr8]

yeah i can say the same, 2 units in a cluster (+ pizero2 as qdevice - that also has nut with gracefull shutdown of the units and wol to restore), both with 32gb ram, one of the unit has 4x4tb nvme recently switched from mdadm to raidz1, all in a unfinished diy mini rack.

 

[alkersan]

the Intel chip is likely an engineering sample, which should not to be resold, that's China for you

Hm, I've just got a similar box with N355 and wondering, if markings on the CPU can tell anything regarding it's origin:

• SRPNT
• 00776
• X506D949
• QR: U5FK447100776

Spoiler: CPU

 

[phil-2024]

Hm, I've just got a similar box with N355 and wondering, if markings on the CPU can tell anything regarding it's origin:

• SRPNT
• 00776
• X506D949
• QR: U5FK447100776

In Windows you can use the Intel® Processor Identification Utility tool to check if the CPU is genuine or perhaps an engineering sample.

 

[audelair]

Hey all… just checking in to this thread as I am having a really hard time deciding whether to jump into the current N100 offerings from CWWK/Topton, etc.

I am currently rocking a Qotom Q355G4 running Opnsense bare metal that is 5.5 years old. The thing is rock solid and the only thing I have had to do is swap the squealing Dajing power brick with a Mean Well. Back then, the research led me to believe it was a rock solid choice and it has proven to be one of the best tech purchases I’ve made.

Fast forward to now and I really need a spare (or turn my Qotom into a spare) and I have read this thread for almost 2 years now and there are too many concerns that I wonder if I am overthinking things or is this generation of products more problematic. The main things are:

• Heat issues. Many seem to have a hard time with heat and point to possibly bad CPU’s, BIOS issues, or bad design.
• The NVME drive seems to get really hot on the bottom of the case and lead to bad thermals.
• NVME drive compatible seems problematic. I had a spare Teamgroup MP33 and unfortunately there are already more than one report of that drive not working. Furthermore, the ones people have success are pretty expensive these days. And most NVME’s these days are fast and run hot which seems bad with the chassis design.
• HDMI issues seem to be common from the discussion here
• Aliexpress prices have gone way up and some of the pricier options seem to not have as much of a markup. Like the HUNSN on Amazon seems like it has decent support, as does the Protectli N150 4-NIC unit. But these are still all based on the same hardware so it probably doesn’t solve any of the above-mentioned issues?

As such, I am thinking after 15 years of pfSense and opnSense on various hardware, I may decide that putting a system together is not worth the effort vs getting a Mikrotik. I run a Mikrotik HEX at my parents house and seem to get around the software okay. I much prefer OpnSense, but starting to feel like it is worth having the simplicity of being able to buy a $200 workhorse (RB5009) and be done.

But maybe my worries are unfounded, especially if I am willing to shell out a little more for a Protectli or HUNSN? Any thoughts are greatly appreciated.

 

[splifingate]

Hey all… just checking in to this thread as I am having a really hard time deciding whether to jump into the current N100 offerings from CWWK/Topton, etc.

I am currently rocking a Qotom Q355G4 running Opnsense bare metal that is 5.5 years old. The thing is rock solid and the only thing I have had to do is swap the squealing Dajing power brick with a Mean Well. Back then, the research led me to believe it was a rock solid choice and it has proven to be one of the best tech purchases I’ve made.

Fast forward to now and I really need a spare (or turn my Qotom into a spare) and I have read this thread for almost 2 years now and there are too many concerns that I wonder if I am overthinking things or is this generation of products more problematic. The main things are:

• Heat issues. Many seem to have a hard time with heat and point to possibly bad CPU’s, BIOS issues, or bad design.
• The NVME drive seems to get really hot on the bottom of the case and lead to bad thermals.
• NVME drive compatible seems problematic. I had a spare Teamgroup MP33 and unfortunately there are already more than one report of that drive not working. Furthermore, the ones people have success are pretty expensive these days. And most NVME’s these days are fast and run hot which seems bad with the chassis design.
• HDMI issues seem to be common from the discussion here
• Aliexpress prices have gone way up and some of the pricier options seem to not have as much of a markup. Like the HUNSN on Amazon seems like it has decent support, as does the Protectli N150 4-NIC unit. But these are still all based on the same hardware so it probably doesn’t solve any of the above-mentioned issues?

As such, I am thinking after 15 years of pfSense and opnSense on various hardware, I may decide that putting a system together is not worth the effort vs getting a Mikrotik. I run a Mikrotik HEX at my parents house and seem to get around the software okay. I much prefer OpnSense, but starting to feel like it is worth having the simplicity of being able to buy a $200 workhorse (RB5009) and be done.

But maybe my worries are unfounded, especially if I am willing to shell out a little more for a Protectli or HUNSN? Any thoughts are greatly appreciated.

I have no concept of what the other Systems for which you talk.

That being said, I employ a Ubiquity ER-4 for Router duties (no-brainer; no error; uptime measured in years).

My CWWK AL-4L-V2 (n305; 48GB DDR5/1+2TB SKHy nvme) runs Prox., and I have one Container that serves as a mirror to my other piHole installation (can't quite call it "fail-safe" as everything inevitably fails (at some point, or another)).

Other than that, it's basically just expending (15/13) watts....

 

[audit13]

Hey all… just checking in to this thread as I am having a really hard time deciding whether to jump into the current N100 offerings from CWWK/Topton, etc.

Fast forward to now and I really need a spare (or turn my Qotom into a spare) and I have read this thread for almost 2 years now and there are too many concerns that I wonder if I am overthinking things or is this generation of products more problematic. The main things are:

• Heat issues. Many seem to have a hard time with heat and point to possibly bad CPU’s, BIOS issues, or bad design.
• The NVME drive seems to get really hot on the bottom of the case and lead to bad thermals.
• NVME drive compatible seems problematic. I had a spare Teamgroup MP33 and unfortunately there are already more than one report of that drive not working. Furthermore, the ones people have success are pretty expensive these days. And most NVME’s these days are fast and run hot which seems bad with the chassis design.
• HDMI issues seem to be common from the discussion here
• Aliexpress prices have gone way up and some of the pricier options seem to not have as much of a markup. Like the HUNSN on Amazon seems like it has decent support, as does the Protectli N150 4-NIC unit. But these are still all based on the same hardware so it probably doesn’t solve any of the above-mentioned issues?

But maybe my worries are unfounded, especially if I am willing to shell out a little more for a Protectli or HUNSN? Any thoughts are greatly appreciated.

I purchased a barebones (no ram, no SSD) Topton n100 4 x i226v from Aliexpress about a year ago for bare metal opnsense after much research.

I am very new to opnsense and have a very simple network with an Asus GT-AX6000 as an AP, Adguard Home, and Wireguard. This is how my n100 has been used for the past year with zero reboots, crashes, etc. The Topton started with opnsense 23.1 and has been upgraded through the opnsense update option to 25.1 with no problems.

I read about the unit running hot so I placed a USB-powered 120 mm fan on top to blow air away from the unit. With opnsense, it idles around 11 watts and 29 celsius in the utility room of my home.

I added an 8 GB stick of OEM Hynix ram as recommended by Topton and an Intel 660p NVMe to test out the unit with Windows 10 before loading opnsense. It was used this way for about a week with no issues.

For opnsense, I removed the NVMe drive and added a 120 GB SATA3 Kingston SSD since it was sitting in my drawer and I could find no other purpose for the drive. This combination of Hynix ram, Kingston SSD, and Topton has been perfect.

Set up of the unit for Windows and opnsense was done with a HDMI cable from the Topton to my DVI monitor, USB keyboard, and USB mouse.

Hope this post helps in some way.

 

[audelair]

For opnsense, I removed the NVMe drive and added a 120 GB SATA3 Kingston SSD since it was sitting in my drawer and I could find no other purpose for the drive. This combination of Hynix ram, Kingston SSD, and Topton has been perfect.

Oh, for some reason i forgot these mostly had SATA available for 2.5" drives. I have a couple spare ones (Crucial MX300/MX500), and I know they will run way cooler than the current NVME, and it also saves me a good chunk of money. Good reminder! I'm warming up to getting one of these now. Thanks for the feedback!

 

[slybunda]

Hey all… just checking in to this thread as I am having a really hard time deciding whether to jump into the current N100 offerings from CWWK/Topton, etc.

I am currently rocking a Qotom Q355G4 running Opnsense bare metal that is 5.5 years old. The thing is rock solid and the only thing I have had to do is swap the squealing Dajing power brick with a Mean Well. Back then, the research led me to believe it was a rock solid choice and it has proven to be one of the best tech purchases I’ve made.

Fast forward to now and I really need a spare (or turn my Qotom into a spare) and I have read this thread for almost 2 years now and there are too many concerns that I wonder if I am overthinking things or is this generation of products more problematic. The main things are:

• Heat issues. Many seem to have a hard time with heat and point to possibly bad CPU’s, BIOS issues, or bad design.
• The NVME drive seems to get really hot on the bottom of the case and lead to bad thermals.
• NVME drive compatible seems problematic. I had a spare Teamgroup MP33 and unfortunately there are already more than one report of that drive not working. Furthermore, the ones people have success are pretty expensive these days. And most NVME’s these days are fast and run hot which seems bad with the chassis design.
• HDMI issues seem to be common from the discussion here
• Aliexpress prices have gone way up and some of the pricier options seem to not have as much of a markup. Like the HUNSN on Amazon seems like it has decent support, as does the Protectli N150 4-NIC unit. But these are still all based on the same hardware so it probably doesn’t solve any of the above-mentioned issues?

As such, I am thinking after 15 years of pfSense and opnSense on various hardware, I may decide that putting a system together is not worth the effort vs getting a Mikrotik. I run a Mikrotik HEX at my parents house and seem to get around the software okay. I much prefer OpnSense, but starting to feel like it is worth having the simplicity of being able to buy a $200 workhorse (RB5009) and be done.

But maybe my worries are unfounded, especially if I am willing to shell out a little more for a Protectli or HUNSN? Any thoughts are greatly appreciated.

the n100 unit i got has replaced a edgerouter 4, which has now become the backup unit. only reason for going to opnsense is for the processing power for full line speed wireguard. if i didnt need 1gig wireguard performance id have stuck to the edgerouter.
heat is a non issue on the cwwk box i got, its the tall fin version with a 12cm fan on top standard pc 12v fan with a usb to fan adapter so fan runs mega slow and silent but since the fins are tall and wide apart ya dont need huge airflow.
nvme i use is 64gb forsee brand from a steamdeck temps are fine even without the heatsink since you dont do full speed constant writes to it, i got a heatsink on the nvme since i got one spare and it only makes things better.
i never used the included power adapter, i had a spare delta adapter lying around and used that.
mine was used and came with 8gb crucial ram and 500gb wd red drive which i repurposed for another machine. well worth the 90 quid.

id have preferred the 2 port version since its smaller but may run hotter.

 

[mini-man]

View attachment 37327

after intense testing this little x86-p5 with the nvme expansion is ready to go in production, successfully running 5 nvme disks.. nvme0 is a temporary drive that will be replaced with another 4tb, nvme1-2-3 are in a mdadm raid5 with lvm and btrfs on top, nvme4 is the host boot drive.

no zfs.

How were you able to get this setup correctly? Currently I am running the same development board with 4 x Sabrent Rocket 4TB and a 128gb sata ssd as the boot. I have the nvmes setup in a raidz1 and whenever I start a copy process from one location of the pool to another all of the drives fail (in the pool at least; they still show up after I run lsblk). From what I can find in dmesg it seems like it is trying to put the drives in a power saving mode

[  701.003075] nvme nvme0: controller is down; will reset: CSTS=0xffffffff, PCI_STATUS=0xffff
[  701.003091] nvme nvme0: Does your device have a faulty power saving mode enabled?
[  701.003096] nvme nvme0: Try "nvme_core.default_ps_max_latency_us=0 pcie_aspm=off pcie_port_pm=off" and report a bug
[  701.037106] nvme 0000:01:00.0: Unable to change power state from D3cold to D0, device inaccessible
[  701.037500] nvme nvme0: Disabling device after reset failure: -19
[  701.047144] zio pool=Mayo_Coffee vdev=/dev/disk/by-partuuid/4de2a31d-5a88-4b89-9bca-984701f99802 error=5 type=2 offset=2492974759936 size=131072 flags=2148533376
[  706.131095] nvme nvme1: controller is down; will reset: CSTS=0xffffffff, PCI_STATUS=0xffff
[  706.131100] nvme nvme1: Does your device have a faulty power saving mode enabled?
[  706.131101] nvme nvme1: Try "nvme_core.default_ps_max_latency_us=0 pcie_aspm=off pcie_port_pm=off" and report a bug
[  706.174097] nvme 0000:02:00.0: Unable to change power state from D3cold to D0, device inaccessible
[  706.174556] nvme nvme1: Disabling device after reset failure: -19

 

[pigr8]

i did not do anything particular and did not add specific parameters to kernel (the one suggested by your log for example), i have aspm disabled by default

i dont really care since i have no unused devices that can go in powersave, surely not the nvmes since they constantly writing, also both eth (one used for traffic, the other for sync), i just disabled the sata controller but it's a combo with usb3 anyway.

Linux n305 6.8.12-10-pve #1 SMP PREEMPT_DYNAMIC PMX 6.8.12-10 (2025-04-18T07:39Z) x86_64

The programs included with the Debian GNU/Linux system are free software;
the exact distribution terms for each program are described in the
individual files in /usr/share/doc/*/copyright.

Debian GNU/Linux comes with ABSOLUTELY NO WARRANTY, to the extent
permitted by applicable law.
Last login: Wed Jul 23 07:59:09 CEST 2025 on pts/0
root@n305:~# cat /etc/kernel/cmdline
root=ZFS=rpool/ROOT/pve-1 boot=zfs intel_iommu=on iommu=pt net.ifnames=0 split_lock_detect=off
root@n305:~# lspci -vvv | grep LnkCtl:
                LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
                LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
                LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
                LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
                LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
                LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
                LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
                LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
                LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
                LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
                LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
                LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
root@n305:~# dmesg | grep ASPM
[    0.145155] ACPI FADT declares the system doesn't support PCIe ASPM, so disable it
[    0.367644] acpi PNP0A08:00: _OSC: OS supports [ExtendedConfig ASPM ClockPM Segments MSI EDR HPX-Type3]
[    0.368226] acpi PNP0A08:00: FADT indicates ASPM is unsupported, using BIOS configuration

 

[mini-man]

After looking around some more I think I have figured out what my issue was/is. While I can't be certain yet--I just did this last night-- I believe updating to the newest bios has fixed the issue with the multiple nvme drives. I'm going to monitor them closely for the next week, but I am hoping that this is what was wrong.

 

[pigr8]

After looking around some more I think I have figured out what my issue was/is. While I can't be certain yet--I just did this last night-- I believe updating to the newest bios has fixed the issue with the multiple nvme drives. I'm going to monitor them closely for the next week, but I am hoping that this is what was wrong.

did you update with the correct one that support 4 nvme? there are 2 bioses

 

[mini-man]

yes I used this one CW-AL-2L_V2-V3(X86 N100-N200-I3-N305_V2-V3 4M.2)2024.05.15.iso

 

[pigr8]

yes I used this one CW-AL-2L_V2-V3(X86 N100-N200-I3-N305_V2-V3 4M.2)2024.05.15.iso

good iirc it's the same i'm using

 

[mini-man]

After less than a week with the new bios, 2 of my drives are going into a fail state with the following being reported by dmesg. This is similar to what I have had happen before the update. This seems to occur when I am performing any big copy operation. it should also be noted i have them configured in a zfs array

[401143.598118] nvme nvme1: controller is down; will reset: CSTS=0xffffffff, PCI_STATUS=0xffff
[401143.598135] nvme nvme1: Does your device have a faulty power saving mode enabled?
[401143.598140] nvme nvme1: Try "nvme_core.default_ps_max_latency_us=0 pcie_aspm=off pcie_port_pm=off" and report a bug
[401143.631233] nvme 0000:01:00.0: Unable to change power state from D3cold to D0, device inaccessible
[401143.631888] nvme nvme1: Disabling device after reset failure: -19
[401144.110108] nvme nvme0: controller is down; will reset: CSTS=0xffffffff, PCI_STATUS=0xffff
[401144.110126] nvme nvme0: Does your device have a faulty power saving mode enabled?
[401144.110130] nvme nvme0: Try "nvme_core.default_ps_max_latency_us=0 pcie_aspm=off pcie_port_pm=off" and report a bug
[401144.140151] nvme 0000:02:00.0: Unable to change power state from D3cold to D0, device inaccessible
[401144.140811] nvme nvme0: Disabling device after reset failure: -19
[402365.147031] nvme nvme0: Identify namespace failed (-5)
[402365.147917] nvme nvme1: Identify namespace failed (-5)
[402415.084696] nvme nvme0: Identify namespace failed (-5)
[402415.085500] nvme nvme1: Identify namespace failed (-5)
[402477.822942] nvme nvme0: Identify namespace failed (-5)
[402477.823779] nvme nvme1: Identify namespace failed (-5)

 

[casulo]

my ones been running rock solid non stop now for many months. really like these cwwk units. as long as you pair them up with a good quality power adapter and genuine ram they do the job.

Indeed. And i also swap the CMOS battery, because some of them come almost empty, then bios settings are lost. Auto power on doesn't work, which can be a big PITA.

 

[mini-man]

@pigr8
So far I have tried Truenas Scale, Fedora Server, and Ubuntu Server to no avail. Each OS gave me the same errors.
What OS and Kernel version are you running that allows your drives to work correctly?
Do you have any special bios settings?

 

[mini-man]

Hi folks,

I have the 4 port N305 version of CWWK with the adapter board for 4x NVME disks. I have Proxmox running with several VM's and for example the 4 NVME's are passed through to one VM. This also works wonderfully. Now I am trying to pass a WIFI card to another VM, but this is failing. As soon as I try to start this VM it crashes. If I try an NVME disk in the same slot instead of the Wifi M2 card, it works without problems. I have also tested a second Wifi card with the same result.

The following note appears in the Proxmox log:
unable to change power state from d3cold to d0 device inaccessible

Are there any BIOS settings that I need to adjust here?

Interestingly I am getting that error with 2 of my nvme drives (Currently have 3 installed in the daughter board). is there anything you did to fix this?
I have my drives running as a zfs pool and they get that error anytime I try a big copy operation from one folder to another

 

[TrevorH]

Did you try the kernel parameters the error suggests on linux?

Try "nvme_core.default_ps_max_latency_us=0 pcie_aspm=off pcie_port_pm=off"

I'd start with pcie_aspm=off and work up from there. Or try them all at once and see if any of them help.