hello,
installed latest AgustinLorenzo nss image for ax9000 from 26.12.2025
how to configure the internal fan of ax9000 ?
is it possible with latest nss image ?
Should be fixed in latest rebase:
7dff92e214
ath11k-firmware: Fix build for QCN9074
`board-2.bin` is already included in `QCN9074/hw1.0/testing/2.12/WLAN.HK.2.12-01460-QCAHKSWPL_SILICONZ-1`
which is already being copied over by the Makefile.
Addresses the build error:
ath11k-firmware-2.12.01460/QCN9074/hw1.0/board-2.bin': No such file or directory
Signed-off-by: Sean Khan <datapronix@protonmail.com>
| date | commit | desc |
|---|---|---|
| 2025-12-29 | 6d99ed6294 | ath11k-firmware: Use date versioning to match upstream |
| 2025-12-13 | 7dff92e214 | ath11k-firmware: Bump ath11k firmware to 2.12 |
Yes, following that discussion too. Figured I'd stick with 2.12 as it was also resolving issues I had with mesh and high frequency of path changes.
if you're not using Github Actions, and have dl/ath11k-firmware-2024.10.14~15f05012.tar.zst saved locally, just copy everything from IPQ8074/hw2.0/2.9.0.1/WLAN.HK.2.9.0.1-02146-QCAHKSWPL_SILICONZ-1 to your router's /lib/firmware/IPQ8074/ and reboot.
➤ l /lib/firmware/IPQ8074/
drwxr-xr-x 2 root root 326 Dec 27 22:27 .
drwxr-xr-x 1 root root 368 Dec 27 22:27 ..
-rw-r--r-- 1 root root 152.8K Dec 27 22:27 Notice.txt
-rw-r--r-- 1 root root 43 Dec 27 22:27 fw_version.txt
-rw-r--r-- 1 root root 148 Dec 27 22:27 m3_fw.b00
-rw-r--r-- 1 root root 136 Dec 27 22:27 m3_fw.b01
-rw-r--r-- 1 root root 320.0K Dec 27 22:27 m3_fw.b02
-rw-r--r-- 1 root root 162 Dec 27 22:27 m3_fw.flist
-rw-r--r-- 1 root root 284 Dec 27 22:27 m3_fw.mdt
-rw-r--r-- 1 root root 340 Dec 27 22:27 q6_fw.b00
-rw-r--r-- 1 root root 328 Dec 27 22:27 q6_fw.b01
-rw-r--r-- 1 root root 4.6K Dec 27 22:27 q6_fw.b02
-rw-r--r-- 1 root root 3.1M Dec 27 22:27 q6_fw.b03
-rw-r--r-- 1 root root 961.8K Dec 27 22:27 q6_fw.b04
-rw-r--r-- 1 root root 252.9K Dec 27 22:27 q6_fw.b05
-rw-r--r-- 1 root root 9.8K Dec 27 22:27 q6_fw.b07
-rw-r--r-- 1 root root 613.8K Dec 27 22:27 q6_fw.b08
-rw-r--r-- 1 root root 432 Dec 27 22:27 q6_fw.flist
-rw-r--r-- 1 root root 668 Dec 27 22:27 q6_fw.mdt
Just pushed one for 25.12. Haven't compiled against it yet, but should build fine.
To https://github.com/qosmio/openwrt-ipq
+ 617b0b36c4...04665e89fc 24.10-nss -> 24.10-nss (forced update)
+ a37b111de4...6d99ed6294 main-nss -> main-nss (forced update)
* [new branch] 25.12-nss -> 25.12-nss
4 Likes
Got this with latest commit
ERROR: info field 'version' has invalid value: package version is invalid
ERROR: failed to create package: package version is invalid
make[3]: *** [Makefile:169: /home/xiaoli/openwrt-ipq/bin/packages/aarch64_cortex-a53/base/ath11k-firmware-ipq8074-2025-12-17-r1.apk] Error 99
make[3]: Leaving directory '/home/xiaoli/openwrt-ipq/package/firmware/ath11k-firmware'
time: package/firmware/ath11k-firmware/compile#0.37#0.08#0.43
ERROR: package/firmware/ath11k-firmware failed to build
Fixed and amended:
| date | commit | desc |
|---|---|---|
| 2025-12-29 | a693dafa66 | ath11k-firmware: Use date versioning to match upstream |
2 Likes
board-2.bin is present in cd /lib/firmware/ath11k/QCN9074/hw1.0 && mv ./
But it is not possible to connect with qcn9074.
i tested different options with different channels and bands… no connection possible ???
i also tested this in wiki:
An issue on the lower 5 GHz band radio (QCN9024) caused by the current BDF file (as 17/12/2022) is leading to non-connecting issues. Proposed by avanletin, a workaround is available, following the instructions below:
-
Paste the entire command into a linux terminal to generate a new
BDFfileboard-2.bincompressed. -
Backup the current
board-2.binfile on the AX9000:cd /lib/firmware/ath11k/QCN9074/hw1.0 && mv ./board-2.bin ./board-2.bin_stock. -
Copy the compressed file
generic-ax9000.tar.gzinto the root directory (/) on the AX9000. -
Uncompress the
BDFfile:tar -xzvf generic-ax9000.tar.gz. -
Reboot the router.
also no connection possible
With the latest commits:
2025-12-30T20:03:36.9983935Z make[2]: Entering directory '/home/runner/work/qosmio-nss/qosmio-nss/feeds/nss_packages/firmware/nss-firmware'
2025-12-30T20:03:37.1122576Z mkdir -p /home/runner/work/qosmio-nss/qosmio-nss/dl
2025-12-30T20:03:37.1159568Z SHELL= flock /home/runner/work/qosmio-nss/qosmio-nss/tmp/.nss-firmware-2025.05.01~e4726900.tar.zst.flock -c ' /home/runner/work/qosmio-nss/qosmio-nss/scripts/download.pl "/home/runner/work/qosmio-nss/qosmio-nss/dl" "nss-firmware-2025.05.01~e4726900.tar.zst" "3ec87f221e8905d4b6b8b3d207b7f7c4666c3bc8db7c1f06d4ae2e78f863b8f4" "" || ( /home/runner/work/qosmio-nss/qosmio-nss/scripts/dl_github_archive.py --dl-dir="/home/runner/work/qosmio-nss/qosmio-nss/dl" --url="https://github.com/qosmio/qca-sdk-nss-fw.git" --version="e4726900138b3e77fbbeaa32e58aa40a66d6b01a" --subdir="nss-firmware-2025.05.01~e4726900" --source="nss-firmware-2025.05.01~e4726900.tar.zst" --hash="3ec87f221e8905d4b6b8b3d207b7f7c4666c3bc8db7c1f06d4ae2e78f863b8f4" --submodules || ( echo "Checking out files from the git repository..."; mkdir -p /home/runner/work/qosmio-nss/qosmio-nss/tmp/dl && cd /home/runner/work/qosmio-nss/qosmio-nss/tmp/dl && rm -rf nss-firmware-2025.05.01~e4726900 && [ \! -d nss-firmware-2025.05.01~e4726900 ] && git clone https://github.com/qosmio/qca-sdk-nss-fw.git nss-firmware-2025.05.01~e4726900 && (cd nss-firmware-2025.05.01~e4726900 && git checkout e4726900138b3e77fbbeaa32e58aa40a66d6b01a) && export TAR_TIMESTAMP=`cd nss-firmware-2025.05.01~e4726900 && git log -1 --no-show-signature --format='\''@%ct'\''` && echo "Generating formal git archive (apply .gitattributes rules)" && (cd nss-firmware-2025.05.01~e4726900 && git config core.abbrev 8 && git archive --format=tar HEAD --output=../nss-firmware-2025.05.01~e4726900.tar.git) && tar --numeric-owner --owner=0 --group=0 --ignore-failed-read -C nss-firmware-2025.05.01~e4726900 -f nss-firmware-2025.05.01~e4726900.tar.git -r .git .gitmodules 2>/dev/null && rm -rf nss-firmware-2025.05.01~e4726900 && mkdir nss-firmware-2025.05.01~e4726900 && tar -C nss-firmware-2025.05.01~e4726900 -xf nss-firmware-2025.05.01~e4726900.tar.git && (cd nss-firmware-2025.05.01~e4726900 && git submodule update --init --recursive -- && rm -rf .git .gitmodules) && echo "Packing checkout..." && tar --numeric-owner --owner=0 --group=0 --mode=a-s --sort=name ${TAR_TIMESTAMP:+--mtime="$TAR_TIMESTAMP"} -c nss-firmware-2025.05.01~e4726900 | zstd -T0 --ultra -20 -c > /home/runner/work/qosmio-nss/qosmio-nss/tmp/dl/nss-firmware-2025.05.01~e4726900.tar.zst && mv /home/runner/work/qosmio-nss/qosmio-nss/tmp/dl/nss-firmware-2025.05.01~e4726900.tar.zst /home/runner/work/qosmio-nss/qosmio-nss/dl/ && rm -rf nss-firmware-2025.05.01~e4726900; ); ) && ( file_hash="$(/home/runner/work/qosmio-nss/qosmio-nss/staging_dir/host/bin/mkhash sha256 "/home/runner/work/qosmio-nss/qosmio-nss/dl/nss-firmware-2025.05.01~e4726900.tar.zst")"; [ "$file_hash" = "3ec87f221e8905d4b6b8b3d207b7f7c4666c3bc8db7c1f06d4ae2e78f863b8f4" ] || [ "3ec87f221e8905d4b6b8b3d207b7f7c4666c3bc8db7c1f06d4ae2e78f863b8f4" = "skip" ] || { echo "Hash mismatch for file nss-firmware-2025.05.01~e4726900.tar.zst: expected 3ec87f221e8905d4b6b8b3d207b7f7c4666c3bc8db7c1f06d4ae2e78f863b8f4, got $file_hash"; false; }; ) '
2025-12-30T20:03:37.1521810Z + curl -f --connect-timeout 5 --retry 3 --location https://sources.cdn.openwrt.org/nss-firmware-2025.05.01~e4726900.tar.zst
2025-12-30T20:03:37.1567826Z % Total % Received % Xferd Average Speed Time Time Time Current
2025-12-30T20:03:37.1568498Z Dload Upload Total Spent Left Speed
2025-12-30T20:03:37.1568851Z
2025-12-30T20:03:37.2101031Z 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
2025-12-30T20:03:37.2101649Z 0 146 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
2025-12-30T20:03:37.2102270Z curl: (22) The requested URL returned error: 404
2025-12-30T20:03:37.2118875Z Download failed.
2025-12-30T20:03:37.2120355Z + curl -f --connect-timeout 5 --retry 3 --location https://sources.openwrt.org/nss-firmware-2025.05.01~e4726900.tar.zst
2025-12-30T20:03:37.2166092Z % Total % Received % Xferd Average Speed Time Time Time Current
2025-12-30T20:03:37.2166725Z Dload Upload Total Spent Left Speed
2025-12-30T20:03:37.2167250Z
2025-12-30T20:03:37.7855523Z 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
2025-12-30T20:03:37.9081535Z 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
2025-12-30T20:03:37.9082115Z 0 146 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
2025-12-30T20:03:37.9092904Z curl: (22) The requested URL returned error: 404
2025-12-30T20:03:37.9109156Z Download failed.
2025-12-30T20:03:37.9111134Z + curl -f --connect-timeout 5 --retry 3 --location https://mirror2.openwrt.org/sources/nss-firmware-2025.05.01~e4726900.tar.zst
2025-12-30T20:03:37.9161697Z % Total % Received % Xferd Average Speed Time Time Time Current
2025-12-30T20:03:37.9162354Z Dload Upload Total Spent Left Speed
2025-12-30T20:03:37.9162998Z
2025-12-30T20:03:38.3593749Z 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
2025-12-30T20:03:38.3594720Z 0 153 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
2025-12-30T20:03:38.3604382Z curl: (22) The requested URL returned error: 404
2025-12-30T20:03:38.3622893Z Download failed.
2025-12-30T20:03:38.3623885Z No more mirrors to try - giving up.
2025-12-30T20:03:40.1225448Z nss-firmware-2025.05.01~e4726900.tar.zst: Download from https://github.com/qosmio/qca-sdk-nss-fw.git failed
2025-12-30T20:03:40.1227151Z nss-firmware-2025.05.01~e4726900.tar.zst: Wrong hash (probably caused by .gitattributes), expecting 3ec87f221e8905d4b6b8b3d207b7f7c4666c3bc8db7c1f06d4ae2e78f863b8f4, got 881cbf75efafe380b5adc91bfb1f68add5e29c9274eb950bb1e815c7a3622807
2025-12-30T20:03:40.1326769Z Checking out files from the git repository...
2025-12-30T20:03:40.1353690Z Cloning into 'nss-firmware-2025.05.01~e4726900'...
2025-12-30T20:03:40.7869017Z Note: switching to 'e4726900138b3e77fbbeaa32e58aa40a66d6b01a'.
2025-12-30T20:03:40.7869513Z
2025-12-30T20:03:40.7869824Z You are in 'detached HEAD' state. You can look around, make experimental
2025-12-30T20:03:40.7870585Z changes and commit them, and you can discard any commits you make in this
2025-12-30T20:03:40.7871342Z state without impacting any branches by switching back to a branch.
2025-12-30T20:03:40.7871791Z
2025-12-30T20:03:40.7872072Z If you want to create a new branch to retain commits you create, you may
2025-12-30T20:03:40.7872747Z do so (now or later) by using -c with the switch command. Example:
2025-12-30T20:03:40.7873149Z
2025-12-30T20:03:40.7873302Z git switch -c <new-branch-name>
2025-12-30T20:03:40.7873564Z
2025-12-30T20:03:40.7873718Z Or undo this operation with:
2025-12-30T20:03:40.7873977Z
2025-12-30T20:03:40.7874078Z git switch -
2025-12-30T20:03:40.7874202Z
2025-12-30T20:03:40.7874410Z Turn off this advice by setting config variable advice.detachedHead to false
2025-12-30T20:03:40.7874708Z
2025-12-30T20:03:40.7874898Z HEAD is now at e472690 QCA_Networking_2022.SPF_12.2: IPQ5018: Add NSS FW 12.2-156
2025-12-30T20:03:40.7891748Z Generating formal git archive (apply .gitattributes rules)
2025-12-30T20:03:40.8604643Z Packing checkout...
2025-12-30T20:03:41.8273679Z Hash mismatch for file nss-firmware-2025.05.01~e4726900.tar.zst: expected 3ec87f221e8905d4b6b8b3d207b7f7c4666c3bc8db7c1f06d4ae2e78f863b8f4, got 881cbf75efafe380b5adc91bfb1f68add5e29c9274eb950bb1e815c7a3622807
2025-12-30T20:03:41.8276571Z make[2]: *** [Makefile:167: /home/runner/work/qosmio-nss/qosmio-nss/dl/nss-firmware-2025.05.01~e4726900.tar.zst] Error 1
2025-12-30T20:03:41.8278178Z make[2]: Leaving directory '/home/runner/work/qosmio-nss/qosmio-nss/feeds/nss_packages/firmware/nss-firmware'
2025-12-30T20:03:41.8283482Z time: package/feeds/nss_packages/nss-firmware/download#2.29#0.40#4.83
2025-12-30T20:03:41.8290607Z ERROR: package/feeds/nss_packages/nss-firmware failed to build.
2025-12-30T20:03:41.8294259Z make[1]: *** [package/Makefile:187: package/feeds/nss_packages/nss-firmware/download] Error 1
Was FW 2.9.0.1 working using the fix you mentioned for ax9000? Can you test with FW 2.13?
Are you building via Github Actions or locally? This is more of an issue with the combo of tar/zstd used in the build process. If you're building from scratch tools/{tar,zstd} should be compiled locally and version 1.35 and 1.5.7 respectively. There could be an instance where not all of your tools were compiled properly, causing the build to use your system tar and zstd, and therefore incorrect hashes being created.
Apply this patch to rule that out.
--- a/include/download.mk
+++ b/include/download.mk
@@ -59,6 +59,8 @@ define dl_pack
$(if $(dl_pack/$(call ext,$(1))),$(dl_pack/$(call ext,$(1))),$(dl_pack/unknown))
endef
define dl_tar_pack
+ echo "TAR binary: $$$$(which $(TAR)) [version: $$$$($(TAR) --version | head -n1)]"; \
+ echo "ZSTD binary: $$$$(which zstd) [version: $$$$(zstd --version)]"; \
$(TAR) --numeric-owner --owner=0 --group=0 --mode=a-s --sort=name \
$$$${TAR_TIMESTAMP:+--mtime="$$$$TAR_TIMESTAMP"} -c $(2) | $(call dl_pack,$(1))
endef
then run
make package/nss-firmware/{clean,prepare,compile} V=s -j4 QUILT=1 PKG_MIRROR_HASH=skip
It should be staging_dir/host/bin/tar and staging_dir/host/bin/zstd. If not, rebuild them.
❯ make tools/{zstd,zlib,xz,tar}/clean tools/tar/compile V=s -j4
i have installed latest nss image from lorenzo.
i think this the best image to use on ax9000 router.
but i also want to run the internal fan for better cooling.
but i have massive problems to install the fan control ..
-
uploadede /install both .ipk files from @devrus91 here via luci System>Software>Upload Package
-
install i2c-tools and kmod-hwmon-emc2305 via luci System>Software
-
Add the below in Luci System>Startup>Local Startup
echo emc2301 0x2f > /sys/bus/i2c/devices/i2c-0/new_device
i2cset -f -y 0 0x2f 0x38 0x18
i2cset -f -y 0 0x2f 0x2b 0x01
echo 60 > /sys/class/hwmon/hwmon4/pwm1
exit 0
problem 1 
echo 60 > /sys/class/hwmon/hwmon4/pwm1 gives error permission denied ?????
gives error permission denied over ssh or /etc/rc.local
error:
-ash: can't create /sys/class/hwmon/hwmon4/pwm1: Permission denied
problem 2:
i cannot install fan control ipk : error sign not trusted...
can anybody take the two fan controll files/source code and compile them with trusted signing for this kernel OpenWrt SNAPSHOT r0-27c2d2f2 / LuCI Master 25.351.34363~3bd538a
Kernel Version 6.12.62
many thanks in advance
Hi @qosmio
I used GitHub Actions. As I understand it, make download is a host-side phase that runs independently of the OpenWrt build environment and always uses host system tools. It does not depend on tools or toolchain and therefore runs before make tools/install and make toolchain/install. The directory staging_dir/host/bin is created and populated only after make tools/install is executed, and it is added to PATH only for subsequent build phases, not for make download.
I believe the original nss-firmware download failure was caused by unavailable or non-functional mirrors.
Do you have any specific reason you need that kernel? Router is working well on not latest kernel - 1gbit per second over wi-fi. I am also using Lorenzo image but not the latest ones
The latest kernel with all the new patches simply runs even better, and I use it as the basis for everything else since it runs so well.
For the future, it would be very helpful if you could recompile the fan controller app for the new kernel.
Which Lorenzo image are you using exactly with the fan controller app?
I finally want to get the internal fan working with the Lorenzo image.
In Option 1, which I described above, the fan can be run even without the fan control app.
But I get the error: permission denied...for creating the PWM file. Without this PWM file, the fan doesn't run at all.
Is there any way to grant permissions for this folder, like chmod or something?
I've tried everything, but it doesn't work... always access denied.
Can any of the developers offer a solution for this?
Thanks in advance.
I am on 6.6.104 kernel custom built image with packages i need. Fan control packages I provide can be installed with that kernel.
Just using new kernel doesnt make it “run better” you should watch changelog and see if there is improvements of bugfixes for your specific hardware. Some times new bugs for old hardware are introduced in new kernels so i dont see any reason to use it)
After echo emc2301 0x2f > /sys/bus/i2c/devices/i2c-0/new_device
new hwmon device will be created. You should look in your /sys/class/hwmon folder to see what index it has and fix /sys/class/hwmon/hwmon{index}/pwm1 path
Fan controll app provides possibility to setup fan curve depending on cpu temps, if you dont need it you may just set fun to static speed and it will work
yes thanks
echo emc2301 0x2f > /sys/bus/i2c/devices/i2c-0/new_device command creates a pwm1 file in hwmon12 folder and not hwmon4…
that was the problem
Time marches on.
Perhaps you could recompile the fan control app for the new kernel.
Currently, the fan runs constantly and is not temperature-controlled.
However, it would be very helpful, not only for me but for everyone else, if the fan control app could also be installed on future kernels.
best regards for 2026
So did some digging, and looks like it's partially true. On first clean run. System tar is used but zstd and libdeflate are always compiled from source.
❯ rm -fr dl
❯ make clean tools
❯ make download -j4
make: Entering directory '/data/code/openwrt/ipq'
make[2]: Entering directory '/data/code/openwrt/ipq'
make[2]: Leaving directory '/data/code/openwrt/ipq'
make[2]: Entering directory '/data/code/openwrt/ipq/scripts/config'
make[2]: 'conf' is up to date.
make[2]: Leaving directory '/data/code/openwrt/ipq/scripts/config'
make[2]: Entering directory '/data/code/openwrt/ipq'
make[2]: Leaving directory '/data/code/openwrt/ipq'
make[1] tools/zstd/compile <------- zstd compiles before downloading
make[2] -C tools/libdeflate compile <----- from source
make[2] -C tools/zstd compile <----- from source
This leads to the specific issue you reported. During the first run, the system tar paired with the build’s zstd produces one hash:
TAR binary: /usr/bin/tar [version: tar (GNU tar) 1.35]
ZSTD binary: /data/code/openwrt/ipq/staging_dir/host/bin/zstd [version: *** Zstandard CLI (64-bit) v1.5.7, by Yann Collet ***]
Hash mismatch for file nss-firmware-2025.05.01~e4726900.tar.zst: expected 3ec87f221e8905d4b6b8b3d207b7f7c4666c3bc8db7c1f06d4ae2e78f863b8f4, got 881cbf75efafe380b5adc91bfb1f68add5e29c9274eb950bb1e815c7a3622807
OK, so then change it to the same hash 881cbf75efafe380b5adc91bfb1f68add5e29c9274eb950bb1e815c7a3622807 and rerun just clean/compile on nss-firmware:
Now it's using build's tar and zstd
TAR binary: /data/code/openwrt/ipq/staging_dir/host/bin/tar [version: tar (GNU tar) 1.35]
ZSTD binary: /data/code/openwrt/ipq/staging_dir/host/bin/zstd [version: *** Zstandard CLI (64-bit) v1.5.7, by Yann Collet ***]
Hash mismatch for file nss-firmware-2025.05.01~e4726900.tar.zst: expected 881cbf75efafe380b5adc91bfb1f68add5e29c9274eb950bb1e815c7a3622807, got 3ec87f221e8905d4b6b8b3d207b7f7c4666c3bc8db7c1f06d4ae2e78f863b8f4
However, if you update the hash and rerun clean/compile for nss-firmware, the build switches to the internal staging_dir version of tar. Even though both versions of tar report as 1.35, they generate different binaries, and the hash reverts to 3ec87....
This explains the conflicting reports from different users. To resolve this reliably, I will prepackage and host the tar file to bypass these environment-specific tool discrepancies. I’ll push an update later this weekend.
3 Likes
Hi @qosmio
Thanks a lot for the confirmation. I actually tried to work around the problem by running make download twice, before and after make tools and make toolchain in GHA. The second make download resolved the nss-firmware download failure because it now used the internal staging_dir version of tar.
Hi all,
I understand that overclocking the processor in the MX4300 might not be necessary, especially with NSS in place. However, does anyone know of a safe way to bump it up a bit, maybe by 100 or 200 MHz?
safe or bump (overclock)...You'll have to pick one. Because
The silicon in these runs significantly less efficiently than the higher-end IPQ8072. My MX4300s idle around 53C at just 1.4GHz, while my faster IPQ8072 units (mx5300/dl-wrx36) run cooler at 43C.
mx5300-1❯ min: 42.5 °C med: 43.5 °C max: 44.5 °C avg: 43.4 °C
dl-wrx36❯ min: 44.8 °C med: 46.5 °C max: 48.0 °C avg: 46.5 °C
mx5300-2❯ min: 42.7 °C med: 44.5 °C max: 45.3 °C avg: 44.4 °C
compared to my 5 MX4300's
mx4300-2❯ min: 52.0 °C med: 53.1 °C max: 54.3 °C avg: 53.0 °C
mx4300-5❯ min: 52.3 °C med: 53.6 °C max: 54.6 °C avg: 53.5 °C
mx4300-1❯ min: 51.5 °C med: 52.4 °C max: 53.1 °C avg: 52.3 °C
mx4300-3❯ min: 51.1 °C med: 51.5 °C max: 52.4 °C avg: 51.7 °C
mx4300-4❯ min: 51.2 °C med: 52.2 °C max: 53.6 °C avg: 52.6 °C
On top of that, the cooling solution on the MX4300 is much smaller and thinner compared to the massive heatsinks in the MX5300 or Dynalink DL-WRX36. There is very little thermal headroom.
I tested overclocking these myself, and while I could get them to boot at higher clocks, they were not stable. Temperatures quickly shot well over 70C. More importantly, the instability caused data corruption on USB-attached drives on at least 3 of my units. It is not worth the risk.
4 Likes