@AndyX90 Do you happen to know if your board have the RTL8211FS or the RTL8211F chip?
@brikr same happens to me, there seems to be something different between boards, not all the V3's are the same.
From my side as un update: The replacement chip ( RTL 8211FS ) arrived, I changed it but it behaves the same, I can conclude that the original chip was ok.
Next steps will be to find a donner board to replace the MT7622 ( seems pretty hard to find such part ) and I will also try to retrofit a RTL8211F instead of the 8211F(S) to the board.
Interestingly enough, the ID seems to be the same for both. I assume you can detect the difference by trying to access the SGMII related registers. If there is any doubt. I don't think there can be? I assume the phy on the 6 LR is using a SERDES link to the mac? So RTL8211F is out of the question. It's a RTL8211FS. Or something entirely different.
@bmork I am way out of my depth here. I took the change that @AndyX90 says works for him. He changed the original value 8 to 0. Both values are out of the range 1..7.
Thanks, but unfortunately doesn't give any output on my device. It seems like the first ~30 seconds of the boot log is missing from the dmesg output for some reason. I am guessing that's where the relevant output would be.
I can provide any sort of hardware support/investigation needed if anyone here knows how to solve this problem ( including scope measurements, comparisons, decoding, etc ... I have all sort of lab equipment ), unfortunately I am not into firmware design and here is stuff I don't fully understand.
As my next steps to follow: after replacing the RTL8211FS chip is to replace the MT7622 after I manage to find a replacement ( because the official Unifi FW doesn't boot also.. A12 error ).
Another thing I take into consideration is to wire a USB network adapter since the USB pins of the MT7622 BGA are close to the border of the chip, I am positive I can run some jumper wires from there to expose the USB3 interface.
In this thread Unifi 6 LR Issues - #23 by darkm4n I described some issues I had which I believe are because of this newer hardware revision and I believe I've now fixed the "bricked" u6 lr, and figured out what happened. After taking it apart and soldering on some pins to access the serial connection, it turned out the bootloader and openwrt were both mostly working fine, just that even the bootloader was unable to initialize the ethernet port, so it was unable to perform the normal tftp recovery. I was able to use the serial terminal to enable the wifi radio, use the wireless network to scp over a backup of the mtd2 and mtd3 partitions (uboot and uboot_env) from my second working u6 lr, apply them, and then the tftp recovery worked as normal. I believe that the issue was I had flashed the stock 6.0.21 firmware. Given that my to u6 lrs came with 6.0.24 and 6.0.25 (neither of which is available for download from the unifi website), I'm guessing that 6.0.21 is from before the v3 hardware rev, and doesn't support it's hardware differences (which I believe is perhaps also why the led wasn't behaving correctly either). It might be good to change the wiki page somehow to warn others of this particular risk and to avoid flashing older firmwares unless you know you have a v1 or v2 device.
Hi. Just found something out. So I got two 6 LRs and they are 2243V LR, so those are fresh and they both come with an AQR112C. Now I've tried performing the usual installation steps but it seems to have been impacted the same way. It lights up blue but nothing happens after that, no interface shows up, soft of odd. The only issue I might thing of is that I've used the kernel image instead of the upgrade image. Could that have bricked it?
Hello, I'm in the same situation with an apparent v3 U6-LR (2246V). I've built from source with the patches applied, and still the ethernet does not come up on the device. Wondering if you or anyone has found a path forward?
I was under the impression that the difference between version 2 and version 3 is the ethernet chip. So I don't have the RTL8211 in my APs but instead have Aquantia AQR112C (as it's supposed to). Date of manufacture is later than what is listed in the wiki. So I assume I have version 2 somehow, even if the manufacture date is relatively recent. The issue is that it still doesn't work. After installing and trying a lot of different combinations of firmware(starter was 5.60.8) the result is always the same: steady blue light and no ssh or web interface available. Nmap shows nothing so I assume the ethernet chip fails to initialise? Perhaps the problem is not in the chip itself?
B
I tried flashing a build using the patches on a confirmed realtek unit (also with code 2234k) with no luck(same MTD layout as in the existing v2 flashing guide). The device appears to boot, and reaches a stable blue light but nothing on the ethernet port. I was thankfully able to tft recover just fine to stock.
@AndyX90 were there any additional steps you took when building with the patches? I just selected the model and non-mod uboot, then applied your changes and built. Could you walk me through your build steps? (Which version did you build off of/which build config you used)
I just selected ARM -> l6r model (no uboot mod) and then applied the patches, and built per the guide.
Okay I would like to issue a correction. Manufacture date of 2243 works fine. Purchased two APs a month ago directly from Ubiquity Europe. They have the AQR112C chip in them. So it seems the process is inconsistent, or they reverted back to V2 chips for whatever reason. Funny thing is either my 2 of my Ubiquity poe injectors burned out or something funny is going on because they work in TFTP mode but do not function for accessing Luci. Regular Poe switch worked just fine. Maybe this helps someone who didn't check their dmesg logs for the exact version of their hardware.
