[lexridge]

Finally received my unit today. The header pins were already there, fortunately. Rev 205 mainboard. Had the latest Linksys firmware installed. Flashed like a charm with tftp. We should try to figure out by serial number when the serial port was disconnected on the latest units built. Would make it much easier and less risky to buy these units used.

[McKenzieEA]

[blkt]

EA8500 here, with retail box Rev.A00 (above UPC code), PCB REV:205(XC).

From http://192.168.1.1/sysinfo.cgi (stock firmware) hw_rev=1 hw_ver=XB

All sorts of nice output from sysinfo.cgi I saved to text while booted into each partition for future reference.

Used DTECH FTDI (FT232RL) 3.3V 3 pin; only the 6 foot version existed when I bought it but good length.

https://www.amazon.com/dp/B07RBKCW3S | https://www.ebay.com/itm/124251758238

Followed mrjcd's walkthrough, minus the Prolific drivers (used driver zip below).

Used PuTTy-TFTP.zip; it contains tftpd32 4.52, PuTTY 0.67 and a known working build for serial flash.

http://mrjcd.com/EA8500_DD-WRT/ | Also, verified COM port and speed 115200 baud in device manager.

Used printenv instead of setenv, then renamed the file to be served by TFTP accordingly to wraith.bin.

Also, used 192.168.1.254 instead of 192.168.1.2 for the Ethernet adapter.

Whichever way you go, make sure it all matches up on both sides.

(IPQ) # printenv image
image=wraith.bin
(IPQ) # printenv ipaddr
ipaddr=192.168.1.1
(IPQ) # printenv serverip
serverip=192.168.1.254
(IPQ) # printenv netmask
netmask=255.255.255.0

If you run "printenv" alone it will show the entire list of environment variables (sysinfo.cgi also provides).

https://www.ftdichip.com/Drivers/VCP.htm | Installation Guides
Windows x64: driver zip, Release Notes, setup exe alternative | Mac OS X: 10.9 and above
https://the.earth.li/~sgtatham/putty/latest/w32/putty.zip | tftpd32.464.zip, tftpd32.452.zip
https://duckduckgo.com/?q=putty+mac+alternative | https://duckduckgo.com/?tq=tftp+mac

[egc]

I use setenv, what is the advantage of printenv?

(btw I had an European version which upgraded fine via the GUI. I use the serial because I am trying to build for this little rascal )

[blkt]

There is no advantage other than show the factory defaults, as I chose to make server-side changes instead.

Using setenv shouldn't matter with correct IP addresses and file name, same as Ethernet adapter, TFTP, etc.

A reboot should revert any environment changes you make, unless you run saveenv.

Before making changes paste printenv output to text file, or sysinfo.cgi output while still on stock firmware.

(pops mentos tablet in mouth, holds up mentos and smiles, freeze frame).

[lexridge]

Just to keep this old thread up-to-date, I just received a lot of three of these routers this week. All were the latest versions EWW604-B1-LS(S or 5) REV:305 (XD) with the newer power supply. The (S or 5) is because this printed character looks much more like an S than a 5 on all I have seen (even in photos with the printed text stating it's a 5). I had to solder serial headers on them all.

So, I preceded to flash them. Or attempted too anyway. I could read the serial line just fine, but it would always ignore my enter key when trying to halt the boot. First router failed multiple times, so I soldered on a header to the 2nd and 3rd routers. Same exact thing. I then went to one of my production routers which is also the newer variety and I had previously flashed using this method with the same USB serial hardware. Popped the lid and connected the serial cables. Again, enter key ignored (I know, but I don't have an ANY key ).

At this point I knew the issue was not in the routers, but instead my serial connection. A few years ago (yes years!) I had taped off the red wire so I would not accidentally use it. I have used this unit in the past 6 months to flash a friends EA8500 router (REV 205), successfully using Green (TXD), white (RXD) and GND.

Now here's the really weird thing. So I tested the voltage of the green wire (TXD) and it read 5v. Wow, it's suppose to be 3.3v I am pretty sure. So then I tested the taped off red wire. It read 3.3v. Now, WTF? So cautiously, I decided to swap the green wire (RXD) with the red wire (VCC) ONLY because of what I was reading voltage wise on my meter. It just made sense to try it. I figured sending 5v over the serial port would have fried it already anyway, which I easily could have done on ALL FOUR routers! Ugh!

But it worked! How did the signals become swapped? Could a kernel driver update have changed the output signals on the USB serial port? I doubt it! I am using Fedora Linux 36. I am happy I was able to flash them all to DDWRT, but so strange how the green and red wires became electrically swapped in the interface.

In conclusion, I am certain that if any of these routers with the same MB model number is missing any components to prevent serial flashing, it would most likely be a very few end run models. Whenever this problem came up for me, I thought for sure I had three routers that had missing serial components and they were most likely going to get resold or junked. I would guess the OP had a defective serial port and NOT any missing components, but who knows for sure, as it was returned.

Just thought I would share my experience with this to help others along the way.

[blkt]

Cheap adapters factory soldered wrong or just wrong color codes, another reason to avoid clones and counterfeits.
Also there are TTL adapters with I/O pins TX RX configured to operate at 5V levels so be sure what you are buying.
TTL-232R-RPI, FTDI USB to TTL serial 3.3V, Adafruit 954 and 1x3 2.54 pitch pin header, or even less without solder.

[lexridge]

Oh yea. Many of us have had bad/cheap/counterfeit TTL adaptors, I am sure. Myself included. The one I am presently using I had purchased about 8 years ago from Adafruit, and have flashed nearly a dozen routers using it. I had the red wire taped off since probably day one and have only used green, white and black, until yesterday. I even had a post-it note in its storage bag with my notes for connecting to various routers which I had previously flashed, never having used the red wire. Just so weird.

Anyway the point of my post was really not to point out the issue with the USB-TTL adaptor, but to show that the later builds of these routers can indeed be flashed, and nothing appears to be missing from the serial hardware onboard. Maybe a few do exist with missing hardware, no way to know for sure without a side by side comparison....but I really can't see Linksys making a change like this with a production model that was about to go EOL. Unless of course, it was because of lack of parts availability, but I don't think that was an issue back in 2018/19 BC (Before COVID).

[blkt]

If your adapter is black PL2303HX or violet PL2303TA, by the end of 2016 Adafruit had switched to SiLabs CP2102.

[lexridge]

I have the black one, PL2303HX. Out of curiosity, I also checked a couple of others that I had in storage and never used. One has -6.2 volts on the TXD wire. Yes, that's a negative 6v. I guess the standard serial specs can use either + or - voltages, but not for TTL use obviously. Anyway, I went ahead and ordered a new FTDI USB-TTL 3.3v from Amazon this morning, with three wires and no VCC. I cannot see any actual need for a VCC line and it's just avoids future confusion and potential fires.

[hebeda]

why you guys do the serial upgrade when there is a firmware only upgrade path ?

you need first to downgrade to this Stockfirmwar:

FW_EA8500_1.1.4.168206_prod.img


then use Kongs old DD-WRT firmware Kong_38570M-020319

from there flash any DDWRT firmware version you like

i added the firmware , only visible when registered


i have flashed dozens of EA8500 without any problem using these firmwares - i am not sure if there are certain stock firmware which do not allow downgrades ...

[blkt]

US region EA8500 auto updates requiring serial. 1.1.4.169978 sets stage, lock, firmwares after are GPG encrypted.
Also factory-to-ddwrt.img has been fixed since r41174, so there is no need to search around for any old firmwares.

[fcj_bg]

Hello I juts created this account for anyone having this issue with a EA8500 US version with PCB 305 (XD), and with no possibilities to update via web due to the signed new firmware.
I do not have any proof but a unit a recently bought with all the above hardware also did not responded the serial inputs for the FTP method.
After a couple of hours and 3 USB to TTL (fully working on other units), which include Ft232rl, Cp2102 and CH340. I discovered there is not connexon between the Tx PCB via and its pull-up resistor, the resistor is marked as RM327 and it is located right in front of the serial vias. The connexon should be to the extreme closer to the via, you can check this with a multimeter, the other extreme is VCC. You can use a pogo pin or solder a wire to the via or your connector, then you can send serial commands and complete the FTP method. I retested the 3 USB to TTL and all of them worked after solder a wire.

[kernel-panic69]

Thanks for this updated information! I presume there's no bridging of pads required - this would then require either soldering a wire in place as you have done or using some method to hold the USB-TTL-UART adapter's lead to the proper side of the resistor in question.

[fcj_bg]