TP-Link SG2216 Support

Hello,

I have an old TL-SG2216 (v1) Switch I would like to use with Openwrt and am willing to help making it supported, if at all possible.
Can't say much about the device yet, waiting for a new TTL-USB Adaptor.

It has 16MByte ST M29W128GH-70N6 NOR Flash and 32MByte ESMT M13S2561616A-5T 200MHz DDR RAM. The Ethernet ports reside on different chips, 8 are connected directly to the SOC, the other 8 are connected to a separate switch chip which is in turn connected to the SOC.

Will provide Pictures tomorrow, an as soon as i can get a serial console I will try to find out what SOC it uses.

Tgt79

Looks like it's listed here and might be Realtek-based:

https://wikidevi.wi-cat.ru/TP-LINK/Switch

Found only the T1600G-18TS V1, which is the follow-on model and equivalent to a SG2216 V3

Pictures!
Imgur
Imgur
Imgur
Imgur
Imgur

Can't get anything on the Testpads for the serial line. Any ideas what I might be missing? The Board traces look as if they are connected all the way through.

R120 to TP3 appears to be missing. If TP3 is floating, you could try just bridging that component to hook up that test point.

2 Likes

You are right.I have no idea how i missed that. TP3 appears to be floating, or at least it has neither a ground connection nor carries any voltage. I'm gonna bridge R120 and try again.

1 Like

In the meantime I managed to decrypt the tp-link firmware update file (thanks to @andyboeh) and binwalk-ed around in it. The SOC is a bcm56218 according to various SDK file path references in the image. Never heard of it, couldn't find any information or datasheets yet :frowning:

The Switch runs VxWorks 5.5.1 btw.

Edit: some of the error messages in the oem kernel suggest that the secondary switch chip is a bcm53314. Or maybe bcm53314 and bcm56218 share the same target on vxworks and the bcm53314 is the main soc, it has an integrated mips cpu too. Hard to say without documentation and/or a working serial console.

Still no serial, can't get the tin to stick when trying to solder the bridge and wires for the connection. either my soldering equipment is broken or there is some kind coating i'm too stupid to remove. Someone else around with this device and more luck getting serial access?
Managed to bridge R120 and put in headers. There is a 3.3V Signal on TP3 now, not getting any console output though :frowning_face:

Anyone has a sg3216 v1 or some other model of the sg32xx line? I would like to know what model the used rs232 transceiver is. The Pinout kinda fits a MAX3232E TSSOP, if the Testpoints actually go to the TTL RX/TX UART

I've seen (clones of) the MAX3232 on other device, so I think it's a likely candidate. But if you aren't seeing anything on the test points, I doubt just adding a transceiver will help much. :confused:

You are right, but knowing the type of transceiver might help determine if the testpoints are really the TTL lines or if i have let myself be fooled by the traces going to the soc and it is actually just an enable signal or something.
The jtag has missing resistors on the vcc and reset lines, don't know how much use that would be.

I'm giving up for now, getting a bunch of linux based TP-Link Routers and Switches next week anyway where I might have more success.

I might be wrong now since I am mixing TX and RX all the time, but isn't TP3 the UART RX line?

If the missing IC really is a max3232e, then the tp3 pin is labeled as "TTL/CMOS Receiver
Output" and tp2 as "TTL/CMOS Transmitter
Input", which would mean you are right. But I am not getting anything on tp2 either, not even the 3.3V level. Maybe I fried the UART during my tests?

Edit: it is as if tp2 or rather its pcb trace is floating, no voltage against anything. Can't see any unconnected resistors in its path though.