Connecting two DIU 3000s together?

[syntrx]

I'm attempting to connect a pair of DIU3000s together using their V.24 ports for a bit of a testing exercise for getting a pair of Quantars, or a Quantar and DIU to talk to each other via IP. The only reason I'm attempting this is with a pair of DIUs is because I haven't got a functioning Quantar available for me to test with at the moment. (More on that exercise another time if it works.)

My question is: Should a pair of DIUs be able to communicate with each other?

If I've got one unit connected to itself in a loopback test per the manual, it works great:

10:49:16  (38)(I) Sending SABM frame
10:49:16  (40)(I) Link Init Requested (Base)
10:49:16  (39)(I) Sending UA frame
10:49:18  (99)(I) Sending XID Command to Base
10:49:18 (101)(I) Receive XID Command from Base
10:49:18 (100)(I) Sending XID Response to Base
10:49:18  (43)(I) Link establishment O.K

The link light goes out, and when I hit the PTT I get the earphone light on the front panel. All well and good.

But if I connect two units together, they don't see each other:

10:59:20  (38)(I) Sending SABM frame
10:59:21  (38)(I) Sending SABM frame
10:59:22  (42)(W) Link establishment failed
10:59:22  (38)(I) Sending SABM frame
10:59:22  (38)(I) Sending SABM frame
10:59:23  (38)(I) Sending SABM frame
10:59:24  (38)(I) Sending SABM frame
10:59:24  (38)(I) Sending SABM frame

Is this a cabling problem that I should investigate further, or does something special happen when a DIU sees itself that causes it to go into some sort of test mode, that it otherwise wouldn't do if it saw another DIU?

 

[d119]

Firstly the equipment was never designed to do that, so the outcome is unknown.

 

[MattSR]

Well yes, we are aware of that, but it ain't going to stop us from trying. The fact that a DIU talks to itself via a V.24 loopback test indicates that it might be possible...

 

[commsec]

I am not sure this applies but we work on other equipment similar to the 3000. We found the only way to get the two devices to talk was to invert the data coming out of one unit. Again I don't know if this applies but it worked for us on other crypto equipment.

 

[d119]

Part of the issue may be that the TX & RX lines (obviously) need to be flipped, and also that the DIU may be looking to the Quantar for clocking. But that still doesn't explain why it works in looparound.

 

[Astro Spectra]

The loopback is part of the built in self test. The setup for which is in the manual along with how to do the various tests as syntrx has found.

As for a DIU to DIU working, while it doesn't do anything for me why not try it?

Assuming you have the correct pin 2 to 3 cross over between units and done something sensible with RTS, CTS, and DCD there are two other things to check:

First up is clocking you will need to take the ETC clock out, pin 24 from memory, and run that back in to the RC input pin 17 of the other unit. Do this each way. Remember that both TC and RC are inputs, so crossing them over doesn't achieve anything.

Second you need to set the DIU and radio addresses. Now just what the correct address combo would be is not obvious because as d119 points out it was never made to work that way. I'd go for 13 as default for both first and see if the unit thinks it's in loopback when you go to do the audio test.

If that doesn't work try one at 14 and one as 13, putting the other address in as the radio address. I doubt that will work as the protocol identifies the infrastructure type.

My feeling is that if it works at all it will only be because the units support loopback so tricking each unit into thinking it sees itself is the most likely option.