I am lucky with the cable I bought on eBay to connect my 3174

 IBM COMMUNICATIONS CABLES

The communications adapter connection on the rear of the IBM 3174 is a DB25 male connector, which can be used for RS232 (EIA V.24/V.28), V.35 or X.21 links by choosing the appropriate IBM cable. These cables have the DB25 female on one side, an appropriate connector on the far end, but also the far end contains 1 or more switches that establish loopback for testing. When in Oper (not Test) mode, they are straight through connections of the thirteen active signal lines and signal ground. 

The connector that plugs into the 3174 has internal jumpers to pins that the microcode uses to determine which type of cable is attached. An ordinary DB25 serial cable would not connect pin 14 and/or 16 to signal ground to identify the cable type. 

The EIA and the X.21 cables have a single switch for loopback testing, while the V.35 cable features three. These cables come in several lengths as well, thus there are a range of IBM part numbers for comm cables. 

ORDERED USED EIA CABLE ON EBAY

I had copied down the various IBM part numbers and begun searching eBay and other sources to see if I could find one. I believed I had a match to a cable, described as EIA, which I bought. The one visual oddity was the existence of two Oper/Test switches, whereas all the cables described in my manuals had one or three. 

The cable picture from eBay

In fact, I had mis-copied the codes and bought a cable for an entirely different IBM device, the 5394. I was faced with the chance that this cable would be wired too differently to adapt to my needs. It had molded connectors, thus if the proper pins weren't wired, if the wires weren't straight through, and if the all important identification pins 14/16 weren't accessible to jumper, I might be out of luck. 

MULTIMETER CHECK ESTABLISHES THAT THIS WILL WORK

I sat down with the continuity tester and the cable, building a map of the pins that were connected through to the other end. I was pleased to see that as an EIA cable, it carried exactly the same thirteen signal lines and signal ground through. 

Next, I checked the 3174 end and found that pin 16 was connected internally to pin 7, although not carried out to pin 16 of the far end. This is correct and will be identified as the EIA cable my the 3174 microcode. I was extraordinarily lucky that IBM carried this method of identification forward for decades and chose the same code for EIA.

I then operated the two Oper/Test switches and checked for any wrap of the signals on the 3174 end of the cable. As I suspected, it wrapped exactly the same sets of pins together as I needed, which makes sense for an EIA cable under test. That is: 

• DTR hooked to DSR
• Xmit Data hooked to Recv Data, Xmit Clock and Recv Clock
• RTS hooked to CTS and Carrier Detect

The two switches separate the loopback into sections, but with both switches set to TEST, this worked exactly like my wrap plug and with the switches in the Oper position would be a straight through EIA cable.

TESTING THE CABLE ON MY 3174

I removed my wrap plug connector and hooked up the cable, with the switches set to TEST. IML of the 3174 proceeded normally to completion, exactly as it had with the wrap plug. This verified my continuity testing and gave me confidence that this cable was going to work for me.

I then set the switches to Oper and tried another IML. The controller stopped with an error 501 in the status display on the operator panel. This with its subcode 0102 told me that the modem was not working as DSR and CTS were not present. Exactly what you would expect to see if the modem were powered off. Another validation of the correctness of my new cable.