I thought I'd document this project for the benefit of anyone who acquires a bit of gear which uses these things, as I found data to be quite sparse and have found out some stuff by experiment.
For those who don't know, a Ledex is a remotely operated multiway rotary switch. The slave switch is driven round by a rotary solenoid until it matches the position of the master switch. They appear still to be made as they offer certain advantages over an electronic solution.
There are two basic types, the simple type which has one control line for each switch position and the more complex type which has less control lines than switch positions. For example, the one in my possession, an HF remote ATU intended for aircraft, has 5 control lines and a 12 way switch, although only 10 ways are used.
In both cases the slave switch has a control wafer which is the complement of the master switch. In the simple case the master switch is a simple single pole switch with as many ways as there are switch positions, the slave switch control wafer is a complete disc with a space at one position. Whenever the master switch is moved, the slave will step round until the space in the slave wafer matches the position of the master pole.
In the more complex case the control wafers form a pattern. My maths is not up to working out how this is derived.
Fortunately I was able to find online the manual for the transceiver which connects to the ATU and a partial manual for the ATU. The TRX manual gives the "truth table" for the master and slave switches. Not having the control head I had to work out an alternative.
The manual states that for any given channel certain control lines are connected to ground and the remainder are connected together.
Based on a simplified diagram, I assumed that the slave switch took care of the "connected together" bit and I could use a single pole 12 way rotary switch connected to a diode matrix as a master switch. (The truth table allowed me to derive codes for the undocumented 11th and 12th positions which don't have ATU components connected but I think can be wired as a bypass)
This actually worked, sort of. As long as you rotate the master switch from position one the slave duly steps in unison. Try to go back, though, and only some positions work, some do nothing and some go to the wrong position. Go back to one and start again, all is well.
Turns out this is due to the slave wafer not looking after the "connected together" bit as I had assumed. Shorting the unconnected lines together makes the unit step round to the selected position.
Initial thoughts were that I would have to put the thing aside and wait for the ham rally season to come round again in the hope of obtaining a big old wafer switch and replacing the contact section with a disc of pcb cut as a master wafer, but the current plan is to use the diode matrix to drive 5 SPCO relays with the N/C contacts tied together, the N/O contacts grounded and the common terminals connected to the control lines. The relays have been ordered from an ebay vendor and I will update this when they arrive.
The other common use for these things is channel switching in remote mount versions of old crystal controlled PMR gear.
In fact the aircraft HF transceiver which is intended to use the ATU I have comprises TRX and linear amp for mounting in the avionics bay, ATU for mounting adjacent to the aerial, and a cockpit mounted control unit which is little more than a stacked series of Ledex masters controlling each unit in unison.