Agilent 54701A active probe power supply

[0xdeadbeef]

I recently bought an Agilent 54701A for cheap which was said to be in good working order ("voll funktionstüchtig").
Since I don't have the official supply 1143A which is hard to get over here and very expensive anyway, I wanted to figure out how to build my own power supply.
From what I can tell from the service manuals of the probe and power supply, the pinning on the Lemo connector should be as follows:

1 red   +17V
2 orange   -17V
3 green   GND
4 black   (Offset +/-5mA)

The colors are the wires of the opened Lemo connector. The pinning order is actually marked on the soldering side of the Lemo connector (circle around pin 1, then a line connecting the following pins in their order).

I created a dual 17V supply from two (isolated) outputs of my bench supply, both set to 200mA (since the manual defined a maximum current draw of 110mA).
However, when I turn on the outputs, there are are 200mA drawn on the +17V line and nothing at the -200mA line. When I measure the resistance between these pin 1 (red) and 3 (green), it's only 500Ohm.
Actually, when measuring the resistance from the shielding to any of the pins, it looks like black pin (4 according to the Lemo pinout) is ground.

So it looks like the pinning is mirrored for pin 1..6 (the center pin 7 is not connected anyway).
If so, the following pinning would be correct:

6 purple   +17V
5 blue   -17V
4 black   GND
5 green   Offset +/-5mA

I'm a bit hesitant to try this as it seems to contradict the service manuals. I don't know how I could misinterpret the pinning order on the connector and I'm unsure if it's dared to even consider that Agilent did.

Side note: according to the service manual, the pins 5 and 6 should be iD0 and ID1. In Figure 16, they seem to be both connected to ground. However, even with the mirrored pinning it seems like only of of them is pulled to ground (500 Ohm) but the other one is not connected to GND at all.

[0xdeadbeef]

OK, when I typed the original posting, I became more and more confident that the pinning must be mirrored (for whatever reason).
I tried the 2nd (purple/blue/black) pinning suggested above and the probe works with that.
There is quite some offset error though, so some offset adjustment would come handy. Currently I left the offset pin open (assuming that 0mA should give 0V offset), but I guess I'll need to come up with some (as easy as possible) +/-5mA offset solution. Suggestions?

[0xdeadbeef]

To continue this monologue of mine: I wonder if anybody ever made attempt to recreate that offset control circuit as simple as possible and with components easily available.
Actually, I couldn't quickly find OpAmps that would support +/- 17V as normal operating range - at least not from LT/AD so I could simulate them in LTSpice.
To simulate the circuit from the 1143A supply manual, I chose the ADTL082 which allows ±18V under "absolute maximum ratings".
With the attached LTSpice simulation, it can be shown that the circuit actually drives +/-5mA into the 500Ohm load of the probe's amplifier box. Well, it can be cranked up to +/-5.4mA or so, but let's ignore that.
Also reducing the 500Ohm load doesn't really change the current, so this is actually a current regulation.
To simply simulation, I replaced that 1.23V Zener/Shunt regulator with a fixed 1.23V supply. I guess in the real world, an AD589 or a similar shunt regulator could do the job.
There are lots of weird resistor values (909k from E48 and what not) used though and I wonder if fine control and offset couldn't be merged to save one OpAmp and one potentiometer.

Anyway, for the moment I guess I leave the offset pin open as there is an internal 500Ohm pulldown anyway which should select an offset of 0V.
I guess I need to calibrate the probe itself though but this seems feasible (described in the service manual).