A friend gave me an ST-5000 cheap chinese transformer step up/down 240/120 which was not working. Taking it apart I found mostly broken wires among a bunch of switches and sockets.

I did not care much for all the crazy wiring so I stripped out the **giant** transformer, which is labelled only NHS-5000VA.

Now I would like to make it work. I figure it is an autotransformer (single coil with lots of taps). I have managed to get it to step down 9VAC to 4.5VAC. But I cannot get the reverse to work. (I use low voltages for safety while testing)

Checking with an ohm meter everything reads very near zero which makes sense for an autotransformer. I am using a Fluke 73. Maybe a higher precision ohm meter would help? Sending in a square wave from my oscope the signal shows up everywhere unmodified.

It has 5 wires all coming out on the same side.

From left to right:

black (larger gauge)
red (larger gauge)
red (larger gauge) (was originally the neutral on 120V output socket)
red (smaller gauge)
red (smaller gauge)

Any ideas on how this thing may be wired or what I might test for?

I have a fair number of standard transformers (by which I mean labelled and isolation type) of various winding and voltages, but this thing has me stumped.

Before I took the unit apart I was also terrified that even with the unit's "switch" and breaker turned off voltage and current still appeared on the some of the sockets.

 

No further clues from how it was origonally connected? And when you say "everything reads very near zero" you mean all the windings are connected together and the resistance is so low you can't tell them apart? A better resistance measurement would help here, it would allow you to at least identify the order of the taps and the two ends. If it can convert 9V to 4.5 then it should work the other way, but your voltage source you test with needs to be able to deliver enough current.

The most useful thing would probably be if you have some photos before it was stripped.

 

Indeed you have read my comments correctly.
Resistance is too low for my meter, all the paths read the same.
I will work on finding a higher precision ohm meter.
Also I like the idea of delivering more current for the step up test.