old Dekabox 3 level concentric (coaxial) stepped attenuator

a relic from the past ;) not shown is the metal cyl case that slips over this and protects it.

 

the first dial seems to step exactly at 10k intervals but the other two seem strange and might need some debugging. it would be a shame of any of those expensive resistors were bad, but I'm not sure yet.

 

this device was under $50 at my local surplus shop. I bet it cost a lot more than that when it was still being made (probably well over 30 years ago, to just take a guess).

 

update: found this online and it seems to describe some of this: www.ietlabs.com/voltage-divider/esi-dp-1211-dp1311-voltag...

 

update2: mystery solved: this is a voltage divider and it works perfectly! I think some of the switches are kelvin-varley style (they shunt a previous stage with more than 1 resistor sub-chain). this was verified when I put a precise 10.000v into the 2 outer terminals and measured from the center 'wiper' to one of the outers. the large diameter is tenths, the middle is hundredths and the knob is thousanths, with its arrow after the 9 being a '10' place. if you dial in 1,2,3 you get 0.123 * the voltage you put in, or 12.3%. purely decimal and fractions of percents as you go down.

 

so, this was definitely not a resistor decade box, it was a voltage divider box. and all the resistors inside seem perfect and there's nothing for me to do to it. it works perfectly and is accurate to many places on my 5.5digit voltmeter.

 

a better inside photo: www.flickr.com/photos/linux-works/7323208500/in/photostream

 

and the device in-use, set to "314" which gives 0.314 * the voltage you put into it (as an example): www.flickr.com/photos/linux-works/7323321384/in/photostream

Done