Gregory's Serveri wasn't sure what sort of power the module needed. maybe 28VAC at 400Hz? i teased out and disconnected a bunch of wires on the large power transformer, fed in 28V to the suspected input wires, and measured the outputs.
|
37 comments
 Marc has an inverter that can turn 28VDC into 115VAC at 400Hz, but he's on vacation and i've been working on the project from home. so what to do? i built my own inverter, natch! this is a junk box 115V transformer with a 12V center tap output. it turns out to work fine at 400Hz. i'm supplying DC to the center tap and there are two MOSFETs each with the drain connected to the remaining two terminals. this is essentially a resonant Royer design, but the MOSFETs are being driven by an Arduino that tries to PWM two half sine waves. it's bad but it works! the basic idea of this device is that you feed in two sets of 3-wire synchro signals, one coarse and the other fine. they are presumably from some kind of sensor. these synchro signals get fed into a pair of autosyns--a device that generates an error signal proportional to the difference between the input electrical angle and the mechanical shaft angle. the error signals enter this transistorized amplifier circuit. there's some filtering, limiting, and a bit of the rate generator signal is mixed in (maybe as a differential gain term?) but basically it just makes a small signal much larger. (red arrow because the signal path travels from right to left) the last stage of the transistor amplifier drives the two mag amps, which are cross connected so that one runs the motor in one direction while the other runs the motor in the opposite direction. depending on when the transistor turns on (due to the phase of the signal from the autosyns) it will turn on one or the other mag amp. (the schematic here might have swapped wires on the mag amp outputs, so ignore that) to sum up: the whole gear train (which is all connected together) gets rotated by the motor so that both the coarse and fine autosyns match the input signals coming from the external sensor's pair of coarse and fine synchro senders. OK--so our external sensor has a mechanical position, and now we've duplicated it inside this avionics box. what now? the gear train connects to several other things: 1) another pair of synchro senders, which produce a buffered version of the sensor input (another set of six output wires). 2) a "Digital Analog Converter" which presumably converts the mechanical position into a digital signal. (i need to take it apart at some point!) 3) a set of two potentiometers, each of which have multiple ganged taps. i haven't tracked down where all this wiring goes yet. i've been figuring out the gear train and counting teeth for...a long time. here's my horrible diagram. looks like the full range of motion between the hard stops is 294.5 degrees on the potentiometers. oddly enough, the fine INPUT synchro turns 1 full turn for 0.8500 of a turn on the OUTPUT fine synchro! full range on the coarse input synchro is 101 degrees (it can't move a full turn). for the output coarse synchro, it is 85 degrees. also strange is that the gear reduction between the coarse and fine synchros appears to be 12:1. that's not really a standard ratio--equipment typically uses 9:1, 18:1, 27:1, or 36:1.  @tubetime Would 12:1 imply that clock time or degrees are involved? I just can't emphasize it enough: jet-age aircraft ran on clockwork. straight up steampunk what with the little brass gears and vacuum tubes and neatly-tied bundles of colorful wire  @tubetime I still find the idea of "landed on the moon wearing mechanical wrist watches" to be kind of strange. @tubetime jet age? I earned my license in 1991 and I still had to use a circular sliding rule (the venerable E6B) on my private and instrument check rides. @tubetime honestly with how modern cars are, i'd be more willing to trust this contraption than something that is 1 30-day period away from the aviaton industry's version of Adobe cloud expiring mid-flight. the anti backlash gearing is a giant pain to deal with. I'm making up a bunch of these springs to hold them in place so I can reinstall gears finally got the "Analog Digital Converter" opened up. it's rather complicated! there are multiple absolute encoder wheels connected with a gear train. since i had the gearbox all opened up i counted more teeth and figured out that the coarse-to-fine synchro ratio is actually 1:11, not 1:12 as i previously thought. @tubetime 1:11 isn't like any encoding altimeters I've seen but I'm still betting on an analog synchro to digital Gillham code converter. @bikerglen the Air Data Computer that Ken has been busy with uses the 1:11 ratio as well. i think this unit is from the same avionics system and is designed to be connected to one of the computer's outputs. maybe altitude. this is going to be a real pain to reassemble. I've put in springs to put tension on the backlash mechanisms. so I got everything lined up and the top plate reinstalled. I just have to pluck out the springs and tape. yep, that was a pain. but it is back together, and I have all the connector pinouts! the potentiometers are not all linear relative to position. fascinating.  @tubetime it's probably difficult to make them completely linear @tubetime its typical to see 'log' pots consist of two or three different gradients like you see on the orange line here. Whats most interesting to me is the 'deadzone' of the green pot I don't think I have seen that before @tubetime there's a parallel digital altitude format used by aircraft avionics. Is it possible this box converts a coarse + fine synchro formatted altitude input into it's digital representation? I'm not at home or I'd look up the name of the digital altitude format.  |
even if the input voltage is supposed to be higher, the ratio of the voltages will remain the same. in this case, the output voltages only made sense for an input of 115V.