Just commenting to say I'm still alive!
The setup I had planned on building has a hardware based sine generator. It doesn't rely on the microcontroller or the DACs for sine generation, only sine attenuation and polarity flipping, through the use of multiplying DACs and an analog switch IC to select and inverted or a non inverted sine wave, generated by hardware. For me, it was easiest to use hardware I had on hand, and it cut sine generation out of the program altogether. These details do seem to get missed by a lot of people posting here. More than once, people have suggested I do exactly what I already plan to do!
• Hardware generated sine wave.
• Multiplying DACs
• Microcontroller only controls polarity of the sine wave and the attenuation.
Since there were still questions, I'll try first to explain synchros:
The best way I can think of to describe a synchro, is if you take a single phase rotor, and mount it inside a three phase stator. That's literally it.
ALL the outputs are synchronized to the single 400 Hz reference that is fed into the rotor. What changes, is the angle of the rotor, in relation to the three phase windings of the stator. When looking at a single winding, for every 90° of rotation, you get a peak, a zero, a peak in the opposing polarity, and another zero crossing... A sine wave. Think of how an amplitude modulated carrier wave works. The carrier stays at 400 Hz, in this case, but the amplitude changes. The positive portion of the sine angle of the rotor's position translates to an in phase output on the stator winding, and the negative portion of the sine angle of the rotor's position translates to an out of phase output on the stator winding. The phase alignment of the outputs always remain in sync with the reference input, but it's the attenuation between peak and zero, and the polarity of the phase that changes. Each stator winding is 120° apart, and so you have to look at the rotor's position, relative to the winding's angle.
Interestingly enough, you could probably recreate a similar effect with some linear potentiometers attached to a crank. Connect the non-inverted reference to one side of the pot, the inverted reference to the other side of the pot, and take each of the three synchro outputs from the three pot's wipers. Attach all three pots sliders radially to a crank and rotate the crank. That might be the best visualization I can come up with. Synchros just do this with inductive windings, so they function like a transformer. Since they are transformers, they require AC to operate.
Second, I'll try to break down my theoretical method (it's not tested, built, or anything) for clarity.
• A hardware generator creates a 400Hz sine wave
• The 400Hz sine wave is fed into a pair of Op Amps: One inverting and one non-inverting. This creates two sine waves 180° out of synch with each other.
• The two sine waves are fed into an analog switch IC. The microcontroller can select the inverted or the non inverted signal to pass through.
• The selected sine wave is fed into the reference of a 12-bit MCP4922 DAC.
• The microcontroller sets the scaling of the DAC, allowing the DAC to function in multiplier mode, attenuating the sine wave on the reference.
•
For reference, there are a total of 9 analog switches, and 9 DACs, three per axis, for each of the three different axes. Each circuit is just repeated 9 times• The Microcontroller receives serial data containing the Yaw, Pitch, or Roll angle. (Each axis will be a separate PC board, and likely have it's own microcontroller).
• The microcontroller will add 120 and 240 to the angle received, then check if any results are 360 or greater, and if so, subtract 360 from the result.
• The three values are used to look up the attenuation value from a lookup table of sine values. (I think that might be easier than trying to calculate sine. Not sure)
• Each of the three values returned are checked to see if negative. If negative, the associated analog switch has it's state set to select the inverted sine wave input.
• The absolute values are sent to the three DACs of each axis. This allows the DAC to function as a multiplying DAC, scaling the analog sine wave on the reference.
• The output of the three DACs are fed into audio amplifiers.
• The three audio amplifiers are fed into transformers to couple into the synchro inputs of the FDAI.
• The non-inverted reference is fed into a power amplifier, and is fed into a step up transformer to power the FDAI and provide it the reference source.
As far as I can tell, that covers just about everything. Since programing is my weak point, I'm taking efforts to make sure as much as possible gets done at a hardware level, so the microcontroller side ends up as low effort as possible. Each axis microcontroller should only need a serial input to read incoming data, an SPI out to command two DAC chips, 2 Chip Selects for the two DACs (the chips are dual 12-bit DACs), an optional pin for synchronized updating, and three outputs to the analog switch ICs to select the polarity of the sine wave input. That's pretty straight forward. 5 (optionally 6) digital outs, 1 serial in (whatever interface works), and one SPI out.
My reason for coupling with transformers was simply so I can both level shift and float the output so it's not tied to any reference. Also, since synchros are inductive devices, I was just trying to keep the coupling inductive. The DACs I'm using fit my budget, but it means I'm dealing with all 0-5 volt, single rail analog. I suppose capacitive coupling might be possible too, assuming the FDAI's controller pack is looking for just voltage differentials. If that's the case, I wonder if I even need the beefy audio amplifier chips on anything but the reference (since that has to actually power the entire unit)? I dunno. I figure, I can possibly "cheat" and use a US power supply transformer to drive the reference. I'd feed the output of the reference audio amp into the secondary, and hopefully, have enough current on the 115 volt primary side to actually power the unit. I can't say i know what the power requirements for a mid 1970s era mechanical ADI even is. I also don't know if feeding 400 Hz into a US 60 Hz transformer is okay to do. I know that at 400 Hz, transformers tend to be smaller, which is one major reason aviation uses so much 400 Hz AC equipment... Smaller transformer means lower mass. I don't know how much losses this will result in, or if I can even find an appropriate transformer to perform the step up. In general though, that's the one area I know I absolutely will need a transformer.
Home
Help
Search
Login
Register
