28vac & 115vac inverter that can be driven by 400Hz analog sine wave

[richfiles]

I'm looking for a low cost means to drive nine 28vac outputs, and a single 115vac inverter with a 400Hz analog sine wave. The source is not PWM... It's an actual analog signal, and this is not something I can get around. I've seen plenty of inverter designs that simply produce square wave, or modified "sine" outputs, or use capacitors on the output side of the transformer or PWMed inputs to create a sine output, but I don't know whether it's safe to feed such a design an analog input.

One thing I've considered is audio driver ICs, particularly for the 28vac outputs. These are not constant outputs, but rather, outputs that can be attenuated from 0-28vac, and are all phase synchronized with the 115vac output. I'm not as experienced with amplifier design and all this analog domain stuff... Something that could take a 0-5v input, feed it into an audio amp, and then capacitively couple that output into a transformer to bring the output up to 28vac would probably be ideal... But then again, I'm not well versed in this stuff. These outputs are simulating the electrical behavior of a synchro control transformer, and at best, are probably operating in milliamp range, so I don't think overall power needs to be too high. That's why I'm thinking audio amplifiers might do the trick.

The 115vac output is different. It's a fixed output, meant to run on a constant 115vac at 400Hz, and it serves as both power supply and phase reference. This is all meant to drive a Lear Siegler ARU-11/A ADI. I've got basic hookup info, but I don't have info on overall current draw of the device, so this is another element I'm not certain of. I don't think current draw is too high though. I've seen devices that do what I'm trying to do, and they are rather compact. I think they may even forgo the transformers and just directly drive the outputs, but I have to imagine such a solution is also very specifically engineered... I'm not exactly at that level.

I am open to anything from general advice, dos and do nots, schematic examples, to "here's a PCB module on Ali Express that'll do what you need"... Also advice on transformers that might meed my needs. I know 400Hz transformers are smaller than 50 or 60Hz ones. That's done to save weight on aircraft. I know 60Hz on a smaller transformer meant for 400Hz will saturate it, but can you go the other way around, or an I asking for trouble doing that. For the 28vac outputs, I was actually thinking about whether I could make audio coupling transformers do the job, provided I can find something remotely suitable. I suppose I could theoretically go through the tedious process of making a transformer as well. My employer makes electric motors, and we are set up to be able to wind stators and rotors, and to stack and fluidize laminations... I suppose I could go that route, possibly, but off the shelf is preferred. Part of designing this is to provide an open source means of controlling aviation instruments like what I have, and right now, all the solutions I see are both proprietary and come with "call for pricing" price tags. The cheaper and simpler and more easily sourced the parts, the better.

I guess I'm looking for advice on two separate fronts here:

The 28vac outputs I imagine can probably be solved with something like an audio amp design. I need nine identical outputs, so cost is a consideration too.
The 115vac power supply is something that I'm more worried over. The instrument this is powering cost a pretty penny, and I want to power it with something that's not gonna blow it up, so I wanna do this right. I've been staring at one schematic after another, looking at inverter design after inverter design, and time and time again, it's square, modified "sine, and PWM... never, ever true analog sine input, so I guess I'm just not sure where to start, or what I can or can't do.

[langwadt]

how much power?

[bdunham7]

There was a rather large 115VAC/400Hz/3PH inverter being sold in another thread that might work somehow, or contain parts that would help you make what you need.

50/60Hz transformers often work just fine up to 1kHz or so, but not the other way around.  You might be able to use a large 115VAC inverter and then step-down transformers for your 9 28VAC outputs, assuming you want them isolated.  How much power do you need?  The inverter I've linked here is pretty big.

https://www.eevblog.com/forum/testgear/rip-ron-barns-estate-test-equipment-liquidation/msg3657853/#msg3657853

[richfiles]

how much power?

First, I do apollogize for not replying after posting this message... I was out of town doing some work.

Unfortunately, I don't actually know the answer to that... Finding the power or current requirements for my ARU-11/A ADI has been very unfruitful. Finding specs for avionics seems to come with barriers. I don't think the device draws an exceptionally high amount of power. The 115vac powers the electronics, 3 internal synchros, and 3 small servos... Mind 2 motors and 2 synchros are INSIDE the barely 3 inch diameter ball. I am trying to go over old Youtube videos, trying to find people who have run these and posted demonstration vids. Will ask what their reference power draw was, or if they have an output rating label on their power source, at the least.

As for the 28vac signals... They are just that, signals. They go into an analog comparator circuit that compares them to the synchros tied to the ball. When an error is detected, the servos are driven though an amplifier circuit to eliminate the error. This is all built into the small electronics package on the rear of the ADI housing.

I also found this today:


from http://www.seekic.com/circuit_diagram/Amplifier_Circuit/400_Hz_AMPLIFIER.html

Developed to increase output power of digital-to-synchro converter systems while providing stable and accurate output and overall gain even with reactive loads, at Includes overload protection. Delivers 95 VRMS 400 Hz continuously into 500-ohm load. Power bandwidth is about 20 kHz. Foldback current limiting drops short-circuit current to 200mA when load exceeds 300 mA.-F. H. Catter-molen and J. A. Pieterse, Digital/Synchro Amplifier Features Overload Protection, EDN Magazine, Nov. 5, 1977, p 107-108.

As for the output... 95vac at up to 300 mA... Says it foldsback current to 200mA when it exceeds 300mA. Suggests to me that it expects most devices to be pretty low current. I have been told by people that these ARU devices will still operate down to 70vac, so 95vac seems a reasonable option to consider. 200-300mA doesn't feel like all that much, but 95vac at 200mA ought to deliver 18w of power. 6-9w per each of 3 high efficiency mil-spec servos, plus accompanying synchros and error amplifier circuits... Possible. I really hope I can get a reply from one one of the youtubers I mentioned, and get an actual measured current draw.

I'm beginning to wonder if I've actually stumbled onto my own solution. It says +150v and -150v in... I also see that -150v is capacitively coupled and tied via a diode to com/gnd. EDIT: BOTH +150 and -150 are capacitively coupled, as well as tied to com via diodes...I suppose these ARE floating com at about the center then. As a mostly digital person, I'm not 100% certain how I'm supposed to read these voltages. Is this a reference to a rectified AC (eg. rectified 120v... I know that normally puts out about 170vdc, due to the RMS conversion). Seems weird to specify +150 and -150, when the com is not tied to a center tap or virtual ground (I think)... (I think I am wrong... I just missed details in the schematic) Am I just seeing a weird nomenclature for labeling a simple 150vdc source's + and - connections here, or is this really a 300 volt supply?

Advice is, as always, welcomed!

[richfiles]

There was a rather large 115VAC/400Hz/3PH inverter being sold in another thread that might work somehow, or contain parts that would help you make what you need.

50/60Hz transformers often work just fine up to 1kHz or so, but not the other way around.  You might be able to use a large 115VAC inverter and then step-down transformers for your 9 28VAC outputs, assuming you want them isolated.  How much power do you need?  The inverter I've linked here is pretty big.

https://www.eevblog.com/forum/testgear/rip-ron-barns-estate-test-equipment-liquidation/msg3657853/#msg3657853

I inquired about that inverter, just for the heck of it. You're right about it likely containing usefull transformers. worst case, I could measure the impedance of the transformers and wind new ones at work (we manufacture electric motors, so we have winding equipment, lamination stacking tools, high precision inductance meters, etc). Honestly, It'd be nice to just have a 400Hz source around for testing. I do have an adjustable digital oscillator capable of single phase and dual 90° phase 400Hz, but the inverter it was part of suffered a catastrophic failure of one of the phase outputs. The other was at reduced capacity, but was parted out over a decade and a half ago. I want to use that oscillator in my project, but needed the amplifier/inverter circuit. I'm actually thinking the circuit in my last post might be just the thing... I hope!

[richfiles]

Progress has been made!
I wanted to post my seeming solution here, as well as get opinions and evaluations of the overall design.
This entire post is a spinoff of this older thread from years ago:
https://www.eevblog.com/forum/projects/trying-to-emulate-three-synchros-using-arduino/



I was on a random chat a couple weeks ago, jokin' around about the number of Chrome tabs I tend to have open. After the initial questioning of my sanity, the next obvious question was why. Part of it was just queueing up youtube vids to play while I work. Part of it was just having frequently used sites or tools open, and probably the biggest part, was pages and pages of parts. Comparing pricing, availability, stocks, electrical characteristics, etc... That led to the question of what I was building or working on with all those parts. That led to a discussion of work, and a discussion of my projects... And that led to the relevant topic of the day. A random person online caught an interest in the high voltage components for my Kerbal Space Program instrument panel project. He took a look at the old 1977 schematic I'd found, I explained what I was trying to do with the project, and explained that I don't have a strong background with high voltage analog circuits. He then tells me to give him a few days, and he'd try to work on a solution! Wow! I am just stunned at this! The dude pulled through and this was the result:



So first things first, This circuit is made of modern parts that are still apparently available, despite shortages, thank goodness! No vintage unobtanium and no in demand current day unobtanium. This circuit creates a second pair of reduced voltage rails for the input op-amp using a pair of resistors and zener diodes. The +22/-22 volt input op-amp has enough swing to drive Q2 and Q4, which in turn, I believe, sit at about 1/3 the + and - primary DC rails, which are derived from the rectified output of a 230v center tap transformer's secondary. Q2 and Q4 then drive the output stages Q1 and Q6, further down. Those two transistors sit at the full DC rails. My knowledge of analog is semi-limited, so the function of Q5, Q9, and Q10 is a bit above me, but I presume these are supposed to be part of the 400mA current limiting built into the circuit.



I'd love to hear some second opinions before I pull the trigger and make this. Everything looks good to me, but as I've said, I have a limited background in high voltage analog. I'm told it ran successfully under a spice simulation, so it has that going for it. power supply isn't shown in the schematic, but it's literally just a bridge rectifier and 8 caps, 4 per rail, so nothing exotic. I took note that U1 had a shutdown pin. The datasheet says that when left open or tied to 1V or less above the negative voltage rail, the part is active. I also noticed that the line can be susceptible to capacitive coupling with the inputs, so tying it or capacitively coupling it to the negative voltage rail can eliminate that possibility. I decided to add two parallel traces to the board to accommodate a capacitor, resistor, or both, for the purposes of tying or capacitively coupling the shutdown pin. Additionally, I noticed the power transistors on the left side of the board had their heatsink tabs facing the wrong way, so that's getting flipped on the board (no changes to wiring, only orientation). Aside from those two details, I have not spotted any other issues. With the design.



A shout out to Dr. Rockzo, wherever and whoever you are. This board is something I've needed to get figured out for 5+ years now, and I kept doing other parts of the project first, pushing this back, knowing it was not my forte. I'd been clowning around, and now through clowning around on the internet, this random rockin' clown has come along to save the day! Hehe, great logo! You've come through with some fine work, and I definitely appreciate it!