How would you build a 20W 400Hz 24V RMS sinewave generator?

I have seen various designs over the years.

Some crude as hell and producing basically a square wave.

A neat design I saw used a Class D chip - LX1725SK - which doesn't exist anymore except on Ebay, but there are loads of new ones. But these will need a BIG inductor as the output filter, which incidentally whistles unless potted.

I also want to avoid a BGA package because they are a pig to prototype and debug.

The original sinewave will come out of a DAC so that's no problem.

I reckon a Class D chip is still the best way today but wonder if there are other ingenious ways.

The load is highly inductive.

Distortion below a few % :)

Any tips appreciated.

Reply to
Peter
Loading thread data ...

Just buy a class d audio box.

Reply to
Sjouke Burry

Microcontroller with high frequency sinus modulated PWM

230V ac into a halfbridge converter, controlled by said PWM via IR2201 / equivalent gatedrivers

Feed that into a transformer if you need isolation, a coil if you do not (dangerous). In any case, you need a LC filter

Close the loop to the micro with a measurement of the output voltage

Reply to
Klaus Vestergaard Kragelund

For a highly inductive load a transformerless half-bridge might be the way to go you can get pushing 90% efficiency and under 5% THD running open-loop:

Reply to
bitrex

A multi-level sine converter avoids any transformer-whistling problem because the signal is baseband and there's little energy at high frequency, works great into inductive loads, and doesn't require a DAC just a clock and some logic, at the expense of higher un-filtered THD and requiring at least two isolated power supplies in the basic version; there are variations that use a single power supply and "flying capacitors" but I've never experimented with that type.

Reply to
bitrex

IIRC the Apollo spacecraft had an inverter that used some crazy arrangement of low-voltage square waves feeding a transformer with many primary windings and the input squares were of the correct amplitude and relative phase such that the transformer spit out a high-voltage stepped sine wave of sufficiently low THD from the secondary but this is probably only of historical interest :)

Reply to
bitrex

Then do you really need a "BIG inductor"?

Reply to
John S

Since you did not specify any efficiency requirements, a simple class-AB audio amplifier capable of delivering 24 V and 1 A should do. An audio amplifier nominally rated as 36 W into 16 ohms (or 72 W into

8 ohms or 144 W into 4 ohms) needing at least a +/-40 Vdc supply and the amplifier must be capable to dissipate 10-15 W .
Reply to
upsidedown

On a sunny day (Fri, 05 Feb 2021 16:55:24 +0000) it happened Peter wrote in :

So you need 48 * 2 * sqrt(2) = 135.764502 Vpp Supply voltage 150V? Assuming (..) you do not want an output transformer,

20W at 48Veff is 0.416667 Aeff and (20/48) * 2 * sqrt(2) = 1.178511 App

A SEPP output audo amp in class AB? Dual bridge configuration?

Or just use 2 more normal (2N3055 like) transistors and a transformer in pushpull. If high efficiency is a must

400 Hz in resonance? tuned transformer?

For 400 Hz you will need some iron in that core.

If your load is REALLY inductive (and a fixed L value) you could try to series tune it to 400 Hz and drive it from say some square wave output chip that only needs to be able to deliver the current:

C Amp --||--- L Load--- R load--- Rsense ---GND | | ---------------

Reply to
Jan Panteltje

Where did you get that 48 V from, the message header talks about 24 V, thus 68 Vpp.

80 V would be sufficient, allowing some Vbe/Vce drops.

In a half bridge, you just need just peak current, not peak-to-peak. Thus

(20/24) * sqrt(2) = 1.2 Apk

With a full bridge, a 40 Vdc supply would be enough. For this, even 4 x 2N3055 would be sufficient.

Reply to
upsidedown

I just did that!

formatting link

formatting link

We originally used a TI class-D amp chip, TPA3255. It worked as advertised, but a FADEC sometimes shorts all three phases of the incoming power, and the TI chips shut down and had to be reset. That annoyed the customer.

The TI eval board is nice.

So I had to redesign the amps with discrete mosfets, and add programmable current limiting, which we did in the main FPGA.

formatting link

formatting link

I have a bunch of the old amps if you want a couple.

John

--
John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.
 Click to see the full signature
Reply to
jlarkin

Anyone a fan of the classics? A 2N3055 in class C could push about 20 watts, yeah?

Turns ratio could obviously be whatever, at 400 Hz it could probably be an ac power transformer.

I doubt the THD will be a few percent, though, but it sure is simple

Reply to
bitrex

Why does it have to be solid-state at all? A motor/alternator solution is good for decades, and easily can handle higher wattages. Such a device gives a comforting growl when in operation, and runs its own cooling fans...

Reply to
whit3rd

You can buy a little amplifier based on the TDA7498E pretty cheap, ($75 or so) including case and an external inadequate 24V supply. They switch at around 300kHz. Ignore the amplfier specs and read the chip datasheet, of course.

Crack it open, change the mode to PBTL (mono) if necessary and jack the supply up to 36V with a standard switching supply (replacing any parts like e-caps that can't take the voltage). The open board ones have caps marked at 50V. Or use a booster based on LM2588 to change

24V to 36V.

If you're willing to wait a few weeks, open board from China is maybe $20 with no power supply.

Then provide a sine wave. You can get little modules with an AD DDS on them, so whack an Arduino Nano, a PIC chip or whatever on there to set up the frequency to 400Hz at power-up and you're pretty much done.

Or you might be able to get there with the stock power supply if you drive into a half the secondary of a centre-tapped 24V mains transformer and use the other half to double the voltage. Be sure to insulate the primary leads to avoid surprises.

I've got a little voice coil shaker that I've been wanting to make a driver for, might get one of those myself.

formatting link

Only 1.3 ohms so voltage won't be an issue.

--
Best regards,  
Spehro Pefhany
Reply to
Spehro Pefhany

The original post doesn't come up for me. One classic solution is Baxandall s Class-D oscillator.

formatting link

It is an old paper - 1959 - and a modern circuit would use MOSFET switches, as I discuss

formatting link
lator1.htm

24V RMS would be a 37.7V rail - say 40V.

The inductive load can become part of the tank circuit, so the centre-tappe d transformer doesn't have to cope with with the whole 0.83 A rms.

It needs a filter inductor with at least a couple of times the inductance o f the tank circuit, which is going to be bulky, and it would be hard to sto p it whistling. The distortion level would be in your ball-park.

I did work up a variant which replaced the inductor with a current mirror. It's only about 50% efficient, so your current mirror circuit would have to dissipate about as much heat as you dissipate in the resistance of your ta nk circuit. This shouldn't be a problem.

formatting link

It is a lower-power circuit than you want. At 400Hz you could use op amps t o drive a decent-sized MOSFET or two to deliver the current required to sus tain the oscillation. Distortion levels could be below the 0.1% level, and you could regulate transconductance of the voltage-to-current converter to get exactly the gain you needed to sustain the oscillation at exactly the a mplitude you needed, which is what I invented the circuit to do.

--
Bill Sloman, Sydney
Reply to
Bill Sloman

On a sunny day (Fri, 05 Feb 2021 22:41:55 +0200) it happened snipped-for-privacy@downunder.com wrote in :

Oops, copied wromg number to calculator, sorry!

Yep

Reply to
Jan Panteltje

I'd wonder whether you were looking for a mass-producible design... or a one-off for testing on-aircraft devices without the aircraft (?)

In the latter case, perhaps just a motor, speed control and an alternator pulled from an aircraft or "army surplus"?

Reply to
Don Y

Peter schrieb:

Synchro?

Regards

Reinhard

Reply to
Reinhard Zwirner

I don't think you have told us one important deciding factor...

What is the input supply ??

Reply to
boB

Thank you all for your inputs.

snipped-for-privacy@highlandsniptechnology.com wrote

Yes indeed; I am trying to generate ARINC407 X/Y/Z waveforms. 400Hz and with the phase relationship denoting the magnetic heading. This is old stuff; the current version of that spec is 1958! Despite that I have not found it online; one has to buy it...

The 400Hz generator is for the reference in the system, which needs to be able to deliver a fair bit of power, without itself dissipating more than a few watts otherwise the whole packge becomes too big.

The other two sinewaves, which get phase shifted, drive much lower power loads. I am still trying to work out how to actually do it. I think the reference is the INPUT to every box in the system and the two sinewaves (X/Y) have a phase relationship relative to the reference which carries the heading. In some systems Z is grounded. But that's another topic.

I will have a 12/14 or 24/28V supply available but that's just a suitable DC-DC converter to get whatever supply rails are needed. Probably +/-24V.

It does look like one of the Class D chips is the way to go...

Reply to
Peter

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.