Amplifiers suitable for low load or no load operation

[spacevector]

Hello,

I have a non-audio application that hopefully I can get some pointers on here. I am in need of a couple of 480V test signals for some microgrid integration testing.

The signal generator we have outputs +/-10V analog signals with up to 15mA output current.

My idea is to buffer this with an audio amplifier and then use a 24V:480V control transformer to get the voltage up to 480V.

The load on the amplifier is going to be minimal - likely <0.5W and will only be operating around 60 +/- 1Hz.

I will need to make 24V rms on the output, so an amplifier that does not clip above 72W (87W for +10% voltage) into 8 Ohms should suffice. And the amplifier should work stable for no load or very low load operation.

I think I should use a class AB amplifier to not have any switching ripple on the output. We can make some PWM filters but would like to keep simple if possible.

Ideally the amplifier should have low phase shift for 60Hz signals but as long as both channels have the same phase shift, I should be OK.

Not very picky on THD. Grid voltage typically has 2% harmonics (-34dB) already so an amplifier that has better than -54dB THD+N at low load should suffice.

Budget is under $200 preferably but we can spend up to $500 if needed.

Any good options you can recommend?

 

[MakeMineVinyl]

Hello,

I have a non-audio application that hopefully I can get some pointers on here. I am in need of a couple of 480V test signals for some microgrid integration testing.

The signal generator we have outputs +/-10V analog signals with up to 15mA output current.

My idea is to buffer this with an audio amplifier and then use a 24V:480V control transformer to get the voltage up to 480V.

The load on the amplifier is going to be minimal - likely <0.5W and will only be operating around 60 +/- 1Hz.

I will need to make 24V rms on the output, so an amplifier that does not clip above 72W (87W for +10% voltage) into 8 Ohms should suffice. And the amplifier should work stable for no load or very low load operation.

I think I should use a class AB amplifier to not have any switching ripple on the output. We can make some PWM filters but would like to keep simple if possible.

Ideally the amplifier should have low phase shift for 60Hz signals but as long as both channels have the same phase shift, I should be OK.

Not very picky on THD. Grid voltage typically has 2% harmonics (-34dB) already so an amplifier that has better than -54dB THD+N at low load should suffice.

Budget is under $200 preferably but we can spend up to $500 if needed.

Any good options you can recommend?

For under $200, a cheap class D amplifier seems like your only practical option. Your control transformer would filter any ultrasonic hash from the amplifier, I'd imagine.

 

[spacevector]

For under $200, a cheap class D amplifier seems like your only practical option. Your control transformer would filter any ultrasonic hash from the amplifier, I'd imagine.

Thanks! I am looking at the Review Index now and looks like the Crown XLi 800 may be a good candidate. Class A/B with 185W at 8 Ohms (~39V) before clipping.

Any thoughts on operation at high output voltage (24V) and very low load? I am worried about instability and protection mechanisms kicking in.

 

[MakeMineVinyl]

Thanks! I am looking at the Review Index now and looks like the Crown XLi 800 may be a good candidate. Class A/B with 185W at 8 Ohms (~39V) before clipping.

Any thoughts on operation at high output voltage (24V) and very low load? I am worried about instability and protection mechanisms kicking in.

That amp would probably work fine if you can stretch your budget a bit. A class A/B amplifier doesn't really care what it is loaded with as long as it is above its minimum load impedance which is generally 2-4 ohms. The amp doesn't care if its lightly loaded or not loaded at all. While I don't know the specifics of this amplifier's protection, the amps we make usually have; high current protection (too much current through the output stage), HF oscillation detection (if the amplifier goes into screaming oscillation or is being asked to pass an out of audio band signal at a high enough level), and DC protection (shuts the amplifier down if there is a large DC offset on the output of the amplifier). The Crown might be somewhat different, but this what we do.

 

[LTig]

I'm far from being an expert in tube amps which use output transformers but the "rule" says not to use them without a load because the transformer becomes inductive which may lead to overvoltages in the last stage. With such a low load at the transformer (less than 0.5 W) you may want to check with an expert whether this is safe for a solid state amp.

 

[MakeMineVinyl]

I'm far from being an expert in tube amps which use output transformers but the "rule" says not to use them without a load because the transformer becomes inductive which may lead to overvoltages in the last stage. With such a low load at the transformer (less than 0.5 W) you may want to check with an expert whether this is safe for a solid state amp.

A solid state amp doesn't care if its loaded.

 

[DonH56]

Most any SS amp should be fine.

Most if not all tube amplifiers may self-destruct.

 

[spacevector]

That amp would probably work fine if you can stretch your budget a bit. A class A/B amplifier doesn't really care what it is loaded with as long as it is above its minimum load impedance which is generally 2-4 ohms. The amp doesn't care if its lightly loaded or not loaded at all. While I don't know the specifics of this amplifier's protection, the amps we make usually have; high current protection (too much current through the output stage), HF oscillation detection (if the amplifier goes into screaming oscillation or is being asked to pass an out of audio band signal at a high enough level), and DC protection (shuts the amplifier down if there is a large DC offset on the output of the amplifier). The Crown might be somewhat different, but this what we do.

Thanks. Going to order this to test out. Is it easier for an unloaded amp to start oscillating? I imagine the nominal load provides damping in the closed loop feedback system?

 

[DonH56]

Thanks! I am looking at the Review Index now and looks like the Crown XLi 800 may be a good candidate. Class A/B with 185W at 8 Ohms (~39V) before clipping.

Any thoughts on operation at high output voltage (24V) and very low load? I am worried about instability and protection mechanisms kicking in.

By "very low load" do you mean very high load impedance, or very low? Low load impedance could be an issue but it seems like you mean a high load impedance so the output is lightly loaded.

Some early class D designs (and A, AB, etc. for that matter) would oscillate when unloaded, but AFAIK all current ones have solved that problem.

@MakeMineVinyl works for a major amplifier design company so listen to him...

 

[spacevector]

By "very low load" do you mean very high load impedance, or very low? Low load impedance could be an issue but it seems like you mean a high load impedance so the output is lightly loaded.

Some early class D designs (and A, AB, etc. for that matter) would oscillate when unloaded, but AFAIK all current ones have solved that problem.

@MakeMineVinyl works for a major amplifier design company so listen to him...

Low load = high load impedance.

Thanks for the input all. I am going with the Crown amp. Worst case, I will have to buy dummy loads to help stabilize but hoping it doesn't come to that!

 

[DonH56]

High impedance should not be an issue. Low impedance and large phase angles can work with the output stage (impedance) to modulate the feedback and make it more likely to go unstable. High impedance does not do that.

The only amps that would go unstable with no load are typically those that have low feedback and high output impedance such that the load is an integral part of the feedback compensation.

A pro amp that goes unstable with no load would be highly unusual.

 

[MakeMineVinyl]

Thanks. Going to order this to test out. Is it easier for an unloaded amp to start oscillating? I imagine the nominal load provides damping in the closed loop feedback system?

If it will make you feel better you can try loading the output with 100 ohms or something like that. But the amp shouldn't oscillate regardless.

 

[DonH56]

If you have an oscilloscope with a MHz or ten of bandwidth you can watch for oscillations. I'd be shocked if you see any.

 

[Head_Unit]

Low load = high load impedance.

That in itself is fine for the amp, but some turn-on/turn-off transients into a massive stepup transformer could be problematic. Having once shut off an entire lab's breakers running a square wave into a large transformer, ramp volume way down for turnon/turnoff Mmmm I'd probably put like 100 ohms in parallel across the outputs as a...I don't even know how to describe it. Basically so the amp is not loaded only by a giant inductor.

 

[Head_Unit]

That in itself is fine for the amp, but some turn-on/turn-off transients into a massive stepup transformer could be problematic. Having once shut off an entire lab's breakers running a square wave into a large transformer, ramp volume way down for turnon/turnoff Mmmm I'd instinctively probably put like 100 ohms in parallel across the outputs as a...I don't even know how to describe it. Basically so the amp is not loaded only by a giant inductor.

 

[spacevector]

For those interested, these amplifiers (Crown XLi 800) worked out well for our purposes. First round of testing has been concluded and these amps didn't miss a beat.

Channel 1 (yellow) is output from the simulator
Channel 2 (pink) is output from the amplifier
Channel 3 (blue) is output of step up transformer (24V:480V)

We did not have to use a dummy load in addition to the controls transformer.

The gain knobs in the amplifier have discrete steps so fine adjustments are not possible. We used some fine tuning in the simulator to get our 480V nominal output.

Interestingly, the amplifier has a phase lead of about 0.5ms at 60Hz (~11 degrees) or may be this is 16.1ms (349 degrees) lag. Good thing the channels (LR across 2 amplifiers) had the same phase error so it didn't affect the testing. Also shown above is the transient response where the simulator output is cut in half at the middle of the waveform.

Thanks to all who provided input!