Boosting a Frequency Generator Output

Hello Andrew,

Some of my thoughts for you.

Do you have the ground from the waveform generator connected to the ground for the op-amp? I would have thought all the grounds should be connected together, they are when I build up an amplifier to increase the output on my waveform generators. From your drawing it looks like the ground for the waveform generator is floating.

I am not sure of the 1 kOhm resistor from the op-amp output to the ground, is this providing too much load for the op-amp?

I usually put a resistor in series from the waveform generator into the non-inverting input of the amplifier, to limit current, but I don't think this stops the circuit functioning basically. Same with de-coupling capacitors on the supply rails.

Have you checked that the PSU and waveform generators have isolated outputs and you aren't creating a circulating current somewhere that is causing your supply to current limit and drop the voltage down to 0V? Not sure why it would then want to go up to 32V though.

At the moment your amplifier is set up with a 2 to 1 ratio, and putting in a 20Vpp signal. With the supply only being 30Vpp, you would be overdriving the amplifier. What happens if you decrease the input signal down to 10Vpp or even 5Vpp. You would then be within the limits of the power supply rails and output of the op-amp.

Are you sure the op-amp is good?

It is the same method as I have used in the past, so it should work for you.

Kind regards

I agree with Donald, what you trying to achieve is outside the capabilities of a 741.  From your description I think you’ve damaged the opamp.

You could try a higher voltage opamp like a TI OPA990, but you probably don’t have one of those at hand.

Thanks Donald,

Do you have the ground from the waveform generator connected to the ground for the op-amp? I would have thought all the grounds should be connected together, they are when I build up an amplifier to increase the output on my waveform generators. From your drawing it looks like the ground for the waveform generator is floating.

No, I have it as you see - not connected.  I didn't think it right to link it's ground to what I have as GNDREF but I can give that a go in the morning if it isn't going to damage anything.  You would connect what I have as Pin1 on the generator to GNDREF?

Have you checked that the PSU and waveform generators have isolated outputs and you aren't creating a circulating current somewhere that is causing your supply to current limit and drop the voltage down to 0V? Not sure why it would then want to go up to 32V though.

It only does that if I connect the non-inverting input and output to the ground of the generator rather than GNDREF.

At the moment your amplifier is set up with a 2 to 1 ratio, and putting in a 20Vpp signal. With the supply only being 30Vpp, you would be overdriving the amplifier. What happens if you decrease the input signal down to 10Vpp or even 5Vpp. You would then be within the limits of the power supply rails and output of the op-amp.

I don't understand what you mean Donald.  The output of the generator is 20Vpp with a 0V dc offset so +10V/-10V hence the 7Vac  (actually Vac = Vpp/2 * 0.7).  Isn't this within the rail limits of +15V/-15V?  I could set the inputs to +-18V which may prevent top-end clipping but I don't think that's what you are referring to.

I am not sure of the 1 kOhm resistor from the op-amp output to the ground, is this providing too much load for the op-amp?

Yes it is!  Removing it will (temporarily) give a x2 output.  But whilst the input stays stable, the output slowly and inexorably drops towards 0V as if the DMM is pulling it down?  At least it's a step forward.

It is the same method as I have used in the past, so it should work for you.

That's good to know, at least I'm not a complete idiot

It would seem then that the load was wrong but a test circuit in the datasheet includes a 2K resistor on the output.  Any more ideas?

It seems ok at the moment - having removed the load - but I wouldn't rule it out.  I'm just not understanding why I'm putting in too high a voltage: +7Vac and I can't measure any DC voltage at the input.  It seems there's a capacitive effect on the output.  If I disconnect the DMM from the output for a few seconds, then reconnect, the output voltage jumps back up to x2 and then starts its decline.

I also have a similar question. but I want to end up with 115vac @ 400hz. I was thinking about using good old tubes push-pull but I would like to do this without tubes. Anybody? Unfortunately, I also need some current. 5 - 10 amps would be really nice.

~~Cris

You definitely need to tie the generator ground to the op-amp ground - otherwise the two circuits are not referenced together (the input is floating).  Do you move your DMM ground when making measurements?

Because you have 2X gain - if you put in +/-10V - the output will swing into the rails and the op-amp is not designed to do that.  It shouldn't damage it (unless it latches up), but if your negative supply really went to 32V that would have caused problems.  I think you're having measurement issues because the grounds are not connected.

I guess I should clarify - by generator ground I mean the output negative reference from the generator should tie to the op-amp gndref.

phoenixcomm

You're aiming for 115 * 5 = 575W minimum with a hope for double that.

When I needed to do this last I bought a no name mega disco audio amplifier for about £200.

It can manage about 1500W out in bridge mode - I couldn't get the parts for the cost of the ready made thing.

I used a standard 750W mains toroidal transformer to get from amplifier to (in my case) 240V.

If you want to make it yourself (and cheap) the best bet is to start off from which of the expensive bits you have,

like output transformer, power supply (if it will run from mains), box, heat sinks.

There are loads of MOSFET power amp designs on the web and plenty of very cheap Chinese sources for them,

so again you can make or buy.

Some of the designs on the web are awful, some look OK, especially for 400Hz, the cheapest use all N channel MOSFETs.

This is a typical offering:

https://www.aliexpress.com/item/1005001874598103.html?spm=a2g0o.search0302.0.0.20461b7056Lye8&algo_pvid=f86a1e08-1205-40…

No idea if it's any good !

If you find an interesting design you should definitely simulate it before buying any bits.

This ready made amp from CPC would do nicely but might be a bit pricey:

https://cpc.farnell.com/crown/xli2500/amplifier-2-x-750w-xli2500/dp/DP33156

MK

Hi Andrew, the 741 (and some other old op-amps like TL0x1 series) behave weirdly enough outside their (often quite limited) specifications personally I try not to use them when experimenting, although they have their advantages like ultra-low cost.

What frequency range do you want to work at?

Another option (don't know if it is relevant though, depends on if you can find the part or not), is to find a transformer. It will solve any ground issues too, since the output will then be floating.

If you're in the audio frequency range then if you can find a audio transformer (vintage equipment! maybe something from some NMOC stash?) that could be usable.

Otherwise (say) R/Z1606 will work for about 500Hz upward in the audio range. It is 1:1, but with a centre tap secondary, so the secondary can be used as the primary, using half of the winding, so that it becomes 1:2 ratio.

You won't be able to load it very much though. I think an audio transformer can be a great problem-solver for working with signal generators.

Donald, Ralph: yes you were both right: 1K was too much of a load and it needed the generator -ve connecting to GNDREF.  I'd wrongly assumed that would cause problems given that they belonged to two different voltage sources.   By upping the PSU to +17V and -17V I can get 12.4Vac out of the op amp which is enough for my purposes.

Shabaz: I'm not doing anything fancy, I literally want to be able to generate a 12Vac, 50Hz output and then reduce that voltage down and change the frequency, which is why I went with what I had to hand.  I do have some SMD op amps but those are +5V/-5V input so I was keeping my fingers crossed for the venerable UA741!  I keep getting warned off them but they're little troopers and always pull through in the end   I want to test the specification of something that works at 12Vac and 50Hz.

Thanks guys

michaelkellett I would hate to use a nice piece of kit like the Crown amp.