Creation of positive AND negative current AC

[treegram]

I need to create two identical value AC current flows. One moving from a more positive terminal to less positive one, and another moving from a less negative terminal to more negative one. What is the most straightforward way of achieving this?

For example, do I need two separate single ended amps, one powered by, say, +18 and +9 volts and the other -9 and -18 volts? Or is there a way to do it with one amp and a special arrangement of transformers or caps at the output?

Thank you for any advice.

 

[Nigel Goodwin]

I suggest you tell us exactly what you're trying to do, rather than how you think it needs to be done.

 

[treegram]

Thank you for your reply. I would like two electrically isolated replicas of a 200mVpp 10Hz sinewave, one with a swing between +9 to +18 volts, and the other with a swing between -9 and -18 volts. Both would have the same amplitude and no common ground connection. I would prefer to power the circuit with batteries.

 

[KeepItSimpleStupid]

Is the circuit making the sine wave or is it an external input?

What is the required output curent?

 

[treegram]

The input signal will be externally generated. I think AC coupled via caps to the op amp inputs to eliminate offset issues. Preferred output current is around 50mA at maximum practical swing. Thanks for your interest.

 

[KeepItSimpleStupid]

It's a tough set of specs:
-18 out
battery powered
isolated
50 mA (assume peak?)
Hopefully the output Z of your signal source is low.

I might start here: https://www.electro-tech-online.com/custompdfs/2011/06/AD202_204.pdf

With the AD204 and the clock generator. This would require a 15 V supply and would produce a swing of +-5V.
You can invert one of the signals easily.
Add an offset with a reference IC (see fig 7) for summing. I didn't look for a reference IC and I didn't figure out how much voltage I have to add.

So, at this point, you can get an asymetric signal within the +-5V output swing with a 15 V power supply and supposedly the ripple is acceptable.

Now we have other issues, the -18 V swing and the battery power.

Just "thinking out loud" to get some idea of how to proceed.

 

[KeepItSimpleStupid]

With the use of a PA75 https://www.electro-tech-online.com/custompdfs/2011/06/PA75U_G.pdf and +-20V supplies, you can pull off what you need. Note one of the amps would not be used and it's unity gain anyway.

Now you have another problem:

You need a DC-DC converter to supply 15 V and two isolated DC-DC converters to supply +-20V. +-20V isn't standard. Now you have to select a battery (12V?) and determone what the possible running time has to be.

 

[KeepItSimpleStupid]

Although probably way overpriced and very power hungry even in the quiesent state, it could create an isolated +-20 V at 33 WATTS from 10-18 volts input. At 18* 50 mA, you only have about a Watt, so clearly you need a better solution.

12 to 15V is should be easy.

Line powered would be easier to do.

 

[KeepItSimpleStupid]

It might also be worth a shot using transformer coupling using an audio transformer. This one: http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=MT4139-ND says it's good at 300 Hz, but ti might work at 10 Hz. Two such transformers may be able to do the inversion of your input and the isolation. You might be able to find one that does some amplification as well.

A better possibility for the output stage is to use the PA75 in Bridge mode with say +-15 V supplies. One amp is set for a gain of -1 and the other is fixed at +1. That's a better choice than trying to get a +-20V supply.

You would do the summing and offset with another OP amp stage.

 

[Nigel Goodwin]

Thank you for your reply. I would like two electrically isolated replicas of a 200mVpp 10Hz sinewave, one with a swing between +9 to +18 volts, and the other with a swing between -9 and -18 volts. Both would have the same amplitude and no common ground connection. I would prefer to power the circuit with batteries.

You're still not really answering my question, what are you actually trying to do?.

Generally, when people post here with obscure requests like this, it's because they are going about their solution the wrong way.

The thread then goes on and on, achieving nothing, and eventually the OP explains what he's trying to do - to find the original question is completely wrong. Often it transpires it's some ludicrous 'free energy' scheme, which is why it's deliberately such a vague question.

 

[treegram]

Thanks to those who have responded so far. I am not sure which part of my question is unclear. But I will try again.

I have a function generator outputting a 10Hz sinewave at 200ma. (It can adjust up to 1Vpp if required.) From this, I need to produce two identical signals of greater magnitude. One is to swing between +18V and +9V. IOW the +9V is the virtual ground. The second is to swing between -9V and -18V. The latter is virtual ground.

The unconventional part, which may be causing confusion, is that there is no common ground or voltage reference between the two output signals, and neither ever crosses zero. This is what I referred to, perhaps incorrectly, as "isolation". Current required from each at output is about 50ma.

Rather than use specialized parts, I am thinking one possible way this can be done is with three 9V batteries and tow fairly standard op amps. However, I am looking for details on any design that would fulfill the above criteria.

I don't know how to be anymore explicit, but would be happy provide further information so long as I know exactly what is needed.

 

[KeepItSimpleStupid]

Would you happen to have 2 function generators? What's the model? Would like to investigate external isolated triggering. It would depend on how far you can cloat the ground terminal on the generator.

Unfortunately 50 mA puts an OP amp into the "HIGH POWER" mode and a swing usually bigger than 15 V p-p puts it into a "HIGH VOLTAGE" mode.

10 Hz is not typically an Audio frequency, but some audio amps can operate in Bridge mode as well. Transformers may work well there too.

Adding an ultrasimple current/voltage boost stage; 2 diodes. 6 resistors and 2 transistors might also be possible.

Any idea on low long this thing, if battery operated, has to work for?

 

[ronv]

+/-

I'm not sure I understand your virtual ground statement, but is this what you are after?
You would need to buffer the outputs with a rail to rail op amp to get your 50ma.

PS. You could get rid of the 3rd battery if you use a rail to rail op amp instead of the one I chose.

 

[Nigel Goodwin]

Thanks to those who have responded so far. I am not sure which part of my question is unclear. But I will try again.

I have a function generator outputting a 10Hz sinewave at 200ma. (It can adjust up to 1Vpp if required.) From this, I need to produce two identical signals of greater magnitude. One is to swing between +18V and +9V. IOW the +9V is the virtual ground. The second is to swing between -9V and -18V. The latter is virtual ground.

The unconventional part, which may be causing confusion, is that there is no common ground or voltage reference between the two output signals, and neither ever crosses zero. This is what I referred to, perhaps incorrectly, as "isolation". Current required from each at output is about 50ma.

Rather than use specialized parts, I am thinking one possible way this can be done is with three 9V batteries and tow fairly standard op amps. However, I am looking for details on any design that would fulfill the above criteria.

I don't know how to be anymore explicit, but would be happy provide further information so long as I know exactly what is needed.

You're still not giving any reason for why you want to do this - only asking about your solution - so why do you want to do this, and why does it have to be so strange?.

 

[treegram]

Many thanks Ron for your interest, and to the otheres who responded. The circuit you kindly provided is almost exactly the type of thing I had in mind.

Just two questions please to finalize. 1) Why are the two output sinewaves out of phase? What change to the circuit is required to have them in phase?

2) What I meant by the "virtual" grounds at 9 and -18 volts is this. The top amp would be powered by 18 and 9V, and the bottom one by -9 and -18V. You can see in both cases the potential is 9V, but the two supplies do not have a common. In this case, I believe the input signal would need to be AC coupled. Can you please advise how to change the arrangement of the batteries and components on your schematic accordingly? Even if LTSpice may not run the sim.

You may say the end result is the same, but trying this variation is a way to improve my abilities with op amps. Perhaps other members will also find it of interest.

 

[KeepItSimpleStupid]

That's not the picture I got at all in terms of what you wanted because the outputs are not swinging from -9 and -18 volts.

1. One Op amp is a non-inverting configuration and the other is an inverting configuration.

2. One potential problem is that the offset is not provided by a reference and would vary as the battery is depleted. Note that his circuit is not isolated either. You can't get an 18 V or -18 V swing with an 18 V power supply. And as Ronv mentioned, it doesn't have enough current output. A transformer could do the isolation and/or inversion, but now you would need more batteries.

I thought you wanted amp A to be at -18 and AMP B to be at 9 and when B is at -18, A would be at 9V. That circuit doesn't satisfy that, if that is what you wanted.

You basically can place the offset at the input or the output by summing. In order for a summer to work properly, it has to be driven by a low Z source. I hope Ronv will confirm my concerns.

 

[ronv]

Hard to do. Here is the picture. Scope probe a is on output a. Scope probe b is on output b. Where is the scope ground?

 

[treegram]

RonV, thanks again for your reply. I understand what you mean with regard to the LTSpice scope. But the circuit could still be built without a common earth, (+18 & +9 and -9 & -18V rails as described). The input signal is AC coupled so the offset voltage of the op amps' virtual ground presumably does not affect it.

With regard to the circuit you previously provided, can you please advise what is the purpose of the 10uF cap between the non-inverting input of the bottom op amp and the inout signal to the top one?

Also, why are the two output signals in the sim anti-phase?

I am getting ready to build one and I would like to understand fully the principles involved.

Many thanks.

 

[ronv]

Sure the circuit can be built without the common ground. But without a reference anywhere you can't define the voltage levels. If you want to use the virtual grounds as a reference that is ok, but then you need to define the output with respect to those reference(s). I used the capacitors to couple the single source to the two circuits without tying the two circuits together. (DC wise) The circuits are out of phase because one inverts and the other does not. This can be fixed with another inverter or maybe a redesign of the current circuit. I can break the two circuits apart and have spice happy with it but like I said in the prior post you need to tell me where to put the scope probes, Spice or real, and what voltage you expect to see because there will be no relationship between the two circuits at that point.
Wait. I think I get it. If the circuits are completely separate I can define the virtual ground on each of them separately? So you really don't care about one circuit in relationship to the other - only to itself?
What do these circuit drive?

 

[treegram]

Yes, as RonV points out, the design can be dealt with as two separate amps. One with a +18 and +9 supply, and the other with -9 and - 18. So the virtual grounds are +9 and -18 respectively, and they are obviously not connected to each other.

Since Ron's latest post I have built up the circuit he kindly provided. I had only 2 NE5532's, not the LT1013's specified. Unfortunately, the output waveforms were clipped, I suspect due to connecting the ground lead of my function generator to the +9 and -18 virtual grounds respectively. Although it does not seem to affect the amps' input signals.

Perhaps someone would be so kind as to set me straight on this. I am beginning to think I need a proper signal ground at 0 volts. I have a attached a schematic diagram including this change in the hope of attracting some further comments. I have also uploaded RonV's original LTSpice for comparison.

BTW using 6 batteries may seem odd, but this circuit is a proof-of-concept thing in relation to my other studies. Thanks for your understanding.