Amplifier, continuously variable from inverting to non-inverting?

[Taylor]

Is there a type of amplifier that allows to have a pot control gain, but which is inverting on the lower half of the pot's range, and non-inverting at the top end of the range? I can think of ways to kludge this from an inverting and a non-inverting amp being crossfaded, but I imagined that perhaps there was already a clever design block that does this.

[Quackzed]

http://en.wikipedia.org/wiki/Phase_splitter

[Taylor]

Hmm, well I knew about the phase splitter, but I was hoping to be able to be able to control gain continuously from a large positive gain to a large negative gain. I'm playing around with something like Doepfer's waveshaper:

http://www.doepfer.de/a136.htm

I guess I can pan between the 2 outputs, but I'm not sure if the results will be quite the same as if controlling gain (particularly w.r.t. clipping). But that's a good place to start, thanks.

[Quackzed]

I just thought of a phase splitter with a 'mix' pot between the 2 out of phase signals, and you could always amp after the splitter with a dual opamp at 12v or something...

[CynicalMan]

D'oh, I remember someone posting exactly this (earthtonesaudio maybe?), but I can't remember where. Oh well, (opens up LTSpice and starts playing around with inverters).





I post if I find anything.  

[Quackzed]

Here it is in falstads... fwiw...

$ 1 5.0E-6 18.278915558614752 64 10.0 53
w 144 48 240 48 0
r 144 48 144 144 0 150000.0
r 144 144 144 256 0 56000.0
t 144 144 240 144 0 1 -4.754346180795753 0.5919052912450837 100.0
r 240 160 240 256 0 4700.0
w 144 256 240 256 0
c 96 144 144 144 0 3.0E-6 -5.077385625359503
R 96 144 48 144 0 1 40.0 3.0 0.0 0.0 0.5
g 144 256 144 288 0
R 144 48 48 48 0 0 40.0 20.0 0.0 0.0 0.5
r 240 48 240 128 0 4700.0
c 240 160 304 160 0 3.0E-6 4.450270897305341
r 304 160 304 256 0 10000.0
w 240 256 304 256 0
c 240 128 336 128 0 3.0E-6 15.59379233607005
r 336 128 336 256 0 10000.0
w 304 256 336 256 0
O 400 144 464 144 0
w 304 160 384 176 0
w 336 128 384 112 0
174 384 112 384 176 0 100000.0 0.5198 Resistance
o 7 64 0 35 5.0 1.953125E-4 0 -1
o 17 64 0 34 1.0229345649675443 9.765625000000001E-155 1 -1

go here,http://www.falstad.com/circuit/e-phasesplit.html
and right-click the app,go to file 'import' and copy / paste the above code to see it in action.

[Taylor]

Yep, I used the Falstad sim to play around with it earlier and got some cool stuff happening. Thanks. Will post a schem if I get a full wave folder circuit happening.

[Nasse]

Think I have seen such circuit, 70´s op amp application National semi "switch hitter" or something. Did quick search but did not find it yet. geofex might have it in some form...

[Taylor]

Think I have seen such circuit, 70´s op amp application National semi "switch hitter" or something. Did quick search but did not find it yet. geofex might have it in some form...

Awesome, the "switch hitter" name gave me enough to find it (fig. 1.18):

http://users.ece.gatech.edu/~mleach/ece3050/sp04/OpAmps01.pdf

[Quackzed]

I went through a period where i was all about rectifying signals  ... definately some strange attack character from a few of my octave 'experiments'...

[Lurco]

http://www.diystompboxes.com/smfforum/index.php?topic=77498.0

[Taylor]

http://www.diystompboxes.com/smfforum/index.php?topic=77498.0

Yep, that's the circuit I linked to above. The funny thing is that I posted in that thread but forgot all about it  - I must have remembered it vaguely enough to think that such a circuit existed.

[Nasse]

1/(2Pi*(R/2)*C)

there is a formula for input dc cap f0 if R= input resistor, feedback resistor and pot are all same value

have seen variations where pot action replaced with switch or jfet

amazing amount of knowledge we have forgotten...

[slacker]

The switch hitter thing is a great little circuit, I've used a few times, but it's only got a maximum gain of +-1. If you want something that actually lets you go from "large positive gain to a large negative gain" have a look at these, shamelessly pimping my own creation

http://www.diystompboxes.com/smfforum/index.php?topic=72404.0
http://www.diystompboxes.com/smfforum/index.php?topic=84810.0

Whip the diodes out and adjust the feedback resistor to taste and you've got a positive/negative booster. Possibly won't get 0 in the the middle like the switch hitter.

[earthtonesaudio]

Another in the spirit of shameless self-promotion:



Note the phase splitter in front makes it so the overall output is always non-inverting.

[R.G.]

The switch hitter is an old circuit. I used it in a home-made magnetometer to concoct an electronic analog compass for one of the early sailboat instrumentation packages back in the 80s. It comes as either the switched version or the continuously variable version.

I think Taylor's needs are met by a two-opamp circuit, one circuit doing the switch-hitter, which gives you continuously variable gain from +1 through 0 to -1, and a second opamp to set the amplification factor from 1 to "high". The trick is that you use the gain pot for the second amplifier in a folded mode. You tie the CW and CCW lugs together and use those two lugs and the wiper as the terminals of a resistor. So this is a resistor which is maximum in the middle and goes to nearly zero at either fully CW or CCW. Then you use a fixed resistor as the feedback resistor for the gain-opamp and the folded pot as the resistor to bias on the inverting input. If you set this up correctly, the gain will be minimum at exactly the middle of travel and maximum at either end.

Then you make both pots be sections of a two-ganged unit. In the middle, the gain is zero. As you turn it either way, the gain increases from 0 to 1 on the first opamp, and minimum to maximum on the second opamp. The whole thing requires a dual opamp, a dual ganged pot, and the usual assortment of resistors, caps, etc. to harness the gain stages, couple signal in and out, and so on.

[earthtonesaudio]

A fixed-gain boost in front of the switch hitter will give you (gain)*(+1) or (gain)*(-1) at either end of the switch hitter pot's rotation, and a gain of (ideally**) zero in the middle.

This approach is inherently more noisy than simultaneously adjusting the gain as R.G. suggests, but simpler.


**Also, the gain at "zero" is actually determined by the common mode rejection of the op-amp, less any imbalances in the rest of the circuit.

[soggybag]

Here it is in falstads... fwiw...

$ 1 5.0E-6 18.278915558614752 64 10.0 53
w 144 48 240 48 0
r 144 48 144 144 0 150000.0
r 144 144 144 256 0 56000.0
t 144 144 240 144 0 1 -4.754346180795753 0.5919052912450837 100.0
r 240 160 240 256 0 4700.0
w 144 256 240 256 0
c 96 144 144 144 0 3.0E-6 -5.077385625359503
R 96 144 48 144 0 1 40.0 3.0 0.0 0.0 0.5
g 144 256 144 288 0
R 144 48 48 48 0 0 40.0 20.0 0.0 0.0 0.5
r 240 48 240 128 0 4700.0
c 240 160 304 160 0 3.0E-6 4.450270897305341
r 304 160 304 256 0 10000.0
w 240 256 304 256 0
c 240 128 336 128 0 3.0E-6 15.59379233607005
r 336 128 336 256 0 10000.0
w 304 256 336 256 0
O 400 144 464 144 0
w 304 160 384 176 0
w 336 128 384 112 0
174 384 112 384 176 0 100000.0 0.5198 Resistance
o 7 64 0 35 5.0 1.953125E-4 0 -1
o 17 64 0 34 1.0229345649675443 9.765625000000001E-155 1 -1

go here,http://www.falstad.com/circuit/e-phasesplit.html
and right-click the app,go to file 'import' and copy / paste the above code to see it in action.

How do you adjust the pot value? Right clicking and editing the part seems to only adjust the max range of the pot?

As I was playing with this and immediately had to add a diode after the cap from each the collector and Emitter. Figuring I would get a Green Ringer type rectifier. I notice the top output didn't change and the lower output showed a small bump rather than a full half wave.

I figure the lower output is due to the signal not being large enough to over come the diode drop. But the upper output has me stumped.

[Taylor]

How do you adjust the pot value? Right clicking and editing the part seems to only adjust the max range of the pot?

It's over on the right side of the screen, labeled resistance, under current speed, etc.

[Taylor]

The switch hitter is an old circuit. I used it in a home-made magnetometer to concoct an electronic analog compass for one of the early sailboat instrumentation packages back in the 80s. It comes as either the switched version or the continuously variable version.

I think Taylor's needs are met by a two-opamp circuit, one circuit doing the switch-hitter, which gives you continuously variable gain from +1 through 0 to -1, and a second opamp to set the amplification factor from 1 to "high". The trick is that you use the gain pot for the second amplifier in a folded mode. You tie the CW and CCW lugs together and use those two lugs and the wiper as the terminals of a resistor. So this is a resistor which is maximum in the middle and goes to nearly zero at either fully CW or CCW. Then you use a fixed resistor as the feedback resistor for the gain-opamp and the folded pot as the resistor to bias on the inverting input. If you set this up correctly, the gain will be minimum at exactly the middle of travel and maximum at either end.

Then you make both pots be sections of a two-ganged unit. In the middle, the gain is zero. As you turn it either way, the gain increases from 0 to 1 on the first opamp, and minimum to maximum on the second opamp. The whole thing requires a dual opamp, a dual ganged pot, and the usual assortment of resistors, caps, etc. to harness the gain stages, couple signal in and out, and so on.

I think I'll probably go for a fixed gain in front of the switch hitter as earthtones mentioned, just because I plan to put this circuit up here for others to build when it's done, and I know that multi-gang pots bum people out (since I'm operating separately on each half of the waveform, we'd need 2 dual-gang pots).

But the "folded" pot is so clever but also so simple, that I can't believe I haven't seen that before. That's something I've got to use at some point.