different opamp gains based on sign?

I've got a differential signal, however, the negatives are small and intere sting while the positives are large and less interesting (but still importa nt).

Is there an amplifier topology that allows me to set a larger gain for nega tive signals and small gains for positive signals?

Need to be able to measure the 'positive' signals which can be as high as 8

0mV, but I am more interested in the 'negative' signals which are over an o rder of magnitude smaller (typically -2mV at most). Basically, gains that allow me to see the positive signal leave the negatives in the noise, and i f I crank up the gain I see the negatives clearly but the positives are cli pped.

I currently handle this with lots of oversampling on the ADC, then in softw are apply gain/scaling so the charts are readable...works just ok (and requ ires a 24-bit ADC I'd like to avoid in the future and stick with 16-bit)... I just can't help but think there is a hardware solution I am missing to ge t better resolution on the negatives.

I realize I'm using 'positive' and 'negative' here arbitrarily, as it is a differential signal like a wheatstone bridge.

Any suggestions?

Reply to
DemonicTubes
Loading thread data ...

It's easy to do if you don't need fast (us or faster) transistions between the pos and neg gains.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Sure thing. One approach is to use inverting and noninverting RRIO op amps. One inverting, one noninverting, subtract the outputs. The one you don't want will rail, so you can set the gains independently. Not every RRIO op amp will work well like this.

Another approach is to use a comparator and analogue switch in the feedback network. That gets slightly ugly if the polarity changes a lot.

A third approach is to use two feedback networks on a single inverting amp, with diodes in series. An instrumentation amp looking at the difference between the two diode-resistor nodes will do what you want--one diode is conducting, so that side gives you the appropriate gain. The other diode isn't conducting, so that gives you the summing junction voltage, which is near ground. (I don't have time for ASCII art at the moment, sorry.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

Well, sure; a PNP transistor, with a little pullup on the emitter, does that.. But, you want this for DIFFERENTIAL output? The familiar tanh response of a differential pair lets you do it by DC biasing near saturation (i.e. imbalancing the input transistors).

Reply to
whit3rd

to:

resting while the positives are large and less interesting (but still impor tant).

gative signals and small gains for positive signals?

80mV, but I am more interested in the 'negative' signals which are over an order of magnitude smaller (typically -2mV at most). Basically, gains tha t allow me to see the positive signal leave the negatives in the noise, and if I crank up the gain I see the negatives clearly but the positives are c lipped.

tware apply gain/scaling so the charts are readable...works just ok (and re quires a 24-bit ADC I'd like to avoid in the future and stick with 16-bit). ..I just can't help but think there is a hardware solution I am missing to get better resolution on the negatives.

a differential signal like a wheatstone bridge.

read the signal with two different adc channels. One with the gain for "pos itive" signals one with the gain for "negative".

Bye Jack

Reply to
jack4747

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.