Author Topic: Exponential Amplifier  (Read 5656 times)

0 Members and 1 Guest are viewing this topic.

Offline eev_carlTopic starter

  • Regular Contributor
  • *
  • Posts: 224
  • Country: us
Exponential Amplifier
« on: July 23, 2018, 02:30:51 pm »
Hi,

I built an exponential amplifier but the circuit doesn't seem to respond to the changing Vin.  The schematic is attached.  I'm using a PN2222A, 741, and 8148.

When I vary the Vin from 10mV-100mV-1V-10V, my Vout remains at 1.5mV.

Thanks in advance,
Carl
 

Offline HB9EVI

  • Frequent Contributor
  • **
  • Posts: 722
  • Country: ch
Re: Exponential Amplifier
« Reply #1 on: July 23, 2018, 02:51:35 pm »
try just with the diode on the inverting input, but with the cathode connected to it
 
The following users thanked this post: eev_carl

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
Re: Exponential Amplifier
« Reply #2 on: July 24, 2018, 02:04:06 am »
Where did you find that example?  The transistor and diode are backwards; the emitter of the transistor is on the input side.

 
The following users thanked this post: eev_carl

Offline exe

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: Exponential Amplifier
« Reply #3 on: July 24, 2018, 08:57:55 am »
Just in case, all analog exponential / log amplifiers I know are quite imprecise (like, drifting with temperature, etc). Those based on diodes drift the most, better ones use BJTs (still not ideal). There are articles on the subject explaining why. Common workarounds are to digitize the signal and process in digital domain. Or approximate log/exp scale with multiple amplifiers (each with its own fixed gain).
 
The following users thanked this post: eev_carl

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
Re: Exponential Amplifier
« Reply #4 on: July 24, 2018, 11:08:53 am »
Common workarounds are to digitize the signal and process in digital domain.

That must be fun when the dynamic range exceeds any ADC ever made.

Accuracy of better than 1% is feasible with proper temperature compensation or if the temperature of the translinear element is fixed.
 
The following users thanked this post: eev_carl

Offline eev_carlTopic starter

  • Regular Contributor
  • *
  • Posts: 224
  • Country: us
Re: Exponential Amplifier
« Reply #5 on: July 24, 2018, 01:30:49 pm »
Quote
Where did you find that example?

"The Forrest Mims Circuit Scrapbook Vol 1."

I build a logarithmic amp from the same article and its transfer function worked perfectly.  I've seen different schematics with the diode reversed too.
 

Offline eev_carlTopic starter

  • Regular Contributor
  • *
  • Posts: 224
  • Country: us
Re: Exponential Amplifier
« Reply #6 on: July 24, 2018, 01:32:27 pm »
Quote
Common workarounds are to digitize the signal and process in digital domain.

This isn't for a practical application.  I have ADCs, MCUs, and DACs here too.
 

Offline eev_carlTopic starter

  • Regular Contributor
  • *
  • Posts: 224
  • Country: us
Re: Exponential Amplifier
« Reply #7 on: July 24, 2018, 01:38:08 pm »
Quote
try just with the diode on the inverting input, but with the cathode connected to it

I've tried this too, but am having trouble figuring out the transfer function.  If I just run a diode into the -V of the amp, what would the values look like?  The input needs to be >0.7V to activate the diode, right?

This is just for a demo so I'm not concerned with errors from temp, etc.  I'm looking for something that I can plot with a rough exponential curve.
 

Offline HB9EVI

  • Frequent Contributor
  • **
  • Posts: 722
  • Country: ch
Re: Exponential Amplifier
« Reply #8 on: July 24, 2018, 02:03:41 pm »
it should look like an antilog curve in the x-y coordinates.

maybe stupid to ask, but are you running the opamp with a bipolar supply? you know, it's actually an inverting circuit, so the output goes to the negative rail of the supply; the circuit cannot work from a single supply
 
The following users thanked this post: eev_carl

Offline Alex Nikitin

  • Super Contributor
  • ***
  • Posts: 1330
  • Country: gb
  • Femtoampnut and Tapehead.
    • A.N.T. Audio
Re: Exponential Amplifier
« Reply #9 on: July 24, 2018, 02:03:51 pm »
Where did you find that example?  The transistor and diode are backwards; the emitter of the transistor is on the input side.

Exactly. It is just an error in the book.

Cheers

Alex
 

Offline eev_carlTopic starter

  • Regular Contributor
  • *
  • Posts: 224
  • Country: us
Re: Exponential Amplifier
« Reply #10 on: July 24, 2018, 02:08:39 pm »
Quote
bipolar supply?

It's split.  I threw in a resistor instead of the diode and ran in a sine wave to verify the 741.
 

Offline eev_carlTopic starter

  • Regular Contributor
  • *
  • Posts: 224
  • Country: us
Re: Exponential Amplifier
« Reply #11 on: July 24, 2018, 02:25:18 pm »
I reversed the BJT and get these outputs.  It seems like i might not be operating in the correct range.  Maybe I'm missing an offset?  Nothing interesting happens until about 10V.

VinVout
10mV1.5mV
100mV1.5mV
1000mV1.5mV
10000mV22.6mV
« Last Edit: July 24, 2018, 02:26:57 pm by eev_carl »
 

Offline Alex Nikitin

  • Super Contributor
  • ***
  • Posts: 1330
  • Country: gb
  • Femtoampnut and Tapehead.
    • A.N.T. Audio
Re: Exponential Amplifier
« Reply #12 on: July 24, 2018, 02:30:31 pm »
I reversed the BJT and get these outputs.  It seems like i might not be operating in the correct range.  Maybe I'm missing an offset?  Nothing interesting happens until about 10V.

VinVout
10mV1.5mV
100mV1.5mV
1000mV1.5mV
10000mV22.6mV

 :palm:

The input should be negative polarity... and please limit the current on the input by, say, 10mA.

Cheers

Alex
 
The following users thanked this post: eev_carl

Offline eev_carlTopic starter

  • Regular Contributor
  • *
  • Posts: 224
  • Country: us
Re: Exponential Amplifier
« Reply #13 on: July 24, 2018, 02:33:51 pm »
Here's what I get with negative polarity

VinVout
-10mV1.5mV
-100mV1.7mV
-1000mV8.22V
-10000mV8.16V

The current is >5mA for Vins >1V.   Should I put a 220ohm resistor in series?


« Last Edit: July 24, 2018, 02:37:16 pm by eev_carl »
 

Offline Alex Nikitin

  • Super Contributor
  • ***
  • Posts: 1330
  • Country: gb
  • Femtoampnut and Tapehead.
    • A.N.T. Audio
Re: Exponential Amplifier
« Reply #14 on: July 24, 2018, 02:53:18 pm »
The current is >5mA for Vins >1V.   Should I put a 220ohm resistor in series?

It could be too late for it - if you've really managed to apply 1V and 10V voltage to the input... .

Cheers

Alex
 

Offline eev_carlTopic starter

  • Regular Contributor
  • *
  • Posts: 224
  • Country: us
Re: Exponential Amplifier
« Reply #15 on: July 24, 2018, 03:20:18 pm »
Quote
It could be too late for it - if you've really managed to apply 1V and 10V voltage to the input... .

Why?  Isn't the 741 good for +-15V and the PN2222A good for 50V or so?
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
Re: Exponential Amplifier
« Reply #16 on: July 24, 2018, 04:50:54 pm »
Quote
It could be too late for it - if you've really managed to apply 1V and 10V voltage to the input... .

Why?  Isn't the 741 good for +-15V and the PN2222A good for 50V or so?

The transistor's emitter-base breakdown voltage is only about 5 volts which is why the diode is included in the circuit.  Emitter-base breakdown is very damaging to transistors which are intended for use in translinear circuits.

For now why not use just a diode instead of the transistor to get an idea of what should be going on?

And move the capacitor to be in parallel with the feedback resistor.  Where it is now causes the operational amplifier to act as an unstable differentiator.  The circuit as shown even with the transistor and diode used correctly will oscillate.
« Last Edit: July 24, 2018, 04:54:15 pm by David Hess »
 

Offline GigaJoe

  • Frequent Contributor
  • **
  • Posts: 524
  • Country: ca
Re: Exponential Amplifier
« Reply #17 on: July 25, 2018, 01:23:51 am »
 
The following users thanked this post: exe, BrianHG

Offline nugglix

  • Regular Contributor
  • *
  • Posts: 209
  • Country: de
Re: Exponential Amplifier
« Reply #18 on: July 25, 2018, 09:51:42 am »
 

Offline eev_carlTopic starter

  • Regular Contributor
  • *
  • Posts: 224
  • Country: us
Re: Exponential Amplifier
« Reply #19 on: July 25, 2018, 01:21:05 pm »
Hi,

I built the exponential amp w. a single diode. Here is my schematic and test results.  The results produce a straight line on a semi-log graph which is what I was looking for.

Thanks to all

(in mV)

VinVout
200-0.7
300-13.6
400-109.6
500-828
600-4590
« Last Edit: July 25, 2018, 01:46:39 pm by eev_carl »
 

Offline exe

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: Exponential Amplifier
« Reply #20 on: July 25, 2018, 01:29:08 pm »
Here is my schematic and test results.

Imho this can't work. But even if does, it's just wrong. Diode is biased only by opamp's parasitic input current, that's not how to build circuits.
 

Offline eev_carlTopic starter

  • Regular Contributor
  • *
  • Posts: 224
  • Country: us
Re: Exponential Amplifier
« Reply #21 on: July 25, 2018, 01:33:15 pm »
Quote
Imho this can't work. But even if does, it's just wrong. Diode is biased only by opamp's parasitic input current, that's not how to build circuits.

What do you think about the anti-log schematics in the TI links posted above?  Are you suggesting that the diode should be replaced with a BJT?
 

Offline Alex Nikitin

  • Super Contributor
  • ***
  • Posts: 1330
  • Country: gb
  • Femtoampnut and Tapehead.
    • A.N.T. Audio
Re: Exponential Amplifier
« Reply #22 on: July 25, 2018, 01:38:54 pm »
Hi,

I built the exponential amp w. a single diode. Here is my schematic


 :palm: :palm:

Your schematics is wrong. It is not what you've built and measured.

Cheers

Alex
 

Offline eev_carlTopic starter

  • Regular Contributor
  • *
  • Posts: 224
  • Country: us
Re: Exponential Amplifier
« Reply #23 on: July 25, 2018, 01:42:16 pm »
Quote
Your schematics is wrong. It is not what you've built and measured.

That's my bad Eagle skills...fixing + to ground

(replaced incorrect Eagle schematic with handdrawn schematic showing the inverting amp)
« Last Edit: July 25, 2018, 01:56:31 pm by eev_carl »
 
The following users thanked this post: exe

Offline exe

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: Exponential Amplifier
« Reply #24 on: July 25, 2018, 02:08:56 pm »
What do you think about the anti-log schematics in the TI links posted above?  Are you suggesting that the diode should be replaced with a BJT?

Ah, now I see the updated schematics :). I'd say, if diode works for you, then don't make your life harder with more complicated circuits. Just know limitations of your solution (mainly thermal drift).
 
The following users thanked this post: eev_carl


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf