Precision Full-Wave Rectifier Distortion at High Frequency

Thread Starter

confuseddesigner

Joined Mar 14, 2017
9
I built a precision full-wave rectifier using two opamps and two diodes in PSpice. The goal is to rectify and amplify an AC signal by 100x.
I did a time domain simulation and found that the output at 1kHz deviates little from the expected results, but at 30kHz there is considerable distortion.

Why did this happen? In this circuit setup, which opamp (U1 vs U2) is the more significant contributor to the distortion? Is there a way to extend the frequency range that gives little distortion?

I also performed a DC bias analysis by setting the input voltage source Vs to 0, but the output of OpAmps is non-zero. Why does this happen? Is it because the DC voltage supplies (Vcc and Vee) to the OpAmp energized the OpAmps and disturbed them a little? Is there way to make the non-zero output go away?

Thanks.

(I attached the circuit diagram and simulation results. In the simulation results, the input voltage was multiplied by 100 for viewing convenience).

MOD: rotated and clipped your circuit image, makes it easier to read with your text.E
AA1 25-Oct-18 08.45.gif
 

Attachments

Last edited by a moderator:

Thread Starter

confuseddesigner

Joined Mar 14, 2017
9
Yes. LTSpice looks right, but if you change the source frequency frequency to 30kHz, you'll see distortion like the attachment "30KHz Response" shows.
 

ericgibbs

Joined Jan 29, 2010
9,037
hi Dana,
Tried the circuit in post #7, it does not perform anywhere close to its claimed performance.
Looks like some ones theoretical concept of a precision rectifier.

Do you have any other source for this design

Eric
 

danadak

Joined Mar 10, 2018
3,682
I am curious, what performance did you get ?

I keep thinking in recess of my mind either Pease or Jim Williams did
a bunch of work on fast rect circuits, just cannot seem to find the info.
I will keep looking.

I also recall there is a circuit using sign changer, single OpAmp, I will
see if I can find that.


Regards, Dana.
 

ericgibbs

Joined Jan 29, 2010
9,037
hi Dana,
Thanks for those links, it the Burr Brown I normally use.
When I get a few minutes I will try the EDN circuit in LTS.
E
I wonder if the TS will come back.?
 

ebp

Joined Feb 8, 2018
2,332
What do the models of the TL081 do with input offset voltage? The amp's offset voltage is pretty dismal unless it is trimmed.

If there is no convenient way to null the offset for the model, a small DC source could be added in series with the signal.
 

ericgibbs

Joined Jan 29, 2010
9,037
hi ebp,
Tried adding a small +/- DC offset of the 10mV sine wave signal, it had some effect, but the output full wave rectified signal was abysmal.
Also tried a range of higher spec OPA's, still considered the output unusable.
The circuit in post #6, gives acceptable results upto around 40kHz.

I cannot vouch for the performance of the TL081 LTS models compliance with its datasheet.

E
 

OBW0549

Joined Mar 2, 2015
3,075
The distortion you're seeing is probably due to the TL081's limited gain-bandwidth product (GBW) and limited slew rate. Precision rectifier designs like this are EXTREMELY demanding on op amp speed because each time the input signal transitions from positive to negative or vice versa, the op amp's output must "jump over" a dead zone equal to two diode forward drops. During the time it takes to do this, neither diode is conducting and the circuit outputs who-knows-what; hence the distortion.

I'd suggest using a MUCH faster op amp.
 

crutschow

Joined Mar 14, 2008
23,727
Here's the LTspice simulation of a simple, no-diode, precision rectifier that will work above 100kHz.
It avoids the diode jump slew problem that OBW0549 mentioned.
The opamp is a 30MHz GBW RR type.
It's operation depends upon having only a single supply voltage. It won't work with a dual supply.

upload_2018-10-25_13-30-26.png
 
Top