precision rectifiers

hi guys,

has anybody tried to use op-amps as precision rectifiers? the reason why i could not use the conventional rectifier is because the signal has to rise above the diode's forward voltage before conduction takes place.

what am trying to do is have the Wein network oscillate at unity gain to produce a sine wave. at less than unity there is no oscillation and a little above unity there is oscillation but with distortion.

pls give links to this problem on using op-amp as precision rectifier.

thanx

 

electronics collections

hi sir,
hope this will be interesting to you. have a look at the precision full wave rectifier. i got this by a google search.
regards
bipin

 

hi bipin,

you must referring to an EDN design mag. by fernando werlang on fast rectification using 2 op-amps. i have protoed that set up way before maybe 3 yrs. ago. that thing won't work for my need. what i need is rectification before conduction. anyway thanx for your sincere reply, i have already visitied the site and have practically read all their featured articles about op-amps. they have interesting topics there for neophytes about op-amps.

 

hi sir,
:mellow: :huh:
I too was thinking you would not be asking for such a circuit, as i have seen your expertization on electronics.
By the way y a normal precision rectifier circuit doesnt solve your problem of small signal rectificatoin? explain little bit, may we get to learn some thing more.

luv regards
bipin

 

This is actually a very interesting topic - one which I have never really thought about.

Maybe you could elaborate on your idea mozikluv

Because I'm curious to see why you are trying to design such a device.

I remember sometime back looking at Silcon Controlled Rectifiers where I utilised a Scholkley diode basically created from two transistors which I could latch using only a very small voltage. This could be used a part of a rectifier network. However I'm not fully sure whether it would solve your problem.

Let us know what you think.

 

hi dave,

my problem is the signal am trying to rectifiy. we must remember that the op-amp responds only to the difference between inputs. the circuis feedback R and input R reaction goes up and down like from a positive value to a negative value with its pivot at virtual earth. again it must be remembered that when you analyse a current to voltage converter the concept of virtual earth is used. at null point it is considered at earth potential but there is still the presence of voltage at microvolt level.

as i have stated before there is a frequency limit to the gain or distortion will set in. so i have set the freq. at 100Khz with a maximum gain of 10, ideal gain is between 9.1 to 9.5. if we transpose Rf to Rn the circuit will now act as an attenuator with a gain or 0.5

since what am trying to rectify is in microvolt level, direct rectification can't be used so there must some way to overcome that. i know it can be done but how?

 

I dont know if this is what you're looking for, but if you supply an op-amp with pos and ground instead of pos and neg DC, using a V divider, the op-amp will see the mid-point of the DC as ground, and any signal incoming will be a pulsating DC of the same values as the input (I mean sine, square, etc.., not values regarding amplification, of course). I know you are quite knowledgable and I presume you knew this, so I apologize if this is not what you were looking for.
Maybe feeding a big step up transformer and rectifying it's output would work, but at uVs, it'd have to step way up. At 1uV, I guess it would take a 1:1,000,000 to get a volt... Still not enough for a bridge. Thats a lot of winding!

 

I just came across this article on a precision rectifier for a tachometer, maybe it will help:

http://www.4qdtec.com/tacho.html

 

Good find

You'll have to work out whether this circuit will be sensitive enough to detect your very small input signal.

 

Have you thought about a germanium diode? Something like a 1N34. They have a contact drop of something under 200 mv. The pre-rectofication is a bit of a mystery. One way out is to put an amp in the signal line, so the diode drops won't cause problems.

Not surprised to hear about the gain figure. By definition, an oscillator has to have a gain of more than 1. Otherwise, it's just an amp in the linear reigon.

A classic Wein oscillator has a non-linear resistance in the feedback. My old oscillator has a low voltage lamp for that purpose.

 

hi,

to qtc3, have already check that out and the output stage has been configured as a Norton amp. it wont do.

beenthere, yes i have tried using that diode but still not to my desired result.

at this time am about to try using a diode as a feedback path connected parallel to the output. am not sure though but i think this would solve the problem. have tried so many approach but they all defy ohms law as far as Rf and Rn are concerned.

will let you guys know of the result. i feel like a grade 1 pupil as far as what am trying to do, its way out of my field of interest.