What's your favorite full wave precision rectifier circuit?
I need something good for sines and triangles of a volt or so, up to 100kHz or so. I'm looking at this
Tim
Hz
ecision_FWR.gif
at
isn't this the standard from some AD appnote:
-Lasse
Firmware.
John
My favorite:
I've been using it since the mid '80's, first for a 400KHz application, back when OpAmp GBW wasn't nearly as good as they are now.
...Jim Thompson
OK, all analog: use a comparator to sense the sign of the input, and then switch a +1/-1 gain stage.
John
I usually use a Gilbert cell driven by a comparator.
Cheers
Phil Hobbs
Uselessly overcomplicated.
Has a certain elegance about it though. Breaking it down to its components, you have a stacked differential amplifier (the mixer) and a high gain differential amplifier (the comparator), which should have a differential output to drive the mixer to saturation (or use diode steering). But that's none other than taking signal x, amplifying the piss out of it to get
100,000x clamped to the supply rails, which is essentially the sign function (in this case, +1 if x > epsilon, -1 if x < epsilon, x/epsilon if |x| < espilon). So it generates sign(x) * x, which is another way of writing |x|. Neato.Tim
Often the only way when you have 500MHz signals to deal with.
...Jim Thompson
If you're doing a discrete (not ICs) design, it's arguably a lot simpler than building the equivalent op-amp-style circuit... we're talking, what, no more than a dozen transistors here (current sources included), right?
I sense a "fewest number of transistors absolute value circuit" design challenge coming on. :-)
than
o more
+10V ! [R]10K ! +-------Vout !/ AC IN------! !\\e 2N3904 ! [R]10K ! -10VWhat do I win?
That ought to produce some real crap designs ;-)
...Jim Thompson
Hz
ecision_FWR.gif
at
H&H AoE. not sure the chapter or page. You add the inverse to the half wave rectified signal. (or something like that) through an opamp. I'm pretty sure they didn't invent it but that's where I saw it first.
George H.
kHz
at
=A0 =A0 ...Jim Thompson
=A0 | =A0 =A0mens =A0 =A0 |
=A0 | =A0 =A0 et =A0 =A0 =A0|
=A0|
=A0 =A0 =A0 |
Hmm, I haven't seen that before. I'll give it a whirl. Thanks, George H.
0kHz
p at
Phil, Do you 'roll your own' Gilbert cell?.
Tim, I'm using an AD multiplier chip, part number 'unsure' (sorry I'm at home with dial-up and it's too painful to go searching the web.) That's claimed to be good up to 10MHz. It works great for detecting noise up to 100kHz. (and beyond)
George H.
It's about 25 cents worth of parts, smaller than a dime, and works great. Useless is in the eye of the beholder, I suppose.
Cheers
Phil Hobbs
Generally I use an array, e.g. an MC1496.
Cheers
Phil Hobbs
Hz
ecision_FWR.gif
at
Hello Tim,
What relative input range is required and how good it must be (linearity, offset drift, etc)?
The larger the relative input range, the larger the gain must be in case of a feedback approach.
Best regards,
Wim PA3DJS
An MC1496 is NOT an array... it's the real deal ;-)
...Jim Thompson
Like-a dis...
...Jim Thompson
100kHz
tep at
t -
Cool, I use the AD630 as a 'lockin' chip. Someone told me that was what I should use and I never bothered lokking for other demodulators. So the MC1496 is 150 times faster and 1/20 the cost. What's the down side? The carrier feed through numbers look a bit scary. 20 - 200 mV with a sqaure wave if the offset isn't adjusted.
George H.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required