I am looking for some help from an Opamp guru that has knowledge of Opamp filters, Precision rectifiers and Absolute value circuits. I have been tasked with reverse engineering an existing working circuit that was designed by a company engineer many years ago and all the knowledge of the design is long since gone. I cannot post a public schematic so I am going to have to find someone I can work one on one with through emails. I need a full detailed understanding of how the circuit works. Compensation for your time may be possible.
If you have the bandwidth to help. please PM me.
Zc
If you have the bandwidth to help. please PM me.
Zc
Now, how could that ever happen? What will become of the western civilization if all knowledge flow away or vaporise?
Why not follow one of these abundent links?
DuckDuckGo
It's not about what you can find, it's about what you can know.
There are many precision rectifier, RMS/Absolute, and AC to DC ICs available that make the design much easier, and usually better. There data sheets usually explain everything. The problems with opamp circuits is there not acurate at less than 10 mv ac, and there bandwith is limited. If your aplication is audio the RMS detects from THAT are specificaly designed and hard to beat. Dont know much about the RF chips except there more money. What range of voltages and freqs will this meter need to read?
Last edited:
I first need to understand exactly how the existing circuit works so that we can then explore newer, better options.
Precision rectifier. See page 19 of the AD8036 data sheet. The 8036 can also be an analog 'OR'.
The only requirement is that VH is always more positive than VL. Did I mention that it is a superb
clipping amplifier out to 250 MHz? If you ever tried to build a clamp amp (clipper) from opamps
you will truly appreciate this chip.
https://www.analog.com/media/en/technical-documentation/data-sheets/AD8036_8037.pdf
G²
The only requirement is that VH is always more positive than VL. Did I mention that it is a superb
clipping amplifier out to 250 MHz? If you ever tried to build a clamp amp (clipper) from opamps
you will truly appreciate this chip.
https://www.analog.com/media/en/technical-documentation/data-sheets/AD8036_8037.pdf
G²
Last edited:
The normal absolute value circuit used one opamp as a zero loss half wave rectifier that the feeds an adder that gets the full input signal. These then add to the absolute or full wave peak value.
True RMS is quite a different beast. dBx came to the conclusion that they could do the rectification after doing a logarithmic conversion to do the math required to get to an accurate RMS value. There have been other methods the oldest of which is to use the signal under test to heat a sensor and have a matching sensor that is fed DC to make their outputs match.
Filtering usually starts with the Sallen and Key paper on filters. Although it is quite worthwhile to get the book "Rapid Practical Design of Active Filters.) I do forget the author's name though.
But at a 15 year old design the questions would be more of did you do a spice analysis of the design? What is the bandwidth and accuracy? Also what kind of signals are you looking at? Most RMS designs have some peak limit issues.
The modern problem is it is easy to find offered solutions and design guides, not a wide field of instructions in the basics.
Unlike Jan, I can't work for free. (Jan have you finished editing my bit yet?)
True RMS is quite a different beast. dBx came to the conclusion that they could do the rectification after doing a logarithmic conversion to do the math required to get to an accurate RMS value. There have been other methods the oldest of which is to use the signal under test to heat a sensor and have a matching sensor that is fed DC to make their outputs match.
Filtering usually starts with the Sallen and Key paper on filters. Although it is quite worthwhile to get the book "Rapid Practical Design of Active Filters.) I do forget the author's name though.
But at a 15 year old design the questions would be more of did you do a spice analysis of the design? What is the bandwidth and accuracy? Also what kind of signals are you looking at? Most RMS designs have some peak limit issues.
The modern problem is it is easy to find offered solutions and design guides, not a wide field of instructions in the basics.
Unlike Jan, I can't work for free. (Jan have you finished editing my bit yet?)
The usual method is to use an op-amp to diminish the diode curve and sum the resulting ~perfect (-ve) 2x half wave with the input so that the sum is 1x for positive and 1x-2x for negative input.
Precision rectifier - Wikipedia
Precision rectifier - Wikipedia
The normal absolute value circuit used one opamp as a zero loss half wave rectifier that the feeds an adder that gets the full input signal. These then add to the absolute or full wave peak value.
True RMS is quite a different beast. dBx came to the conclusion that they could do the rectification after doing a logarithmic conversion to do the math required to get to an accurate RMS value. There have been other methods the oldest of which is to use the signal under test to heat a sensor and have a matching sensor that is fed DC to make their outputs match.
Filtering usually starts with the Sallen and Key paper on filters. Although it is quite worthwhile to get the book "Rapid Practical Design of Active Filters.) I do forget the author's name though.
But at a 15 year old design the questions would be more of did you do a spice analysis of the design? What is the bandwidth and accuracy? Also what kind of signals are you looking at? Most RMS designs have some peak limit issues.
The modern problem is it is easy to find offered solutions and design guides, not a wide field of instructions in the basics.
Unlike Jan, I can't work for free. (Jan have you finished editing my bit yet?)
Not so much edited as cut up ;-)
I think it is a waste to try to reverse engineering something 15+ years old, chances are you will find parts that are obsolete anyway.
And after you are done, you still don't know the exact performance. So then when you try to recreate it you may overlook something, because you don't know the specs, and up with something unusable.
The smart engineering thing to do is to measure the performance, list the requirements and then make a replacement, form/fit if needed.
And I don't work for free either, except for you Ed ;-)
Jan
I think you guys have missed the point. there is an existing machine. I need to document how it works for the client. therefore designing something new is not the point of this at this time. they first need an understanding of how the existing tester works exactly. so they can document their process and in the future design a new machine based on a newer design.
I found an excellent video that got me 99% of what I needed.
YouTube
I found an excellent video that got me 99% of what I needed.
YouTube
The narrator did a great job on the YouTube above explaining the subject.
I have the book he showed. I bought it at a bookstore in Chicago in 1974. It is entitled "Applications of Operational Amplifiers" Third-Generation Techniques by Jerald G. Graeme, M.S.E.E. The author's work-title was: Manager, Monolithic Engineering at Burr-Brown Research Corporation.
Best
Anton
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.