I need to find or build an ac voltage reference. It only needs to read +/- 3 or 4 volts max peak-to-peak. I need to be able to measure from 10 Hz to 20kHz. I have a couple of DC voltage reference chips from Maxim that seem to work well for DC but I don't see any for AC.
johngalt47 said:
Sorry I dont remember good...
Try at google "function generator ICs"
I think that Maxim do some of this chips...
Old manufacturers are Exar and Intersil.
Best regards,
-boggy
Yes, try " 8038" or try this:
http://www.apogeekits.com/function_generator_kit_k8016.htm .
or this:
http://wiredworld.tripod.com/tronics/signal_source.html
Jan Didden
http://www.apogeekits.com/function_generator_kit_k8016.htm .
or this:
http://wiredworld.tripod.com/tronics/signal_source.html
Jan Didden
I think you misunderstand what he is asking for (or is it me getting it wrong?). I think he asks for a signal source that has a very precisely defined amplitude. I hardly think they are any chips or similar simple solutions to buy though. Maybe one could build an oscillator where that amplitude is somehow compared (peak value? rectified LP-filtered value?) to a DC level from a voltage reference, and feedback used to adjust the amplitude?
Christer said:
Wow!
If this IS true, then forget about "AC reference voltages" in that range of stability (couple of ppm). Oscillators don't have this range of precision and stability. Feedback regulated is only (if I remember) Wien bridge topology, but you can obtain better results with Maxim function generator than with Wien bridge (Maxim/Exar/Intersil chips don't have feedback, and generate sine wave with precise nonlinear network)
In this case, best available solution for this is to build signal/waveform generator with precise and stable D/A converter and lookup table in EPROM, but this solution is much more expensive than precise DC references. You must have precise and gain-stable amplifier for this etc.
----
I remember something! Long ago Micro Linear do some 8pin chips, working with lookup tables and digital clock. I don't know what is happen with Micro Linear and their small chips. I remember also that this chips are very rare and only obtainable from Micro Linear (single manufacturer). This is something like digital controlled sine wave generator.
----
Best regards,
-boggy
Yes, but how exact? Remember that you are interested in dB's I'll suppose? What are you going to use it for?
An ICL8038 or similar has pretty stable amplitude.
An ICL8038 or similar has pretty stable amplitude.
A 8038 amplitude is intimately depending on the supply voltage. Use a regulated supply and you get good stability. Would square wave be OK for you, or does it need to be sine?
What do you need in terms of dB or PPM?
Jan Didden
What do you need in terms of dB or PPM?
Jan Didden
I want to verify the accuracy of the ac scale on my Fluke 8800A multimeter. I have already checked the resistance and dc voltage scale to my satisfaction.
What exactly are you going to do?
You could try some FFT software like Rightmark Audio Analyzer, very good russian freeware together with your soundcard. An example of the report is here.
http://audio.rightmark.org
You could try some FFT software like Rightmark Audio Analyzer, very good russian freeware together with your soundcard. An example of the report is here.
http://audio.rightmark.org
Hi johngalt47,
You should send your meter to a calibration lab. Ask for calibration / optimization with data. Indicate the frequency range you want if outside the normal spec for your meter.
The equipment the lab uses will hopefully be a Fluke 5500A or 5520A calibrator. (look up how much these are worth). I do not care for the Wavetek calibrators, they are better for clamp meters.
Make sure you go to a big lab with a proper calibration area. Generally your meter may take a week, might be less. This service is well worth it. So is the calibration data. You are lucky you have a Fluke. Some other brands can not be brought into calibration and have terrible high frequency performance.
This is the only worth while way to do what you have asked.
-Chris
You should send your meter to a calibration lab. Ask for calibration / optimization with data. Indicate the frequency range you want if outside the normal spec for your meter.
The equipment the lab uses will hopefully be a Fluke 5500A or 5520A calibrator. (look up how much these are worth). I do not care for the Wavetek calibrators, they are better for clamp meters.
Make sure you go to a big lab with a proper calibration area. Generally your meter may take a week, might be less. This service is well worth it. So is the calibration data. You are lucky you have a Fluke. Some other brands can not be brought into calibration and have terrible high frequency performance.
This is the only worth while way to do what you have asked.
-Chris
I think Transcat has a lab in your state in Houston. I used to work for their lab in Canada (the only one, in Mississauga Ontario).
Try www.transcat.com.
-Chris
Try www.transcat.com.
-Chris
Hi,
can the heating equivalence be useful here?
Is there a way to accurately compare (near instantaneously) the heating effect of an unknown AC signal and measure the DC voltage of another calibration circuit that has the same heating effect?
Could some chopper across a bridge using AC and DC alternately give sensible numbers?
This would depend on a close approximation to sinewave signal, but whether distortion at -20db or -40db or -60db is needed?
can the heating equivalence be useful here?
Is there a way to accurately compare (near instantaneously) the heating effect of an unknown AC signal and measure the DC voltage of another calibration circuit that has the same heating effect?
Could some chopper across a bridge using AC and DC alternately give sensible numbers?
This would depend on a close approximation to sinewave signal, but whether distortion at -20db or -40db or -60db is needed?
Hi Andrew,
Heating equivalence can of course be used. You have to calibrate it first.
Can you say arbitrary waveform generator?
Cheaper to send the meter for calibration.
-Chris
Heating equivalence can of course be used. You have to calibrate it first.
Can you say arbitrary waveform generator?
Cheaper to send the meter for calibration.
-Chris
Best true-rms meters even today still use heating-based measurement. No bandwith or crest factor problems, but I think that getting good accuracy can be tricky.AndrewT said:
You have to calibrate it against a DC reference, which isn't so difficult.
Having said that, getting better than 50ppm AC "accuracy" is a can of worms.
Having said that, getting better than 50ppm AC "accuracy" is a can of worms.
New true RMS IC's are out. Linear Tech have some nice ones. The old LT1088 is the heat type.mzzj said:
- Status
- Not open for further replies.