Need advice on low-voltage, precision signal generation

Background: I recently started a project to produce a certain type of equipment. The (multiple) units sense and amplify small voltages on the order of 1 mV and then process the information digitally. My intention has always been to *digitally* calibrate the units, to each other (important for this application) and to an absolute standard by using a precision signal generator of some sort. The calibration will, at first order, be a simple storage of the actual (as opposed to the theoretical designed) gains and offsets of the individual units. These numbers can be stored in a file, used to calibrate the information used, and periodically updated to compensate any drift over time. Later, if it turns out to be important for the application (doubtful) I may want to even digitally compensate for weak nonlinearities.

Question: Can anyone help me find a *precision* signal generator, commercial or home made, or any other suggestions to solve this problem. On the low end of things if I could get just a precise 1 mV signal (sine or square wave) that would be a great start. If there is a *precision* way to step down a 1 V signal to a 1 mV one then I could use an ordinary commercial signal generator. My impression is the commercial signal generators are not designed to be precise (or work at all) at the 1 mV level. There are cheap signal generators that generate test signals at this level such as ECG model simulators, but these are toys, and there is *nothing* precise about them. I need a gold standard. I only need very low frequencies < 1000 Hz. Surely there must be an economical and reasonable way to either step down a commercial signal generator in a very *precise* way or to generate the signals some other way. But obviously most home made solutions I can think of would have themselves to be calibrated, and so defeat the purpose.

If anyone has any ideas about this I would really appreciate it.

Certainly you can generate a precise small signal from a large signal using a resistive divider. It's accuracy will depend upon the resistor accuracy. You can buy resistors to 0.1% or better accuracy if needed.

If your needs are under 200KHz, this is my daily spoon that should work for your application:

Audio Precision Audio Analyzers : 2700

Prepare the wallet; or rent one.
Miguel

crutschow said:

Certainly you can generate a precise small signal from a large signal using a resistive divider. It's accuracy will depend upon the resistor accuracy. You can buy resistors to 0.1% or better accuracy if needed.

Click to expand...

My main concern was not having to buy the very high tolerance resistors in a batch of 5000, especially as I will need at least two resistances, but I just checked and it does look like there are a few SMT ones on digikey that can be purchased in small quantities. This is probably the way to go - I can spend all the money saved to buy an extra nice function generator.

Use an oscilloscope amplitude calibration instrument such as the Tektronix PG506 (needs a TM5xx power module to operate). They generate square waves with precise amplitude (about ±0.1% or so as I recall) in a 1-2-5 sequence from 100v p-p down to the microvolt region.

Dean

Dean Huster said:

Use an oscilloscope amplitude calibration instrument such as the Tektronix PG506 (needs a TM5xx power module to operate). They generate square waves with precise amplitude (about ±0.1% or so as I recall) in a 1-2-5 sequence from 100v p-p down to the microvolt region.

Dean

Click to expand...

Sweet! Now we're talking. That is exactly what I need. It says it go down to 200 uV - perfect! I'm much more keen on the precision than multi-functionality. Funny thing, I called tektronix just yesterday and tried to ask one of their "technical support". The lady I talked to was totally useless, and she certainly never mentioned an actual tektronix product. Oh well. Thanks for the tip. Now if we can just get tektronix to send you a commission.

Well, you may have a harder time buying something new like that from Tektronix. I've not worked for them since 1982 and don't know what's in their current product line for calibration equipment. The "original" amplitude calibrator was their 067-0502-00 and 067-0502-01 amplitude calibrator which was replaced by the TM500-series plug-in, the PG506. Then there is the newer CG5001, a TM5000-series plug-in which covers even more bases. Check out ebay for the earlier models as the newer stuff may get pretty pricey.

For that matter, an old Tektronix 500-series oscilloscope such as the 545, 545A, 545B, 547, 535, 535A, etc. has a calibrator built in that does most of what the 067-0502-01 amplitude calibrator did, just not as accurately. Of course, that's a pretty danged big box for just the calibrator!

Dean

Dean Huster said:

Well, you may have a harder time buying something new like that from Tektronix. I've not worked for them since 1982 and don't know what's in their current product line for calibration equipment. The "original" amplitude calibrator was their 067-0502-00 and 067-0502-01 amplitude calibrator which was replaced by the TM500-series plug-in, the PG506. Then there is the newer CG5001, a TM5000-series plug-in which covers even more bases. Check out ebay for the earlier models as the newer stuff may get pretty pricey.

For that matter, an old Tektronix 500-series oscilloscope such as the 545, 545A, 545B, 547, 535, 535A, etc. has a calibrator built in that does most of what the 067-0502-01 amplitude calibrator did, just not as accurately. Of course, that's a pretty danged big box for just the calibrator!

Dean

Click to expand...

It's true, I don't remember seeing those units on the tektronix web site - but I might have missed it. However there seem to be plenty available online, so I should be good. Gotta love internet comerce - its like the world greatest garage sale right at you fingertips.