function generator - need advice

[rize]

My interest is in analog & digital audio circuits. I'm thinking of purchasing a used function generator - something like an HP 3314A or 3325A.

Some eBay sellers claim their units are calibrated. Short of sending the unit to a calibration lab, are there things one can check to determine whether the equipment needs calibration?

The prices on these HP units I'm looking where the seller has a return policy vary between $240 and $400. There are some asking even higher prices but those are out of my range. Since the older HP units are no longer supported and around 30 years old, I'm wondering if it would be wiser to invest in a more modern generator. I can get a no-name 20 mHz unit with sweep functionality for about $250, calibrated with 1 year warranty.

Thank you for your input.

[Damon Hill]

Most function generators will have relatively high THD, no better than .1% at best, with a lot of high-order harmonics. Therefore, no very useful for checking distortion. The sine wave is usually derived from a filtered triangle or square wave.

[bob91343]

A function generator is usually a solution without a problem. I have resisted getting one for many years because they aren't much good at anything. I don't need a square wave, and certainly have no use for a triangle, and that leaves low grade sine.

You can make a square wave generator easily from a single IC and a sine generator or even without the generator.

For as much money as you are willing to spend you can get a great audio generator. One of mine is an HP651A that goes from 10 Hz to 10 MHz, very clean. If you have room, the predecessor, the HP650A isn't bad either. Even low end stuff like EICO and Heath and Knight aren't so bad.

[jackinnj]

Originally Posted by Damon Hill

Most function generators will have relatively high THD, no better than .1% at best, with a lot of high-order harmonics. Therefore, no very useful for checking distortion. The sine wave is usually derived from a filtered triangle or square wave.

Charles Hansen wrote an article for AX using a sq wave and a pair of LT1063 switched cap filters, followed by an analog filter to eliminate the switching noise -- if you know the fundamental frequency, you can take a function generator and turn it into a high quality sine generator. I used this methodology to filter the crude sine signal from the Parallax Basic Stamp and got a signal with THD% well under 0.01%

Take a look at the Tektronix FG generators like the FG501 -- they fit into a TM* mainframe. Square and triangle are very helpful.

FWIW, the Linear Tech AN67 Ultra Low THD Oscillator -- take a careful look at the schematic.

[Printer2]

It really depends what you are going to do with it. I picked up a couple of XR-2206 chips fo $5.00 from Ebay, good up to 1MHz. Sine wave distortion 0.5%, nothing to write home about but for most applications it could be fine.

I had some high end signal generators around when I was working for an aerospace manufacture and I have used $200 generators when I worked at a college. On a scope they did not look any different.

[error401]

Anyone have speculations on how something like an AD9951 would perform at such low frequencies? I know the RF guys really love these DDS chips for their spectral purity, and I've considered building a programmable oscillator out of one (not necessarily just for audio). I think it should be possible to program them to operate at audio frequencies, anyway.

[h_a]

they aren't much good at anything.

Mmmh, let me see: it's needed to measure frequency response, clipping behaviour, square wave response to fine tune compensation....

If you don't need frequency generators, send them to me

For distortion measurement even 0.01% is too high as many solid state amps produce less THD at 1W and other low power levels well before max power.

Have fun, Hannes

EDIT: you can roughly check the calibration of a FG by using your scope (if that's properly calibrated) or a DMM. Multimeters usually work well at low frequencies (50-100Hz, some up to several kHz) and large enough voltages (above 1Vp).

EDIT2: don't forget while checking that multimeters display RMS, while function generators often set peak-to-peak voltage.

[Stefanoo]

I just got a Sound Technology 1410A signal generator that will hopefully arrive in working condition next week.
I was unable to find specifications for this generator anywhere on internet.

Do you guys have any info on this item (distortion, bandwidth, features)?

thanks

[rize]

Originally Posted by bob91343

For as much money as you are willing to spend you can get a great audio generator. One of mine is an HP651A that goes from 10 Hz to 10 MHz, very clean. If you have room, the predecessor, the HP650A isn't bad either. Even low end stuff like EICO and Heath and Knight aren't so bad.

Thanks for everyone's input.

Book by Morgan Jones - Building Valve Amplifiers, recommends audio oscillators instead of function generator or typical bench oscillator for audio testing with the following specs;

- low distortion < .05% sine valve oscillator 20 Hz - 20KHz.
- wide band meter calibrated to read sine waves
- simple form of THD measurement
- peak programme meter (PPM) for measuring noise

I live in the US and I'm looking at HP and Tek since their vintage gear is known for their reliability. So far I have noted 651A from the mid-60's. The specs for HP 239A from 79' seem impressive for a product 30 years old.

HP 651A specs
http://www.tucker.com/images/images_spec/00000986.pdf

HP 239A specs (bottom)
Audio & Test

Audio oscillators must have seen lot of improvements since 1979 but my budget is below $300. Is HP 239A the best option short of building testing circuits from scratch, assuming I can find one under $300?

[dangus]

What's wrong with using a PC sound card or USB sound device? Add a buffer amp, and you've got a very low distortion generator. Add an input buffer with protection and switched attenuation or gain, and you've got a powerful test tool for frequency, distortion, impedance, etc. There's free software for all that.