How good are the computer function generator?

kelticwizard

diyAudio Moderator Emeritus
2001-09-18 2:33 am
Connecticut, The Nutmeg State
RS dhar:

Well, RS, I haven't checked the function generators for accuracy or amplitude. But a sine wave seems to me to be the sort of thing that a computer would do very well. I see no reason to doubt it.

Computer generated sine waves are composed of square waves, and if a digital instrument cannot do square waves, what can it do?

As far as testing Thiele-Small parameters, that would tend to indicate something like Speaker Workshop by Audua, which I have downloased but have not gotten going yet.

However, I have downloaded an NCH tone generator and a freeware oscilloscope that seem pretty fine.

By the way, we are setting up a frequency response test in the thread labelled "Low On Cash" where I give some appropriate URL's. You might wish to join.
 
Computer Fujction Generators

Kelticwizard,

Thanks for reply.

I have done some theory here which says that computers produce synthesized frequency, precise in frequency and amplitude. But then individual softwares may not do this well. That is why I wanted opinions.

Besides, using generator softwares would entail drawing longish wires from and to the computer which may contaminate input and ouptput. In another thread of this forum I had wondered about my DMM registering 20-30mV in the 200mV range with no input.

I have tried Audua too and never have been able to get anything but gibberish. I used Wallin's jig and the rest record of the 500Hz sine wave came out precisely. But in the impedance calibrator window I could never get my computer to register the correct values.

Thanks again. Do please write if you have anything more.
 

kelticwizard

diyAudio Moderator Emeritus
2001-09-18 2:33 am
Connecticut, The Nutmeg State
LOL, I have experienced the same phenomenon with Digitial Multimeters as yourself-readings with no input coming in.

About the only thing I can say is, when input does finally come in, they tend to settle down.

Digital Multimeters tend to break down over time. Ironically, the best one I have owned is my little miniature unit I bought for $20 at Radio Shack, which seems to keep on going year after year. Some more expensive units don't seem to last that long-but they do have other features such as frequency readouts.

I think you are correct in saying that the cables probably have a lot to do with this phenomenon.

That is why I like to have both analog and digital multimeters around. The analogs don't seem to have that "false reading" problem, at least the ones I have. If you can get a good buy on a new or used analog meter, grab it. One problem, of course, is that analog meters might not have a scale in the 200 mV range.

I have recommended Audua Speaker Workshop to people on this forum, even though I have never gotten it going myself. Several have posted days later raving about how good it is. Apparently, once you get it going, it really is very good. I just hate to invest the time to get it set up. I like intuitive programs.
 
Software Based Function Generator

Kelticwizard,

Thanks.

More on Audua. The underlying principle of Audua is very convincing. I struggled with this software for months. I think my problems are purely hardware in nature. Would you believe, I could record a test signal of 500Hz without even connecting the Wallin's jig to the soundcard input?

There is a helpline to Audua and someone answered my queries. The Audua wanted me to send a sample record file and I had problems in doing this. They suggested that my computer (Pentium 166MHz, 32MB RAM) might be too slow for the business. I don't think so. My system is not over installed and there are no TSR programmes to eat into the RAM.

Working with a 8-10 Ohm series resistance is rather delicate. the contact resistance of the jacks from and to the soundcard can have significant effect, not to speak of the workplace being saturated by the SMPS.

If anyone out there has tried Audua successfully, I should be glad to hear a word or two of elderly adivce.

Thanks again.
 

MRehorst

Member
2002-05-17 8:48 pm
Try Spice for generating test tones

On the topic of computer generated test tones, and low budget ones at that, take a look at Switcher CAD III, a complete spice simulator with schematic capture FREE for the downloading from Linear Technology.

Besides being able to do all your circuit simulations, it has a couple special features that make it really neato for audio circuits: you can input and output voltages and currents as .wav files. That means you can generate tones, sweeps, and warbles easily, and play them back using winamp or any other software that can play .wav files (literally ALL audio programs). You can even burn the files to CDRs and make you own test CDs.

You can also read a .wav file as input to your simulation. This means you can feed a piece of music or a waveform into a crossover circuit in the simulator, then output the results to .wav files, and even listen to them. All without ever soldering!

You can also do FT on the output and look at signals in frequency domain.

The program comes with a large library of LT op-amps and switching PS IC models, and you can easily import other models.

You can't beat the price! Their tech support is great. I have emailed about bugs and minor nit-picks and always get a response within a couple hours. They update the program frequenctly, too, and there is a one-click update feature in the program.

I have attached a simple schematic for a sound generator that you can use to see how the program works and what it can do.
You must change the name of the file to UFO3.ASC before you try to run it. The output file is 15 sec long in glorius 16 bit, 44.1 ksps stereo. It takes about 3 minutes to run on my PIII 450 laptop.

Listen to the output with headphones!
MR
 

Attachments

  • ufo3.txt
    4.1 KB · Views: 33
Computer PWM sine waves

If you can get into the math, Don Lancaster has an interesting PDF file on low distortion sine wave generation at: http://www.tinaja.com/magsn01.asp

The easiest, low distortion sine generator I have ever made used an inexpensive clock crystal, multiple counters and Linear Tech LTC1063 5th order filters -- the original appeared in Audio Amateur, but there were some problems with the schematic appearing thereof (and you do need an active LPF to remove the vestigial clocking signal.)
 

MRehorst

Member
2002-05-17 8:48 pm
Originally posted by kelticwizard
LOL, I have experienced the same phenomenon with Digitial Multimeters as yourself-readings with no input coming in.

About the only thing I can say is, when input does finally come in, they tend to settle down.

Digital Multimeters tend to break down over time. Ironically, the best one I have owned is my little miniature unit I bought for $20 at Radio Shack, which seems to keep on going year after year. Some more expensive units don't seem to last that long-but they do have other features such as frequency readouts.

I believe most Flukes come with lifetime warranty now. I had an older unit that died and checked repair price. It was about the cost of a new unit, so I bought a new unit with the lifetime warranty. I suppose if it dies and they have run out of parts, they'll just give me a new model with the same sort of functions.

I think you are correct in saying that the cables probably have a lot to do with this phenomenon.

The hi-Z input picks up noise and maybe some triboelectric charge from movement of the input cables over dissimmilar material. Old VTVMs did the same thing. No big deal.

That is why I like to have both analog and digital multimeters around. The analogs don't seem to have that "false reading" problem, at least the ones I have. If you can get a good buy on a new or used analog meter, grab it. One problem, of course, is that analog meters might not have a scale in the 200 mV range.

Analog meters are still best for adjusting things where you can watch the meter peak or null. The digital meters are still just a little slow for that sort of use.

I have recommended Audua Speaker Workshop to people on this forum, even though I have never gotten it going myself. Several have posted days later raving about how good it is. Apparently, once you get it going, it really is very good. I just hate to invest the time to get it set up. I like intuitive programs.

The real problem with computer generated audio signals is the sound card in the PC tends to leak power supply and digital switching noise into the output. Newer cards and machines seem to be a bit better than old ones, but if you want a real low noise and distortion signal, it's probably best to dump the digitally generated signal onto a CDR in standard 16 bit, 44.1 ksps format so you can play it into your system using a CD player.

There is also the possibility of some innaccuracy in the calculations in the synthesis software, though I suspect that in most cases this is much less of a problem than the power supply noise.

MR