Log in
Log in
Forgot your password ?
or
Create an account
or
Log in using a Facebook account
NotificationsNo new notification
< Back to The Gentleman's Lounge
Enlarge
The Gentleman's Lounge
Thematic Forums
The Gentleman's Lounge
Actu 9 9
Articles 3 3
Media 66,026 66k
Tutorials
Forums 1,051 1,1k

Thread New to the game

  • 6 replies
  • 2 participants
  • 651 views
  • 0 follower
1 New to the game
First I'll come clean.
My real interest is not sound recording, it's testing audio equipment using a PC to control the test signals and detect the results.
I'm starting a project to generate audio tones, pass them through a system, digitise the output and record it. Sounds like a sound card application to me but there is one snag; the digitised data will be demodulated by sampling it synchronously with the generated tone. That means the digitising and the generating have to be locked together at some (preset by me) sampling rate.
None of the commercially available publicity material for sound cards tells you if this is possible and its also impossible to get at the card designers to find out whether the cards can be addapted. Also I'll have to program the thing when its set up and the card drivers never tell you whats possible before you buy.
Anyone out there who can help, please?
2
Richard-

The most obvious solution I can think of would be to generate the tone, split it in the analog realm, send one side to the gear under test and into the A/D converter and the other side straight to the A/D converter, then use the computer to compare the resulting wave forms. Any differences should be the result of the gear under test, assuming your cables are all good and high quality/low loss.

You were not real clear as to exactly what the desired end result is. Sounds to me like you're trying to build some sort of Test Tone Generator/Spectrum Analyzer. How about Tektronix or some other test equipment manufacturer?
The Axeman (##(===> Cuts From My New Blues CD
3
Axeman-

Absolutely right, it is a kind of spectrum analyser. I want to inject one frequency and investigate the level of second harmonic distortion produced, so your suggestion of splitting the analogue and subtracting the fundamental signal from the distorted signal before digitising is a good one; except that I'm not expecting any of the fundamental to get through so subtraction wont work. To minimise noise I'm relying on synchronous demodulation which is why I want to lock the sampling frequency (48kHz) to the generated fundamental (6kHz sine wave).

I've already built an analogue version from discrete components and defined the operating principle and I could go the route of assembling something similar from test instruments (eg Tektronix or National Instruments) but I thought I'd save a few pennies and make life interesting for myself by trying a sound card.
4
If you know the original frequency, and all you're looking for is the level of second order harmonic (assuming a pure sine wave), can't you set a filter up in software based on the math? either that, or work it out so that the original frequency DOES come through and treat it like a carrier?

I'm just a dumb non degreed technician, Navy trained. What you're wanting to do is really over my head, bro!! :lol:
The Axeman (##(===> Cuts From My New Blues CD
5
Hi there Axeman,

Not so dumb I think, you've seen whats happening so far. The project is something I've had going round in my head for a long time and I'm trying to make a spare time project out of it to investigate the science before I go public with what I've found. It's something to do with non-linearity in wound components (transformers and the like) when they saturate. I'm puting complex waveforms through the component under test so the level of the harmonics generated is varying on a short timescale.

The final intention was to gather data and then perform the maths in the host computer, using the DAC in the sound card as a data aquisition unit.
For me thats going to be the complex bit, I'm a hardware man through and through and completely blind to software. Thats why I was hoping to use a sound card driver to get the data into the computer and commercial analysis software to regenerate a waveform representing the distortion. Meanwhile the simple task of generating the carrier is handed over to the cards output channel.

My biggest problem is getting to talk to the hardware guys that design sound cards. All the cards I've looked at have been designed specifically to provide an easy route to getting audio information in and out of a computer and any spare functionality on the card is either not acknowledged or made inaccessible. I know all the components I need are in there, I just cant get a designer to tell me whether they can be reorganised to do what I want.
6
You mean a saturated transformer is not a flat line DC out?? Brother, you've lost me!!!! Or are you saying that the rate of stauration is not linear across all frequencies?

Not that it would make a difference in my ability to write software to analyze such a supposition!!!

Richard, you have fascinated me and peaked my interest, but I fear the subject material may be beyond me!!!!!
The Axeman (##(===> Cuts From My New Blues CD
7
The output from a fully saturated transformer is only flat if it remains fully saturated. If it's fed with signal that only saturates over part of a cycle then it genrates a distorted waveform that includes lots of harmonics.

Same process as limiting in a transistor or valve amplifier, only the distribution of harmonics is different. Transformers also have the problem of standing DC current causing partial saturation of the core and limiting the throughput of energy. Thats how saturable reactor voltage regulators work.

My software skills are VERY limited and I'll not be attempting anything other than sticking together standard function modules to get the effect I want. Not very efficient but nobody writes efficient code now that we have abaundant processing power in a laptop!

The subject matter is not beyond anyone with basic school physics, I'm just being a little obscure because it's a pet project of mine and I don't want to say too much until I think I know what I'm doing. I'm only going to get there if I stick to my original enquiry about using a sound card as a cheap way of doing data acquisition. Maybe I can't :oops: