Troubleshooting your Symasym

[keantoken]

And with the mod:

[anatech]

Hi keantoken,
Notice the higher levels of high frequency nasties?

-Chris

[keantoken]

Actually, I accidentally switched the pictures when I was saving them. I saw the difference when I first looked at my attachments and redid my simulation and resaved the files. Refresh your browser and look at the attachments again. Maybe there is something wrong with my software.

I figured out about odd and even harmonics after looking at a buch of threads and I believe that I can safely say that I know a little about how to gauge distortion from an FFT. FFT's are cool

[anatech]

Hi keantoken,
Now. If you can build one, you can confirm your predictions and then actually try your circuit changes and measure the results.

-Chris

[keantoken]

Any way of turning an oscilloscope into an FFT?
Or should the changes be audible? If I build one, it will have to be the one I use because I don't have that much money

[anatech]

Hi keantoken,
There are programs that can turn your sound card into an audio test rig (does FFT). RMAA comes to mind.

You need to condition the input signal for your sound card (ie attenuate or amplify the signal)

The Symasym sounds very good. It's worth building.

-Chris

[MikeB]

Keantoken, you might need to learn how to use FFT...
First, you have to skip the 1st cycle, for 1khz this would be skipping the 1st ms. (No print delay)
Second, the fft-window needs to be a multiple of cycles, for 1khz a multiple of 1khz.
Third, get rid of the onsemi models of MJL3281, they are useless (forbidden s-word). Try MJL0281 instead, these are functional.

Mike

[keantoken]

Oh, yeah! I didn't know about skipping the first cycle because it seems to be the "warmup" time of an amp. just a little while ago I discovered that I could make the timestep blank, I had it on 10u. Could this have been one of my problems? I will get the MJL0281 model. I will also look for that RMAA program (or is it a program?) I can also record sounds with Audacity and do an FFT on them, will this work?

Also, for conditioning the output, would 3.3 and 4.7 ohm resistors in series with a resistor connected from the center to the line-in work? (see ASCII)


| in
|
|
4.7
|
|-----R?----->out to line-in
|
3.3
|
|
==
=

[anatech]

Hi keantoken,
The program, RMAA is available on this web site.

You need to scale the output of the amplifier down to a signal voltage, so 10 : 1 or higher may be required. You need to correct this divider for high frequencies as well. Think of an oscilloscope probe.

-Chris

[keantoken]

Got RMAA...
By correcting it for high frequencies do you mean RF? I suppose this would involve capacitors. Is this because of bad resistors or something? Wouldn't a 10:1 voltage drop be around 7.5 ohms on top and .5 ohms on the bottom? I have two 10W .51 ohm resistors but I don't know about 7.5 ohms...
My oscilloscope has a probe impedance of 10M and a parallel capacitance of 47pF, are you talking about putting a capacitor in parallel with the whole divider?
I am still not quite enlightened on these technostuff terms