Square not as hollow as it used to be

[crx091081gb]

I have the x0xi0 front panel mod and I was comparing the second VCO to my original one and I've noticed the square wave on VCO 1 (original) is distinctly less hollow than VCO 2.

I've thought for a while my square wasn't as hollow as it used to be, any ideas how to debug this?

http://crx091081gb.net/squaretest.wav

Here's me playing F twice on VCO2 then on VCO1 and back again, 2 sounds hollower.

I found this

http://forums.adafruit.com/viewtopic.php?f=12&t=17935&p=91964&hilit=hollow#p91964

But I don't want to start swapping transistors till someone tells me I'm on the right track.

[antto]

you cannot judge by looking at just one frequency, because the 303 square wave is a very special waveform
basically it "morphs" depending on frequency, and also which rapid change of frequency (for example in the begining of a new note, where the previous note was a very distant pitch)

what you hear in this wav is that the pulse width on this specific note is not equal between the two VCOs
i'm really not familiar with x0xi0 (because i am not a fan of modifying the hell out of a 303) but i can tell you, it's very easy to make the square wave sound "different" because it is a waveshaper circuit very sensitive to the parameters:
1) sawtooth bias voltage
2) sawtooth peak-to-peak voltage
3) "12V" line and C11 / R45 time constant
4) "5.333V" line and R36

there is no "right" or "wrong" here
if you take 10 303s and 10 x0xes and measure and listen to the square waves on one specific frequency you'll probably hear a small difference too
your 5.333V line should be 5.333 if you want to use CV/Gate out
you can't change the 12V line much (i think in many x0xes it's actually 11.87V, no idea about the 303s)

a bunch of months ago i made the square shaper circuit on a breadboard, then rerouted the x0x sawtooth thru it, and back into the x0x, but i used potentiometers for R45 and R36
https://www.box.com/s/26zoiu21p06763usbkqt

a better thing to do is to record a reference pattern which has a range of notes, including low, mid, and high pitches

[crx091081gb]

Thanks for the detailed reply Antto. I haven't got my scope handy, the desks a mess and I should be making music instead of obsessing over details. Here's a sequence on VCO1 and then on VCO2 then repeated with the filter fully open.

http://crx091081gb.net/squaretest2.wav

Now that I'm listening to a sequence and not the raw oscillators I can't really tell the difference.

I think my problem with x0xb0xes is I built it myself and then obsessed over whether I got it sounding right.

Does this square sound normal?

[antto]

there is nothing wrong with them

[veracohr]

FYI, a square wave sounds most 'hollow' when the duty cycle is exactly 50%, and consists of only odd harmonics. As the duty cycle deviates from 50%, the harmonic content changes, including even harmonics.