RQ: someone with an oscilloscope...

[guest]

i double checked my calculations
and you are right
there is no 75hz highpass before the vcf

the reason for the asymetrical waveform
on the square wave circuit
is that c11 begins to droop at low frequencies
and c10 doesnt have as much effect at low frequencies
the c10 effect also accounts for the curved rising edge
of the square wave at low frequencies

[antto]

i'm still trying to come up with a model of this - and nothing works
every time the algorithm gets too complicated - smells like a wrong approach

btw something inverts the sawtooth in the shaper
the sharper edge on the square is the sharper edge in the sawtooth, but on the square it is inverted

[guest]

yes
the shaper is an inverter

ill write up an equation for it later this week
is it ok if its a function of jw and V
or does it need to be only of V

[antto]

yes
the shaper is an inverter

ill write up an equation for it later this week
is it ok if its a function of jw and V
or does it need to be only of V

i have no idea what you mean with "jv"

the biggest problem is - i don't know how it looks raw
i plugged a probe into the square output, this time thru a 100K resistor, to my soundcard's LINE-IN
it seems it doesn't mess with the square too much, but the main problem is - it's f*ckin high-pass filtered, and i have no clue if the actual bias is changing either

if i know how it looks at a few different frequencies - i'll have a better chance of thinking up a closer algorithm

another option is to try and "undo" the high-pass filter from the recording, but that's tough..
and there's a problem, in the recording - there is noise and the nasty power hum.. this makes it even harder to "undo" it
and if my recording is "clipped" somewhere (for example in the audio card) - the "undo" will explode

[Kritical Audio]

I have a Tektronix scope Maybe I can help? Tell me where to set the probe/probes and I'll try tomorrow.. I think jw stands for jumper wire?

[antto]

I have a Tektronix scope Maybe I can help? Tell me where to set the probe/probes and I'll try tomorrow.. I think jw stands for jumper wire?

probe at R62 (right before the filter input) or if it's easier - on the waveform switch itself, it should be the same

take pictures of the square at (around) 16, 24, 32, 40, 80, and 160Hz
there's no need to be exact, just when you take the pictures write down the exact frequency it detects (if possible) as a reference

you can use Keyboard mode and the TUNE knob to tweak the frequency

i'm also concerned about the VCO circuit being very "fragile" i mean.. it's very easy to interfere with it's normal behaviour
i am not familiar with oscilloscopes but, as far as i know, the "probe" must have an amplifier and drain as little as possible from the circuit
this is why i used a 100K resistor when hooking it to the soundcard, it still interferes but not as much as when i plug it directly, i would add even more resistance but then the signal is burried in the noise floor (it's a on-board sound card after all, it has plenty of noise)

anyway, it'll be great if you can do this.. thank you

[antto]

i just ran a recording thru my "undo" filter which should theoriticaly restore the original signal before the DC-coupling of my audio card
here's a screenshot:

i recorded both the sawtooth and the square wave from the waveform switch at the same time, both thru 100K resistors, right into the LINE-input

anyway
now as i look at this - it looks so simple and i already got some ideas
what i was doing before, now seems totaly wrong
from the picture, it looks like the bottom side of the square wave is simply the top side of the sawtooth, not inverted! it's "tilted" with the same angle of the sawtooth, no matter at what frequency
so for this part of the square - it seems the transistor and the other RC elements do NOT affect anything

guest: can you verify this?

also, it looks like the sawtooth has no bias here
notice how my "undo" filter captured the bias in the square wave, but the sawtooth is centered around 0
i'm guessing it _is_ ran thru a HP filter somewhere after all

[guest]

the bottom of the square wave should be flat
and both the square and saw are dc coupled before c17

the square wave has four sections
flat bottom
rising edge with variable attack time
the lower the frequency the longer the attack
a slowly drooping top section
with more droop at lower frequencies
and a sharp falling edge

[guest]

here is my falstad export
it seems pretty accurate

$ 1 5.0E-6 10.20027730826997 78 5.0 50
t 320 224 288 224 0 -1 3.9060717001723395 0.5672672547949205 300.0
r 288 208 288 160 0 22000.0
r 288 240 288 304 0 10000.0
r 320 224 368 224 0 100000.0
r 320 224 320 176 0 10000.0
c 320 176 368 176 0 1.0E-8 0.12171641550847845
c 288 208 240 208 0 1.0E-6 8.671804446560614
g 240 208 240 224 0
R 288 160 288 128 0 0 40.0 11.8 0.0 0.0 0.5
R 368 176 416 176 0 4 130.0 -2.6 8.5 0.0 0.5
w 368 224 368 176 0
R 288 304 288 336 0 0 40.0 5.333 0.0 0.0 0.5
O 288 240 256 240 0
O 368 224 416 224 0
o 12 16 0 34 16.77994891013191 0.0 0 -1
o 13 128 0 34 20.0 9.765625E-5 1 -1

[antto]

in my previous picture - i forgot to run the SAW channel thru the "undo" filter so it was actually the original recording
i re-processed both channels with the undo filter and here's how the sawtooth looks like now:

pretty odd..

anyway

both the square and saw are dc coupled before c17

you mean there is another HP filter _before_ c17/r62 ???
where? is it also before the square shaper?

in my recording maybe the waveforms are inverted or something
here's how i *think* it works:
the first part is the beginning of the sawtooth
it looks like it's simply passed thru, as it's "tilted" the same way
it's flat too
secont part:
this is the "soft" transition to the top, there is a LP filter involved here
it looks like it's the transistor now being active and "inverting" the sawtooth with some bias
the "softness" turns into yet another "flat" side at lower frequencies
third part is the hard flat top side

i got an approximitation now and it pretty much looks like the thing, but i'm only plotting it right now, who knows will it behave right on different frequencies..
here's some pseudo code:
consider "x" being a sawtooth of 5.5V peak to peak (without the bias, so it's centered around zero)
if (x > 0)
{
y = x/2
LPF1 = y
LPF2 = y
}
else
{
z = LPF2((x * -1) + 3)
z = LPF1(z)
if (z > 4.5) { y = 4.5 }
else { y = z }
}

the first part is the "tilted" (bottom) side of the square, it's pretty much the sawtooth itself
a special thing here, i set both LP filters to the output value "y"
this is equivalent of discharging the capacitors in a LP setup (i think so)

then, at one point the other part of the code is used
i add +3 bias to an inverted copy of the sawtooth, this would now "tilt" it in the opposite direction compared to the first portion of the square
the transition between the previous part of the code and this part is softened a little by LPF2 (i set it to ~160Hz)
then the output of this is also filtered by LP1 which is the main reason for the smoother "attack" (again ~160Hz)
and finaly i clip (limit) this to +4.5
this is the "flat" side (top) of the square
it ends when when the sawtooth resets, with a sharp transition and everything repeats

this is just a very basic approximitation, and i'm not really happy with LPF2 there.. it doesn't have the desired effect..

here it is ploted at ~14Hz:

[antto]

guest: you're a genius
and you barely put a few pieces

yes, this indeed has the properties of the square, at different frequencies too
hm, it looks so simple
now i'll be staring at it for hours, thanks!

[guest]

you can play with the transistor beta
and the saw tooth wave offest and amplitude
to really fine tweak it

[antto]

okay, now it seems the duty cycle has nothing to do with the pitch-CV

when i put a scope on the transistor - i can actually see the same waveform i recorded from the x0x, with the first part "tilted"
i'm still staring at it and trying to figure it out, the transistor is still a mistery to me..

[antto]

what triggers the transistor? whatever it is - it is the reason for the duty cycle and it has an exponential behaviour across frequency, at lower frequencies it triggers it sooner than the "middle" of the sawtooth, while on higher frequencies it goes towards the middle (and iirc past the middle)

[guest]

the sawtooth triggers q8
by pulling current through its base
because c10 and r134 act as a highpass filter of sorts
more current will get drawn out the base at higher frequencies
perhaps triggering it sooner

but r45 and c11 act as a lowpass filter
to find the average value of the saw wave
which is how it finds the center to make the flip
so this counteracts the above action
im sure it also gives some interesting effects on transients