Q: VCO pitch range

[antto]

uhm, wait
C21 and C20 are part of the VCA indeed, but they have an effect with the filter
and, it's not a "simulated TB-303"
with 2.2uf for C21 my calculator says 0.32Hz
and for C20 -> 0.72Hz
i'd say it'll be more bassy, but it will probably be ugly, i can try it..

[antto]

http://www.box.net/shared/1p4fzf8gaq

the first part is with the "2.2uF"
as you can see, it has more low freqs, the waveforms look ugly

second part - original values 72Hz for C21 and 159Hz for C20
sounds good with the distortion too, unlike the previous one

[bcbox]

C20 & C21 are the coupling caps to the VCA. Check out page 14 of the x0xi0 I/O manual for actual frequency response measurements. 0.1uF will give an extra octave. The VCA can distort a bit more because of the extra low freq. energy but it is a very worthwhile mod. C38, C59 etc are fine with stock 1uF values. C15 is the cap in the resonance feedback loop that was mentioned a few posts back.

[Brassteacher]

C20 & C21 are the coupling caps to the VCA. Check out page 14 of the x0xi0 I/O manual for actual frequency response measurements. 0.1uF will give an extra octave. The VCA can distort a bit more because of the extra low freq. energy but it is a very worthwhile mod. C38, C59 etc are fine with stock 1uF values. C15 is the cap in the resonance feedback loop that was mentioned a few posts back.

Ok, maybe I'm not as crazy as I thought. Still, I've got to stop trying to do things that require somewhat more complex thinking at 4 AM . As for the VCA distortion, is it due to the opamp running out of headroom, i.e., clipping? If it can handle higher frequencies ok it shouldn't be an issue with rise/settling time. Does it distort with the output taken directly off the VCA, or could the LA4140 possibly be the culprit?

@antto, when I typed "simulated TB-303", based on how quickly you ran a simulation with different cap values and checking waveforms, I assumed you were running a Spice model. While I'm thinking about it, in your software synth, have you modeled the Johnson noise (The rest of you, get your mind out of the gutter! ) and Shot noise that would be present in a hardware synth? Just curious.

[antto]

Brassteacher: it's not a spice simulation, it's a SynthEdit VST
i don't have "capacitors" in there, i have HP and LP filters that mimic some Resistor/Capacitor parts
and i use a calculator to get the Frequency (in Hz) depending on the RC values
i'm not an electronic geek, i can't even solder.. i bought the x0xb0x to be able to improve my synth (like probing the oscillator to see the actual waveform, which DID help now, thanks to guest)

i don't know what you mean with these noises, maybe i have to google it first

[bcbox]

As for the VCA distortion, is it due to the opamp running out of headroom, i.e., clipping?

correct.

Does it distort with the output taken directly off the VCA, or could the LA4140 possibly be the culprit?

good question. yes it's the VCA distorting, will happen mostly on accents because of the extra control current that's sent to the VCA. However, the mixer stage can also distort, and it distorts asymmetrically. It does this even with a stock 303 when volume is max and accents are used.

[Brassteacher]

Brassteacher: it's not a spice simulation, it's a SynthEdit VST
i don't have "capacitors" in there, i have HP and LP filters that mimic some Resistor/Capacitor parts
and i use a calculator to get the Frequency (in Hz) depending on the RC values
i'm not an electronic geek, i can't even solder.. i bought the x0xb0x to be able to improve my synth (like probing the oscillator to see the actual waveform, which DID help now, thanks to guest)

i don't know what you mean with these noises, maybe i have to google it first

Yeah, I realized it was your softsynth after your last post, I was just saying that I had assumed at first you were talking about a spice model.

Johnson noise is noise generated by current passing through a resistor, or resistive element. The higher the resistance value, the more noise is generated. Also, in general metal-film resistors tend to generate less noise than an old-fashioned carbon composition resistor. How audible is it? Well, I had a friend and fellow repair tech that learned the hard way to NOT replace the vintage carbon resistors in an old Fender tube guitar amp with "better" more modern metal-film resistors. He learned this after the amp owner came back to the shop a few hours after picking up the "repaired" amp, and was quite clear in describing just what methods he was planning to use to remove very specific parts of the tech's body if the original resistors were not replaced in an expeditious manner. What does it sound like? To me it is like really subtle wider-band pink noise, hard to describe really, it's more noticeable when comparing two identical or very similar circuits, one built with all metal-film resistors and cermet trimpots, and the other built with all carbon composition resistors and carbon trimpots.

Shot noise is the noise generated by electrons moving through the various layers of a semiconductor. The more you have, the more noise. In general, the higher the gain of the device is asked to produce, the more noise you get. Also, the type of semiconductor has an effect on how much noise is generated. For instance, (in general) JFETs tend to be fairly quiet, and zener diodes are far, far noisier. This is why a lot of analog synth noise generators use zener diodes to, well, generate the noise. Shot noise can be fairly easy to hear in certain conditions, and hearing just what it sounds like is easy. Just find a synth with an analog noise generator and sample/model that, or find an older stereo amp or receiver that has a phono preamp built in, set it to the phono input (if the phono input jacks have shorting plugs, remove them), and crank the volume with no signal present.

BANNED adding that noise in may help you get to your goal of making your softsynth sound even more realistic, if you haven't added a bit of that noise in already. How much to add may take some experimentation, since the noise is more noticeable when it is successfully lessened in a piece of gear (at least to my ears). It's sort of like being in a building with fluorescent lighting and not noticing how loud the lights are until someone turns them off.

[antto]

trying to make the synth actually sound and behave like a 303 was the chalange, as it's not easy to do
if you just put a saw/square for an OSC, and some filter, and some envelopes - this won't work
every part of this little synth is unique, maybe only the sawtooth is near a simple boring sawtooth
the hugest part is the sequencer, which i'm still working on, but thank's to my n0nx0x firmware, i got inspired and rewrote my softsynth's sequencer to a point where i can copy/paste code FROM n0nx0x, i mean huge pieces of code
the noise.. it's not so important at the moment
i DID plan to put some main's hum into the oscillator (i once played with it http://www.box.net/shared/7924v2q504 )
but that's for later
noise could be needed only in the filter
as the real filter DOES have noise too, and it has an effect - when the filter has lots of feedback (hacking the R97 value) this noise also triggers selfoscillation

[antto]

[bump]
well, i just tryied this..
in n0nx0x i had an offset of 23 for the notes (which was equal to how the stock firmware was written)
to have one full octave more, i changed this value to 15

then, i retuned the x0x (TM4 only)
the higher pitches are fine (up to the "upper C key" + Up+Up + play-mode-transpose-Up that's gotta be C-6 or 987.7Hz according to my calculator)
unfortunately, as you already have warned me - the very low notes got detuned

someone earlier mentioned replacing an AN6265 (or whatever it was) with a LM358
which IC is that exactly? (i guess the one near the VCO section)
[/bump]

[guest]

ic11 is the one to replace
the lm358 might give some improvement
but if not enough
try tlv2372 or tlv272 or tlc272
probably in that order

[antto]

ahum, i do have 2 or 3 spare LM358(N) parts (i'm not quite sure if they actually work)
unfortunately, my IC11 is directly soldered onto the PCB (it's a 6562)
i have such a socket which i can try to rip off from my old mainboard PCB
so the tlc272 is probably the best i can try?

[guest]

with some solder wick and flux
you should be able to get the old one out
without damaging it

you can try the lm358s you have on hand
and see how they do before getting the other ones
the tlv2372 would be my first try
although they are pretty similar opamps

how far down does the an6562 go
before it starts sounding off

[antto]

well, in my firmware, i've set the base pitch to 15 (it was 23, equal to the stock x0x OS) ... that's the normal "c" key..
put a Down on it and you get 15-12 = 3 (the lowest c key possible from the firmware) .. i mean, i haven't pushed this all the way down to note zero
so, the first 3-4 or 5 notes sound off (i haven't measured it precisely)

i don't have the wick thing, i'll have to do it with the iron only (i guess it'll be almost impossible but i have to try)
i'll try to desolder the socket from the mainboard first, to see if there is any chance to desolder this..
if i can't, or if it's risky - i won't touch the 6562 on the x0x
wish me luck

and, thanks guest ;]

[guest]

i dont think the lm358 will give much improvement
but the tlv2372 should be able to get the whole range
right on down to note 0

if there is a local electronics shop
it might be worth picking up some wick
or a desoldering pump
they are both pretty inexpensive

[antto]

yeah, well it wasn't much "local" but about 30 minutes away from home
sadly, i went last week to that shop to get some sockets and pots and other stuff - the shop was gone
gotta find another one
they didn't had any wick or pumps

i desoldered the socket from my motherboard, but i broke 2 legs
then i ripped apart my old VCR.. no other socket
i'll have to buy one from somewhere (my nose tells me it's gonna be hard to find both wick/pump and sockets around here.. grrr)