the x0xb0x sequencer is quite easy. if you buy a mainboard and ioboard (or etch your own/produce your own), you can just cut the mainboard in half. the analog and digital sections are entirely seperate. the digital section doesnt have a lot of ICs on it, but has all the switches and LEDs, so compone...

all of the exponential converters ive seen use tempco resistors to balance out the VT error (or use feedback heaters to keep a constant T). are there ones out there that use transistor absolute temperature generators (Vptat)? its definitely less complicated to use a tempco resistor, and probably jus...

i missed the part about the bleed only being at high frequencies. one thing you can try, is soldering a matching 47pF cap across the other 130k resistor on the opamp. without this, there is an imbalance at high frequencies (maybe 2kHz and up).

check the 130k resistors around the opamp for matching as well. pretty much anything being imbalanced with put it off.

ok, so new vcf almost done. why is it that the vcf is always the biggest pain? at any rate, here is the old vcf, if anyone was curious. also, i believe its just as important to publish mistakes as successes, as there are things to learn from each. http://www.openmusiclabs.com/files/vcf.png nothing t...

so youll be happy to know i am fixing the VCF to get rid of the noise, and a few other things. i spent this week going over everything with a fine toothed comb, and decided the VCf just wasnt good enough. i had a 3-pole diode ladder filter (low-pass, self resonant) with resonance knob, cut-off knob,...

yeah, thats a debate i have with myself constantly, which is one reason why things are taking so long. i spend more time debating than doing layout. my goal here was to see how inexpensive i could make a modular synth that was usable. i would say a fair bit of the noise was the recording. it was goi...

yeah, the clicks were actually intentional on the testers behalf. for some reason he was feeding a square wave lfo straight to the output.

the noise floor is a bit of the recording, and a bit of the vcf. at high resonance, there is a lot of noise in the vcf because of the low-end opamps im using.

ok, just to prove im actually still working on this, here is small song played on it by one of my beta testers:

http://www.openmusiclabs.com/files/modular_jamz.mp3

its mostly the vco/vcf/random sequencer.

im working on the next pcb rev right now.

the wav trigger might do the trick:
https://www.sparkfun.com/products/12000

always good discussions over at mutable. i saw that you were talking a bit about the tradeoff between PWM frequency and bit depth. did you consider dual-PWM as a solution to this? what bit depth would you say your mutli-timer method achieved?

A is log and B is linear

wednesdayayay wrote:I haven't used it in a while but I really enjoy my openmusiclabs repatcher in combination with max/shbobo shnth/aalto/illucia application suite

thats great to hear!

unfortunately, we retired the rePatcher. i think there are one or two still kicking around somewhere if people are interested.

baloo wrote:x0x-heart is out of stock.

they are probably getting back from the factory next week, and then a week of QA or so, and then off to synthcube. sorry for the delay.

one thought, is that with the waveshaping, you can get all the other waveforms simultaneously, without any extra processing or DACs. although it does require the sync pulses and such to get pinned out from the micro. perhaps there is also some preference to how it sounds?

after digging around a bit (the LFO_speedCV buss changes name somewhere in the schematic), i cant find any other connections except to the DAC as well. so it looks like it is digital.

why digital pots over dacs? there are 6 usable PWM outputs on an atmega328, so if you need more than that, an external chip is required (which you need anyways with PWM - opamps). but digital pots also require amplifiers to convert the resistance to a voltage. some DACs already have these amps built...

it nominally supports 1V -> 5V (same as 303), but can usually go down to around 0.5V and up to 10V, although the tracking isnt that great above 6V.

this is a description of using a CDTA as a universal filter: http://mwrf.com/passive-components/single-cdta-leads-multiple-filters you have to rewire where you send the input, so not any sort of magic, but still interesting. and it has a good transistor level schematic of how one might build a CDTA ...

definitely interesting, thanks for pointing them out. it looks like they took a differential current to voltage converter, and used it to drive a differential output OTA (with appropriate linearization for Vgs). they will be good at high speeds (RF), but ill have to think about it a bit to understan...

thank you, may i have another!

ive actually made some progress, and i think almost all of the circuit is finalized. i just need to do a rev on the pcb.

ive seen this vary pretty dramatically between instruments, but usually 50mV to 100mV is a good range.

you could just go all discrete then. make a 4-bit folding ADC out of a quad opamp. or maybe one of those led bar graph drivers and a priority encoder.

this sounds interesting. what sort of "modular system" are you thinking about? hardware/software? controller/synthesizer?

it doesnt sound like you need a really fast conversion, or audio-quality clock jitter performance, which makes things pretty easy. the arduino would definitely suffice for this. and you wouldnt be reliant upon someone else if you wanted to mod things. if you end up with an lsb that toggles with each...