Author |
Message |
adamstan
Joined: May 23, 2008 Posts: 34 Location: poland
|
Posted: Tue Jul 16, 2013 11:39 pm Post subject:
New DIY project - TOS polysynth :-) Subject description: A crossover between Polymoog and Polykorg |
|
|
Hello
I've always wanted to have the Polymoog, however getting one isn't easy - and getting working one is even more difficult , and that beast is HUGE. So I decided to build something similar myself. As with my first synth, this is mix of many designs. Currently I'm designing last boards, and going to order them today or tomorrow morning. They should arrive in two weeks, and then - soldering mayhem But, back to the design itself - what will it be?
- 61 keys
- Master oscillator section based on Polymoog - two VCOs driving two divide-down ranks using PLL
- Dividers built with CMOS counters - 4040 with diodes & 4024 for each note, like in TOG thread by AnalogCustom http://electro-music.com/forum/topic-45583-0.html&postdays=0&postorder=asc&highlight= - this is the biggest board
- Waveshapers based on Polymoog too
- Digital envelopes and LFOs
- there will be 4 LFOs, like in Polymoog - one for each VCO, one for PWM and one for filter
- VCF & VCA from Korg PS-3100. Why? Because it has lowest parts count, that's why
- Polymoog resonators and maybe some 4-pole master VCF at the output.
I will post current schematics when I get back home today. Here is waveshaper simulation done in falstad circuit simulator:
Capacitors have to be selected for each note to provide consistent saw level & PWM range across entire keyboard.
The VCF/VCA modules will sit on small "articulator" cards, that will plug vertically into 5 motherboards (one for each octave). Each mobo will also have associated waveshaper and envelope card. So it will look like this:
Green bars are the articulators, red one is waveshapers, and blue ADSR.
The nice thing about this design is that it will have no ASICs/proprietary parts - just off-the-shelf op-amps, CMOS and TTL cirucits. And a bucketload of transistors of course (almost 600) |
|
Back to top
|
|
|
adamstan
Joined: May 23, 2008 Posts: 34 Location: poland
|
Posted: Thu Jul 18, 2013 12:02 pm Post subject:
|
|
|
OK - here are some schematics, for those who like looking at such awful things
1. ADSR/LFO
2. Master oscillators/HFO
3. Frequency dividers
4. VCF&VCA-"voicecard"
|
|
Back to top
|
|
|
xpmtl
Joined: Aug 10, 2007 Posts: 162 Location: Brussels, Belgium
|
Posted: Thu Jul 18, 2013 10:55 pm Post subject:
|
|
|
Following this thread with much interest. Good job so far _________________ http://sdiy.xpmtl.net |
|
Back to top
|
|
|
Taylor
Joined: Jul 13, 2006 Posts: 170 Location: us
|
Posted: Sat Aug 03, 2013 7:35 pm Post subject:
|
|
|
Very interesting. Any chance you could post higher res schematics? The ones above are thumbnail size. _________________ musicpcb.com |
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004 Posts: 24079 Location: The Netherlands, Enschede
Audio files: 278
G2 patch files: 320
|
Posted: Sat Aug 03, 2013 8:22 pm Post subject:
|
|
|
Taylor wrote: | thumbnail size. |
Click on them .. and then click on that again .. blows 'm up here beyond imagination _________________ Jan
also .. could someone please turn down the thermostat a bit.
|
|
Back to top
|
|
|
Taylor
Joined: Jul 13, 2006 Posts: 170 Location: us
|
Posted: Sat Aug 03, 2013 8:45 pm Post subject:
|
|
|
Blue Hell wrote: | Taylor wrote: | thumbnail size. |
Click on them .. and then click on that again .. blows 'm up here beyond imagination |
For some reason for me this brings me only to
http://wstaw.org/m/2013/07/18/VCFVCA_png_300x300_q85.jpg
which is a 300p x 300p file. If I remove "_png_300x300_q85.jpg" and replace it with ".png I get the full file. Strange.
So is there a full voice card for each of the 61 notes? ....!!!!! _________________ musicpcb.com |
|
Back to top
|
|
|
adamstan
Joined: May 23, 2008 Posts: 34 Location: poland
|
|
Back to top
|
|
|
adamstan
Joined: May 23, 2008 Posts: 34 Location: poland
|
Posted: Wed Aug 14, 2013 11:45 am Post subject:
|
|
|
I started putting it together, and ran into big obstacle. The VCO doesn't oscillate at all It just ramps up to +15V and stays there... And the weirdest thing about it is, that even shorting the timing capacitor (C102/C202) with wire doesn't bring integrator output down... The circuit behaves exactly the same when I remove capacitor and discharge transistor (T101/T201).
The HFO part works, producing 5V square wave @~2MHz.
I've made another test, and it seems that integrator output is stuck on +Vcc, but I don't know why... That rise or ramp I've mentioned before is just rise time of my PSU - I've connected oscilloscope directly to PSU positive rail and got exatcly same result as when it was connected to integrator output. Damaged chip? In both channels? |
|
Back to top
|
|
|
adamstan
Joined: May 23, 2008 Posts: 34 Location: poland
|
Posted: Mon May 26, 2014 2:03 am Post subject:
|
|
|
I've recently returned to this project after long break. I've managed to fix VCO/HFO, and now I have whole oscillator section working, with TOS outputs following master VCO frequency
VCO wasn't working, because TL082 doesn't like to work on a single supply (U103/U203). Dave Luce used 3130 in this place in Polymoog. So I rewired this IC to split supply (+-15V) and put diode and pull-down resistor after comparator outputs, to block negative voltages. I've also replaced discharge transistors with ordinary NPN.
Next problem was, that even when I got VCO running, HFO wasn't following it. There was no signal on PLL return inputs (pin 3 of 4046). Why? Well, 4024 outputs are fully buffered, not OC, so pull-up resistors (R311/312/411/412) weren't enough to convert TTL levels to 15V, and 5V was too low to trigger 4011 gates powered from 15V. I had to add transistor there. |
|
Back to top
|
|
|
adamstan
Joined: May 23, 2008 Posts: 34 Location: poland
|
Posted: Tue May 27, 2014 11:49 pm Post subject:
|
|
|
I've just tested the filter - it works
http://www.youtube.com/watch?v=ZfIkoWUpIrE
Upper trace - square wave from divider, lower trace -"voicecard" output. I do some manual filter sweeps with and without resonance.
The high-pitched whine in background is crosstalk from VCO - it needs some proper decoupling |
|
Back to top
|
|
|
DES
Joined: Feb 28, 2003 Posts: 794 Location: New Jersey
Audio files: 8
|
Posted: Sat May 31, 2014 8:54 am Post subject:
|
|
|
Wow -thats an ambitious project! I might have to dig out my Mostek TOGs . Knew I was saving them for a reason!
Good luck with the build! _________________ Dave
www.davesneed.com |
|
Back to top
|
|
|
adamstan
Joined: May 23, 2008 Posts: 34 Location: poland
|
Posted: Fri Jan 16, 2015 1:24 pm Post subject:
|
|
|
My "motherboards" that had to house "voicecards" turned out to be too small - card sockets were too close, and adjacent boards wouldn't fit. While redesigning them, I've took the opportunity to make some other design changes.
I've redesigned filter (using op-amp buffer instead of FET) and waveshapers. Sawtooth one is similar, but simplified (removed diode and serial resistor from input, as they were unnecessary), while PWM is achieved by feeding differentiated square into comparator. I've also added pseudo-ringmod made with two diodes (it's essentially logical AND of divider outputs, which are TTL).
Waveshaping circuitry has been distributed between motherboards (differentiating network for making sawtooth reset pulse, and differentiating capacitor for PWM, and "ringmod") and voicecards themselves, leaving out separate "waveshaper" boards.
Now, while waiting for new PCBs to be manufactured, I work on software for EG and LFO.
I'll probably leave out PR1 and PR2 calibration pots, and replace first one with short, and second one with calculated resistor. As for PR3 - it was present in PS3100, and it may be useful to set uniform filter response across keyboard, ironing out some transistor mismatching.
This version of filter works much better (at least in PSPICE ) than the one with FET buffer, and TL082 in SO8 package uses similar space as TO-92 transistor, so even with added waveshaping circuitry I've managed to keep board pretty small - 1.4x1.1in (36x28mm).
Description: |
|
Filesize: |
111.33 KB |
Viewed: |
235 Time(s) |
This image has been reduced to fit the page. Click on it to enlarge. |
|
|
|
Back to top
|
|
|
adamstan
Joined: May 23, 2008 Posts: 34 Location: poland
|
Posted: Mon Jan 19, 2015 2:40 am Post subject:
|
|
|
Yesterday I've got ADSR module running, so I tested it with VCF/VCA module, and must confirm, that this simple VCA, as well as filter, when modulated by EG, indeed has remarkably low CV bleedthrough. And I haven't matched transistors at all - they're just from one batch.
BTW there is horrible error on ADSR schematics - DAC has its power pins mixed up, so there was some trace-cutting involved. Also, during tests, I've destroyed three perfectly good AVRs, replacing them due to non-working PORTC.2-PORTC.5 pins, 'cause I forgot, that on new chips you have to disable JTAG to make these four pins operational... |
|
Back to top
|
|
|
|