MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

OBX voice card (with non-custom CEM3310 equivalent circuits)
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page Previous  1, 2, 3, 4, 5 ... 67, 68, 69  Next [all]
Author OBX voice card (with non-custom CEM3310 equivalent circuits)
EMwhite
Quick question for Old Crow;

'ad a butchers at your BOMs and wondered what the total cost would be for 8 * voice cards + the host board parts. I'm not pushing for a price on boards, just interested in cost of parts at this point 8_) And a + or - $50 estimate is fine, just looking for a ball park so I can start to set $$ aside.

And, a question: I don't see any microcontrollers or CPUs on the host board, and I like that (ALOT). Can you confirm that this beast will 'just' be all knobby/button-ey, no code, memory, etc.?

Thank you!
kynsi
The host board contains a PIC...
The Real MC
There is no PIC for system control. I helped with schematic peer review on the host board.

The host board is pure CV control from knobs. No microcontroller, no code, no memory, no patch storage.
Nordcore
w/o controller it would never get polyphonic.
Minimum controller needs 8 CV outputs for the pitch + 8 gates, everything else might be controlled by pots and/or external CV inputs.

The control board in the obx contains the modulation sources:
- one LFO (tri, square, S&H)
- pitch-wheel
which are routed to the modulation destinations:
- pitch
- filter pitch
It contains the noise source.
All of this can be replaced by simple inputs, where you just patch a modulation from some modules and adjust the mod amount with a pot.

The control board contains the CV generation for the autotune. This could either be a battery of 16 trim pots - or you rely on properly calibrated voice boards an just ground that signal.
So all of this can be done in "simple" analogue just with pots - the only thing you loose is patch storage.

Whats left is the "glide" circuitry. This applies an adjustable slope to each of the eight note pitches. Than the generation of the eight note pitches themselves, and the eight gate signals.
This could be done by a very simple midi to cv converter:



P.S.: the software for a similar design is already done (only 4ch but with some more bells 'n whistles, see my avatar picture... )
oldcrow
You are probably looking at about US$1000 in parts for 8 voice boards and 1 host board. This does not include the bare boards or a front panel.
I ordered $620 in parts this week, but I had some parts on hand already.

There is one PIC part on the host board for the noise generator. Specifically, this noise source: http://www.electricdruid.net/index.php?page=projects.noisegen

Everything else is analog save for a 4011 gate chip used to select if a VCO is muted or set to a waveform.

Crow
/**/

EMwhite wrote:
Quick question for Old Crow;

'ad a butchers at your BOMs and wondered what the total cost would be for 8 * voice cards + the host board parts. I'm not pushing for a price on boards, just interested in cost of parts at this point 8_) And a + or - $50 estimate is fine, just looking for a ball park so I can start to set $$ aside.

And, a question: I don't see any microcontrollers or CPUs on the host board, and I like that (ALOT). Can you confirm that this beast will 'just' be all knobby/button-ey, no code, memory, etc.?

Thank you!
oldcrow
My initial "rev1/prototype" setup is designed to operate as a 4-voice system, with an externally-provided MIDI to CV/gate unit (I have an MOTM-650 for this). At the moment, I am only setting it up to test a single voice.

After that test phase concludes, I will be making a "slot adapter" (see the L-shaped board farther down this thread) and a 4-card carrier board. The carrier board bolts onto the back of the host board where a single voice card can currently mount, and by using the slot adapters 4 voice cards can mount to the carrier. The carrier board will have the four glide/portamento slew circuits as well as a panning mixer and volume VCAs similar to the original OBX. There will need to exist a breakout panel for the 4x CV and 4x gate inputs, as well as the left, mono and right audio outputs. I may try to incorporate the host board's local porta/glide circuit and volume VCA, but it depends on how much cross-wiring this entails. The host board was originally meant to test single voice cards after all. ;)

Crow
/**/
home_listening
Thanks for all the updates, following keenly.

The BOM is handy too, cheers.

keep up the good work!
oldcrow
I update the BOM pdfs as I find errata (wrong resistor values, swapped reference designators, etc) so check it every few days ;) --Crow

home_listening wrote:

The BOM is handy too, cheers.
oozitron
Is there any reason to use 5% resistors (as indicated in the BOM) instead of 1% everywhere?

Will 5% resistors add to the (slight) note-to-note differences (imperfections) in tone that make older synths sound better?

Or is that a silly question meh

Drew (who has a lot of the required resistors, but they're all 1%)
oozitron
CS80 Website note:

The links to the two BOM files are backwards. Figured it out after spending several minutes wondering why the Voice board needed 21 pots!

Dang this is exciting screaming goo yo

Drew
oldcrow
Oops, fixed. Thanks SlayerBadger!

oozitron wrote:
CS80 Website note:

The links to the two BOM files are backwards. Figured it out after spending several minutes wondering why the Voice board needed 21 pots!

Dang this is exciting screaming goo yo

Drew
oldcrow
It won't make a difference worth noting. I just use 5% because they're cheaper. --Crow

oozitron wrote:
Is there any reason to use 5% resistors (as indicated in the BOM) instead of 1% everywhere?
The Real MC
BTW if you want the "Tom Sawyer" rezz, that sound requires at least four voices in unison. You won't get that sound with just one voicecard.

A big part of that sound is the inherent filter scale error and EG timing tolerances between the voicecards. The filter scale doesn't match between voicecards due to the cheap expo converter in the filter cutoff CV summing circuit. Slight EG timing differences between voicecards due to tolerances in 3310s and/or Cx charge caps also add to tracking error.

The service manual also says that filter tracking is linear only within a three octave range - exceed that range and there is no guarantee that tracking will match between voicecards. This is not hard to do with EG filter modulation, which certainly is what the "Tom Sawyer" rezz is doing.

So because the filters between voicecards aren't tracking in scale, this creates a rather unique flanging effect that is the soul of that rezz sound. It took some fiddling around trying to get "that sound", but once I hit the UNISON button it was WHOOAAAHHH! I tried the same effect on my Memorymoog and Andromeda, but the tracking is "too perfect". Well!

AFAIK, Scott has retained the cheap expo converter at my urging. This is one of those circuits where imperfections can be a good thing.

The later Oberheims got more accurate tracking with the move to CEM3320s and software envelopes. Whether that is a positive improvement is highly debatible...
The Real MC
This is the guilty party that inspired Scott's project. Restoring one of these is no walk in the park!
EMwhite
The tolerances (read, varied tolerances) is presumably what gave it that great sound. Hard to mimic in a VST but these guys tried: http://www.sonicprojects.ch/obx/separatevoicedesign.html

I was so intrigued that I actually dumped the $150 or whatever it was to buy it, painstakingly fought with Wine an an ASIO drive to run under xwindows on my MacPro and succeeded. How does it sound? Well, it sound good to me but surely not the same.

The 'better' versions of OPX like Pro and Pro II have fancier modulation and other features like up to 12 voices and built in effects but I don't like it.

Funny thing... I'm such a snob that I won't use a period inaccurate version of a virtual synth's representation of an actual synth and certainly won't use it's digital reverb hihi

Really looking fwd to this build but I think I said that already. By the way, I was in touch with Olivier @ Mutable and was told that MIDIpal will do > 4 voices, in fact, it will do 16 voices worth of Note distribution. It will even 'wrap' so if your starting channel is 10, it will go 10,11,12,13,14,15,16,1. Not sure that any of the Midi->CV converters will do this but the thought (my evil plan) is either to pickup a second MOTM-650, or to build two quad ACXSynth interfaces.
oldcrow
Giant bag of resistors arrived. Now I have to go to Target and get a couple Stanley Sortmasters as I am out of bin box space. Guinness ftw!
aladan
oldcrow wrote:
Giant bag of resistors arrived. Now I have to go to Target and get a couple Stanley Sortmasters as I am out of bin box space. Guinness ftw!


I've started using these:

http://www.cheapaschips.com.au/estore/style/ha6095.aspx
http://www.cheapaschips.com.au/estore/style/ha6094.aspx

Life is so much easier with storage drawers thumbs up
oldcrow
Oh I have plenty of drawers, never mind I have this thing:

http://www.uline.com/Product/ProductDetailRootItem.Aspx?modelnumber=H- 1428

I just find these days I like to organize things by project, and the Stanley Sortmaster is great for this, given they can be locked shut and toted around. USD$9 at Target. Just came back with two. 8_)

http://www.stanleytools.com/default.asp?CATEGORY=ZAG+ORGANIZERS&TYPE=P RODUCT&PARTNUMBER=014026R&SDesc=Stanley%26%23174%3B+SortMaster+Organiz er

Crow
/**/

aladan wrote:
I've started using these:

http://www.cheapaschips.com.au/estore/style/ha6095.aspx
http://www.cheapaschips.com.au/estore/style/ha6094.aspx

Life is so much easier with storage drawers thumbs up
Nordcore
EMwhite wrote:
Not sure that any of the Midi->CV converters will do this but the thought (my evil plan) is either to pickup a second MOTM-650, or to build two quad ACXSynth interfaces.

If there are about ten people wanting one, I could make an pcb for this: https://www.muffwiggler.com/forum/viewtopic.php?t=83363&start=78
(Size 50x100mm (2" x 4" ), pcb picture is just a draft. )
As said, firmware for four channels *does* run, and changing that to 8 is easily done. (It was written with flexible channel numbers in mind. )
There is an boot loader /flasher for the atmel, so the firmware is field up-gradable. (Computer with capability of sending midi sys-ex required. )
Costs would be round about 15€, including programmed atmel (first programming had to be done with a hardware programmer) and world wide shipping (registered mail).
mOBiTh
I'm definitely in for building an 8 voice and midi-cv converter for it.


I wonder what 16 voices would sound like hmmm.....

hihi
demian
It sounds better.
oldenjon
hyper
EMwhite
Nordcore wrote:
If there are about ten people wanting one, I could make an pcb for this: https://www.muffwiggler.com/forum/viewtopic.php?t=83363&start=78
(Size 50x100mm (2" x 4" ), pcb picture is just a draft. )


I'd be interested in seeing a list of 'features'/capabilities and certainly would want the 8-voice version. Is this based on a piece of code that you already have tested and running? (may take two as I have a separate SEM 4-voice project that I'm cobbling together).

Tell me more...
Isaiah
oldcrow
Apologies if I'm asking something obvious, but is there a 2-/4-pole switch for the VCF, or did you decide not to implement that?
The Real MC
*raises hand for eight voices plus MIDI->CV converter*
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page Previous  1, 2, 3, 4, 5 ... 67, 68, 69  Next [all]
Page 4 of 69
Powered by phpBB © phpBB Group