Author |
Message |
Taylor
Joined: Jul 13, 2006 Posts: 170 Location: us
|
Posted: Tue May 11, 2010 1:24 am Post subject:
Developing my CMOS organ concept... Subject description: Where Farfisa meets Lunetta? |
|
|
I started this thread a little while ago. I've now started fleshing out my idea more seriously. The basic plan as I envision it goes like so:
So I start with an inverter generating a square wave. The resistor will be a pot to allow live retuning (actually probably 2 pots: coarse and fine tuning). I'll have 12 of these inverter oscillators. Each one will go to its own 4040 divider chip to generate between 4 and 6 octaves. The outputs of these dividers, which will still be square waves, will go out to an individual 2-pole opamp filter per note, tuned to give a sine wave. Then all these sine waves will be mixed with active mixers.
Anyone care to critique this concept? In particular I'm wondering about the mixing: I'm assuming I can do a certain number of notes into a single summing amp, but probably not all 50+ notes into one opamp, so what's the limit here?
And the keying: I realize this could be improved upon, but I'd like to keep it simple since whatever I do here will be duplicated 50 or more times. I don't mind a slight click, but will this setup have more than just slight clicking?
The Lunetta begins to come in in the details: the 40106 from which I derive the oscillators will have a pot to control power voltage, and I'm thinking of using SPDT DIP switches to allow the unfiltered square waves through without the filtering if I so choose. The tuning pots will allow me to tune this organ to just intonation or other unusual tunings. I'm just getting into synth DIY so I'm very excited about this project but also not sure how well it will work. |
|
Back to top
|
|
|
v-un-v
Janitor
Joined: May 16, 2005 Posts: 8933 Location: Birmingham, England, UK
Audio files: 11
G2 patch files: 1
|
|
Back to top
|
|
|
slabman
Joined: Sep 01, 2005 Posts: 102 Location: UK
|
Posted: Tue May 11, 2010 9:26 am Post subject:
|
|
|
If you want to solve all the usual organ problems (noise, crosstalk, bleed) then you'll end up building a conventional organ with all that that implies in having keying circuitry and so on. If you want an experimental drone machine to produce cluster type sounds then this approach will work. But be warned - it's a lot of wiring! |
|
Back to top
|
|
|
Rykhaard
Joined: Sep 02, 2007 Posts: 1290 Location: Canada
|
Posted: Tue May 11, 2010 10:13 am Post subject:
|
|
|
I had a similar idea a couple of years ago for building myself a string machine. I did some tests on the pitch stability of the CD40106 with 1% metal film resistors and poly (something - not styrene) caps.
Over a 48 period, testing 3 oscillators on a single chip, the maximum frequency chance from them was about 0.5%.
However - that COULD have been because the power driving the chip, was a +9V battery. That could have been the change.
The air temperature in the room would have been changing over that period as well, whether it were tested in Summer (open windows) or in Winter (furnace cycling).
I DO think that it could be possible with that chip. As were just mentioned above, just a heck of a lot of wiring and trimmer adjustments in setup. |
|
Back to top
|
|
|
Taylor
Joined: Jul 13, 2006 Posts: 170 Location: us
|
Posted: Tue May 11, 2010 12:37 pm Post subject:
|
|
|
v-un-v wrote: |
You need to take a look at the Davoli Davolisint (as used by Cluster/ Harmonia and also, I believe, by Soft Machine). Sound generation was based on TTL chips (a bit like CMOS). |
Interesting, however my goal here is full polyphony.
slabman wrote: | If you want to solve all the usual organ problems (noise, crosstalk, bleed) then you'll end up building a conventional organ with all that that implies in having keying circuitry and so on. If you want an experimental drone machine to produce cluster type sounds then this approach will work. But be warned - it's a lot of wiring! |
Could you elaborate a little, or point me towards an explanation of these pitfalls? To what extent would these issues be likely to crop up with this arrangement? I don't mind a little noise but I would like to be able to play it in a "conventional" organ style in addition to experimental drone stuff.
It is of course going to be a large circuit, however, among the very few things at which I consider myself pretty skilled is PCB design. I would build this on a professionally manufactured PCB I'll design, with just about everything, including the momentary SPST keys and the pitch pots, mounted on the PCB. So there will be almost no offboard wiring at all.
So, ok, it sounds like this setup should work for my needs. I guess I'll get going on it and work out the details as they come up.
Rykhaard wrote: | I had a similar idea a couple of years ago for building myself a string machine. I did some tests on the pitch stability of the CD40106 with 1% metal film resistors and poly (something - not styrene) caps.
Over a 48 period, testing 3 oscillators on a single chip, the maximum frequency chance from them was about 0.5%.
However - that COULD have been because the power driving the chip, was a +9V battery. That could have been the change.
The air temperature in the room would have been changing over that period as well, whether it were tested in Summer (open windows) or in Winter (furnace cycling).
|
I wonder if temperature compensating resistors could be used to provide a little more pitch stability? In my case, with the tuning pots on the front panel, tuning won't be expected to be perfect, since I'll probably be retuning all the time anyway. I come from a string instrument background, so we're used to tuning up before playing anyway. Last edited by Taylor on Tue May 11, 2010 10:11 pm; edited 1 time in total |
|
Back to top
|
|
|
slabman
Joined: Sep 01, 2005 Posts: 102 Location: UK
|
Posted: Tue May 11, 2010 1:16 pm Post subject:
|
|
|
I guess the single biggest source of problems is the keyboard - this is where biggest culprit in generating clicks, induced AC hum, crosstalk & so on. That's why more modern organ designs only put control signals through the keys and use those signals to switch the signals through to the output via keying circuits. Once you introduce keying circuits you have the possibility of attack/decay control. Juergen Haible's polyphonic Korg clone is a good source of info for that - one transistor VCAs, etc.
Also, filtering each note to get a sinewave - if the filter is tightly tuned enough to produce a sinewave, and you detune your divider chain to a different intonation, the filter will no longer be centered on the note and you'll get a volume drop. Might be worth considering waveshaping rather than filtering. There's probably a simple CMOS circuit knocking around that will give you 6 shapers from one IC. |
|
Back to top
|
|
|
Taylor
Joined: Jul 13, 2006 Posts: 170 Location: us
|
Posted: Tue May 11, 2010 1:34 pm Post subject:
|
|
|
slabman wrote: | I guess the single biggest source of problems is the keyboard - this is where biggest culprit in generating clicks, induced AC hum, crosstalk & so on. That's why more modern organ designs only put control signals through the keys and use those signals to switch the signals through to the output via keying circuits. Once you introduce keying circuits you have the possibility of attack/decay control. Juergen Haible's polyphonic Korg clone is a good source of info for that - one transistor VCAs, etc.
Also, filtering each note to get a sinewave - if the filter is tightly tuned enough to produce a sinewave, and you detune your divider chain to a different intonation, the filter will no longer be centered on the note and you'll get a volume drop. Might be worth considering waveshaping rather than filtering. There's probably a simple CMOS circuit knocking around that will give you 6 shapers from one IC. |
Ah, yes you're totally right. Don't know why that didn't occur to me, but yes, changing tuning will either change waveshape or volume... I'll check into CMOS waveshapers. |
|
Back to top
|
|
|
droffset
Joined: Feb 02, 2009 Posts: 515 Location: London area
Audio files: 2
|
Posted: Tue May 11, 2010 2:33 pm Post subject:
|
|
|
This is something I've thought about too, I think it's a great project. Here are some thoughts:
I've been looking for ways to do a diode matrix keyboard scanning circuit using a salvaged toy keyboard(full sized keys!). Ray Wilson has a solution that I haven't tried yet:
http://www.musicfromouterspace.com/analogsynth/KEYENCODERPCB/KBDMATRIX022005.html
For the key switching check out the CMOS chip 4066, could be very useful since it's a set of digitally controlled dip switches.
Is there a good crystal that can produce the proper frequency to divide down into your tones? If you can find a crystal that gets one note close enough you can use something like this to divide the various semitones:
http://electro-music.com/forum/post-245362.html#245362
For mixing all those signals... I'm not sure, maybe groups of keys can have their own opamp mixers? Is there a way to vary the output level based on the intensity of the input levels? Maybe a microcontroller can control that? _________________ ==================
Check out the FREE Intro to Lunettas Document
https://docs.google.com/document/d/1V9qerry_PsXTZqt_UDx7C-wcuMe_6_gyy6M_MyAgQoA/edit?usp=sharing
Edit: Spelling mistakes. |
|
Back to top
|
|
|
parasat
Joined: Aug 13, 2009 Posts: 33 Location: chicago
|
|
Back to top
|
|
|
Taylor
Joined: Jul 13, 2006 Posts: 170 Location: us
|
Posted: Wed May 12, 2010 2:47 pm Post subject:
|
|
|
That's very interesting, but too complex to duplicate for all 50-60 notes. I think I'm going to go ahead with the single-opamp filter on each note and just accept the drawbacks.
Originally I envisioned the ability to tune each oscillator across a large range for lots of weirdness, but as noted above this will not work because of the filters on each note. Instead, I'll just have each oscillator slightly detunable, so I can do different intonations, but these slight tuning shifts won't make a huge difference with the volume or waveshape. |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Wed May 12, 2010 3:31 pm Post subject:
|
|
|
You could always tune each filter to it's own note and then accept the key-follow distortion as one of this instruments inherent "features". Sounds great to me. People often add some type of key follow filtering anyway, this will just be a different type. Maybe look at changing to an LPF? _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
Taylor
Joined: Jul 13, 2006 Posts: 170 Location: us
|
Posted: Wed May 12, 2010 3:35 pm Post subject:
|
|
|
Uncle Krunkus wrote: | You could always tune each filter to it's own note and then accept the key-follow distortion as one of this instruments inherent "features". Sounds great to me. People often add some type of key follow filtering anyway, this will just be a different type. Maybe look at changing to an LPF? |
Right, that's basically what I'm thinking.
I believe that the opamp circuits shown in my picture above are 2-pole lowpass filters. Or did you mean leaving the oscillators as squares and then tacking a single LPF on the output of the entire circuit? |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Thu May 13, 2010 1:06 am Post subject:
|
|
|
Sorry, I assumed they were BPFs. What I really meant was to just look at different topologies if the one there doesn't "do it" for you. Maybe even use different ones across the keyboard? _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
Taylor
Joined: Jul 13, 2006 Posts: 170 Location: us
|
Posted: Thu May 13, 2010 1:24 am Post subject:
|
|
|
Yep, that may turn out to be necessary. Would love to get the parts count down a bit but I guess this is just how it goes for an instrument like this.
What's SRE, by the way? |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Thu May 13, 2010 10:41 pm Post subject:
|
|
|
"Special Religious Education" _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
Top Top
Joined: Feb 02, 2010 Posts: 266 Location: California
|
Posted: Fri May 14, 2010 1:40 pm Post subject:
|
|
|
I have had a plan to do basically the same thing, minus the filtering (don't care about it being sine waves, would be fine with just some filtering on the end).
Having the 12 master oscillators tunable by pots with coarse and fine tuning should not be too much of a problem really if you don't mind tuning the organ up at the beginning of a show/recording. It should not drift too much if you have a stable power supply (I've looked at Rykhaard's tests of 40106 oscillators).
My plan would be basically 12 oscillators into 12 octave dividers, and have all notes (for me probably just a few octaves, 2-3) just switched on/off via home-made spst switches (keys), and at least to start, I will just use resistors to mix and a pot for attenuating output volume (which would be plenty high if running off 9 or 12V). If it is messy sounding, I will go from there in terms of cleaning it up with propper mixing, etc...
I am interested to see what you come up with if I don't beat you there... |
|
Back to top
|
|
|
Taylor
Joined: Jul 13, 2006 Posts: 170 Location: us
|
Posted: Fri May 14, 2010 1:52 pm Post subject:
|
|
|
Cool Top Top, I've seen your stuff over at DIY Stompboxes and it's always interesting. I'm "Taylor" over there.
You'll probably beat me to it. My project will be rather huge, and right now I'm trying to figure out how to get in onto a PCB that's reasonably priced. I'm thinking of doing the front panel as a professionally fabbed PCB as well, so I don't have to do all that drilling and labeling, and I can get metallic lettering, etc. I would be happy to offer PCBs of this project to anyone interested, however they will end being at least $50 each because of the size unless we can get more than 2 or 3 people interested. I also think they may end up using surface mount chips to get the size down, which I know a lot of people are turned off by the prospect of soldering SMT chips. |
|
Back to top
|
|
|
Rykhaard
Joined: Sep 02, 2007 Posts: 1290 Location: Canada
|
Posted: Fri May 14, 2010 2:21 pm Post subject:
|
|
|
Taylor wrote: | Cool Top Top, I've seen your stuff over at DIY Stompboxes and it's always interesting. I'm "Taylor" over there.
You'll probably beat me to it. My project will be rather huge, and right now I'm trying to figure out how to get in onto a PCB that's reasonably priced. I'm thinking of doing the front panel as a professionally fabbed PCB as well, so I don't have to do all that drilling and labeling, and I can get metallic lettering, etc. I would be happy to offer PCBs of this project to anyone interested, however they will end being at least $50 each because of the size unless we can get more than 2 or 3 people interested. I also think they may end up using surface mount chips to get the size down, which I know a lot of people are turned off by the prospect of soldering SMT chips. |
Initial interest is there, from me, so far.
In my experiences at work and slightly, at home with SMD's ... keep your soldering tip at an absolute max of 700ºF as well as using a VERY small pointed tip. (I use a 1/32nd pencil point, normally anyways.) Flux the pads first; a miniscule drop of solder on 1 pad (for a 2 pad SMD); place the part with tweezers over top of the solder; flux the part and pads and then heat the solder portion whislt holding the part. Cool it down with your breath and apply solder to the other pad.
In my using this method, except for a couple of times at work where I held the iron there for longer than about 2 seconds, I've never had a problem.
550-600ºF is a bit safer. |
|
Back to top
|
|
|
Taylor
Joined: Jul 13, 2006 Posts: 170 Location: us
|
Posted: Fri May 14, 2010 2:40 pm Post subject:
|
|
|
Yeah, I actually don't find SMD chips too difficult to solder. I use some in my commercial effects. I don't use any SMD resistors or caps, usually 1/8 watt resistors, which are nearly as small as SMD Rs but a little easier to keep track of.
I'll keep you guys updated of the PCB situation, perhaps we can get something reasonably priced if we can convince a few others that they need a weirdo analog organ... |
|
Back to top
|
|
|
Top Top
Joined: Feb 02, 2010 Posts: 266 Location: California
|
Posted: Fri May 14, 2010 11:56 pm Post subject:
|
|
|
I've only dealt with SMD when modifying commercial effects. I think am too clumsy to deal with it. As soon as I touch the resistor it seems it is stuck to the end of my iron.
PCB is probably not for me... I need to chase down something much simpler or I will never get to it.
Sounds like a really cool project though! |
|
Back to top
|
|
|
adambee7
Joined: Apr 04, 2009 Posts: 420 Location: united kingdom
|
Posted: Sat May 15, 2010 12:24 am Post subject:
|
|
|
when you solder SMD components its best to prepare the surface first so its flat. Say your soldering a resistor, Put a small bit of solder on one of the pads and let it cool. With a pair of tweezers, pick up the resistor and with the iron, heat up the pad with solder on it and position the resistor in place. Your not actually touching the component with your iron. When in place solder the other end as normal and if need be quickly reheat the other end. You'll get used to it pretty quickly. Use a flux pen if you need to aswell. Its not as bad as you think it just takes practice. |
|
Back to top
|
|
|
Rykhaard
Joined: Sep 02, 2007 Posts: 1290 Location: Canada
|
Posted: Sat May 15, 2010 6:29 am Post subject:
|
|
|
adambee7 wrote: | when you solder SMD components its best to prepare the surface first so its flat. Say your soldering a resistor, Put a small bit of solder on one of the pads and let it cool. With a pair of tweezers, pick up the resistor and with the iron, heat up the pad with solder on it and position the resistor in place. Your not actually touching the component with your iron. When in place solder the other end as normal and if need be quickly reheat the other end. You'll get used to it pretty quickly. Use a flux pen if you need to aswell. Its not as bad as you think it just takes practice. |
Aye! I forgot about that! It IS much easier to _NOT_ touch the SMD with your iron. It CAN get 'carried away'. Otherwise as you do it, aye. |
|
Back to top
|
|
|
synthmonger
Joined: Nov 16, 2006 Posts: 578 Location: flada
Audio files: 3
|
Posted: Wed May 26, 2010 3:24 pm Post subject:
|
|
|
Similar idea here too! I decided on using 4 gates from the 40106 though. Using any inverter next to one another will cause annoying crosstalk. I decided on using my VCO version of the 40106 oscillator and just simple wave shaping for each vco. They'll also have a summed cv input for coarse and fine tuning. _________________ Youtube!
modular demos!
Whacky tunes! |
|
Back to top
|
|
|
Taylor
Joined: Jul 13, 2006 Posts: 170 Location: us
|
Posted: Wed May 26, 2010 4:14 pm Post subject:
|
|
|
Sounds cool, do you have a preliminary schematic or is this all in your brain at the moment? |
|
Back to top
|
|
|
Rykhaard
Joined: Sep 02, 2007 Posts: 1290 Location: Canada
|
Posted: Wed May 26, 2010 7:35 pm Post subject:
|
|
|
synthmonger wrote: | Similar idea here too! I decided on using 4 gates from the 40106 though. Using any inverter next to one another will cause annoying crosstalk.. |
Bingo! I forgot about this exactly. Not - is there crosstalk between the inverters across the chip, so that 4 oscillators would not be a good idea?
Another way that I've gone with this, that I would if it would still provide crosstalk is 2 oscillators (single inverter) on 1 side of the chip, and 1 on the other, with each of them, using the next inverter as an output buffer. Could this have the same crosstalk effect?
I've got cross talk all OVER the place in my Atari Jazz Console. (40106 clocks and 4023 clocks, with all of them using 40106s as output buffers.)
I believe it would. A possible solution if so then, would be 3 or 4 clocks on 1 x 40106, with the buffers on a 2nd one.
Testing to be done ... |
|
Back to top
|
|
|
|