Are there any analog phase modulation VCOs?

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!
Dogma
Super Deluxe Wiggler
Posts: 4324
Joined: Wed Aug 28, 2013 10:46 pm

Post by Dogma »

Man that looks beautiful. I'm becoming addicted to these panels...as a a primarily a euro user these panels conformity is comforting..

Be amazing if you could do a TZPM/TZPM osc - possible from the Tillman/Hutchins schems?
User avatar
guest
Super Deluxe Wiggler
Posts: 8243
Joined: Mon Aug 19, 2013 11:49 am

Post by guest »

just wanted to add some information that i recently found. im finishing off my OTA sine waveshaper paper (finally, its only been a year of procrastination), and collecting all the sineshaper papers i can find, when i came across a really sweet circuit by barrie gilbert. it was used in the AD639 trigonometric function IC:

http://www.analog.com/media/en/technica ... /AD639.pdf

there are 2 cool things about this: 1. it has -74dB THD, which is better than all options except OTA with cusp cancelling, and digital techniques. 2. it can modulate to +/-500 degrees in its current state, and the circuit can be expanded with more stages to whatever number of degrees you want.

what is particularly interesting here, is that it opens up the possibility of doing PM without messing with the integrator in any way (no TZVCO required). you can merely add your phase after the VCO and send it to the waveshaper to deal with the wrap around. there may be some caveats here, i havent fully thought it through yet, or built one up.

the AD639 is long out of production, but the circuit isnt too horrible to do in discrete, and an OTA version is possible as well, but requires a lot of OTAs. not sure if its any easier than a TZVCO, though.
openmusiclabs.com
User avatar
wsy
Super Deluxe Wiggler
Posts: 2728
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy »

guest wrote: ...... what is particularly interesting here, is that it opens up the possibility of doing PM without messing with the integrator in any way (no TZVCO required). you can merely add your phase after the VCO and send it to the waveshaper to deal with the wrap around. there may be some caveats here, i havent fully thought it through yet, or built one up.

the AD639 is long out of production, but the circuit isnt too horrible to do in discrete, and an OTA version is possible as well, but requires a lot of OTAs. not sure if its any easier than a TZVCO, though.
The AD639 seems pretty amazing.

Can you share your discrete schematic? :-)

- Bill
"Life is short. But we can always buy longer patch cords" - Savage
User avatar
guest
Super Deluxe Wiggler
Posts: 8243
Joined: Mon Aug 19, 2013 11:49 am

Post by guest »

heres the basic layout, from his paper on the chip:

Image

there are 2 different ways of going about it (plus variants on these, which are listed in the patent). the top one is the most straightforward, and makes the operation easier to understand. basically, each diffpair is seperated by pi*Vt volts from each other, which is a large voltage, so only one diffpair is active at a time, except for a small period during changeover. each diffpair does the usual tanh function, approximating 180 degrees of a sinewave. for each diffpair, you get another 180 degrees. the diffpairs can be replaced with OTAs if that seems easier. the lower schematic is a reduced transistor count version that only uses 1 transistor per 180 degree section. seeing as only 2 diffpairs are ever interacting at any point in time, there could be a circuit which would take a look at the current position, and move the less active diffpair to the side its most likely to transition to. there would inevitably be some crossover distortion in this scheme, and you would need comparators for all the breakpoints (or a folding ADC of some kind).
openmusiclabs.com
User avatar
wsy
Super Deluxe Wiggler
Posts: 2728
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy »

guest wrote:heres the basic layout, from his paper on the chip:

Image

there are 2 different ways of going about it (plus variants on these, which are listed in the patent). the top one is the most straightforward, and makes the operation easier to understand. basically, each diffpair is seperated by pi*Vt volts from each other, which is a large voltage, so only one diffpair is active at a time, except for a small period during changeover. each diffpair does the usual tanh function, approximating 180 degrees of a sinewave. for each diffpair, you get another 180 degrees. the diffpairs can be replaced with OTAs if that seems easier. the lower schematic is a reduced transistor count version that only uses 1 transistor per 180 degree section. seeing as only 2 diffpairs are ever interacting at any point in time, there could be a circuit which would take a look at the current position, and move the less active diffpair to the side its most likely to transition to. there would inevitably be some crossover distortion in this scheme, and you would need comparators for all the breakpoints (or a folding ADC of some kind).
That.... is breathtakingly beautiful.

- Bill
"Life is short. But we can always buy longer patch cords" - Savage
User avatar
guest
Super Deluxe Wiggler
Posts: 8243
Joined: Mon Aug 19, 2013 11:49 am

Post by guest »

yeah, that is one of the coolest circuits ive seen.

ive reworked it into a 4 transistor version that has less than 0.05% error over +/-180 degrees, or less than 0.3% at +/-360 degrees. there is still the issue of temperature drift, which is fixable, but starts the push the circuit complexity into the realm of not really worth it.
openmusiclabs.com
User avatar
J3RK
Super Deluxe Wiggler
Posts: 7673
Joined: Tue Dec 22, 2009 12:46 pm
Location: Seattle

Post by J3RK »

Isaiah wrote:Any news on this, Dustin?
There is now actually. The test PCBs are working quite well considering it's my first try with a sawtooth core, and with the saw animator, and trying to improve on things a bit.

I've got nearly perfect sawtooth waves now. There are some VERY tiny spikes post-animator, but I actually think I can get those to almost completely disappear. I'm sure they'll still show a tiny bit on the scope, but they should be inaudible with a few more tweaks.

The triangles are looking good, but since I was dumb, and didn't add trims, I need to tweak a few resistor values to get the symmetry just right.

The sine shapers are working very well when one considers that they're receiving asymmetrical triangles. :hihi:

The VCOs track VERY well, and the phase modulation is working very well overall. However, post-animator, post summing sines, while relatively smooth, are a little bit misshapen. I'm hoping to clean that up a bit too.

Even with the current levels of "dirt" I'm getting some pretty decent PM sounds out of it. Using the multiple envelopes via the VCAs, and cross patching Phase inputs between the two VCOs, I've gotten some pretty nice sounds.

I'll be working on getting the waves cleaned up as much as possible over the next few days, and then I think that's about it for this one.

What I intend to do, is compare it with the Linear FM of my newly updated Mirror Core VCOs (which arrive next week.) I'm really curious how closely the two methods match up sound-wise. Especially since the LFM will sound a bit cleaner, and use much less components.

I'd like to try some of the other PM ideas mentioned in this thread too, but that will have to wait for a few other projects I've got going on right now.

I'll post some sounds after I go through the "cleaning" process. :party:
Are you lemon? Does your head come to a nub?
User avatar
neil.johnson
Super Deluxe Wiggler
Posts: 1961
Joined: Thu Apr 08, 2010 3:51 pm
Location: Cambridge, UK
Contact:

Post by neil.johnson »

The AD639 was designed a long time ago. If you think that's beautiful then try it from the horse's mouth as it were:
[video][/video]

I tried out that circuit in LTSpice and it is rather good. Barrie has the advantage of having all transistors on the same piece of silicon.

Neil
Random ramblings :: http://www.njohnson.co.uk
SSI parts and more :: https://www.cesyg.co.uk
User avatar
J3RK
Super Deluxe Wiggler
Posts: 7673
Joined: Tue Dec 22, 2009 12:46 pm
Location: Seattle

Post by J3RK »

I'm working on eliminating one last little buzzy artifact right now, but things are coming along decently well now. The sines post phase shift aren't the cleanest in the world, but the shapes are good. I don't think it's going to be possible to get super clean sines after shifting the sawtooth, shaping to triangle then shaping to sine then summing the sines. I think they're nearly as good as they can be given this. Anyway, this is a quick phone-video, so not the greatest.

I have one phased/summed sine out modulating the phase of the other VCO, and that VCO's non-phased sine coming back to modulate the first. One is indexed via VCA, then the remaing phase inputs have various function generator outputs bouncing them around. The notes are coming from the arpeggiator on the keyboard controller.

[video][/video]
Are you lemon? Does your head come to a nub?
User avatar
J3RK
Super Deluxe Wiggler
Posts: 7673
Joined: Tue Dec 22, 2009 12:46 pm
Location: Seattle

Post by J3RK »

I think I've fixed the last of the glitches.

[video][/video]

[video][/video]
Are you lemon? Does your head come to a nub?
User avatar
fma
Common Wiggler
Posts: 239
Joined: Sun Oct 18, 2015 11:25 pm
Location: Seattle

Post by fma »

That sounds great!
User avatar
J3RK
Super Deluxe Wiggler
Posts: 7673
Joined: Tue Dec 22, 2009 12:46 pm
Location: Seattle

Post by J3RK »

Thanks!

There were quite a few little component tweaks to get the glitches to go away, as well as an additional couple of trims added. I'll be laying out one more version of the PCB that has all this in place. I also want to lower the trim values to "proper" trim ranges with series resistors, so that only the useful range of the trim is adjustable.

I also plan to add an inverting/summing amplifier to the Linear FM input, so multiple CVs can be used there. The clips above are only phase modulation. It can get very complex with a bit of linear FM added in, so I want to make that as flexible as possible, so different combinations can be used. You can get similar sonic results using one or the other, but I find when both are used in varying amounts it adds a bit more color than either alone provide.

I'm also going to add a summed sawtooth output, and possibly use the unused half of the LM393 comparator to add a PWM output.

The original plan was to completely focus this on being like a DX/OPL operator, but I've found a few situations where some additional wave shapes might be kind of fun.

One thing that's cool is that using the bipolar sawtooth output, you can get a nice 360 degree looping modulation out of the summed sine output. This is really smooth and pleasant sounding IMO. I'm curious what using the phase modulated and summed sawtooth would sound like doing the same modulation while also being modulated using the static saw from the other VCO.

I actually got the sine waves to come out a lot cleaner than I thought I would. They're still not exactly "world class" :hihi: (well, the ones after the phase modulation section anyway) but they sound quite good now. The non phase modulated sine is actually very clean.

It's a pretty big board at 3x6, and I'll be adding one or two more ICs, but I think it provides a lot of variety in the sounds it makes, so it's worth the space. I'd actually like to bump it up to 4x6 to spread things out a bit since it's so dense, but I'll probably try to keep it at 3x6 to save space behind the panel.

I've made a list of all of the refinements that I've done, and am going to update the schematic today. I'm more than happy to post it here once the updates are in, since I've borrowed most of the elements from Terry Michaels, Ian Fritz, and Bernie Hutchins / Lester Ludwig. What I've done is mostly refinements in the wave shaping sections.
Are you lemon? Does your head come to a nub?
User avatar
J3RK
Super Deluxe Wiggler
Posts: 7673
Joined: Tue Dec 22, 2009 12:46 pm
Location: Seattle

Post by J3RK »

Update:

Fixed everything. If you grabbed the previous schematic, please grab this one, and use it instead.
Attachments
operator.pdf
(286.43 KiB) Downloaded 139 times
Last edited by J3RK on Tue Feb 28, 2017 1:18 pm, edited 7 times in total.
Are you lemon? Does your head come to a nub?
User avatar
wsy
Super Deluxe Wiggler
Posts: 2728
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy »

J3RK wrote: I really like it, though I have started to think that this amount of work to emulate one voice of a $1 IC that most people hate makes me giggle a bit. :lol:
That's exactly how I feel about the FM Ogre ... except that I had to puzzle out DMA interrupts on the PIC33 to get it to work
with full magic enabled. :-)

Great sound, isn't it? :guinness: :guinness: :guinness:

NB: It seems that finally, close to 40 years after it was introduced, that TZFM is finally heading to modular. It's
about time, too.

- Bill
"Life is short. But we can always buy longer patch cords" - Savage
User avatar
J3RK
Super Deluxe Wiggler
Posts: 7673
Joined: Tue Dec 22, 2009 12:46 pm
Location: Seattle

Post by J3RK »

wsy wrote:
J3RK wrote: I really like it, though I have started to think that this amount of work to emulate one voice of a $1 IC that most people hate makes me giggle a bit. :lol:
That's exactly how I feel about the FM Ogre ... except that I had to puzzle out DMA interrupts on the PIC33 to get it to work
with full magic enabled. :-)

Great sound, isn't it? :guinness: :guinness: :guinness:

NB: It seems that finally, close to 40 years after it was introduced, that TZFM is finally heading to modular. It's
about time, too.

- Bill
It is indeed. Your Ogre is one of the few modules on my list to build at some point. I'm thinking I'd like to do a Loudest Warning format 4U module with it. (along with a few others) I love the sound of this brand of FM. I used to have a ton of tracker music that was done in EdLib (DOS based OPL tracker ) and some of it was REALLY good. Same with some old arcade games that used the YM chips.
Are you lemon? Does your head come to a nub?
User avatar
MrNezumi
Veteran Wiggler
Posts: 536
Joined: Sun Jun 19, 2016 8:12 am
Location: Out of Bounds

Post by MrNezumi »

J3RK wrote:Here's the schematic for the VCO. The Phase CV input resistors may need to be tweaked by a few ohms this way or that to get a perfect 360 degrees using a +/-5V sawtooth input. On my boards 3K does the trick. I can't really put in a trim for the feedback side as that resistor will interact with the others in the phase circuit...... (more tech talk)
I read and reread this whole thread hoping I could grasp the theory, but failed. That in itself doesn't bother me too much, but it stops me from understanding how to (physically/actually) use it.

Does there need to be two of the phase oscillators for this thing to work it's magic or could one phase oscillator be paired with another oscillator (the Mirror one or Ultrafade or Stroh Pico or MOTM-300?) Is the section in between the osc sections a fancy ringmod or does it serve a purpose specific to PPM oscillators?
User avatar
wsy
Super Deluxe Wiggler
Posts: 2728
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy »

Have you tried patching the PMed sine out right back into PM In (with an attenuator, of course).

That should make some really really _big_ sounds that are pretty much ready-to-use into your final ENV + VCA.

EDIT: This is a trick that WILL NOT WORK on TZFM, only on TZPM. Reason is that there's a stable equilibrium point at
about 45 degrees in the differential equation that accurately describes TZFM with self feedback (and this happens in real
life too; try it with a Rubicon. If you patch the FM modulated output back into TZFM in, and crank up the modulation pot, the
Rubicon will *stop* *oscillating*. It's not a flaw; the Rubicon is ddoing exactly what the math of TZFM says to do.

TZPM has no such stable equilibrium, so that's actually a little bit preferable for musical applications.

- Bill
"Life is short. But we can always buy longer patch cords" - Savage
User avatar
Isaiah
Super Deluxe Wiggler
Posts: 1825
Joined: Wed Jun 02, 2010 6:31 am

Post by Isaiah »

J3RK
Fantastic work, Dustin!
Sounds ace!

Is there any chance of including a Phase-shifter Bipolar Saw output and the pair of Triangle outputs?
That'd be ace if possible.
User avatar
J3RK
Super Deluxe Wiggler
Posts: 7673
Joined: Tue Dec 22, 2009 12:46 pm
Location: Seattle

Post by J3RK »

MrNezumi wrote:
J3RK wrote:Here's the schematic for the VCO. The Phase CV input resistors may need to be tweaked by a few ohms this way or that to get a perfect 360 degrees using a +/-5V sawtooth input. On my boards 3K does the trick. I can't really put in a trim for the feedback side as that resistor will interact with the others in the phase circuit...... (more tech talk)
I read and reread this whole thread hoping I could grasp the theory, but failed. That in itself doesn't bother me too much, but it stops me from understanding how to (physically/actually) use it.

Does there need to be two of the phase oscillators for this thing to work it's magic or could one phase oscillator be paired with another oscillator (the Mirror one or Ultrafade or Stroh Pico or MOTM-300?) Is the section in between the osc sections a fancy ringmod or does it serve a purpose specific to PPM oscillators?
No, you don't need to have two of them. It is kind of cool to be able to take a phase modulated sine, and modulate the phase of the second VCO with it, but it's definitely not necessary. I would say that whatever you use in conjunction with it should have a 10V P2P Sawtooth and Sine at least.
Last edited by J3RK on Thu Feb 23, 2017 11:22 am, edited 1 time in total.
Are you lemon? Does your head come to a nub?
User avatar
J3RK
Super Deluxe Wiggler
Posts: 7673
Joined: Tue Dec 22, 2009 12:46 pm
Location: Seattle

Post by J3RK »

wsy wrote:Have you tried patching the PMed sine out right back into PM In (with an attenuator, of course).

That should make some really really _big_ sounds that are pretty much ready-to-use into your final ENV + VCA.

EDIT: This is a trick that WILL NOT WORK on TZFM, only on TZPM. Reason is that there's a stable equilibrium point at
about 45 degrees in the differential equation that accurately describes TZFM with self feedback (and this happens in real
life too; try it with a Rubicon. If you patch the FM modulated output back into TZFM in, and crank up the modulation pot, the
Rubicon will *stop* *oscillating*. It's not a flaw; the Rubicon is ddoing exactly what the math of TZFM says to do.

TZPM has no such stable equilibrium, so that's actually a little bit preferable for musical applications.

- Bill
I just discovered this last night actually! While tuning in a few sounds, I noticed that there was a very defined zero-point, which made me determine that I need to raise the value of my fine tuning summing resistor. Very cool sounds in there. I didn't actually patch the PM sine back in, but ran the same V/Oct signal to both VCOs then fine tuned both. Similar, but having the exact same timing in the patch you described would be perfect. I'll give that a whirl tonight.
Last edited by J3RK on Thu Feb 23, 2017 11:22 am, edited 1 time in total.
Are you lemon? Does your head come to a nub?
User avatar
J3RK
Super Deluxe Wiggler
Posts: 7673
Joined: Tue Dec 22, 2009 12:46 pm
Location: Seattle

Post by J3RK »

Isaiah wrote:J3RK
Fantastic work, Dustin!
Sounds ace!

Is there any chance of including a Phase-shifter Bipolar Saw output and the pair of Triangle outputs?
That'd be ace if possible.
Without switching to SMT or adding an inch to the board, that may be difficult. The board is already pretty dense at 3x6. The shifted sawtooth might be kind of handy to have. The triangles less so IMO. The main point for me was to get some DX/YM/OPL-ish types of sounds, as there are a ton of VCOs that do everything else, and better in some cases. If it will fit, I'll toss another buffer on there somewhere.
Are you lemon? Does your head come to a nub?
User avatar
J3RK
Super Deluxe Wiggler
Posts: 7673
Joined: Tue Dec 22, 2009 12:46 pm
Location: Seattle

Post by J3RK »

Another phone video... sorry... I'll do some proper recording after I get it into the case. Anyway this one uses WSY's suggestion of patching the phased sine back into the same VCO, plus sync with a second VCO with all manner of cross phase modulation. I've also got a four quadrant multiplier patched in. :cloud: Most of the changes are just the fine tune of the modulating oscillator.

Anyway:

[video][/video]
Are you lemon? Does your head come to a nub?
User avatar
fma
Common Wiggler
Posts: 239
Joined: Sun Oct 18, 2015 11:25 pm
Location: Seattle

Post by fma »

:yay: :yay: :yay: :yay:

drool worthy.
User avatar
wsy
Super Deluxe Wiggler
Posts: 2728
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy »

Ayup, you found the TZ phasemod magic. 8-) 8-) 8-)

And you have an excellent front panel too.

Are you going to GPL your schematic, or go commercial?

- Bill
"Life is short. But we can always buy longer patch cords" - Savage
User avatar
J3RK
Super Deluxe Wiggler
Posts: 7673
Joined: Tue Dec 22, 2009 12:46 pm
Location: Seattle

Post by J3RK »

wsy wrote:Ayup, you found the TZ phasemod magic. 8-) 8-) 8-)

And you have an excellent front panel too.

Are you going to GPL your schematic, or go commercial?

- Bill
Thanks!!

Actually, I'll be making PCBs, and using it in some custom work, but the schematic is available for anyone to use however they like. All of the component functions are more or less readily available and common. I just fine tuned them to work together here.

:party:
Are you lemon? Does your head come to a nub?
Post Reply

Return to “Music Tech DIY”