gate level with pwm

[jrp]

As described in my other posts, i want to use the seq3 for my analog drumbox (and more, of course).

The drums are to be triggered and velocity controlled.

Do you think there is a (reasonable) way to do velocity controll with the trigger signal itself?

Instead using a logic trigger and a cv for velocity im thinking of pwm to get different voltage levels of the triggers.

I have no idea if this is possible. The trigger is short, so the pulsing would have to be fast.

Say velocity 128 gives you a full positive 10ms trigger.

Velocity 64 gives you a trigger with an 50/50 pulsing (for 10ms). When charging a cab (in an release envelope generator) the resulting cv would have half the level of the first case. This would make velocity controlled drums possible, without the need of an aout.

This could work i think (it´s like in the switched cab filters, eg) but i´m no programmer. I don´t know if this would lead to changing one line in the code, or reprogramming 1000s of lines...

Cheers, jens

[TK.]

No, this is not possible. Generating PWM with an adequate resolution is one of the most CPU intensive tasks you can think of. Therefore most microcontrollers provide special timer peripherals, which are doing this in parallel to the CPU. The PIC18F4620 has an integrated timer which can generate two independent PWM, but the timers and IO pins are already allocated for other functions.

Btw.: I'm very busy with a lot of other stuff and I'm especially not searching for programming tasks. My motivation to spend my sparetime with implementing alternative solutions for people who want to save some EUR is very low...

just consider this when asking questions which are starting with "do you think it is possible" and ending with "but I'm no programmer" ;)

Best Regards, Thorsten.

[DrBunsen]

I am interested in this use case as well.  Recently I did a search and found an IC which takes a single serial input and gives 12 channels of PWM.  I will have to re-trace my steps and dig it up again if you're interested.

[TheAncientOne]

How does your drum box work now JRP? Some take what is effectively a variable width single pulse as a result of the strength of 'hit' on a piezo sensor, and some have a velocity voltage done as a sample and hold on the pulse.

Since you're not doing fine tuning here, the AOUT_LC would be fine.

If you only need an emphasis effect rather than a full range, perhaps there might be a way to double up the 48 auxiliary trigger outputs, to give 2x24 emphasised triggers, (in fact it might be possible to do trig1, trig2, or trig 1+trig2, thus giving 3 levels). This would need some coding, but perhaps be less messy than trying to hack something more ambitious.

That way you could have 32 triggers, 8 with 8-bit CV, and 24 with 2-bit CV.

Another thought, though TK would have to pronounce on this one, is tha it might be feasible to split the existing AOUT_LC into 4 bit sections, giving a max of 16 4 bit D->A converters, which might be enough for a lot of percussion work, though I say this is just a mad brainstormed idea, and I've no idea how feasible it might be.

If you are planning a MIDI controlled Gamelon Orchestra, I realise this may still be a problem....

Mike

[jrp]

That sounds like a cool idea. Yes, two or three levels would be enough. Easiest would be to just hit two notes at the same time, or build the drummodules so that they have different trigger ins, for different levels.

Thorsten, I can really understand what you are saying. I was more or less thinking loud - hoping that some other people share the same interest.

Of course i wouldn´t expect someone to write me a personall version.

And i also see that most of the troubleshooting, adding new features and stuff ends up with you doing the writing.

My biggest respect for that, for this webpage. For sharing all of this. Your DIY Philosophy is really great.

Again, i can understand that posts like this sound to you like: Hey, i´m new, i have a special problem, please tk do the coding for me...

Nix für ungut!

BTW, it´s not about saving a euro, but about saving the cvs for where they are really needed: Controlling filters, vcos, etc

[jrp]

I have been reading through the manual again.

As faar as i understand, it would be possible, without changing any code, to get the desired effect. 3bit velocity controll. Or 1bit, anyway, those 48 triggers are more than enough to "do something"

A Track running in mode3 - predefined chords (can be user defined, right?)

That track is routed to Aout, ch16.

Just as an example, lets say the first 3 triggersnotes of the 48trigger dout chain are midinote 10,11,12

A chord of these three notes will make the lodest hit, a chord of say 8,9,10 with note 8,9 not defined will make the softest note.

I´ll have to see how good or bad, how intuitive the handling of this method is in reall live.

If we take a step back and say two trigger levels are enough it should really be ok.

Two midikeys for each sound, one hard, one soft.

And thgen there still are the cvs, so if a sound "needs" more controll i can always plog in another vca (in a semimodular systerm)

[TheAncientOne]

Sounds like a plan to me. BTW I have a drawing for the PAIA "Hip Bass DRum", though I've forgotten where I got it from. If I can find it, I'll post the URL, or could email a copy of the PDF. It's the cheapest 'moddable' bass drum module I've ever heard, and it has about 3 Euro's worth of bits in it, and nothing you can't find easily.

Have you checked out the http://www.midibox.org/forum/index.php?topic=6331.0 analogue drum modules thread?

Mike

[TheAncientOne]

Just had another idea. Paul Maddox had a design called MIDI2SDS, which provided variable height triggers from MIDI data, It was an 8 channel unit.

I think all the info is on: http://www.elby-designs.com/midi2sds/midi2sds-about.htm and it doesn't look that expensive to build.

Hope this helps - you could just run one or more from the one of sequencers normal MIDI ports.

I might try this myself.

Hope this helps

MIke

[stryd_one]

Hmm I don't toy with this kinda gear, but couldn't you mod up the IIC_MIDI firmware to drive pins for analog gear instead of serial MIDI? It would be transparent to the seq of course...

[jrp]

the midi2sds is just perfect for the job, problem is the controller...

On the other hand, a midibox cv can be configured to do the same (i think).

On one midi chanell you would have to set split zones for one key each (or two keys, it´s easier to play say an hihat on the keyboard if you have two keys for the same sound - play with two fingers)

Guess i´ll build two midibox cvs for my setup.

But thanks a lot to all of you for giving me some thoughts on this!

cheers

[DrBunsen]

I've had the 2-bit AOUT_LC idea as well. Just a series of 2-resistor "ladders".  That would give 24 channels of 4-level (well 3 and off) triggering.  You could use trimpots or tweak the resistor values to match each module.

[jrp]

Yes, this is a good idea. ok, full level controll is a nice thing to have, but in most cases it´s not really needed.

What i don´t like about the two bit solution is the handling. Programming would have to be done with a combination of different note ons for each sound and volume. Not that intuitive.

When playing the drumbox from a midi keyboard or pads it´s much more handy to have it respond to one (or two) note-ons and the midi-velocity data.

So the midibox-cv is the most flexible solution.