Question about arduino polyphonic MIDI interrupter

[bozidar]

Hello,
I wrote the code for arduino based polyphonic MIDI interrupter, for my DRSSTC, that can play 2 notes. But I don’t know if the pulses can overlap or be close to each other (like the pulses in the picture)?
Is it safe to run the coil with this pulses?

[TMaxElectronics]

Not necessarily. The issue is that some drivers (all recent-ish UDs afaik) use a fixed frequency start oscillator (or at least "kick" the bridge to start it), meaning that they will switch regardless of what the feedback reads.
This doesn't matter if the primary circuit has no current flowing in it, as the first few cycles will have a low enough current that the hard switching isn't going to kill the IGBTs. But if you have two pulses very close to each other there might still be significant current flowing in the primary circuit, resulting in hard switching and some sharp thermal transients in the Transistors, which they may or may not survive.

I recommend implementing an off-time after every pulse, so the current in the bridge has time to decrease. I don't know how you have coded your pulse generator, so I can't tell you how best to do this in you case. But in my interrupter I turn off the interrupts of the timers creating the pulses, and re-enable them with a different timer after the off-time has elapsed.

[Max]

Polyphonic interrupters are really getting popular these days...

While I completely agree with TMaxElectronics, there's an important point to consider: AFAIK the quite popular Onetesla MIDI interrupter as well as loneoceans MIDI interrupter do not implement any protection for this case. And those have been used by many people without major issues.

This brings me to the interesting question if anyone ever scoped his primary current during MIDI playback to see if he can actually find hardswitching events.


Kind regards,
Max

[bozidar]

I'm using almost identical Loneoceans UD2.7, just with through hole components. In place of optic fiber receiver I'm using arduino nano with optocoupler . Arduino mega sends MIDI messages via bluetooth to an arduino nano (which I use as an interupter).
This setup works perfectly when I play only one note

This is a SimpleTesla interrupter that plays MIDI with multiple notes (6, if I understood correctly-   1:40 - 2:30 min)

And this is mine playing 2 notes 

I guess Onetesla MIDI interrupter works the same way, right?
Can  someone post some scope pics of Onetesla output when playing 2 notes MIDI files?

Maybe the OCD will not allow the current to rise too much when the two pulses are close to each other, so could I just reduce on time?

[davekni]

UD2.7 could make a single hard off-to-on transition if started during ring-down from a previous enable pulse.  That may not be enough to cause problems.  Still nice to avoid.

Last year I purchased a MIDI interrupter kit from Eastern Voltage Research.  I don't know who's design they use.  They said it wasn't their own.  It had a minimum pulse spacing (~200us if I recall correctly) that appeared to be more an artifact of coarse timer intervals in code rather than an intentional feature.  It also claimed to handle 4 notes, but output garbage if more than 2 were played simultaneously.  I've since replaced it with my own design.

[Max]

In my tests with Syntherrupter I noticed that the min offtime has a surprisingly high impact on the audio quality. I remember that 100us minimum offtime sounded noticeably worse than 30us. Shouldn’t be a surprise that the impact is even bigger with more simultaneous notes. That was however with speakers. Well possible that it is less audible with a tesla coil.

Kind regards,
Max

[davekni]

I noticed the same result using a speaker:  Sound quality is much better without limits on gaps or pulse widths.  However, at least for my DRSSTC, sound output isn't linear with pulse width.  It wouldn't be linear with short spacing either, although I avoid close spacing.  SSTCs are likely much more linear.

So, at least for my low-impedance DRSSTC, distortion happens either way.  I experimented with algorithms for spacing and combining pulses and pre-distorting pulse widths to undo the coil's roughly square function of pulse width to sound output.  Did improve sound, but it's not yet where I'm really happy with the result.

[bozidar]

Thanks for the answers.
I added off time to the code and I don't think the sound quality is much distorted so maybe I'll try it on DRSSC .

[TMaxElectronics]

How have you implemented it and how long did you set your holdoff time?

[bozidar]

After each pulse I just added delay function 
Holdoff time is 45 us.

[TMaxElectronics]

So how do you trigger the pulses? If you use interrupts I'd recommend changing it to work differently as you always want to minimize the time spent in interrupts, especially since you might shift the other pulses back which can add quite a bit of noise to the signal
I have a default setting of 30us which only slightly reduces the sound quality.

[bozidar]

I use timer 1 for one note and timer 2 for the other note. For now the signal has no noise.
 If you want I can post my code 

[bozidar]

Here is the code (which I know is not perfect) 
And here are a few videos