polyphonic MIDI problem / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

polyphonic MIDI problem

1 2 next

Move Thread

LAN_403

RateReducer

Mon Aug 26 2013, 12:08AM

Registered Member #4603 Joined: Wed Apr 25 2012, 07:33PM
Location: Austria
Posts: 159

Hi guys. I have a problem playing polyphonic MIDIs on my bigger DRSSTC. I have a MIDI interrupter which consists of two NE555 interrupters playing into an OR gate. This works fine for my small DRSSTC with crappy primary feedback but my other coil with this driver:

does not like it at all... It makes popping, distorted sounds. Does not sound nice at all and i am afraid to kill my bridge with it

How can i solve this problem?

greetings

RateReducer

Mon Mar 24 2014, 06:46PM

Registered Member #4603 Joined: Wed Apr 25 2012, 07:33PM
Location: Austria
Posts: 159

A few month later - still the same problem...
I built Steve Wards polyphonic MIDI interrupter to get rid of the problem but nothing changed. :(
When i use just one MIDI channel it workes just fine - perfect sound. When i use 2 channels the problem starts.

What could be wrong?? I use toslink LWL transmitter and receiver, could that be the problem?

Uspring

Tue Mar 25 2014, 10:11AM

Registered Member #3988 Joined: Thu Jul 07 2011, 03:25PM
Location:
Posts: 711

Just an idea:
Usually drivers synchronise the falling slope of the interrupter to the next zero crossing of primary current in order to avoid hard switching. If, during that wait time, the interrupter signal goes up again, the driver might do something unpredictable.
Also: The primary tank might take some cycles to run down after the interrupter turns off. If the interrupter goes on again before that, the driver might also hard switch.
A cure for that would be to avoid very short off times from your interrupter. Perhaps your driver handles these situations gracefully, though.

RateReducer

Wed Mar 26 2014, 08:13PM

Registered Member #4603 Joined: Wed Apr 25 2012, 07:33PM
Location: Austria
Posts: 159

Hmmm Okay... I had some similar thoughts on that problem...
What could i try to shorten those "wait times" that the primary circuits need to ring down?

Edit: I use this driver: stevehv.4hv.org/newdriver08/ DRSSTC_pndriver1_3b.pdf

Uspring

Thu Mar 27 2014, 10:16AM

Registered Member #3988 Joined: Thu Jul 07 2011, 03:25PM
Location:
Posts: 711

Primary ring down times depend on coupling arc load and other things difficult to control. Not any real options there.
I'd first check, if low interrupter off times are really the problem before you put effort in interrupter timing.

Steve Conner

Thu Mar 27 2014, 10:23AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Maybe you're triggering the OCD? This normally works by ending the burst. It won't reset until the interrupter commands a new burst. This could cause distortion.

I've always used cycle-by-cycle current limiting that holds the primary current at the limit instead of shutting the drive off completely. Arc Attack recently started doing this too.

Also, some Toslink receivers are internally AC coupled and have limits on the duty cycle they can transmit. Try viewing the output of the Toslink receiver on a scope and comparing it to the signal you are sending to the transmitter, to see if the pulse width is getting distorted. The HFBR-24xx series are a better bet for serious Tesla coil work.

RateReducer

Thu Mar 27 2014, 04:13PM

Registered Member #4603 Joined: Wed Apr 25 2012, 07:33PM
Location: Austria
Posts: 159

Hi, Steve. I do not think that the OCD is the problem. The distortion even starts at minimal power... I have to upload a video...
My oscilloscope is a very old device - it is hard to see the polyphonic bursts on it ^^ but instead i will try to use the same toslink receiver on my other DRSSTC (working properly with polyphonic MIDI) to see if the same error occurs.

Edit: I tried to bypass the Toslink receiver and fed the signal of the interrupter without LWL into the DRSSTC - same problem.

Here is a video:

First i use the 2 channels seperated and then combined. You will notice the distortions when both channels are active. When i use more power it sounds really frightening....

Any ideas?

RateReducer

Fri Aug 22 2014, 08:37AM

Registered Member #4603 Joined: Wed Apr 25 2012, 07:33PM
Location: Austria
Posts: 159

Still got the same problem...
I tried adding a resistor across the bridge output like Kizmo did:

I thought that would be a good idea since i also had some weired "bzzz to prptptptptptptptt (sounded like machine gun)" sounds like Kizmo @ a certain interrupter setting. The resistor solved the problem - but it did not solve the Midi problem....
I do not know what to try next. It is sooo strange, the two Midi channels alone work perfectly - clear sound, no distortion at all but when i crank up the duty cycle of the second channel it does not sound good at all but makes really distorted bursting noises

Since i got a new scope (Rigol DS1052E - hacked to 100MHz

) i scoped the primary current and the bridge output - looks nearly the same like the picture in Kizmos thread (1st page second last picture) so i guess that fits?
The only difference is that the current does not look like a sine wave but more like a square wave - maybe because i scoped the feedback CT, Zener diodes....?

Do you guys have any idea what i could try next to get rid of this bullsh... ?
Now that i have a proper oscillocope i could do some decent measuring on my DRSSTC.

DerStrom8

Sat Aug 23 2014, 03:10AM

Registered Member #3704 Joined: Sun Feb 20 2011, 01:13PM
Location: Vermont, U.S.A.
Posts: 92

How are you limiting the on-time of the notes? An A4 note at 50% duty has an on-time of 2.273mS, which is very high for a Tesla coil, and could cause the OCD to trip even at lower powers. Easy way to know for sure would be to add some sort of visual indicator that your over-current detector is tripping, like an LED.

Also, how are you mixing the outputs?

I'm new to SSTCs, so I may not be on the right track, but it would be something to look into.

loneoceans

Sat Aug 23 2014, 05:29AM

Registered Member #4098 Joined: Fri Sept 16 2011, 09:26PM
Location:
Posts: 236

Hey RateReducer,

It looks like using the scope is the best way to go about looking more closely at what your coil and driver are doing. For starters, I would try not to use the simple OR method for polyphony. As mentioned by Uspring and Steve, you might need to make sure you pause for a while after each pulse. They way I like to think of it is simple - a new note pulse cannot begin when an early pulse is on-going or for some time interval after it ends. This allows for ring-down and helps in duty cycle management.

There are several ways to implement this in practice - for example, simply dropping over-lapping notes, shifting the over-lapping notes by a random amount capped at a certain shift, and so on. What I do is that if two notes collide, the first note (or if at the same time, choose one say based on frequency) is allowed to play completely and a short wait for the hold-off, where the next note is allowed to continue to run after. This also allows easy implementation in a microcontroller with interrupts, where a counter that triggers the 2nd note is allowed to continue to run after a delay due to a collision. So it seems like the best way to go is not a simple naÃ¯ve OR but a more considered approach especially for DRSSTCs. I've tried a simple OR approach in a conventional SSTC with very long pulse-widths and it works great though.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob