Flaming Pipe Organ / Midi HT Spark Controller

[itanc999]

Hi there anyone/everyone...

ISSUE IN BRIEF

'HT spark causes ATMEGA 328P pu to crash"


The project as stated is a midi controlled pyrophone or flaming pipe organ (see tancs pyrophone on youtube). Currently this system uses a pre bought midi - 32 outputs decoder board that drives relays that operate valves and turns on a 555 oscillator to drive a mosfet to drive a 12v automotive ignition coil to produce the spark.

The next phase is to make more tubes (24 in total). with this many tubes the wiring becomes unmanagable (240 pairs of wires over 10m) so i want each tube to have its own decoder and battery thus reducing the wiring to a simple short midi cable between each tube.

I have designed an ATMEGA 328 circuit (arduino derived) that listens for NOTE ON/OFF and PROGRAM midi messages and when it hears its note it executes a sequence of valves and sparks based on the current program number. this enables the machine to have different voices on different midi programs much as a standard midi synth.

The issue is the spark causes the decoder to crash

When the spark circuit is not connected it works fine receiving and passing midi data, recognises program change and throws relays fine.

the system flows thus....

midi in
6n137 opto isolator
74hct14 hex inverter (for midi thru)
atmega 328 ext osc
uln 2803 darlington array for output
relay
osc
mosfet
coil
plug

any hints/ help gratefully received

thanks

tanc

 

[kpatz]

Can you post a schematic?

Sounds to me like your spark circuit is creating back EMF or a drain on your power source causing the micro to crash. Do you have decoupling capacitors in your circuit? At minimum, 0.1uf caps across the Vdd and Vss (or Gnd) pins on the micro and other logic ICs, as close to each IC as possible, for starters. Add some larger capacitors (10-100uf) across the power rails near the spark circuit and the digital circuits. If your spark circuit draws a lot of current, even larger capacitors (1000uf+) may be needed.

A snubber circuit on the spark circuit will protect the mosfet and the rest of your project. Depending on how the coil is driven, a reverse-biased diode across the coil will absorb the surge generated when the coil current is cut off.

If you still have issues, you may need to use a series choke and a parallel capacitor (LC filter) in the power rail to the spark circuit.

 

[itanc999]

hey kpatz

thanks for the reply. I am trying to find a good simple free program for osx to draw circuits with...any suggestings? till then i am using pen and paper but have not made a pretty version yet so cannot post.

i have tried the decoupling capacitors (.1 uf as you say to no effect as was the case with those i tried across the out of the regulator although i did not try any as large as 1000uf

i have an rc snubber across the coil terminals but have not tried the diode yet

the only sucess i have had since last post was to alter the flow thus;

POWER SUPPLY

midi in
6n137 opto isolator
74hct14 hex inverter (for midi thru)
atmega 328 ext osc
uln 2803 darlington array for output
6n137 opto isolator

BATTERY

bc337 inverter
mosfet
coil
plug

and use a pwm output from the ATMEGA instead of the 555 oscillator.


this is ok for the moment but i will be eager to try all of your suggestions asap and will post a schematic as soon as i can

thanks ever so much again

tanc