Digital highvoltage distributor

[libbresse]

I want to build a distributor to my ignitionsystem in my dragrace snowmobile.

Would like to have input on the sketch I have done if it is good enough or if I should build it different.


circuit:
http://www.pixbox.se/pic_show_id26325744.html

Igniton driver:
http://search.digikey.com/scripts/DkSea ... C40LPBF-ND

[Zig]

Ok..

Your schematic shows an AVR micro driving optos which in turn drive high voltage transistors and primary of ifgnition coil.

First of all rearange ignition coil primary to be in the path from 400V to the collector of the transistor.

Secondly see if You can locate some logic level MOSFETS instead of bipolar transistors.

Also include a free wheeling diode in parallel with ignition coil primary winding, your power switching device will thank You.

[libbresse]

Zig: I have change the scheme and will change the opto coupler to the component in the link below

http://www.pixbox.se/pic_show_id26362451.html


http://www.fairchildsemi.com/ds/FA%2FFAN3111C.pdf

Now I'm thinking about how to take care of the signal from the hall-sensor, thinking of some comparator that reads the voltage and send a digital signal to the microprocessor. do you have any other idea ?

[Zig]

The output of the hall sensor is proportional to the magnetic flux density generally speaking.

I assume You will be using one of the AVR micros.

I guess Your software will be interrupt driven, with the interrupt being driven by the Hall signal.

I think using something like an LM311 comparator with a bit ( say a hundred mV hysteresis) to clean up the Hall output will be OK.

The output of LM311 is Open collectorso it needs a pullup resistor. you might even get away with programing the AVR input to turn the internal pullup on to minimise external component count.

If you do plan to go with an interrupt driven approach rather than a polled kind of a thing, make sure Your ISR is very short, perhaps setting a bit or a flag.. or may be a timer counter to determine duration of trigger pulse and to calculate ( measyre) engine RPM and select apropriate advance timing for the ignition.

You may have to precaculate a LUT ( look up table) so your micro has less to do in real time.

In any case .. good luck.

[libbresse]

Zig: Done a circuit with your help Can you tell me if I have forgotten something ?

http://www.pixbox.se/pic_show_id26529066.html

Remember it's only a distributor and not the ignition, for that I have a MSD ignition system.

[Zig]

Looks about right.

You seem to be using a SPICE kind of a package.

Please try to include a Noise source in series with Your pulse source.

Run a simulation and observe the Instability in output pulse train which goes to Your butterfly board ( i think it is AVR based Butterfly).

Note instability ( jitter in leading edge ).

Then add say a 100K resistor between comparator output and non inverting input ( pins 5 and 7) and observe the jitter in leading edge.

Please comment after the experiment.

[libbresse]

Zig: Thank you for the input I'm working really much right now. I will test the setups you gave when I have time

[libbresse]

Zig: I have now tested the resistor from the outputpin to pin 5 with the noise source on. It seems like it become a hysteresis with the resistor that can avoid some really nasty behaviour from the comparator output I will post the plots from the simulations later today

[libbresse]

Ah the plot-pictures became really bad but here they are.


http://www.pixbox.se/pic_show_id26672764.html
http://www.pixbox.se/pic_show_id26672765.html

I could not generate a true noise signal its only one frequency. But I still get the idea of putting in a hysteresis