Re: Testing timing of Red Atom Modules

[oldfogeyuk]

I’ve been working on a home-built version of the ‘Brown’ module I’ve rigged on an outboard built from a 1939 Evinrude Midget Racer drive with 1954 J.A.P. 34cc Model ‘O’ attached.

It’s set up with a switch so I can switch from contacts and condenser to electronic and make comparisons immediately. I’m also running it in water to provide a realistic load.

I struggled with getting high speed and low speed running, it would do one but not the other. When I changed the value of C1 it would work better at one end and be worse at the other. There was no single value that would work.

Now I’ve had my timing strobe on the job, I’m shocked to discover that it does not behave as stated. The amount of advance actually reduces as speed increases.
At idle the timing is around 18 degrees BTDC, and it drops to 10 degrees at full throttle. Timing on contacts is 15 degrees at all speeds.

I rigged a lash-up with a 10K pot and a 0.47uF capacitor in parallel with C1 so I could add more capacitance at low speed by dialling down the pot to zero, and I used a 0.1uF for C1 to give better running at high speed. This gave good results. Better than running on contacts in both cases.
Now with the pot at zero C1 = 0.47uF+0.1uF, and at idle the timing is around 16 BTDC
with the pot at 10K C1 is pretty much just 0.1uF and in this configuration running at speed the timing is 20 degrees BTDC. That would be a workable configuration, but I still wanted to get to the bottom of it.

My thoughts are as follows:
– the reactance of C1 in parallel with R3 changes the voltage ratio of the R2 R3 potential divider. As speed increases the voltage on the base of Q2 is lowered more because increasing dv/dt cause the capacitive reactance to fall.
– Q2 still needs 0.7 volts to turn on,
– so as speed increases the voltage applied to the module has to rise higher to get 0.7 volts at the base of Q2.
We might make the assumption that the peak voltage from the magneto increases as speed increases, but if it increases at a lower rate than the potential divider falls, then the timing will be later not earlier. That’s what I’m finding.

I have to repair my Oscilloscope, so it’ll be a few days before I can get some traces and see what the actual voltages and rise times are, but at the moment, I’m looking at dumping C1 altogether and just setting the trigger point by adjusting the values of R2 and R3. Actually I’ll probably replace C1 with a small disc ceramic to decouple any spikes that might cause spurious triggering.

It’s possible that the effect I’m noticing is unique to the Wipac ‘Bantamag’ magneto due it’s relatively high inductance. In which case this work won’t be of much value to anyone else, but I’d welcome comments and suggestions.