Variable frequency delay

get 23 more sensors and a 24 way switch?

possibly something could be done with a microcontroller something with a timer with PWM and event capture and a few lines of code.

there's an analogue solution too

in ramp peak |\\ ---generator-+-detector---rheostat---| \\ | comparator -- edge --- out `-----------------------| / detector |/

You can do this with a dual-slope ramp scheme. The input pulse switches on an input current to an integrator that ramps up at rate dependent on the phase delay you want. The integrator output goes to a comparator, and when it hits threshold it disconnects the charging current and connects a

*constant* discharge current. That ramps the integrator down until the next input pulse starts the cycle over again. The comparator output is a rectangular wave that is delayed by an amount controlled by the up-ramp current.

I used this approach back in the mid-1970s while working for GM's Cadillac division, to provide a simple (cheap) way to control ignition timing without an on-board computer. Various switches (manifold pressure, temperature, engine speed) combined in a simple diode logic array to select the charging current for the desired spark advance at various running conditions. Worked great, cost only 10% what a computer would have cost but gave

90% of the benefits. But I think they only used it for one or two years, and only on a few models, before computers took over.

Best regards,

Bob Masta DAQARTA v4.51 Data AcQuisition And Real-Time Analysis

Scope, Spectrum, Spectrogram, Sound Level Meter FREE Signal Generator Science with your sound card!

The simple, foolproof method would be to move the magnetic sensor or the magnet. I assume that is not an acceptable option?

An even simpler option would be to simply buy a timing light which has this functionality built-in. Not a very fun solution, though.

Obviously, 15 degree delay is not possible at zero speed. You will have to define a minimum required speed.

Depending on how low you have to go, this problem should not be too difficult with a microcontroller and a bit of arithmetic. You'd have to time one complete revolution (the time between two pulses), and then delay the next pulse by 1/24 of that value. This method assumes that the speed does not change between two consecutive revolutions, so it would only be accurate at constant speeds.

Use a microcontroller. I can't think of a solution using discretes which isn't significantly more complex.

The obvious discrete solution is a phase-locked loop (PLL) with a 12-stage twisted ring counter as the divider. That gives you 24 signals at the same frequency as the input, with phase shifts in 15-degree steps.

A much simpler solution is an absolute delay which can be varied via a knob. You manually adjust the knob until you're looking at the right part. This assumes that the speed doesn't vary while you're looking at it.

So, use a phase-locked loop (CD4046 type) with one or two decoded counters (CD4017 type) in the reference branch. That will give you ten to twenty taps at various phases with respect to the input.