CDI Ignition spark, direct measurement.

A

Angelo

Well-known member
Hello,

I am building a trigger generator to test some old school CDI ignition box. Old, yes, but 45kV for 11'000 rpm racing V8.
I will use a standard plug and check visualy the spark. But this will not tell me the quality of the spark.
The main defect on these old boxes are in the 450V capacitors in the primary circuit. The capacitance is lower after years, and temperature abuse. The energy/voltage is still present to generate a spark in air, but much lower. And when running, the engine will stutter.

Is there any well known circuit connected directly to the secondary coil, in order to measure the spark quality ?
Clearly a voltage divider to lower the value from 45kv to some volts.
A series resistor to measure the current ?

I have found close to nothing. There are some automotive voltmeters, but they only give an aproximate value.

Any guidance welcome

Angelo
 
Short of taking out the capacitor from the unit and measuring it I can't think of an easy way - caution the sparks from these units are easily powerful enough to kill. The higher the pressure in the cylinder and the wider the gap the more voltage is needed for a spark, so it will first show at full throttle.
 
Disassembling the unit is a real nightmare. Lots of resin is pourred around components and wires.
And I don't want to disassemble and create a possible future fail in cables and components, just to discover this is a good unit.
And that's why I need to check spark quality. A "perfect" unit on the bench started to produce misses from mid-range at full throttle.
Same thing in the cars on the track.
 
I suppose you could try a variable length gap and measure the longest spark at atmospheric pressure.
 
How about running the test on the primary side? Could you measure the voltage on the capacitors prior to discharge? If the caps are the predominant problem, might this help? The primary side is at a lower voltage, which could make the electronics easier. I understand that what counts is at the spark plug, but if the dominant error are the caps, then monitor them. Voltage vs time?
 
That's much more tractable, but won't detect failing transformer issues - its not only the capacitor that could have degraded when you have 10's of kilovolts flying around.
 
The variable gap "plug" is a good idea, easy to implement. Will give a try.
As MarkT wrote, there is also 2 transformers, lots of smaller resistor and capacitors, all beeing abused by temperature. Or storage, as they are from 1970 to 1985.
The target is not always to repair the unit, which could become expensive for the customer. We have a kit to update it.
We keep the old box, and we install a modern CDI unit, a new coil and a new alternator rectifier.
 
MarkT said:
That's much more tractable, but won't detect failing transformer issues - its not only the capacitor that could have degraded when you have 10's of kilovolts flying around.
Click to expand...
Having made a 50KV resistor divider, which needed to be in oil, I'm aware of the issues. You need 100's of megohms and ways to avoid creep and corona. Not to mention HV insulators.

Solving the whole problem is obviously better, but 1/2 a loaf is better than nothing.

I seem to recall that in advanced engines that they use some sort of ionization detection to help with engine management? A good spark generates ions and makes sufficient "light". (Ultraviolet ?). Think that might be used on F1 engines? This avoid direct HV measurements which makes life easier.

Spark gap measurements are tricky. They depend on radius of curvature. Pointy edges etc break down sooner than rounded, so unless you use spherical electrodes, you can't estimate voltage. Also dust, and moisture affect the break down voltage.
 
Some people are using the spark plug ionization current. There's a lot of papers on the subject. Properly interpreting the current tells you a lot.
 
Angelo said:
I found this very interesting article: https://cdn.imagearchive.com/homemo...ch/97/97691-THE-SPARK-ENERGY-TEST-MACHINE.pdf
Click to expand...
Interesting article. In that circuit the ground of the plug is technically not at ground. In a real engine the threaded section is grounded, since the block is ground. I'd isolate the case from the plug, but that's me. Been shocked more times than I'd like to admit.

The circuit shown can measure interesting stuff, but not under load, like in a real engine. The ionization technique is used under load, including under boost, which is tougher to spark. High end engine management systems are using ionization current to manage spark these days. It can measure misfire cylinder by cylinder.
 
clinker8 said:
Having made a 50KV resistor divider, which needed to be in oil, I'm aware of the issues. You need 100's of megohms and ways to avoid creep and corona. Not to mention HV insulators.
Click to expand...
A pure resistive divider like that will have poor bandwidth, probably lose detail of a fast transient, as stray capacitance may dominate over shorter timescales.
 
MarkT said:
A pure resistive divider like that will have poor bandwidth, probably lose detail of a fast transient, as stray capacitance may dominate over shorter timescales.
Click to expand...
It was used for a DC measurement, was charging a laser pulse capacitor. Used a lot of resistors in series to minimize capacitance. Did the job that was required.

Measuring HV transients isn't that easy, due to the isolation requirement. Measuring the ionization current sounds easier to me. But I'm a guy that mega squirted his vehicle. Don't know, maybe they've added that in by now.
 
You must log in or register to reply here.
Back
Top