Coil output voltage risetime

[Circlotron]

CDI ignitions often tout fast coil voltage rise time as a beneficial feature, better able to fire fouled plugs. And it may well be. Not arguing that at all. What I did tonight was to compare the rate of rise of coil voltage compared to an inductive discharge ignition. I used an MSD #8253 coil https://www.msdperformance.com/products ... parts/8253 and an MSD6 ignition box. This coil is quoted as having "lightning quick" rise time... https://www.google.com.au/webhp?complet ... g+quick%22 To measure the rate of rise of coil voltage I simply put the probe against the coil secondary wire insulation when it was sparking into a 10mm gap. This won't measure the actual voltage but it will show how quickly the voltage rises until it sparks. Then without changing the measuring arrangement I replaced the whole setup with an inductive discharge ignition with its own coil.

And guess what? The rate of rise of voltage is pretty much identical in both cases! See the pics. Notice the slope of the increasing negative-going voltage is the same in both cases, and then the plug fires as per the sudden upward stroke of the trace. In the past I have said this MSD6 + coil has only about 40mJ output from the coil and only 0.12mS spark duration. The final nail in the coffin is the same-as-inductive rise time. Obviously this rate of rise is quite good enough, but it cannot be trumpeted as a benefit only CDI has, at least not with this "lightning quick" coil anyway.

Inductive dv/dt

Inductive risetime.png

CDI dv/dt

MSD6 + 8253 risetime.png

[mag2555]

So this test was not done with the spark having to jump a 10mm gap with cylinder pressure working against it?

[Circlotron]

Just 10mm gap in open air. Cylinder air pressure will affect the voltage which the plug fires but not the rate at which the voltage goes up. That is what we're comparing.

[groberts101]

I think what mag was eluding to in the above is that it's probably going to be very tough to get an accurate measurement of rise time when the coil isn't having to push the spark to ignite in a pressurized combustion space. Similar to an engines ignition lead tolerance and combustion efficiency when comparing static rpm rise to loaded rpm rise. This would be a MUCH tougher environment for the coil to function.

I wonder if increasing resistance using a very long wire.. or multiple wires linked together would give enough resistance to help flush out any additional differences between the two boxes?

[ijames]

Kind of just wondering out loud here, but you are capacitively coupling to the probe so you have a frequency (or rise time) dependent voltage divider, right? I agree that that means that you can't measure the secondary voltage, so to me that means that you can't read anything into the observed slopes of the rise time. The gap should need about the same voltage to fire with either source, and it looks to me like the CDI needed half the time to go from first inflection to firing, compared to the inductive setup (3.3 divisions vs. 6), so I would think that that means that the CDI has half the rise time or is twice as fast. Am I missing something?

[upinthehills]

I think you are mis-reading the scope a bit. I count the rise time to the coil as being 6 divisions in one case and 3.25 divisions in the other case. The slope is the same just because the signal is lower magnitude in one case. Am I reading this wrong? If the spark happens at the same voltage, the CDI has nearly twice the slew rate or half the rise time..

[mk e]

I think you are mis-reading the scope a bit. I count the rise time to the coil as being 6 divisions in one case and 3.25 divisions in the other case. The slope is the same just because the signal is lower magnitude in one case. Am I reading this wrong? If the spark happens at the same voltage, the CDI has nearly twice the slew rate or half the rise time..

I think that's just a difference on the trace for where the test was started? I think that piece is missing so were all assuming something different.

Even is the rise rate is actually different I'm struggling to understand exactly why I should care? I guess if there is variation in the voltage required to spark cycle to cycle or cylinder to cylinder a faster rise rate would mean more more consistent timing but faster in and of itself doesn't seem to offer any advantage?

[ijames]

I think you are mis-reading the scope a bit. I count the rise time to the coil as being 6 divisions in one case and 3.25 divisions in the other case. The slope is the same just because the signal is lower magnitude in one case. Am I reading this wrong? If the spark happens at the same voltage, the CDI has nearly twice the slew rate or half the rise time..

I think that's just a difference on the trace for where the test was started? I think that piece is missing so were all assuming something different.

Even is the rise rate is actually different I'm struggling to understand exactly why I should care? I guess if there is variation in the voltage required to spark cycle to cycle or cylinder to cylinder a faster rise rate would mean more more consistent timing but faster in and of itself doesn't seem to offer any advantage?

Starting at the end, I don't see why the rise time would make much difference in performance, either, but I don't know that it doesn't, either, so measuring it is certainly useful so we can find out one way or the other, and if a mfg brags about it then checking up on them is a good thing.

upinthehills, you are comparing how tall the triangle is in each case, in the y direction. My point is that the y axis is uncalibrated so you can't compare the heights of the two traces, because of the capacitive coupling and different rise times of the secondary voltage. You can compare the x direction, however, because that is given by the beginning of the rise as a start, the time base of the scope, and the place where the gap fires as the end time. So in the x direction you can tell how long it takes to go from the start of the event to the firing of the gap, and if you assume the firing voltage is the same you can now compare the two rise times. You don't know the firing voltage but you do know that the CDI got there in half the time.

[upinthehills]

upinthehills, you are comparing how tall the triangle is in each case, in the y direction. My point is that the y axis is uncalibrated so you can't compare the heights of the two traces, because of the capacitive coupling and different rise times of the secondary voltage.

That's what I was trying to say. I didn't really see you had answered this already. Just didn't think we needed to talk about the exact why the Y was different. The spark happens much quicker with the CDI. This is probably just showing the CDI coil has less inductance. Less inductance and you drive it with a much higher voltage, so you get a faster rise time.

The point for a fast attack on the spark is to try and reduce various leakages you may have. I'm not sure it's important. In the seventies the average car needed everything all the time - plugs, points wires were always wearing out. So working with marginal stuff it might help, cracked insulation on your wires, plugs fouled because people didn't know how to use a choke, etc

[user-23911]

Rise time is directly proportional to the circuit capacitance.
In particular the secondary capacitance due to the times ratio of the coil.

Putting your scope probe next to the HT lead to take off the signal, that increases the capacitance.
The combined capacitance of the scope probe plus the scope is far more than in the circuit.

So what you're really doing is loading the circuit in order to take a reading and in so doing, you're screwing up the result.

So yet another example of GIGO.

[upinthehills]

Putting your scope probe next to the HT lead to take off the signal, that increases the capacitance.
The combined capacitance of the scope probe plus the scope is far more than in the circuit.

Not in this case. We have 12V and high currents in this circuit. It's the building of the magnetic fields that is absorbing the power and causing the rise time. You would actually see this same type of signal trace on the fuel injectors and their windings with no secondaries. You can take a foot of wire and wrap it around a nail and put it on a scope and see this same thing. It's amazing the power of iron for this, a straight piece of the same wire will have a much, much faster rise time closer to the speed of light.

[blackford]

Where in the trace is the spark occurring? Before the spark, there would be no current flow and thus no voltage where you were probing it. Then some voltage would be measured on the wire, but most voltage being across the plug gap. Is the level tapering off in the 2nd trace? In the first trace, the level before the slope is nearly identical as the level after the spike, but in the 2nd trace the 2 levels are noticeably different. The slope is the same but the voltage change is different. Rise time is 10% to 90% and fall time is 90% to 10%. One trace has a faster fall time because its starting out at a lower level. Just trying to understand where things are occurring in the trace. Thanks

[Circlotron]

upinthehills, we are actually looking at a representation of the secondary voltage rate of increase here, not the primary amps rate of increase. And joe 90 is in principle correct about capacitive loading affecting the measured result but in this case the effect of bringing a scope probe with 100 megohms resistance and something like several pF of capacitance merely laid on the coil HV lead I think is totally negligible. And the capacitive voltage divider effect of the probe near the wire reduces the effect of the internal capacitance of the probe even more.

[upinthehills]

upinthehills, we are actually looking at a representation of the secondary voltage rate of increase here, not the primary amps rate of increase.

I understand. You can look at the same thing on the primary side too. The primaries of these coils and also the fuel injectors will all show the same sharp rise in voltage until something breaks down and gives.

The ECU I'm building watches this spike and if it doesn't drop quickly at some point it means there was an ignition mis-fire. It's nice to know if a spark plug fails to fire so it seems a good feature to me.

If your scope has two probes you can plot both and see what I mean...

[Circlotron]

Where in the trace is the spark occurring? Before the spark, there would be no current flow and thus no voltage where you were probing it. Then some voltage would be measured on the wire, but most voltage being across the plug gap. Is the level tapering off in the 2nd trace? In the first trace, the level before the slope is nearly identical as the level after the spike, but in the 2nd trace the 2 levels are noticeably different. The slope is the same but the voltage change is different. Rise time is 10% to 90% and fall time is 90% to 10%. One trace has a faster fall time because its starting out at a lower level. Just trying to understand where things are occurring in the trace. Thanks

The first pic shows things clearest.
Zero voltage is 1 division down from the top, starting from the very LHS.
The voltage begins to increase in the negative direction (goes downward.)
When the voltage is great enough it jumps the 10mm test gap and pulls the voltage back to the sustaining level. That is the horizontal part of the trace on the RHS.
The spark occurs at the moment the line suddenly goes vertically upward.

Rise time is 10% to 90% and fall time is 90% to 10%.

Dead right, but not as many people here would be familiar with the term dv/dt so I used the term rise time but a bit loosely.
The MSD spark pic was one of the middle sparks in a chirp, and they can look a bit different. I was triggering the box by hand and if you trigger it slowly this way the first spark can look a bit weak - that's why I didn't use it. I'll take a pic again, this time of the first spark but triggered at a more realistic rate.