What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

capacitive vs inductive discharge ignition

kaweeka

Well Known Member
I currently have a Lightspeed Plasma III ignition system however, after 3 coil failures, I am considering alternatives and have been reading on the difference between a capacitive vs inductive discharge ignition. It seems to me that with a large bore cylinder and my typical flight being LOP it might favor inductive, given the longer duration of spark. As with everything there are trade-offs and there is only so much I can learn from Google. So many folks here have extensive engineering and racing backgrounds that I would yield to advice from knowledge and experience.

Thanks,
David
 
Both work fine in racing and both will light off a stock Lycoming with no problems. Some coils designed to have 14V on the primary windings may not take kindly to having 400V put on them.

Lots of cheaply made coils out there produced offshore too. They should be lasting thousands of hours if used as designed, mounted correctly and have quality construction.
 
Last edited:
Both work fine in racing and both will light off a stock Lycoming with no problems. Some coils designed to have 14V on the primary windings may not take kindly to having 400V put on them.

Lots of cheaply made coils out there produced offshore too. They should be lasting thousands of hours if used as designed, mounted correctly and have quality construction.

Where does the 400v come from??
 
Sometimes when watching an electrical storm you will see lightning bolts that are pale and thin, sometimes they are fat and very bright. Similarly 12volts feeding your ignition coil versus feeding it with 400vdc. Huge difference in the voltage and temperature of the spark produced hence hotter spark. That's the best I can describe it as.
 
Illustration below from the Bosch handbook...

CDI = capacitive discharge
TI = transistorized inductive
CI = contact breaker (points) inductive

The defining feature of CDI is the rapid rise time, here charted as about twice as fast as a typical inductive EI. Rapid rise translates to better resistance to fouling. Think of it this way...the voltage shoots up to the level needed to jump the plug gap before it can leak away to ground via surface fouling. Slow rise = more leakage.

The flip side is very short spark duration. That is why, for example, the Lightspeed Plasma III is a multi-spark system, one right after another for (per the website) about 20 degrees of crank rotation, or 1.23 milliseconds at 2700 RPM. A long duration spark, or a series of sparks, has a better chance of lighting non-optimum mixtures. Given enough time, something flammable is likely to pass near or through the hot spot.

A conventional inductive coil builds a magnetic field when current flows through the primary windings at 12 volts. A high voltage spark is generated when the primary current is interrupted, and the magnetic field collapses through the many, many turns of secondary winding.

A CDI is different. The control electronics amplify the supplied 12 volts and charge a capacitor to around 400 volts. When it's time for a spark, the capacitor is discharged to the coil primary. In this case the coil acts as a transformer; if the primary/secondary turns ratio is 100, for example, the 400 volts in the primary might generate 40,000 in the secondary...but it only lasts as long as it takes to discharge the capacitor.

Note that inductive makes a spark when the primary current is turned off, while CDI make a spark when primary current is pulsed on. The 12 vs 400 thing has no direct bearing on "hotter spark". They are simply the common primary voltages for two entirely different kinds of spark generators.

Although vendors might claim some large secondary voltage figure ("...as much as 45,000 volts for a hotter spark!"), it doesn't mean voltage rises to anywhere near it in practice. Secondary voltage rises until high enough to form a conductive plasma path between the plug electrodes....after which it rises no higher. Note vendors routinely caution us to never trigger the system with the plug wires disconnected. Without an easy path to ground, the voltage may indeed rise to the advertised maximum, and burn through the insulation on the coil windings.

So much for differences. Either will light a Lycoming. I'd be more interested in practical aspects; a record of reliability, choice of user adjustable advance or no advance at all, price, service, etc.
 

Attachments

  • Bosch Fig 37 Voltage Rise.JPG
    Bosch Fig 37 Voltage Rise.JPG
    61.6 KB · Views: 231
Last edited:
If there isn't a more powerful or hotter spark I don't understand why Dan. There is approximately 33 times the power being applied to the primary of the ignition coil if there is actually 400 volts across it.
 
If there isn't a more powerful or hotter spark I don't understand why Dan. There is approximately 33 times the power being applied to the primary of the ignition coil if there is actually 400 volts across it.

Voltage does not equal power.
 
Not Necessarily

Very true but if you have 400 volts across the coil versus 12 volts the power is up by 33 times

This assumes that the Primary/Secondary ratios of a CDI coil and an Inductive coil are the same, but they might not be...

Skylor
 
Very true but if you have 400 volts across the coil versus 12 volts the power is up by 33 times

Norm, that's a straw man argument.

I spent a few minutes digging around on my bookshelves. Turns out I have many references discussing practical ignition application, but very little in terms of design equations.

A quick spin on the web is interesting. For example, consider this equation for inductive coil stored energy. Primary voltage isn't even mentioned.

E = (1/2) * L * I^2

where

L = primary inductance
I = primary current

Coils designed for CDI applications have far less inductance than Kettering type coils, which makes sense. And there is a lot more, like differences in coil efficiency.

I'll leave you to your own reading, and stick with the prior statement...a comparison of CDI and inductive primary voltage has no direct bearing on spark energy. And let's remember, spark energy above that required by the application may be fun to brag about, but it's of no practical use.

Break.

Had a good laugh while at the bookshelf. Among others, I pulled down a copy of Petersen's Basic Ignition and Electrical Systems, which I hadn't cracked open in a long time. How long? The cover price was $2.00, with a copyright date of 1971. Apparently I bought it when I was in high school. Pretty useful if you want to know all about round coils and mechanical voltage regulators ;)
 
Last edited:
Coils designed for CDI applications have far less inductance than Kettering type coils, which makes sense. And there is a lot more, like differences in coil efficiency.

I'll leave you to your own reading, and stick with the prior statement...a comparison of CDI and inductive primary voltage has no direct bearing on spark energy. And let's remember, spark energy above that required by the application may be fun to brag about, but it's of no practical use.

Dan makes a couple important points here.

Many CDI systems simply use whatever coil is convenient and inexpensive. Most coils are not designed for use with CDIs as the vast majority of OEM ignitions are inductive discharge. This MAY explain the short coil life on some CDIs.

The "Marketeers" make lots of noise about spark energy and that CAN be important on very high specific output, high rpm racing engines but it's essentially meaningless as applied to stock Lycomings where compression ratios and rpms are low.

We've been asked the question about spark energy before and I answer that we don't know on our systems as it's irrelevant to this application. I simply tell them it reliably fires 900hp Lyconentals running 100 inches of manifold pressure so it will certainly fire your stock Lycoming.
 
12v vs 400v

Dan's posts were quite good.

To the question of the difference of 12 volts verses 400 volts on the coil primary, BOTH inductive and capacitive systems produce a similar spike of a few hundred volts on the coil primary, maybe 400 volts.

In the case of an inductive system, the 12 volts is the CHARGING voltage for the coil which will serve as the inductive storage element. When the points open, the magnetic energy is released, producing a ~400 volt spike on the primary and a corresponding many thousands of volts on the secondary and can be thought of as a transformer as well.

In a typical capacitor system, a capacitor as storage element is electronically charged to 400 volts and then connected (dumped) across the primary using the coil only as a transformer to make the spark.

So, both systems are quite similar in a way. BTW, the coil in a cap sys can often be much smaller since it does not have to store anything, acting only as a transformer.

Sermon: maybe related to the OP. Almost any ignition system can be wrecked by allowing too high a secondary voltage to develop. This can happen from an open spark lead, very wide plug gap, or any combination that does not load the secondary circuit properly.

Ron
 
Last edited:
Dan's posts were quite good.

"Almost any ignition system can be wrecked by allowing too high a secondary voltage to develop. This can happen from an open spark lead, very wide plug gap, or any combination that does not load the secondary circuit properly. "

Ron

That's really interesting. When I replaced all 4 coils after one of the 10 year old ones went bad, Klaus sent the "new style" (Dynatec Mini) he said was more powerful and suggested I widen the gap to .040". The coil only lasted 20-30 hours. I checked the wires and they're fine. Maybe .040 is just too large a gap?
 
That's really interesting. When I replaced all 4 coils after one of the 10 year old ones went bad, Klaus sent the "new style" (Dynatec Mini) he said was more powerful and suggested I widen the gap to .040". The coil only lasted 20-30 hours. I checked the wires and they're fine. Maybe .040 is just too large a gap?

Seems unlikely. I looked at a plot of spark volts, amps, power, and energy yesterday. IIRC, a one millimeter (0.040") gap at 4 bar and 80F only needed volts in the teens to initiate the spark. That textbook is at the office. I'll post the plot Monday.

For now, Paschen's Law:

V = 3000 x P x D + 1350

where

V= breakdown voltage (first phase of spark initiation)
P= pressure in atmospheres
D= plug gap in millimeters

V = 3000 x 4 x 1 + 1350
V= 13350

A 0.030" gap would put voltage at 10350. 0.030" and 0.040" gaps should not stress a coil whose secondary is claimed to be good for 40K.

Speaking of which, the limited information on the Dynatec website says both the half ohm and 3 ohm models are for ignitions with dwell control, which suggests they were not intended for CDI. Go back and read post #2. Might be interesting to see if the primary resistance is now out of spec on the failed coils.
 
75KV

.....
For now, Paschen's Law:

V = 3000 x P x D + 1350

where

V= breakdown voltage (first phase of spark initiation)
P= pressure in atmospheres
D= plug gap in millimeters

V = 3000 x 4 x 1 + 1350
V= 13350

A 0.030" gap would put voltage at 10350. 0.030" and 0.040" gaps should not stress a coil whose secondary is claimed to be good for 40K.

Speaking of which, the limited information on the Dynatec website says both the half ohm and 3 ohm models are for ignitions with dwell control, which suggests they were not intended for CDI. Go back and read post #2. Might be interesting to see if the primary resistance is now out of spec on the failed coils.

You're keeping us hopping :) - had to review Paschen's Law. Wondering, though, where your 4 - for 4 atmospheres comes from. A 8.5 CR Lyc at full throttle should be up near 20 atmospheres. A CR of 8.5= pressure ratio of 20 (8.5^1.4).

That would make it V= 3000 X 20 X 1 +1350 = 61350 volts !!! Sum Ting Wong - coil is gone !!!
Maybe Paschen means pressure at standard temp. That would be more like gas DENSITY.

Then in the Lyc case, since compression ratio is density ratio, V = 3000 X 8.5 X 1 + 1350 = 26850 volts . This seems more reasonable. And it would seem to be a safe margin to a 40K volt limit.

Re: primary ohms. I think the "needs controlled dwell" is warning away use in an inductive system without proper current or dwell control, not limiting it to inductive systems.
inserted:
But now that I think about it, there is a concern, and it is not related to the damage of the primary of the coil that I mentioned in the para below. It is this: The .5 ohm coil likely has a very very high turns ratio. All things equal, if the primary of a 30kv coil is changed from 3 ohms to .5 ohms - same amount of copper, that implies an increase in turns ratio of 2.5. Then, if you hit the primary with the same whatever hundreds of volts, the output will be 75kv and near certain damage to the secondary turns by arc over if not limited by load.

It seem to me a CDI sys is unlikely to fry the primary of a coil. There is inherent limiting. And the voltage is controlled by the load - the gap and any resistance in the secondary circuit.

The cause I favor is an original coil defect or minor damage during initial testing. This can easily happen during testing where the secondary is NOT connected for a moment . While the internal damage might be very slight it may increase stress at that location which grows over many hours to final failure. Much like fatigue in a structure.

To make these systems as reliable as possible a built-in spark gap to limit secondary voltage might reduce the likelihood of internal damage. Spark gaps are common protection devices in many high voltage activities.

A quote in a performance auto site I read today on the subject: "Coils don't normally die - they're murdered." :)

Ron
 
You're keeping us hopping :) - had to review Paschen's Law. Wondering, though, where your 4 - for 4 atmospheres comes from. A 8.5 CR Lyc at full throttle should be up near 20 atmospheres. A CR of 8.5= pressure ratio of 20 (8.5^1.4).

The cause I favor is an original coil defect or minor damage during initial testing. This can easily happen during testing where the secondary is NOT connected for a moment . While the internal damage might be very slight it may increase stress at that location which grows over many hours to final failure. Much like fatigue in a structure.

A quote in a performance auto site I read today on the subject: "Coils don't normally die - they're murdered." :)

Ron

I'm not sure where you are getting 20 atmospheres from. An 8.5 CR engine will have a PR of around 8.5. The density ratio will be less however as a result of the temperature rise from compression and losses from having a VE less than 100%. Density would be the key factor here in required spark energy though the compression temperature rise would have a slight offsetting effect as the spark would have to input less energy into raising local temperature of the mixture between the the plug electrodes to initiate combustion.

A compression tester will illustrate what pressure ratio your engine develops at cranking rpm at least and this will be lower at 2500 rpm BTW, due to VE loss. You certainly don't see 294 psi on a Lycoming with a compression tester.

I totally agree on the last two paragraphs. Introduce a 2-3 inch spark gap somewhere in the system is bad news for a coil. It will arc internally which can damage some coils in short order. Have seen that more than a few times.
 
Last edited:
You're keeping us hopping :) - had to review Paschen's Law. Wondering, though, where your 4 - for 4 atmospheres comes from.

The figure came from the above noted graphs in Heywood, and a 4 bar reference in a Mike Busch article, but that doesn't make it correct for a given Lycoming at full throttle. You're right, it's too low in the context of this discussion.

As previously noted, the Heywood text is at the office, so I'll get back to it tomorrow. My guess? The 4 bar figure is probably typical for a late model automobile operating at part throttle cruise, as emissions and fuel economy research would be foremost at the MIT-Sloan lab.

Cylinder pressure at spark initiation would is not be a simple product of geometric compression ratio. All the model inputs would be in Gordon Blair's "Design and Simulation...", another text on the office shelf.

On the practical side, a Champion plug tester is 125 psi, 8.6 bar.
 
Last edited:
Back to the original poster's thread, which is very interesting- why three coils have failed. It would be interesting to see a schematic for the Lightspeed Plasma III. From the schematic, it would be easy to calculate the voltage and currents versus time- the coil specs, and knowing temperature, would get us closer to the answer of why the coils are failing. It's a binary search of probable causes to any solution.
 
I'm not sure where you are getting 20 atmospheres from. An 8.5 CR engine will have a PR of around 8.5. The density ratio will be less however as a result of the temperature rise from compression and losses from having a VE less than 100%. Density would be the key factor here in required spark energy though the compression temperature rise would have a slight offsetting effect as the spark would have to input less energy into raising local temperature of the mixture between the the plug electrodes to initiate combustion.

A compression tester will illustrate what pressure ratio your engine develops at cranking rpm at least and this will be lower at 2500 rpm BTW, due to VE loss. You certainly don't see 294 psi on a Lycoming with a compression tester.

I totally agree on the last two paragraphs. Introduce a 2-3 inch spark gap somewhere in the system is bad news for a coil. It will arc internally which can damage some coils in short order. Have seen that more than a few times.
Thank you, Ross, for your sage observations. Both you and Dan among others here are artisans in both talk and walk. Very few with that combo. Thx.

Re: the 20 bar pressure for 8.5 CR.
The justification for that figure is just straight thermodynamics.

Verbally, when you trap a volume of air in a cylinder and then reduce the volume, two separate things happen. First, the pressure increases as the volume is reduced. Second, the temperature increases as well, which adds additional pressure.

Getting into a little math, the exponent of the first is 1, the exponent of the second is .4 for air. Combined, it is 1.4

So, The relation of CR to PR is CR^1.4, or in the Lyc case, 8.5^=20.0072 PR

That pressure derived is of course, only a starting point, but is the proper starting point when we then roll in all of the other well known factors that will tend to reduce the number.
------------
As an interesting reverse example, those amazing jet engines that push the tubeliners across the sky have pressure ratios(PR) near 40. In terms of compression ratio(CR) it would be 40^(1/1.4)= 13.9CR

Ron
 
Last edited:
This is the illustration I was thinking about Saturday, from Internal Combustion Fundamentals, Heywood. Note the line "The 4 bar pressure roughly corresponds to engine conditions at spark onset". The reference calls out a book called Flow and Combustion in Reciprocating Engines...which magically, is available here:

https://r11tc.files.wordpress.com/20...ng-engines.pdf

Ahh, more reading. Looks interesting, and hopefully contains the rest of the story
-
 

Attachments

  • Spark Power Energy Volts, Current.jpg
    Spark Power Energy Volts, Current.jpg
    166.4 KB · Views: 160
This is the illustration I was thinking about Saturday, from Internal Combustion Fundamentals, Heywood. Note the line "The 4 bar pressure roughly corresponds to engine conditions at spark onset". The reference calls out a book called Flow and Combustion in Reciprocating Engines...which magically, is available here:

https://r11tc.files.wordpress.com/20...ng-engines.pdf

Ahh, more reading. Looks interesting, and hopefully contains the rest of the story
-
Hi Dan, could you please double-check the URL? Seems the automatic post URL formatter changed it. Thanks!

Never mind - found it here: https://gctbooks.files.wordpress.co...ustion-engine-fundamentals-by-j-b-heywood.pdf
 
Last edited:
Its density

My original assumption is right. Just density.
Found a straightforward document about "pressure" as it relates to spark ignition. Dan might like it. NACA 1920 report number 54.

https://core.ac.uk/download/pdf/42795274.pdf

Bottom line: pressure is only relative to spark requirement as it effects DENSITY.

BTW, I worked out the geometric CR for an 8.5 CR Lyc. at ignition time.
Included rod angle's small effect as well in the calc:
20 deg BTDC = 6.54 CR
25 deg BTDC = 5.80 CR
30 deg BTDC = 5.12 CR

my conclusion : don't give up ANY reliability just to get a "super spark", its not needed.

Ron
 
I currently have a Lightspeed Plasma III ignition system however, after 3 coil failures...

Three things I do to avoid Lighspeed coil failures (gleaned from my conversations with Klaus):
- Replaced the spark plug wires at 500 hours (per the Lightspeed manual). Klaus will make them for you to the lengths you need. Or you can make your own.
- Use the IK27 iridium plug, keep gap below .040"
- Spark plug wire supports - replaced Adel clamps with nylon cable clamps (allegedly reduces chance of spark plug wire shorting to ground).

My Lighspeed coil failures stopped when I took the first two steps.
 
Bottom line: pressure is only relative to spark requirement as it effects DENSITY.

BTW, I worked out the geometric CR for an 8.5 CR Lyc. at ignition time.
Included rod angle's small effect as well in the calc:
20 deg BTDC = 6.54 CR
25 deg BTDC = 5.80 CR
30 deg BTDC = 5.12 CR

my conclusion : don't give up ANY reliability just to get a "super spark", its not needed.

Ron

Yup. Agreed.
 
My original assumption is right. Just density.
Found a straightforward document about "pressure" as it relates to spark ignition. Dan might like it. NACA 1920 report number 54.

https://core.ac.uk/download/pdf/42795274.pdf

You bet, thanks. I like the old NACA papers.


Looks like an older edition of Heywood.

"Flow and Combustion..." here, add the usual https:// to the URL:

r11tc.files.wordpress.com/2010/05/flow-and-combustion-in-reciprocating-engines.pdf

Here's the 2007 Mike Busch article referencing Paschen's Law and 4 atmospheres. First time I had heard of Paschen and made a note on my hard drive.

https://www.avweb.com/ownership/the-savvy-aviator-43-high-altitude-misfire/
 
Last edited:
Back
Top