Tests of CDI Ignition Modules

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Some clarity is in order here. The paper is a compilation of notes from a class and textbook. The reason for posting it was it contains the mathematical proof of where the energy is stored. It clearly states at the top of the notes HF.(high frequency) It is series of organized notes and is much different then a published paper.

HMEL, you keep on mis-stating your own reference, there is no proof that all of the energy is in the air gap, here is a direct quote from your reference "f. Air gap in magnetic core media. we review the solution to the magnetic circuits we face in inductor and transformer design on a DC basis. The bulk of the energy is primarily stored in the air gap not in the core itself" ---- there is no way that "the bulk of the energy..." is a mathematical proof that "all of the energy...", in fact the amount of energy in the air gap is proportional to the size of the air gap so it can be less than or greater than the core, it all depends on the design. Peter.
 
Don;
Other's may be interested in the technical aspect of making these measurements, so I'm replying to your post.

For laughs, I asked openAI (ChatGPT) this question;
I wondered when someone was going to bring chatGPT into the discussion. I think it'll be interesting to see some of the ideas (and failures) it comes up with. It's trained on a lot of PHD content. Asking it to create different signal filtering using python would be interesting too.
 
I wondered when someone was going to bring chatGPT into the discussion. I think it'll be interesting to see some of the ideas (and failures) it comes up with. It's trained on a lot of PHD content. Asking it to create different signal filtering using python would be interesting too.
I wanted to see what it would come up with.
My career was doing the qualification testing of commercial avionics products (Shake, Bake, EMI per RTCA DO-160) as part of the FAA and EAA certification process for TSO or PMA products. I spent hours trying to fix EMI emissions and susceptibility. Grounding and coupling capacitors (to a dead quite ground, noisy grounds were the norm) were the simple ways of achieving compliance to the desired emission limit line.
I also learned that relays clicking +28 volt discrete signals onto aircraft wiring (e.g. autopilot trim), would cause broadband noise that the spectrum analyzer used for the testing, displayed as inband noise. It was not continuous wave so we could explain that failure away, as it was non repeatable, but I had to learn that the hard way.
I dealt with some ESD issues, but they are nothing like a spark plug energy discharge. I understand Don's measurement issues.
 
  • Like
Reactions: Zeb
I wanted to see what it would come up with.
My career was doing the qualification testing of commercial avionics products (Shake, Bake, EMI per RTCA DO-160) as part of the FAA and EAA certification process for TSO or PMA products. I spent hours trying to fix EMI emissions and susceptibility. Grounding and coupling capacitors (to a dead quite ground, noisy grounds were the norm) were the simple ways of achieving compliance to the desired emission limit line.
I also learned that relays clicking +28 volt discrete signals onto aircraft wiring (e.g. autopilot trim), would cause broadband noise that the spectrum analyzer used for the testing, displayed as inband noise. It was not continuous wave so we could explain that failure away, as it was non repeatable, but I had to learn that the hard way.
I dealt with some ESD issues, but they are nothing like a spark plug energy discharge. I understand Don's measurement issues.
I spent some time at that sort of thing myself. It takes a good fundamental skill set, a rare instinct for EM behavior, passion, and a bunch of good luck to succeed.

Off topic anecdote: I once tested a batch of autopilot servos at Lightning Technologies in Pittsfield, MA, by hitting the capstan bearings with repeated 100,000-amp discharges to see if the bearings would weld and freeze the controls surfaces. It did, but it didn't. The welds were real but they were between the high-carbon steel ball bearings through grease to the races and could be broken free with minimal effort. Most people may not realize that an airliner is struck by lightning twice a year on the average. Not hazardous; passengers rarely know it happened.
 
  • Like
Reactions: Zeb
I spent some time at that sort of thing myself. It takes a good fundamental skill set, a rare instinct for EM behavior, passion, and a bunch of good luck to succeed.

Off topic anecdote: I once tested a batch of autopilot servos at Lightning Technologies in Pittsfield, MA, by hitting the capstan bearings with repeated 100,000-amp discharges to see if the bearings would weld and freeze the controls surfaces. It did, but it didn't. The welds were real but they were between the high-carbon steel ball bearings through grease to the races and could be broken free with minimal effort. Most people may not realize that an airliner is struck by lightning twice a year on the average. Not hazardous; passengers rarely know it happened.
You make an interesting point and why it's good to listen to both the pulleys and servos speak while they're freewheeling during inspection. Bad static wicks and braids on flight controls do some damage to bearings from microarcing. Lightning traveling through instead of along the surface sounds scary.

A locked bearing on a capstan would be pretty bad, since the clutch rotates with it. In the unlikely case of a failure, someone will probably blame the manufacturer for it, like the operators that blamed Boeing because they hired shady maintenance and failed to hire pilots who are keen on disabling AFCS on any plane they fly.
 
Here is something I found with RADAR equipment and ignitions, Don you and Ignator might know about this, VSWR (Standing Reflected Wave). It's been my experience with ignitions that this increases with the plug gap, wire resistance, and the air/fuel mixture. Basically anything that impedes the current flow of the spark will give me a reflected wave and cause either a little to a lot of noise. Even to the point of destroying components.

Now remember that the first radio transmitters used spark gap generators and then to the antennae. I've never seen one but, those suckers must have made a hell of a lot noise.

Probing (100:1) the spark I use a 1 meg ohm resistor at the end of my scope probe against the bare plug wire which, allows me to see the voltage. As for current that is another problem. I use a current pickup on the plug wire with a voltage divider. The big problem is to get the arc to behave just like in a combustion chamber. For this I'm thinking of using a gas discharge device like the ones we use to use on land lines for lightning strikes, just need to figure out which one. They should provide the high breakdown resistance and the low plasma resistance once an arc is established.

My experience with building and installing MegaSquirt fuel injection we normally get a lot of processor resets from the ignition noise. To get around this I use a noise kit for stereo equipment, normally a big inline inductor coil and E-cap to bypass the noise to ground.

Analog Devices VSWR

Ray
 
Here is something I found with RADAR equipment and ignitions, Don you and Ignator might know about this, VSWR (Standing Reflected Wave). It's been my experience with ignitions that this increases with the plug gap, wire resistance, and the air/fuel mixture. Basically anything that impedes the current flow of the spark will give me a reflected wave and cause either a little to a lot of noise. Even to the point of destroying components.

Now remember that the first radio transmitters used spark gap generators and then to the antennae. I've never seen one but, those suckers must have made a hell of a lot noise.

Probing (100:1) the spark I use a 1 meg ohm resistor at the end of my scope probe against the bare plug wire which, allows me to see the voltage. As for current that is another problem. I use a current pickup on the plug wire with a voltage divider. The big problem is to get the arc to behave just like in a combustion chamber. For this I'm thinking of using a gas discharge device like the ones we use to use on land lines for lightning strikes, just need to figure out which one. They should provide the high breakdown resistance and the low plasma resistance once an arc is established.

My experience with building and installing MegaSquirt fuel injection we normally get a lot of processor resets from the ignition noise. To get around this I use a noise kit for stereo equipment, normally a big inline inductor coil and E-cap to bypass the noise to ground.

Analog Devices VSWR

Ray
You've obviously "been there, done that" Ray. Standing waves and VSWR are used to describing reflected power in systems of fixed impedance transmission lines and single-frequency RF power flow. With broadband pulses, reflected power and energy are just as real, but much harder to describe and measure and analyze. Still the idea of VSWR helps to visualize reflected power. I visited a spark gap transmitter installation in Nova Scotia that was one of the original transcontinental Marconi stations used for radiotelegraph with England. Primitive and a very dirty signal, but hey; it worked. The had a water-cooled spark gap and keyed it on and off to send Morse Code.

I've made some progress figuring out how to make some of these spark plug measurements, and am about to write a summary of where I am, but it will take me some more time today to put that together. I'll describe what I'm using for a voltage divider and a current sensing resistor.

Don
 
In post #196 I told about the difficulty I have had making accurate oscilloscope spark measurements because of extreme noise generated by the avalanche discharge at the very beginning of the spark. I’ve managed to design some significant filtering for both the voltage and current channels on the scope. I’ve taken filtering about as far as I can without substantially distorting the waveforms I’m trying to capture. Just for anyone who may be interested, here’s what I have for the 1000:1 voltage divider:

1677099864656.png

I use a 10:1 probe on the scope to give further voltage attenuation to 10,000:1 (80 db) The filtering on the voltage divider itself is shown here, with 1000:1 attenuation (60 db) at dc.

1677099968635.png


I'll put the spark plug current sensing network in the next post.
 
And here’s the spark plug current sense network:

1677100137916.png


It’s basically a 10-ohm resistor, i.e. 10 volts per amp, or -20db attenuation at dc. Its filtering shown here:

1677100319429.png


This has helped a lot, but to get really reliable results it’s necessary to measure spark plug current to 1 ma or lower in an environment where peak currents can be an amp or more (as they appear even after filtering). They say that perfection is the enemy of good enough, though, so I’m going to go ahead with what I’ve got. I’ll repeat the early CDI measurements already given in post #1 and see if I get about the same results.
 
  • Like
Reactions: Zeb
Here is a typical spark waveform from my measurements on an automobile coil. It shows a little bit about the nature of sparks. First, here is one that is “normal” strength, like those you’ve seen on ignition analyzers. The yellow trace is voltage, blue is current. Voltage scale is 500 v/cm; current scale is 10 ma/cm. After the initial spike the arc voltage settles at about 400 volts until the arc quits after 2.7 ms, with ringing of the spark plug voltage at about 2.5 kHz for a few milliseconds with no current flow in the plug. Arc current after the initial spike decays linearly from about 28 ma, ending at about 3 ma. Doing the math on those numbers gives a 6 watt average over the length of the spark, for about 16 mjoules of energy delivered to the arc.

1677115033781.png


The arc is relatively stable during this discharge. One thing you can see from this is that the arc has a negative incremental resistance. Voltage stays nearly constant, even rising slightly, as the current decays nearly 10:1. All of this agrees with measurements I have made over the years and with what I have read of the theory of plasma arcs. I plan to show another set of measurements with a lower energy level as soon as I can put it together.
 
  • Like
Reactions: Zeb
Those look very clean of noise. Looks like the low pass RC worked per design.
So is that ringing frequency a function of your transmission line length, assuming it's the reflected wave bouncing back and forth?
 
Is there an issue using a contactless current sensor? Latency would need to be accounted for, as well as calibration, but maybe it might reduce circuit interaction?
I used to test SWR on an IFR4000. It was my first practical intro to reflections, but more like paint by numbers (per manual) in a neatly controlled environment. An ignition system seems like the wild west in comparison. It's nice though how those concepts transfer over to other topics.
 
HMEL, you keep on mis-stating your own reference, there is no proof that all of the energy is in the air gap, here is a direct quote from your reference "f. Air gap in magnetic core media. we review the solution to the magnetic circuits we face in inductor and transformer design on a DC basis. The bulk of the energy is primarily stored in the air gap not in the core itself" ---- there is no way that "the bulk of the energy..." is a mathematical proof that "all of the energy...", in fact the amount of energy in the air gap is proportional to the size of the air gap so it can be less than or greater than the core, it all depends on the design. Peter.
I am sorry you can not understand the mathematical proof presented. There are other proofs available in various texts if you like but you will have to do your own research. Pretty much done with this subject.
 
Is there an issue using a contactless current sensor? Latency would need to be accounted for, as well as calibration, but maybe it might reduce circuit interaction?
I used to test SWR on an IFR4000. It was my first practical intro to reflections, but more like paint by numbers (per manual) in a neatly controlled environment. An ignition system seems like the wild west in comparison. It's nice though how those concepts transfer over to other topics.
Good question. I'm not aware of a contactless current sensor that is truly broadband covering from dc through 1 GHz without frequency or phase distortion.
 
Those look very clean of noise. Looks like the low pass RC worked per design.
So is that ringing frequency a function of your transmission line length, assuming it's the reflected wave bouncing back and forth?
The noise problem is not so apparent in measuring strong sparks because the noise is all in the initial discharge spike and is short compared to the whole discharge. But who needs to measure strong sparks? You can see, hear, and even smell a strong spark. I'm working on a follow up post to show why measuring weaker sparks becomes a major problem with noise spikes.
 
Fantastic work Don. I appreciate it. Very nice clean waveforms especially since they are directly from the secondary.

Question:
What type of resistors (carbon/metal film/wire wound) are you using and their wattage rating.

Thanks
 
The noise problem is not so apparent in measuring strong sparks because the noise is all in the initial discharge spike and is short compared to the whole discharge. But who needs to measure strong sparks? You can see, hear, and even smell a strong spark. I'm working on a follow up post to show why measuring weaker sparks becomes a major problem with noise spikes.
And if you really want to get a sense of the spark, you can taste it. A nice spark across the tongue ... let's just say that it will be memorable! :)
 
Post #210 showed how to measure the energy in a large, stable spark. There was only a single spike at the beginning. For making energy measurements it could pretty much be ignored. With weaker sparks, however big spikes get in the way of measuring the spark energy. This picture shows that less current does not sustain the arc as well, and it tends to break up into instability. To measure the total energy you have to also calculate the energy delivered by each one of the pulses as well as that in the stable discharge.
1677182681933.png

With even yet less current a stable discharge is never achieved and the energy in the spark takes the form of a series of pulses (called a relaxation oscillation).
1677182801495.png

The screen shot below is an expansion of scale to show one single pulse. The voltage scale is 5 kv per dotted division. The peak voltage (yellow, negative) is at least 15 kv, but in less than a microsecond that drops to a level within the noise level of the scope. You can’t calculate the energy without a good measurement of the voltage during the current pulse, so you need more gain.
1677182884846.png

The screen shot below shows what happens when you turn up the gain one notch to 2 kv per division. This voltage trace doesn’t look anything like the previous one. The voltage spike at the leading edge of the pulse has saturated the oscilloscope amplifier, and the grotesque curve is the recovery time of that amplifier coming out of saturation.
1677182931676.png

So that’s the measurement problem with big spikes on the waveform.

The leading edge of that initial voltage spike, expanded 1000:1 in time, shows an abundance of ringing at a couple frequencies at once, about 60 MHz and 35 MHz. Peak voltage excursion is about 12 kv with the filters I now have in place.
1677183018418.png

As I see it, there are two paths open now to continue measurements. One is to do even more filtering on the scope inputs. A second is to try testing using a resistor spark plug or a series resistor in the spark plug wire. I’ll try the resistor first and report on how that works.

Don
 
Fantastic work Don. I appreciate it. Very nice clean waveforms especially since they are directly from the secondary.

Question:
What type of resistors (carbon/metal film/wire wound) are you using and their wattage rating.

Thanks
Just using 1/2 watt carbon comp Allen Bradleys, from my 50-year-old stock, except in the voltage divider I ordered carbon films with a 10 kv rating each. Spike times are so short they don't seem to ever break down or change even with these fantastic overvoltage stresses.
 
Don,

I want to thank you for working on this project! I always strive to learn something new each day and never really thought that I would be so interested in the spark of a sparkplug, but your presentation has peaked my interest.

They say a picture is worth a thousand words and after viewing the two on the 5 kv versus the 2 kv images all I can say is Wow! The 5 kv looks perky and alert while the 2 kv looks like a sad old Mama. ;)

As a layman in this field, I'm curious as to how the difference between sparkplug types may affect your work? Don't mean to put the cart before the horse here though. Again, thank you for your efforts.
 

Latest posts

Back
Top