Ignition coils for small engines

Hi

 

Just been searching re high pressure v voltage breakdown and found this pdf file with data regarding pressure v voltage breakdown. It looks like a presentation document and page 16 gives a graph for breakdown voltages from low pressure to about 1000 bar; it also gives the low breakdown voltages for a variety of gases. Page 15 refers to the calculations and considerations re the breakdown processes.

 

I assume the data is valid as it has been produced at/for CERN , usual disclaimers etc.

 

Cheers

 

 

Martin

this was just placed on top of my scanner a moment ago to show you.

those silver eyelets are 5cm centre to centre.

the coil is designed for kettering type points.

it is epoxy filled. they are about $59 Supercheap auto in australia.

Hi Martin,

 

Thanks very much for the link. The breakdown properties of air are much clearer to me now. I'm also an official idiot for not seeing what was in front of me! I didn't spot that the X-axis of the graph is a product, pressure times distance.

 

So, let's say we have a spark gap of 0.5mm. According to the Paschen curve that equates to a breakdown voltage of about 2.5kV at one bar (close to normal atmospheric pressure). Without bothering to measure it I assume that my hit 'n' miss engine has a relatively low compression ratio, let's say 6. If we start with air at 1bar and 20°C and assume that the compression process is isentropic, and take the heat capacity ratio of air as 1.4, then according to my calculations we end up with air at 12.29 bar and 324°C at top dead centre. If we now go back to the Paschen curve we come up with a breakdown voltage of 25kV for a spark gap of 0.5mm and a pressure of 12.29bar. According to my Bosch Automotive Handbook spark voltages are normally about 25-30kV. At least this is in the right ballpark.

 

The question is though, do our ignition systems really generate 25kV? And if they don't, how come they work? ETW talks about ignition voltages for model engines ranging from 2kV to 8kV. I suppose that what it all boils down to is that at some point one has to stop fiddling around with the maths and go and build something to see if it works!

 

Regards,

 

Andrew

Andrew,

 

I suspect that the answer is in the air temperature. I would guess that the breakdown voltage falls off with increasing temperature, so perhaps 2 to 8 kV is enough?

The Paschen curves are not really applicable to a spark plug because of the geometry of the gap, they are for large parallel and smooth plates in Dry gas. Dielectrics ( fuel-air mix in this case) break down in a complex way. For example UV will cause the gap to fire at a much lower potential as will ionising radiation ( eg geiger tube) or carbon particles from previous cycles. Water vapor will also lower the breakdown voltage as will any  combustion products.

 

 For a spark plug the electrodes have relativly sharp edges and accumulated very small carbon particles so field emission will set in long before simple breakdown as the voltage rises accross the gap. The ionisation  will then avalanche the channel accross the gap.

So you don't need so many volts but rough points help a lot. That is one way in which abrading the points really helps.

 

Regards,

Alan.

Hi Hugh and Alan,

 

Good points, thanks! A further search on the internet found an approximation to the breakdown voltage that takes into account gas temperature. This brought the breakdown voltage for the previous parameters down to 12kV. However, as Alan has pointed out there are many other factors that fortunately conspire to lower the breakdown voltage.

 

So, I suspect the best we can say is that we need a few kV and a few mJ to produce a spark that will run a model engine. It is probably time to stop wielding the calculator and start building something. Bit of a relief actually, as the high voltage probe I bought for the oscilloscope is good for 15kV, so I don't need to find a higher voltage probe.

 

The first thing to do is knock up a small adjustable spark gap and see if my ignition coil setup will create a spark in air at STP.

 

Regards,

 

Andrew

I think your over complicating things.

Points and coil systems operating from battery voltages have been in use for over 100 years and work!

Now we have transistorised units that perform the same basic functions as points at battery voltages but you still need the high primery to secondery ratio to get a good spark.

If you want lower coil ratios use a power transfer system like a magnito or CDI, but then you need a higher primery voltage, and so the cycle goes on.

 

Steve

Andrew,

The spark gap voltmeter is easy to build and a good rough and ready indicator of what an ignition system is doing. If you use balls from a race that are not too chewed you will get around 2.5KV/mm for balls around 5mm. The advantage of balls is that they are more reproducable than points and less subject to wear. ( you get field emmission with a few volts and a sharp enough needle) Some people put the gap inside a glass tube to  a. jazz it up  b. improve the stability  c.pretend that they are Dr Frankenstein.

 

I would not use a scope probe unless you run into a gap to avoid the risk of flashing the series resistor of 1000M ohms ( if it's a 1000:1 probe). Most  "Car Ignition Scopes" use a coupling capacitor of a few pF ( turn or so of well insulated wire around EHT lead or a fancy clip) into a few M of coax with -say- a reliable 2.2nF capacitor across the coax to make a 1000:1 capacitative divider. You don't use a 1000M resistor because there are no low frequencies to see, the -3dB point being around 70Hz.  You calibrate the thing with the scope cal sig or an audio oscillator  or anything else handy- eg a multivibrator hooked up to the EHT lead ( coil disconnected) then adjust for 1000:1 division or whatever.

 

Normally you see a fair overshoot spike before the gap fires, you get to see what the mixture sees. The overshoot before ignition will jitter, it's the random delay before the channel in the gap ionises so the bigger the spike the less delay and the smoother she runs. Usually people trig the scope from the points/ Hall device. For more than one plug you bug the coil lead so you can see the different plug breakdowns sequentially( with a distributor). For seperate coils you run seperate taps to the coax inner.

 

It is a nice idea to always use DC scope coupling with a hv probe. If you go AC the series cap can charge toward   the  full direct voltage level ( 25KV on a CRT display) before it ionises and takes out the input stage of the scope.

 

regards

Alan.

Hi Steve,

 

It may be a bit complicated, but I like to try and understand what I'm doing before building stuff. I have very little workshop time, and I don't want to waste time building stuff that will not work, where I could have deduced that by thinking about it in the first place. Just because coils and contact breakers have longevity doesn't mean I want to use them. They might work but they are not reliable. I remember being out there in the wet fiddling about with eroded contacts when the car wouldn't run properly! On my hit 'n' miss engine I am planning to use a magnet and hall effect device. Neither do I particularly want to buy an ignition system that uses a Darlington transistor and a few wire-ended resistors. A Darlington transistor is a bipolar device and therefore has both majority and minority carriers. The mobility of minority carriers in the semiconductor lattice is slower than the majority carriers, so bipolar devices are relatively slow to turn off. I'm using a MOSFET and I let the flyback voltage from the coil avalanche the drain-source junction diode. This occurs at about 800V. So, 800V on the primary times a 10:1 step-up gives several kV. Which is roughly what I see on the oscilloscope. Once I've got a microcontroller in the system, it becomes easy to implement things like speed related advance and retard, multiple sparks and other things I haven't thought of yet.

 

I did think about a CDI type ignition, but I didn't feel like designing the step up converter. It would probably have to be a flyback style converter, and it's a pain getting small quantities of gapped ferrite cores.

 

Finally I think it is a nice juxtaposition to have an old style hit 'n' miss engine together with a surface mount microprocessor ignition system.

 

Regards,

 

Andrew

On some of the full size hit & miss engines some of the blokes use the (I think its) ballast transformer for fluoresent lights,using the back emf, no transistors allowed.Ian S C

Hi Ian, when I was at the Internal Fire Museum earlier this year the owner told me that they use fluorescent light transformers for some of their old engines, unfortunately I wasn't able to press him for any circuit information. I have looked at a transformer from an old light fitting and find only a two-wire connector so not sure how that would be used. We need someone clever enough to be able to describe how such a unit might be used in ignition systems, however, the transformers I have seen are really quite large so may not be suitable for smaller engines.

 

John

John, The flourescent transformers are chokes and would be used on an engine with low tension ignition. I have some of these on some of my models (the one as my avatar is one).

If anyone is not familular with this type of ignition let me explain. There is no spark plug but in its place is a set of points inside the cylinder head. The circuit consists of a just a battery ,the points ( called an ignitor) and the coil all connected in series. There is a mechanical arrangement that pushes the points together and breaks them at the right moment.

David

 

I must admit, I am tempted by a three-cylinder steam engine with solenoid operated valves under microprocessor control - not just reversible and speed controlled - but angular control and inching like a stepper, or even waving back and forth.

 

Neil

Strange but true, I did start building a three cylinder radial steam engine (Cygnet Royal, I think) during 'O' level metalwork. Of course I never finished it, but I do still have the bits somewhere. I also suspect that I was the only person in whole history of the school to use the shaping machine, while making parts for Cygnet. I had to use the shaper, as we didn't have any milling machines.

 

Of course any steam engine also deserves a microprocessor controlled steam whistle!

 

Regards,

 

Andrew

Hi David, thanks for the gen re low tension ignition. I have of course come across mention of such systems but hadn't found a decent explanation which you have now supplied. Presumably then the mechanical striker operates externally and is what you can see on early magnetos.

 

Good gen - thanks

 

Regards   John

Could of coarse go to a hot tube and forget electric, but its difficult in little engines. The original design of the motor I'm sort of putting together had this type of ignition, but I'v gone sparky, the engine has a bore of 2" and a stroke of 4", The design is about 120yrs old. Ian S C

Flourescent chokes are NOT designed with Kilo Volts in mind, they have pile windings so a varnish insulation breakdown is on the cards. If you want to wind an inductor for High Voltages then you wind in a layer then insulate before winding the next layer or you wind in a former that has slots i.e. small partitions along the core to insulate the windings. Impregnation with a good epoxy is not a bad idea.

 

Backlight invertor transformers ( from LCD displays) produce about 2.5KV pk-pk in a very small high frequency transformer ( 30-60KHz). They might be a starting point for a small ignition system. There are certainly enough of them thrown away by people who service displays.