Author Topic: Simple Transistor Ignition System  (Read 2261 times)

Offline DD805

  • Full Member
  • ****
  • Posts: 17
Simple Transistor Ignition System
« on: December 02, 2017, 12:47:52 AM »
When I first started making hit-miss engines, about 20 years ago, I got this schematic from Tom Stuart, another model engineer. I have used this circuit on all of my engines and have never had a problem getting it to work. You can ignore the distributor if you are using this on a single cylinder engine.



 I do not know who drew it originally but it looks a lot like the one on page 168 of Ignition Coils and Magnetos In Miniature by Bob Shores. He states he got it from Bruce Satra. The main difference I can see is that the Shores diagram includes a 10k ohm resistor between the coil and the spark plug where the circuit I use shows none. (I cant show the Shores circuit due to copyright.) I do wonder if the 10k resistor would cause a hotter spark?? Also the resistor between the transistor base in the Shores circuit is 20 ohms versus 10 ohms in mine.



The second picture is a close-up of the transistor connections. As I understand it DC current from the battery feeds to the transistor and is connected to the Emitter. Current passes from the Emitter to the Collector depending on the bias voltage of the Base. The connections are marked E C B on the transistor. As noted on the diagram different transistors use different arrangements for the connections. It could be B E C or C E B or whatever. Be sure to connect them properly.

Think of the Base as the handle on a faucet, turn it on current flows, turn it off it stops. The Base is turned on and off by connecting it to ground through the 10 ohm resistor. There is only a very small amount of voltage and current passing from Base to ground so no arcing occurs.

That is why you do not need special points. Any piece of wire will work. So where the diagram shows a set of points, substitute any kind of simple contact. I personally prefer small paper clips although I once used a piece of 22 g electrical wire. Any metal contact should work.

Below is one of my hit-miss engines. In the picture the yellow wire is the lead to the plug. The white wire is the Base to ground wire and the black wire is ground to the body of the engine. You can clearly see the connection to the paper clip contact. The blue thing is just a piece of plastic to hold the paper clip and keep it from grounding constantly. The paper clip extends down to a brass contact on the rod that operates the valve. The rod is electrically grounded to the engine body and then the ground wire. At top dead center the rod jumps forward and touches the Base ground contact, which turns on the current flow thru the transistor. It instantly moves back, breaks the base ground connection which turns off the transistor and of course causes the magnetic field in the coil primary to collapse and pop goes the spark plug.



Below is a drawing showing a couple of methods I have used to operate contacts. A is the one I described above and is specific to the engine. B is one where a shaft rotates a cam to make contact. In my application the cam is simply a round head screw on a small disk. There is probably a thousand ways to operate contacts.


A couple of comments on parts. The transistor specified is a NTE 332 which I used several times. When the store ran out they sold me a different but electrically equal substitute. Worked just fine. That made me think, why not other similar transistors. I happened to have purchased a junk bag of various transistors(about 400) for two dollars and started testing them in my circuit. Found that most power transistors work. I now have almost 50 transistors ready for use.
Power transistors are the bigger, sort of square ones versus the small round cans or round plastic with a flat area. Power transistors are everywhere. Tear apart almost any electrical device and you will find them. Go for the ones that are connected with long leads and just cut them off. De-soldering is a pain.
Be sure to mount the transistor on a heat sink. They can get very hot and even cause a fire. I once charred a piece of oak when I let the engine stop at top dead center too long. More than a few seconds can destroy the transistor. A piece of 1/8" thick aluminum 1"x2" will do. Oh yeah, don't grab hold of it until it cools off. Don't ask me how I know.

I have only used this on single cylinder, low speed engines so I do not know if it will work on fast multi-cylinder engines.
Be aware there are two different types of transistors, PNP and NPN. I can't explain the differences but the NTE 332 is a PNP type. I have been told that an NPN will work just as well and the connections are the same. Its just that with an NPN you would reverse the power connection.

Use slide on connectors to attach the wires to the transistor rather than solder. It is easy to ruin a transistor by allowing the engine to stop at a point where the contact is closed causing a constant current flow. Only takes a short time to overheat and burn out. With slide-on connectors replacement is a snap.
Lastly, I hope some of you with a better understanding of transistors and electronics will jump in and correct any errors or misconceptions I may have made.

Karl

Offline 10KPete

  • Full Member
  • ****
  • Posts: 1561
  • Nordland, WA, USA
Re: Simple Transistor Ignition System
« Reply #1 on: December 02, 2017, 02:39:09 AM »
Great information, Karl! Thanks for posting this. I have heard of 'battery saver' contacts being used that interact with the intake valve so no current can pass unless the valve is closed...?????

Pete
Craftsman, Tinkerer, Curious Person.
Retired, finally!
SB 10K lathe, Benchmaster mill. And stuff.

Offline MJM460

  • Full Member
  • ****
  • Posts: 894
  • Melbourne, Australia
Re: Simple Transistor Ignition System
« Reply #2 on: December 02, 2017, 06:40:45 AM »
Hi Karl, thanks for posting that.  I am not up to an ignition engine yet, but I find the circuit interesting, and solves that issue for when I am up to trying a build.  It is not too hard to analyse, so I may be able to add some comments which might help with your burning out issue.

As you say, the spark occurs due to the collapsing magnetic field in the coil when the contact opens.  But first you need the contact closed to build that field.  That can happen slowly, so long as it is up to strength before the contact opens.  As drawn, when the transistor, which is just being used as a fast electronic switch, is 'on' due to the base being connected to ground, the current flows through the primary coil,  and the magnetic field builds up.  The coil generates a back emf while the field builds that limits the current.  In the 'on' state, the transistor voltage drop is only about 0.2 V, so the current is controlled by that back emf.  As  the field builds, the back emf fades away, until the field is the maximum for the applied voltage, no back emf and the current is only limited by the DC resistance of the coil  which is quite low, so high current and the transistor heats.  That 10,000 ohm resistor you mentioned does not affect the spark at all, although it will slow the time needed to form the magnetic field, so the 'on' time of your transistor may have to be longer.  The spark is due to the emf caused by the collapsing field as you have described.  The voltage is proportional to the rate of current change, which for a fast opening switch is obviously extremely high.  The resistor plays no part, as the only circuit to discharge the coil emf is through the spark, not through that resistor. 

However, if your points stay closed, the transistor stays 'on', the current continues to flow, but as the field is max, there is no current change and no resistance in the circuit, effectively a short circuit on your battery, no wonder the transistor heats and the battery goes flat fast.  That 10,000 ohm resistor now comes into its own and limits the current to something the transistor and battery can stand all day. 

If you feel the spark is not as strong with the resistor in the circuit, a 5,000 ohm resistor would probably do as well.  I have never seen the Shore circuit, but I would suggest a slightly more sophisticated circuit would include more resistance than the 10 ohm in the base circuit.  I would try 1,000 ohm, and if it does not seem as good, try 100.  It's main purpose is to protect the other components if the transistor fails, normally it's value is not of great importance.  But the circuit might also include a higher resistor say 100,000 ohms between the base and the positive, so the base is firmly pulled away from ground to positively turn off the transistor, rather than being left floating when the contact is open.

I would still use the heat sink you mentioned, but the cam may need modifying to lengthen the 'on' time, it is only being able to precisely time when that sudden opening occurs that is really  important.  However, the contact closing should also be quick when it occurs, just not a 'when' issue.  Using an NPN is no issue, probably preferable even, with the right circuit inversions, but just a bit more complicated then 'switching the connections'.  We had better come up with a full circuit for peer review if someone wants to do it that way.

That's my hobbyists analysis of the circuit, so still plenty of room for those with more electronics knowledge to come in and help or correct.

MJM460
The more I learn, the more I find that I still have to learn!

Offline jadge

  • Full Member
  • ****
  • Posts: 369
  • Cambridge, UK
Re: Simple Transistor Ignition System
« Reply #3 on: December 02, 2017, 11:07:30 AM »
As said the output voltage from a coil based ignition system is all down to change of magnetic field with time. So the faster it changes per unit time the higher the output voltage. It's not easy to measure changes in magnetic field, so di/dt in the primary is often used instead.

Obviously the circuit shown works, but it has some disadvantages. As hinted at, if the engine stops with the contacts closed then the transistor can overheat. Base current is controlled by a low value resistor, and since collector current versus base-emitter voltage is exponential, variations in supply voltage lead to large collector current changes.

I suspect this circuit will not be particularly effective at high engine speeds. In saturation the transistor will have an excess of charge carriers, which take time to clear when the device is turned off. So turn off is relatively slow and hence di/dt is also slow. In this respect the choice of a PNP transistor is a little odd, as holes have lower mobility than electrons. So in general a PNP transistor will be slower switching than a NPN transistor.

I have a copy of the book by Bob Shores. It's an interesting read, but suffice it to say I don't always agree with his technical analysis.  :thinking:

When designing an ignition system for my hit 'n' miss engine I set a design goal of a 1:10 ratio of primary to secondary turns on the coil. That means di/dt needs to be fairly fast; I aimed for a transistor turn off time of less than 100ns. I chose to use an avalanche MOSFET as the drive transistor. Since MOSFETs are majority carrier devices they have faster turn off times than some bipolar transistors. When  the current is turned off the voltage on the end of the primary winding increases in an attempt to keep the current flowing. The MOSFET I chose had a Vds rating of 800V, so the voltage on the drain rose to 800V before the inherent drain-source diode broke down. In an avalanche MOSFET this diode is characterised to withstand repetitive breakdowns within a given current limit. It's not clear how the circuit shown deals with flyback voltages? Practical results from a test circuit gave about 8kV peak output voltage on the secondary as seen on the oscilloscope, before I overheated the drive transistor, as it had no heatsinking.  :embarassed:

The results were encouraging, although I may have to use a turns ratio of 1:20 to get the output voltage I want. The next job is to run the circuit with a spark gap on the secondary, rather than open, so I can see what happens with a simulation spark plug. The other thing I need to do is make the spark plug I have designed and see if it works.  ;)

Andrew

Offline DD805

  • Full Member
  • ****
  • Posts: 17
Re: Simple Transistor Ignition System
« Reply #4 on: December 03, 2017, 05:21:57 AM »
10kpete, Yes some modelers use a Battery Saver technique on hit miss engines. I have one on my Fairbanks Morse 25 HP. It has been many years since I made this and I honestly cant remember how it works. I suspect there are many ways you could prevent it from firing while it is coasting.

jadge, Hey thanks for the great information. You confirm my belief that this circuit would not be good with a multi-cylinder high speed engine. However with the slow speed hit miss engines I make it works great. By the way, I just can't get my mind around the idea that the holes move. Electrons moving yes. Holes no. I got my electronics training in the NAVY in 1960 and it was all vacuum tubes. Missed getting transistor theory by about 6 months so in my world there ain't no holes.

MJM460, You and jadge certainly know a lot more about electronics than I. Thanks for your input. I am very interested in the info that the 10k resister will help prevent the transistor from burning out. I will have to experiment with this. Due to the slow speed of the average hit miss engine I doubt one needs worry about dwell time being too short. Electrically the amount of time one of these engines holds the contact closed is like forever.

Karl

Offline MJM460

  • Full Member
  • ****
  • Posts: 894
  • Melbourne, Australia
Re: Simple Transistor Ignition System
« Reply #5 on: December 03, 2017, 08:54:02 AM »
Hi Karl, for a standard DC circuit you are correct about the time being forever, but as that coil is an inductor, a reactive element that stores energy in a magnetic field, and actively opposes a change in current as jadge says, the circuit then has an associated significant time constant.

When the inductor is in the circuit, instead of ohms law, which gives a current which is independent of time, and is established at about the speed of light, you need an exponential function.  The current is given by
  I = V/R x (1 - e^-tR/L)
where V is your battery voltage, R is the circuit resistance which is the sum of the DC resistance of the coil and the extra resistor you insert.  L is the inductance of the coil in Henries.  Then t is the time in seconds since the circuit was switched on. 

After a long time, the formula becomes essentially I = V/R, or standard ohms law, but a long time is greater than 3 times L/R.  You can see why L/R is called the time constant, and it has units of seconds.  It works like a half life of radioactive decay, absorption of your medication or cooling of a boiler once the burner is extinguished.

I don't know the value of the inductance of your coil, but a good modern digital multimeter will measure it directly.  My meter has it, and it was not particularly expensive.

So for your every slow engine, your brief momentary contact may be enough.  For a higher speed engine, you may need the full revolution to have 3 x t available.  For a multicylinder engine you may need a separate cam and coil for each cylinder before you run out of time, but with those small cheap coils you use, multiple coils and separate cams is probably a better option than a distributor.

I suspect there is diminishing returns on increasing R to decrease t, which is what the maths tells us to reduce the time constant, as, though you get to max current quicker, the max current is lower, so there is less stored energy to create that spark.  Even for the slow engine, the value of the resistance is a trade off between current when the contact stops with the circuit 'on' and enough energy for a good spark, so with bigger resistance, more time is needed to provide a good spark.

I hope jadge will come back if this explanation needs some correction or even just a little tweak.

MJM460

The more I learn, the more I find that I still have to learn!

Offline jadge

  • Full Member
  • ****
  • Posts: 369
  • Cambridge, UK
Re: Simple Transistor Ignition System
« Reply #6 on: December 10, 2017, 12:02:50 PM »
I think the above is a fair description, but does illustrate some of the issues with a simple ignition circuit.

The first thing I looked at when designing my ignition system was how much energy do I need? From the literature and talking to people in automotive research a stoichiometric petrol/air mixture requires about 150-200J for ignition. In practice most ignition systems supply much more, from about 2mJ upwards. Current high performance coil-on-plug systems (Mclaren) provide about 20mJ and some after-market systems reach 100-200mJ. Coil based systems can have a closed magnetic core, or not. Having a closed core usually results in a higher inductance for a given number of turns, but on the minus side the core may well saturate at lower currents, limiting the amount of energy available. As a rough guide primary inductance for an air core ignition coil is around 100H while that for a closed core is more like 1mH.

I chose to use an air core. A good approximation of the inductance of an air core coil can be calculated from the appropriate Wheeler formula. I chose a value of about 100H. In order to avoid problems with destructive currents, I included an analogue current limit, based on a sense resistor in the output transistor source. Timing is such that I operate in the first part of the current equation, so the current rise is pretty much linear with time. It also means I avoid any saturation issues.

The intention is to use a simple processor to control the ignition. That way I should be able to tune the on time to the engine revolutions and avoid issues if the engine stops with the contact closed.

In theory adding some soft iron in the core of the coils should improve performance. The energy is still stored in the air gap, but the lower reluctance of the iron should encourage the lines of magnetic flux to flow through the core and improve coupling between primary and secondary. In practice it didn't seem to make much difference.  :???:

Andrew