Universal small engine ignition & Science

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.

74Sprint

Well-Known Member
HMEM Supporting Member
Joined
Apr 21, 2013
Messages
276
Reaction score
71
Location
Winnipeg, Canada
I been this forum for some time and one of the things I noticed is that there seems to be quite a bit of head scratching when it comes to internal combustion ignitions. Mostly how to make and use them. I guess I should mention that I have been studying internal combustion engines for 40 years and building ignitions for 35 years now. I have a very strong electronics and racing engine (automotive) background. I'm still trying to get equipment to make minaturers and I will. I recently down graded my small business to hobby scale and will be offering electronic modules for sale when they are ready.

Some of things that were talked about that I came across are coil size, triggering, power supply, and spark strength. I want to start by talking about what is going on inside the combustion chamber and then what is needed to make a bang.

How much spark do you need?
Well that depends on 3 things, how much air, fuel, and compression you use. Air is one of the best electrical insulators you can use, just look at any power pole and look at the insulators. They use nothing but porcelain or glass and air for insulation. Fuel (hydro-carbons) are also a dielectric (insulator), in fact NHRA & IHRA both measure the dielectric strength of the fuel in your tank to make sure the fuel is legal. So when you combine air & fuel together you get a pretty good insulator, now when you add in compression and squeeze that mixture you the resistance of the air/fuel mixture is an even higher spark resistant insulator. So the more air, fuel (and type of fuel), and the higher the compression the higher the spark voltage required to jump the sparkplug gap. You can make the spark easier to happen by closing up the gap or increasing the spark voltage. This next part is VERY important and most people don't know about; once a spark has been created the resistance in the sparkplug gap falls to near zero. When this happens the secondary winding is shorted out and all that high voltage is converted into current. It is the current that actually does the burning. Trust me on this one I have been burnt and seen the difference between a current burn and a voltage burn on my body and others. This is also why I like an inductive ignition over a CDI ignition. Note that all the top classes in drag racing use magneto ignitions. If you still think that high voltage does the burning then you'd get burnt every time you got shocked by static electricity. A low compression (4:1) engine using hydrogen can run on a spark as little as 8,000 volts (8Kv) or a high compression (17:1) engine using 120 octane racing gas would need more than 60,000 volt (60Kv) spark, I have worked with both.

The ignition coil:
The ignition coil is probably the most understood part so here comes the science. I'm sure everyone here knows that an ignition coil has 2 coil windings, a primary and a secondary. The primary makes a magnetic field which collapses and induces current flow in the secondary, stepping up the voltage and making a spark. The important parts are the wire gauge used, how it is wound, how many windings in each coil (ratio), and the iron core. The wire gauge will tell you how much current can flow in each winding before it burns up, in both the primary and secondary. The inductance (#of coil windings), current flow, wire resistance, and operating frequency (reactance/impedance) will dictate the size of the magnetic field in the primary winding and how much comes out of the secondary. Ignition coils have always been a balancing act and are usually designed for a certain voltage and current (both sides) and how they are to be used (size, rpm, and cost), no magic bullet here. Ignition coils are wound to form an auto(otto) transformer can't remember proper name but, you wind the primary around the iron core then wind the secondary around the primary winding, anyway that's what I was taught. Now when current flows through the primary the magnetic field starts to build and it will cause a small inductive spike in the secondary. It takes time to build that field (dwell time) and we can calculate the time constants for a given coil size (charge time) but, not now. Once the field is built we can collapse it, when this happens the field not only cuts through the secondary and inducing a current flow but, it also cuts through our primary inducing a large voltage spike in the primary. It doesn't matter if we are using points or a transistor it is still there. Mind you the faster you can cut the current flow in the primary the faster the magnetic field collapses the stronger the spark, so component choice is important. For points we use a capacitor/condenser to absorb this spike and stop the points from burning up. Have you ever noticed that when you pull a cord out of a wall socket and noticed the spark? Well that spark is created because current wants to always continue to flow in the same direction even when the current is removed. That current in the wire builds a magnetic field and when this field collapses it tries to keep the current flowing. But there is a sudden high resistance when the blade of the cord leaves the socket, current suddenly bunches up the voltage rises even further and a discharge spark happens. No magic and nothing is wrong. As for ignition coil size, well that's a little hard to say. How much input voltage are you going to use? What turns ratio do you need? How much current can your battery/ies put out? How much spark do you need? What is the average operating RPM that you expect? These all need to be answered before I can tell you what coil or cap to use. I had custom coils made for RC engine conversions that work with 6-12 volts but, I still need to package them for retail sale.

Points and the capacitor/condenser size.
Yes I read all kinds of explanations of how/why we need a capacitor but I think I explained above why, now for sizing. To size the capacitor for points you need to know the size of the primary coil in Inductance (Heneries), resistance of both the cap and coil (primary & secondary), and the average operating RPM. Unless you have a way to measure (scope) the voltage of the kick-back spike try to use a voltage rating of the capacitor of about 400v or as high as can afford. I like to use polyester metal foil caps because they are self healing. There are 2 things we would like to know but, the most important is the charge time which, relates to the stored energy amount. Both the coil and the cap need time to charge up & discharge and this is expressed in Time Constants, we only use the first 5 time constants normally.

Ok the calculations:
L = inductance in Heneries (no milli or micro in decimal only)
C = capacitance in Farads (no milli or micro in decimal only)
T1 = 63.20%
T2 = 86.50%
T3 = 95.00%
T4 = 98.20%
T5 = 99.30%

Coil time constant:
T = L/R, in seconds
So let's use my coils:
T = 0.00001701 (Heneries)/ 0.1 (Ohms)
T = 0.0001701 total time in seconds
So:
T1 = 0.000105972 seconds
T2 = 0.000147137 seconds
T3 = 0.000161595 seconds
T4 = 0.000167038 seconds
T5 = 0.000168909 seconds
So it takes my coil 0.17 milli-seconds to fully charge (implied).

Now we need to find a capacitor that's not to small or to big. Too small and the points will still arch, too big and the cap will slow down the coil discharge (that's bad).

Capacitor time constant: An electrolytic capacitor of .22uF to .47uF is usually used.
T = C x R
T = 0.00047F (which usually has a resistance of 0.5 ohms) x 0.5
T = 0.000235 seconds
So:
T1 = 0.00014852 seconds
T2 = 0.000203275 seconds
T3 = 0.00022325 seconds
T4 = 0.00023077 seconds
T5 = 0.000233355 seconds

So the 0.47uF capacitor is a good match for the coil size, not too big and not too small.
We can next check if they are a good match at the average RPM.

average RPM = Resonant Frequency
L = inductance in Heneries (no milli or micro in decimal only)
C = capacitance in Farads (no milli or micro in decimal only)
π = 3.1416
resonant-frequency-formula.jpg

Fr = 1/2 x 3.1416 x 0.00008941
Fr = 1,780 hz / 60
Fr = 296 rpm avg. RPM

Now this may sound low but, I wanted a coil resistance that would give excellent engine starting. When the cap and coil are at resonance current flow will be at max. Besides the value of the cap will work well beyond any engine RPM of any engine I'm familiar with.

Ray

See Part 2 below
 

Attachments

  • Schematic Ignition Using CD4047B.png
    Schematic Ignition Using CD4047B.png
    213.8 KB · Views: 484
  • Ignition Using CD4047B.pdf
    97.1 KB · Views: 107
  • Schematic Ignition Using CD4047B.png
    Schematic Ignition Using CD4047B.png
    213.8 KB · Views: 453
  • Ignition Using CD4047B.pdf
    97.1 KB · Views: 71
  • Metal Detector.jpg
    Metal Detector.jpg
    20 KB · Views: 407
  • Metal Detector Options.jpg
    Metal Detector Options.jpg
    29 KB · Views: 398
  • Board Bottom Layer.pdf
    77.2 KB · Views: 84
  • Board Top Layer.pdf
    77.4 KB · Views: 65
  • Ignition Using CD4047B Board Use.pdf
    88.1 KB · Views: 80
Hall Effects:
If you do a search on DigiKey.com for magnetic switches you'll find it returns 4,291 different kinds. But we are only interested in 41 of them. These are listed as "MAGNETIC SWITCH UNIPOLAR 3SIP" we want this type because it will only react to a south pole on the front face, it acts a a switch and is not linear, and the 3SIP (3 long leads) is easy to solder to a board or connect to wires. Using a magnetic switch means that the Hall Effect turns on when the magnetic field reaches a upper (+) level and turns off at a lower (+) level. It doesn't need a second reversed magnet field (-) or to see a zero (0) level.
This is IMPORTANT; if you use a south pole Hall Effect you can epoxy the SOUTH face of a magnet to the backside of the Hall Effect so that it detects ferrous metals such as iron or steel. What this means is that you can remove the points and use the points cam lobe to trigger the Hall Effect. You can also use a woodruff key, square key, or even a rocker arm to trigger the Hall Effect. Allegro is probably one of the largest Hall-Effect makers https://www.allegromicro.com/en/Products/Magnetic-Digital-Position-Sensor-ICs.aspx and they have tons of different types.

For tooth or lack of.
63377-82409101fe28e201a2a5edf3b54f2f8b.jpg

These are usually linear and require additional circuitry like an op-amp.

63362-20554f67ac8b0c75533c48039f015dea.jpg

This is an example of tooth triggering, there is more to this and I recommend reading this application note: https://www.allegromicro.com/en/Des...ations/Hall-Effect-IC-Applications-Guide.aspx or for more info https://www.allegromicro.com/en/Des...ments/Hall-Effect-Sensor-IC-Publications.aspx. Allegro even offers moduals that already have the magnet embedded https://www.allegromicro.com/en/Pro...nsor-ICs/Wheel-Speed-Sensor-ICs/ATS19200.aspx.

The Ignition:
Ok enough science and now for my universal circuit. My parameters for this ignition are, cheap, simple to make and understand, work with points or Hall Effect, and work with most available coils from 5v -15volts. Coil current can be anything less than 26 amps! The key to my circuit is the CD4047B Multivibrator, used here as a one shot multivibrator. As a one shot the multivibrator CANNOT be retriggered until it times out, this eliminates point bounce, noise, and false triggering. The CD4047B allows for either a high level, a low level, a rising edge, or a falling edge for triggering. It also has either a negative or positive output so one can use either a PNP, NPN, IGBT, or a MOSFET for charging the coil and it allows one to adjust the dwell (Charging Time of the coil). The CD4047B can also be used with a step-up transformer to provide in a CDI the high voltage needed for the CDI capacitor. The CD4047B has been around for over 35 years and is very much overlooked for it's versatility. You can use whatever Hall-Effect you want but I have chosen the Diodes Incorporated AH3362Q because it has excellent overall features for automotive use. It is a South Pole Switch type and can be used to detect the very small (1/8" deep x 1/16" width) neodymium magnets I use or ferrous metals with the magnet behind it. For use here I recommend the ON Semiconductor TIP50G High Voltage NPN transistor which can handle 1 amp and has a Collector − Emitter Voltage rating of 400 vdc. For an IGBT I recommend the ON Semiconductor FGP3040G2-F085 in a TO-220AB package, this also has a 400 vdc rating and can handle up to 41 amps or 26 amps continuous (when pulsed / properly heatsinked). There are also purpose made transistors, IGBTs, and MOSFETs for ignition purposes that can be used with this circuitry.
I haven't built this particular circuit but, I'm confident almost everyone can get it to work. If you have any questions or need help either post here, PM me, or email me at [email protected]. I'm quite busy so expect up to perhaps 5 days for me to answer, sorry about that.

Ray

63360-cde13a7cf2daf693d66b9cdeb648656b.jpg


Files:
View attachment 109537 View attachment 109539 View attachment 109541 View attachment 109556
 
Plenty of theory there for use to use !
With the circuit using a npn transistor i would have the coil on the collector side of the transistor so the transistor is low side switching and use a flywheel diode to protect the transistor - same for an IGBT or mosfet although these will need a dedicated driver circuit to ensure they turn on and off fully due to the gate capacitance .
If using a pnp transistor the coil is shown positioned correctly in you schematic .
 
[QUOTEand use a flywheel diode to protect the transistor][/QUOTE]
A diode will protect the transistor but will not allow any voltage to develop at the secondary.
A properly sized capacitor will control the rise of the voltage on the transistor and the peak, but usually the spark shorts the secondary before the peak is reached.
 
Plenty of theory there for use to use !
With the circuit using a npn transistor i would have the coil on the collector side of the transistor so the transistor is low side switching and use a flywheel diode to protect the transistor - same for an IGBT or mosfet although these will need a dedicated driver circuit to ensure they turn on and off fully due to the gate capacitance .
If using a pnp transistor the coil is shown positioned correctly in you schematic .

Normally and using a normal transistor you would be correct. I have done it both ways depending on the transistor, IGBT, or MOSFET. With HV transistors I find I get faster turn-off times with the coil on the emitter (reverse biasing voltage from the coil primary when the field collapses which, I have found to be very important). With IGBTs and MOSFETs the coil should usually be on the collector/drain. As for a flywheel diode I've never had to use one in this configuration (even with 2n3055's) depending on the primary coil resistance and/or with HV transistors. You have to remember that I drew the schematic based on That Transistor and With My Coil, there are so many different configurations that I could not possibly draw all of them. The 2 main things I wanted to point out was the way a Hall-Effect can be used and The main item the CD4047B multivibrator. If someone wants me to use a certain transistor, IGBT, or FET that they have I'm sure I can redraw the schematic and redo the board layout for them, no biggy. Oh not all IGBTs or FETs need a driver circuit to turn them fully off or on, it all depends on the specific device and yes there are times I've had to use drivers or make my own circuit mods. When I get a chance I'll build the circuit and post the results along with images of scope, voltage, and amps results and/or modifications, well lunch time is over.

My automotive test bench. I'm also building a small engine ignition workbench.
High Voltage Automotive Workbench.jpg


Cheers
Ray
 
Hello Ray greatly appreciate your expertise this helps tremendously :)
it looks like you missed the decimal point above.
T = 0.00047F = .47mF = 470uF

Your right, it must have been a brain fart. Sorry about that, also I don't know how to edit that post, can anyone help me out?
T = 0.00000047 Farads = 0.47uf

Also sorry about the picture of the testbench being so messy but, it's the mad scientist thing.

Ray

P.S. This is for anyone that wants to wind their own ignition coils. https://www.nutsvolts.com/magazine/article/high_stepping_automatic_coil_winder.
I'll be making one in a modified form to make mag coils, pickup coils, and CDI HV transformers. I can't find a supplier domestic or foreign that can make them at the price point I need. Nice part about this winder is that I can modify and upgrade the code to suit my needs.
 
Last edited:
P.S. This is for anyone that wants to wind their own ignition coils. https://www.nutsvolts.com/magazine/article/high_stepping_automatic_coil_winder.
I'll be making one in a modified form to make mag coils, pickup coils, and CDI HV transformers. I can't find a supplier domestic or foreign that can make them at the price point I need. Nice part about this winder is that I can modify and upgrade the code to suit my needs.

I built the Side Winder by Bob Shores and am very happy with it. Here is a video of it.
 
>>>I haven't built this particular circuit but, I'm confident almost everyone can get it to work. If you have any questions or need help either post here.......

As I understand it you haven't built this circuit yet??
I encourage you to build it as presented and report back on your results.
Using a scope see if the transistor fully saturates driving a real coil.
Can you get a reasonable spark at say 1500 sparks per minute (3000 rpm single cylinder 4-stroke engine) = 25Hz trigger freq.
If so does the transistor stay cool.
A video of it working with a typical spark plug would be nice.
I'm suspect of your transistor and ability to drive it as presented. The 40447 only has an output drive capability of 2ma. The transistor gain is about 20 so you only have enough drive to support 40ma of collector current.
Even if the 4047 could drive the transistor properly the 1k resistor on the base of the transistor is, at best only going to allow 12v/1k = 12ma of base current. With the transistor gain of 20 the collector current would only be 0.24 amps.
(transistor gain from spec sheet - depends on voltage, expected collector current and manufacturing tolerances. Your mileage may vary)
 
Last edited:
There are plenty of circuits out there that work great. Problem is when you start building v8's and 9 cylinder radials. The circuit needs to function in the 25 to 30 thousand sparks per minute. If someone could pull that off For a small engine with 6:1 compression then we would have something great.
 
There are plenty of circuits out there that work great. Problem is when you start building v8's and 9 cylinder radials. The circuit needs to function in the 25 to 30 thousand sparks per minute. If someone could pull that off For a small engine with 6:1 compression then we would have something great.
If you split your ignition circuit into two you can spark alternately. Divide even further and thats why you see modern practice as coil on plug.
 
dsage - your right I should get this done but, right now I have people waiting for me to get some other projects done, these will generate me income which, I very much need at this time. But yes I will get it done. My next project which I have just started this week is related to ignitions and I'll try to do both in parallel.

stevehuckss396 - Actually 25 to 30 thousand sparks per minute is no problem, my best low power single coil system is 52,000 per minute (13,000 RPM for a 4 cycle V8) and and high power single coil is 60,000 per minute which is the max allowing for charge/discharge time of the coil. Doing a V8 is easy doing a 9 cylinder radial is a bit more complicated but it's the size of the distributor/signal generator that is is the big problem there.

The biggest problem here (this forum) and other's that I'm on is that everybody has a different idea of what they want for an ignition like - I don't want to use a battery bigger than 3.7 volts or I only want to use a 9 volt battery or no (NiCad or NiMh or Lithium based) or it can't be bigger than 8 sugar cubes or no CDI or has to be CDI or it's for points only or no Hall Effects ..... the list goes on and on.

So unless a person wants to make their own ignition design what would make a good universal ignition? Not a perfect ignition but one that works on almost anything and works with a 6-12 volt battery or power supply. I have made an ignition that's for automotive use, it draws 48 amps during coil charging, and puts out a 1 amp @100,000 volt peak spark.

But yes I will get the universal ignition done and sell them to whoever wants one. I'm even working on a programmable one also.

Ray
 
If you make a board that can deliver 50000spm I can adapt to what ever you make. I use a lawn tractor battery so if its 12v or lower I'm good. Roy's units I just put a 5 volt regulator in the base of the models along with the ignition board.
 
I currently use a 11.7 volt 2,000 mah Li-Po battery, it's fairly cheap and lasts for 6 hrs. on the low power ignition. That low power ignition is for my .65 cid 4 stroke RC airplane motor which is still sitting on the bench collecting dust. Here is a link to a early development video of a 2 stroke CDI running on 12 volts 37,200 RPM, the 5 volt version can do 21,000 RPM, voltage droop is the limiting factor. I haven't set it up for an IDI ignition (plain transistor inductive ignition) so I don't know what that can do.

Steve I found the linear voltage reg from 12v to 5v creates a lot of excess heat but, this heat can be decreased by using a 6-7.2v battery or supply but, hey that's an even cheaper battery. The nice thing is one can use a cheap ($10 or less) 6-12 volt 500ma AC to DC wall plug-in adapter to power it. The board is about the size of a pack of cigarettes so it would be excellent for static displays.

Steve if you think that this little ignition would be something that would be of interest and I can make a little of money that I can put towards getting a lathe/mill machine then I will put more of my attention towards it.

Things I have in the works towards testing is:
- An explosion proof encasement/box that needs the lid to be milled flat and a window cut into the lid and different sized sparkplug holes drilled into the side of the lower box.
- I have an old scuba diving tank for supplying high pressure air to the explosion proof box (300 psi) for spark strength testing. I have to figure out pressure regulation.
- I need some really thick clear plastic for the window (Lexan, acrylic, or plexiglass).
- A good multi setup mechanical signal generator (Hall-Effect, points, and inductive pickup). I have a GM V8 HEI inductive pickup test stand as shown in a previous post, that thing can simulate 15,000 rpm and throw 12" sparks!!! using IDI not CDI. That IDI is wicked. I had it on the race car and was setting the timing when I got a shock from it through one of the 8.8mm wires. It made me throw the timing light across the garage (20 feet). The zap come in my little finger and out my elbow. My little finger was numb for 2 hours and my forearm was cramped up for half a day, good thing it didn't go through my heart.
- Microchip PIC16F1619 discovery kit, still waiting to hear back from Microchip on a couple of questions. This is a chip that is specifically designed for single cylinder operation but, I feel it can do 8 or more cylinders. It's design would half the parts needed for a programmable ignition. It can work with Hall-Effect, points, and inductive pickup.

Yup a lot of irons in the fire. Well back to work for now.

Ray
 
A 12" spark would be getting up towards 1 million volts. That's a LOT of voltage.

Yup but it was across the wooden bench top and about 120,000 ionized volts (so much for safety), by the time I realized it, after about 10 seconds the spark started to fail. I blew half the output transistors. That was $60 worth, about $120 in today's value. Yup R&D - Research & Destruction LOL.

Ray
 

Latest posts

Back
Top