Transistor ignition

[GordonL]

I have been playing around with transistor ignition systems lately and have not been having too much luck. I have tried the simple with just a single transistor and the more complicated circuits with two transistors and resistors etc like the TIM-4. A couple of years ago I purchased the Jerry Howell kit. On the single transistor circuit I can get a weak spark sometimes. On the TIM-4 I get nothing. On the Howell I can make the LED light but no spark. I have traced the circuit several times and cannot find any wrong connections. I have tried several different coils from one I purchased from Howell to coils from automotive coil over plug etc. Nothing seems to work. I am using 4 AA batteries for power.

I realize that no one can analyze a problem without seeing the equipment but does anyone have suggestions on what I can/should be checking for? I am not an electronics expert but it seems like sooner or later something should work. Others talk about using these circuits and have spark jump 1/2".

Thanks: Gordon

[Roger B]

Hello Gordon,

Can you give a few more details? Are you using contact points or a Hall Effect sensor? What sort of automotive coil did you try? Can you add some pictures/diagrams? A photo of a hand drawn sketch is fine.

[gbritnell]

Hi Gordon,
I have built the Howell ignition from the supplied kit and have 4 of them and they all work great. Like you said it's hard to diagnose on a forum but I would go back and check all your solder connections and locations. What are you using for a trigger, points or a Hall transistor? If you hook up a Hall wrong they will burn out in an instant.
gbritnell

[Roger B]

Are you using 'normal' or rechargeable AA batteries? AA batteries are certainly not sufficient for an automotive coil. I am using a 6V motorbike coil with a 6V 6AH sealed lead acid battery.

[Don1966]

The TIM 4 has a large diode with a band on it, make sure that this band is facing the correct direction by looking at the direction you were given to assemble it.
Gordon you need to place a voltmeter across the ignition coil with clips to measure the voltage across it. Use a jumper to trigger the circuit after removing any trigger devices you have connected and look at your voltage reading. The directions show how to trigger the module with points, place your jumper across these terminals. the meter will not be 6volts but at least 5 or more only when you trigger it. This will tell you if your ignition module is working.
If you have voltage all the time the large transistor could be shorted.

Regards Don

[PStechPaul]

Here is a simulation of an electronic ignition that I am (still) planning to build, and the schematic for yours should be similar. This circuit uses stored energy in the inductance of the coil which is released when the points open, and the voltage is limited by the primary capacitor and a TVS limiter:



There are also capacitive discharge types that charge up a capacitor to several hundred volts (using a circuit similar to photoflash units), and then dump it into the coil primary. If you can provide a schematic or pictures of what you have it will help greatly.

[GordonL]

Thank you for the replies. Now I really feel dumb. I did not realize that there are transistor ignition systems as well as CDI ignition systems. I have used the S/S system as well as the Hobby King system. I am trying to get a system which is compact and eliminates the points.It looks like the CDI system is actually the closest to what I am looking for. I guess that it is time to do some further investigation into the CDI. It looks like the coil is the main problem. This system is used on small engines so perhaps one of those coils would work.

The Hobby King works well but the spark plug cable is pretty clunky. Perhaps the outer shield could be removed and just grounded to the engine.

Gordon

[Don1966]

Gordon glad you found you problem.

Paul I think you best check your circuit because the secondary of that ignition coil should be connected to ground. The transistor will burn up from the high voltage as will the zener diode and cap.

Regards Don

[PStechPaul]

I found diagrams of the TIM4 through TIM7 ignition systems, and they use a different arrangement for the coil so it can be connected to ground rather than +12V. But other circuits are virtually the same as mine, and it should be no problem for the transistor and zener if properly rated, and if the capacitor is chosen correctly. The secondary of the coil should be connected to a spark plug or spark gap, which will also limit the primary voltage, but ultimately the circuit should be able to withstand an open secondary.

Some more information and schematics:

http://constructionmanuals.tpub.com/14050/css/14050_96.htm
http://www.wiringcircuit.com/auto/Transistor_Ignition_1185.html
http://www.jerry-howell.com/IgnitionModules.html (Jerry Howell's TIM-6x, but without schematics)
http://www.patents.com/us-5111798.html (patent)
http://www.auto-wiring-diagram.com/3511-typical-transistor-ignition-system-circuit-diagram-of-1965-ford-and-mercury.html
http://homemadecircuitsandschematics.blogspot.com/2013/01/make-this-enhanced-capacitive-discharge.html (capacitive discharge type)
http://newautoaa.blogspot.com/p/primary-circuit-of-ignition-system.html (detailed general ignition information)
http://chevythunder.com/ignition_systems_hei_operation.htm

Hmm... Most of those show one coil terminal grounded. The last link shows a separate secondary which is grounded.

I think I will have to build my circuit and see if if it blows up! 

[Allen Smithee]

I would expect the actual coil to be an auto-transformer rather than an isolating transformer, so in reality the transistor is grounding a tapping in a single coil, "shorting" the primary to induce the flux change in the whole coil that generates the HT voltage. It's drawn as two coils by convention, but it's probably a single coil with a tapping.

AS

[Admiral_dk]

Sorry Don - but the ignition coil is wired correctly - you're right about the transistor and zener, they will not survive 

Paul change the transistor and zener into one of these special ignition IGBT's :   FGB3040CS,  IRG4PC30,  IRGS14C40 or NGB8202N.

There are more to choose from, but these special devices are extremely rugged and protected against too high voltage from an unloaded ignition coil and other mistakes. The transistor you picked will only withstand 160 volts and only 600mA. A normal ignition coil draws 3-6 A. and the voltage just after switching off reaches close to 300 volts if you have the right capacitor across the primary.

[Don1966]

I am not going to argue with you, but this is the correct setup for an ignition system.
You will notic the HV winding grounded.

Don

[PStechPaul]

The ignition coil I have (#1303 from NAPA) has only a (+) and (-) terminal and the HV terminal for the distributor wire. The case is painted and there is no connection to the internal coil(s), which are connected as an autotransformer. The primary is 5.37 mH with 1.38 ohms resistance, while the secondary is connected to the (+) terminal with 62.4 H and 9.9 kOhms resistance. The turns ratio is sqrt(62.4)/sqrt(0.00537) or about 340:1.

The components shown in the simulation are only the closest I could find in the standard libraries. I will use a 600V or higher power transistor and a high energy (500-1500W) TVS in place of the zener, and a 600V capacitor. The 1N5378B is rated 100V and 5W, so it might be able to withstand the excess energy stored in the coil if it is not dissipated in the spark. The energy stored is 0.5*I^2*L, and I depends on the dwell, applied voltage, and inductance. In the simulation, peak current is 2.4 amps, so energy is 28.8 mJ (W-sec). The duration of the spark is probably on the order of 100 uSec, so the peak power could be 28.8/100 = 288W. The primary of the coil and the capacitor form a series resonant circuit with a frequency of 5 kHz, so each half-wave would be 100 uSec.

It really doesn't matter much if one side of the primary is connected to ground or +12V. The voltage on the other end will still rise to several hundred volts due to inductive flyback effect. For grounded coils, a PNP transistor is needed and will be connected to +12V. But for my circuit, an NPN is connected to ground. There is a much wider assortment of high voltage NPN BJTs and MOSFETs, and even IGBTs, that can be used. And my circuit maintains most of the wiring of the convention points-activated ignition systems. I think I will need to add a capacitor across the 12V supply because the battery and external wiring probably do not have low impedance as is required to keep its voltage constant under high frequency pulse conditions.

As a finished project, I plan to add a PIC, which can be used to provide an adjustable dwell and spark advance/retard, as well as a tachometer based on the time between point openings. Hopefully I can build this and demonstrate it at Cabin Fever in April.

[Don1966]

Paul it's your circuit do what you wish, but I am glad that your taking on, as per our discussion the timing advance. I look forward to see your progress and hope to see you at Cabin Fever this year.

Regards Don

[PStechPaul]

The problem with timing advance is that the circuit can only act on the closing and opening of the points or the sensor. So it would need to be set up mechanically so that the points activate at the maximum advance, as it can only retard the spark. The simple circuit I show has no mechanism to change the timing. The more "advanced" unit might just maintain a certain current (and energy) in the coil, and then break the current to release the energy as high voltage upon a signal from the PIC. The degree of advance would actually be a delay determined by the RPM, so if the points actually open at 60 degrees BTDC, an actual advance of 15 degrees at 600 RPM would be a delay of

45 degrees * 100 mSec / 360 degrees = 12.5 mSec

To measure RPM, I could count the number of pulses in a one second period, but there are only 10 pulses per second at 600 RPM, so the resolution would be only 60 RPM. A better method may be to read the time between pulses, which can be done to a precision of 100 uSec for a count up to 65536, or 6.55 seconds, corresponding to 9.2 RPM. It would have a resolution of 1/65 or better than 2%  at 9200 RPM. I think the CCP capture mode would be perfect for reading RPM. It does involve floating point math, but that's not a problem when programming with C.

[Allen Smithee]

Most commercial (car) ignitions retard the spark by more than 300 - 360 degrees so that the trigger point is safely in the normal operating range for fail-safe fallback modes

AS

[Don1966]

Paul I think your right about using the CCP method. Woodward control uses this on there governor controls.
Although the use of a stable burst freq is used to fill the off time capture between pulses. You could use the burst freq count to calculate the time. Not having any mechanical method of advance the delay action is the only way to do the advance and retarding. Keep at it I am real curious as to your outcome.

Don

[Roger B]

These 'predictive' advance/retard systems can cause problems. A year or so back Jo and I were struggling trying to use a Rexel system designed for high speed two strokes on lower speed four strokes (my vertical and Jo's R&B). Amongst the many problems this ignition system expected the sensor to be on the crankshaft and set the advance in crankshaft degrees, we had the sensor on the camshaft so, I think, we had twice the advance at half the revs (or maybe the other way round    ).

[PStechPaul]

There would need to be a setting for the ratio of crank to cam, and perhaps number of cylinders. Probably best to mount the sensor on the crankshaft to get a direct reading. The other problem with the predictive delay used for advance is the case where the speed is changing significantly from one revolution to the next. It changes a lot more over 300 degrees than it does over 60 degrees. 

But first I need to get something working, and then the enhancements can be added. 

[Don1966]

Agree Paul, but I think a speed trigger would be the way to go. Adjustable set for a given speed so it could cover a set range of speed before being pressed into use. Or even multi trigger for small incremental advances for different speeds. Just thinking out loud here.  Guys like George B and Steve H could give more input as to what speed runs the best without advance or retard on these small engines.   

Don

[Roger B]

My current, limited, experience suggests that variable advance/retard comes a long way after getting it to run and getting the carburation sorted.  Maybe if you are working at strictlybusiness's level (unbelievable stuff) the firing point is critical. I could move the ignition timing of my vertical engine 30° with no noticeable effect 

[Don1966]

I thought some of our readers would appreciate this.

Here is a short video I made of back EMF or fly back voltage as some might call it. The demo I built years ago to teach a class. The LED is connected backwards so the battery will not light it. The magnetic field will collapse and generate and EMF keeping current flowing in the same direction with the source now being the inductor so the LED light. This is the voltage that will hit your transistor or whatever switching device you use to turn it on and off with. This voltage can reach very high voltages and distroy what's driving it. Precaution are taken to damping this whenever possible in electronic circuits notice how the LED light only when the switch is open.

Regards Don

<a href="https://www.youtube.com/watch?v=uUrYt4rNOiA" target="_blank">http://www.youtube.com/watch?v=uUrYt4rNOiA</a>

[Admiral_dk]

Ok Don

Perhaps my statement overdid it a bit, but almost all ignition coils I've come across that wasn't dual spark (a sparkplug in each end of the high tension coil) has the secondary connected to the primary internally and doing this at the switch end (and not at the power end) this adds the primary voltage to the secondary voltage (if wired correctly). I haven't seen an old style ignition coil like the one you showed in your reply, since my childhood - that's not to say the they can't be found .... and I do know that vehicles made in the USA are different to most made elsewhere in a lot of small details 

Paul

I really meant it - you will not find a transistor or Mosfet as the switching element in modern ignition systems and for a good reason - there's no way they can compete with the reliability nor simplicity in those modern IGBT's I referred to earlier

[Don1966]

Ok

Don

[Allen Smithee]

Yes, but that circuit is still showing the coil as an isolating transformer, where all the ignition coils I've ever seen have been autotransformers with no isolation between primary and secondary (because the secondary uses the coils in the primary as well, with a considerable copper saving and and consequent reduction in size and copper losses).

AS

[Admiral_dk]

Sorry - I just realized that I should have added "among the brands I've seen" - the statement without could be interpreted as if I'm claiming to know all, and I'm well adware of my limitations   

Allan put it nicely - all the ones I've seen has been autotransformers - except from the dual spark ones, on my previous four banger Jap bikes.
That did not prevent those who drew the wiring schematics from putting an transformer in it, even though that actual component on the vehicle was an autotransformer when you measured it.

[PStechPaul]

Don's circuit is almost identical to mine, except for the duty-cycle-limiting capacitor, and the isolated secondary. The bodies of the plugs are connected to the low side of the secondary and through the engine to ground. This could also be a junction between primary and secondary, and it could be just as well connected to 12V or ground. This makes little difference when the current stops and the collapsing magnetic field causes the high flyback voltage (not quite a back EMF, which is usually used to describe motors and transformers). No matter what the interrupting device is, the voltage will be limited only by the properties of the resonant LC circuit, or by the breakdown voltage of a TVS, zener, or the device.

With simple mechanical points, the voltage will be limited by the air gap, which is typically about 20 to 50 thousandths of an inch, and the general "rule of thumb" of 10,000 volts per inch means that they will arc at 200 to 500 volts. But a problem with mechanical points is that they open slowly, and will start arcing even when only 0.001" apart, which could be as low as 10 volts. Once an arc has formed, it becomes a plasma consisting of vaporized metal from the points, and this lowers the breakdown voltage of the air between the points. While this arcing is happening, energy is lost, and the output voltage of the coil is low because it is determined by the turns ratio of about 350, so it may start at only 3500 volts at the moment the points open. The plasma reduces the voltage that would otherwise be created according to the widening gap, but if there is enough energy, eventually there will be enough voltage to produce a sufficient arc across the spark plug gap (typically 0.030 to 0.060) to ignite the air-fuel mixture. This environment may change the 10,000 volts/inch approximation, and I think the ignition system typically relies on 30,000 to 100,000 volts for operation. This translates to a primary voltage of 100 to 300 volts.

Once the spark plug fires, the voltage will likely drop because of the continued reduction of energy in the coil, but (I think) it doesn't much matter once ignition has started. However, some old ignition systems (like the Model T) used a continuous high voltage AC that provided timing via the distributor. Modern ignition systems are mostly capacitive discharge which require electronics to sense flywheel position and time the spark.

One problem with the mechanical points used in simple model engines is the fact that the slow opening may actually retard the spark significantly from the position at which they first open. I think this is why some engines need as much as 40-60 degrees of advance, even (or especially) at low speeds. It does not seem logical for the spark to occur that early in the compression cycle, so it must be that the actual spark of sufficient voltage must occur a considerable time after, and close to TDC.

The LC resonant circuit may also affect the point at which maximum voltage occurs. When the points open, the current in the primary of the coil goes through the capacitor (condenser), and the voltage will start at zero. As it charges up, it will reach a point where it holds the coil's energy as a voltage, and then the current flow reverses and an exponential decaying sine wave will occur. But if the spark plug fires, it will take out much of the energy and the waveform will be shorter and lower in amplitude. For my example of a 50 mH coil and 0.1 uF capacitor, the resonant frequency is about 2.55 kHz and the first peak should occur at the quarter cycle point, or 100 uSec after point opening. At 600 RPM, this translates to 360*0.1/100 or 0.36 degrees, so it may be insignificant.

BTW, I was not familiar with the specialized ignition IGBTs you mentioned. It seems they have their own clamping circuit as well as a current sensing lead, which may be quite useful. The data sheet for the FGB3040CS:
http://www.mouser.com/ds/2/149/FGB3040CS_F085-347381.pdf

[Don1966]

Paul I believe the guys are missing the point here. I am not worried about how the coil is built. You don't drive a coil by the ground terminal with a transistor you drive the positive side when it has a common connection. The high voltage is created when the coil collapses this is when the switch is off. The path for the high voltage ground is now going through the source not a good idea. This voltage can distroy the control circuits as well as the driving circuit.

Don

[Admiral_dk]

Ah now I get your objections Don and I do understand them 

Again I can only refer to the systems I worked on and none of them had the power to the coil come from the box .... The ignition coil is connected to the ignition switch (and from there to fuse and battery). The "midpoint" on the "autotransformer" is connected to the switch (old fashioned point style). In a few cases I've seen the other end of the switch having a separate ground to the rest of the box, going directly to the battery - most has one (or more parallel) ground wires going to the chassis ground. And I'll give you, that a problem with those ground wires from the box is a disaster waiting to happen if or when our old "friend" Murphy "pays a visit"      .... on the other hand - I've never had one of them go bad yet - knock on wood etc.

The only systems I've seen where the primary of the coil is connected to ground have all been CDI systems and all of those had the output for the box connected to the midpoint again (non dual spark).

Yes Paul - they are quite clever aren't they, those devices 

Best wishes
Per

[GordonL]

As usual it does not take long for the discussion on things electronic to get far beyond my understanding. I guess I should just stick with the machining of engines and not try to move beyond my ability.

It does show that we have some pretty impressive talent here and on other similar forums.

At this point in my life I am content to just enjoy my hobby and my shop and not try to learn a whole new career at 75.