Model Engine Ignitions

[74Sprint]

Ray:
That is not the output from a Tesla coil. It's a seriously wicked spark (as Steve indicated) perhaps 5/8" long. Have another look at the video.
My challenge stands. SHOW us a similar output from one of your circuits.
The only reason the module Steve has employed in his big spark plug cannot be used in a car ignition system is that it is an epoxy encapsulated device that self generates sparks at a fixed rate. Also it will only run for several seconds before it self destructs. If it could be controlled it would produce way more ignition spark than anyone needs.

No I won't be showing you that, at least not right now maybe in a year or so. Why well I don't have the time or money to complete that project just for your fun. You definitely need to explore and research more about what's going on besides this forum. Did I hit a nerve with you? Seems like it. Are you afraid your sales will drop? I haven't criticize your ignition now did I. Your ignition works be happy. As you said though a spark like that is not useful for IC engines and I agree.

Cheers
Ray

 

[74Sprint]

Steve video in the works right now. I just have to figure out how to shrink it's size.

ray

 

[dsage]

For starters I didn't term the phrase "smart coil" the automotive community did. On this coil just like the original GM HEI 4 pin module it starts charging on a positive going pulse and then discharges as soon as the pulse goes towards negative.
Nothing happens because it needs to see a positive going pulse not a level one. They call it a smart coil because the driver circuit is built into it. You only need battery positive, negative and a 5v-12v positive fire/signal pulse.

Ok. Thanks for the info. I'm not familiar that one. I wasn't aware that there were coils that had an edge triggered "module" (so to speak) inside. The typical multi-wire coils I am familiar with have a simple transistor driver inside. Which saves you from having an external driver. The input 5v signal is a simple on/off drive for the power transistor inside the coil. The transistor driver is on (and the coil driven) as long as the 5v drive it high.
Interesting. I'll have to research that some more.
Thanks

 

[74Sprint]

Ok so here is my first video showing how easy it is to get a LS1 or 2 smart coil to fire with a Hall-effect. I forgot too mention that the coil draws a quick 6-7 amps when firing and is why I changed the 2n4403 PNP to a MOSFET and IGBT but even these will get hot and the spark will fade, best to use a heat sink. The video
LS1,2 Smart Coil Triggering was done with a MOSFET 3N06L22 obsolete now. I also tried a IGBT GB14C40L also obsolete but both worked ok. I also made a few tweaks for longevity. Oh sorry about the lighting.

The changes, once again if anyone wants the Gerber files to make your own boards just ask me.

Cheers
Ray

 

[dsage]

I don't have any "sales" of my ignition driver.
In fact I have built more than 20 of them and given them away. They are built and used by a lot of model builders on the West Coast and it was featured in Model Engine Builder Magazine as a DIY project.
It's not an ignition "system" per-se. It's just a dwell protected coil driver. And was meant to replace the TIM-6 driver that had a bad habit of burning out (very expensive) model coils and the transistor.
It's not fancy but it's all that is required if you have a good basic 2-wire coil around. Good for any sized engine really and limited only by the coil.
You can find the circuit for it on this forum.

 

[dsage]

Ray:
FWIW it seems to work. Thanks.
Where was the current measured?
I have to wonder why the transistor ever gets hot? The maximum power would be 12v x 20ma (led current plus a couple more for the 5k and 1k resistors) or 0.25 watts (give or take).
Are you sure the transistor is turned on properly? The ON resistance of the IGBT or a MOSFET should be milli-ohms and should dissipate almost no power. (the whole reason for using them)
If it's a MOSFET you're using then your bias is possibly correct but transistor dissipation at say (20ma and say .01 ohms on resistance)
I squared R = .02X .02 X .01ohms = about 4 MICRO WATTS.
The biasing is not correct for an IGBT.
All that aside a common bi-polar transistor (in a different arrangement) is all that should be required if all you are doing is signalling the coil to activate and it's doing all the heavy lifting. Apparently the activation signal only requiring 18ma or so.

 

[74Sprint]

Ok. Thanks for the info. I'm not familiar that one. I wasn't aware that there were coils that had an edge triggered "module" (so to speak) inside. The typical multi-wire coils I am familiar with have a simple transistor driver inside. Which saves you from having an external driver. The input 5v signal is a simple on/off drive for the power transistor. The transistor driver is on as long as the 5v drive it high.
Interesting. I'll have to research that some more.
Thanks

The only problem with them is they are expensive. For Pro cars MSD wants $1,000 USD for a set of 8, I can get 8 used ones for $75 at the wreckers. Mind you I can get the Ford coils for $10 each new. Or from Rock-Auto new from $18 CAD plus shipping. I've seen a lot of people in the old days, not so much now use the old 4 pin GM HEI for a coil driver and that one I know has a 40msec. pulse at 650 RPM and goes down as the RPM goes up. I didn't have the CD4047 hooked up in the video. I still have to make a wheel that will fit on my electric motor but, that will be my next video using my big 2 channel scope. It's like I said before Dave how much spark does one need for the models on here. I think your design will do just fine, where as mine are a lot more complicated and more expensive. Yes I read your build, it was a good read. As for sales, on the RC Groups forum a lot of people were trying to out do each other and some got really p_ssed thinking they were going to get rich and criticized anyone that wasn't in their camp. Sorry if I implied something like that. All I'm showing here is alternatives, that I've come across. All of this is leading me to a processor controlled distributorless ignition system for my racecar. I would rather use the $3,000 to buy a lathe or mill.

18ma with no magnet and 9ma with the magnet and 6-7 amps firing.

Cheers
Ray

 

[stevehuckss396]

Thanks for the video Ray. Would this system work for a 9 cylinder up to 4000ish rpm?

 

[dsage]

Ray:
FWIW I had a look at the specific transistor you are using. I didn't realize that the MOSFET you are using is an N-channel enhanced device. 3N06L22 .
Since you are using it as a high side driver (not recommended) it will never be properly turned on. The gate will need to go at least 1.7 volts above the source (Vgs ot Vth - See the spec sheet). Which means if the transistor is going to be fully ON the gate and source will be at almost the same voltage as the supply (12v) and the gate will need to be 12+1.7 - 13.7 volts minimum to keep it turned on. So this helps explain why the transistor is getting hot. It is not fully on (20milli-ohms resistance according to the spec sheet)
You should be able to measure this with a meter across drain and source when the transistor is on (no magnet on the hall sensor). The voltage from drain to source should be very low (like near zero).
A better solution would be to arrange for a low-side driver or use a P-channel device.
Even a regular bipolar transistor will suffer from this same problem. (i.e the base needs to be above the Emitter (and collector in this case when it's on) by .7 volts. An IGBT has another set of problems but suffers from the same issues.
Just FWIW.

 

[74Sprint]

Ray:
FWIW it seems to work. Thanks.
Where was the current measured?
I have to wonder why the transistor ever gets hot? The maximum power would be 12v x 20ma (led current plus a couple more for the 5k and 1k resistors) or 0.25 watts (give or take).
Are you sure the transistor is turned on properly? The ON resistance of the IGBT or a MOSFET should be milli-ohms and should dissipate almost no power. (the whole reason for using them)
If it's a MOSFET you're using then your bias is possibly correct but transistor dissipation at say (20ma and say .01 ohms on resistance)
I squared R = .02X .02 X .01ohms = about 4 MICRO WATTS.
The biasing is not correct for an IGBT.
All that aside a common bi-polar transistor (in a different arrangement) is all that should be required if all you are doing is signalling the coil to activate and it's doing all the heavy lifting. Apparently the activation signal only requiring 18ma or so.

Your right the biasing is not correct but, I just threw it together with what I had laying around. I really didn't think anyone would be interested in it because the coil is 12v and expensive to buy compared to other coils. If there is enough interest maybe I'll further research the use of this coil. I'd really like to do something like you did for a 4-6 volt ignition. I do have a CDI ignition for 4.5v-15v that I'll show later when I have time and that one does work. I have a video of it on my YouTube channel. But it was designed to be used with a PIC12F1840 microcontroller as the signal provider.

The current was measured coming out of the power supply so that was overall draw. Again your right there really shouldn't be that much draw on the signal for the transistor, IGBT, or MOSFET to get hot which is why I went with the 2N4403 PNP first. I found this schematic today that someone made up, I don't know if it is correct. I don't really see GM giving this info out. It's signed by Bowling & Grippo who are the designers/controllers of the MegaSquirt project.

I guess the best thing to do is bias and connect things properly and measure the current draw on the signal line. Maybe later today and get I'll back to you on that. I can also go to the MegaSquirt website and see what they are using to signal the coil.

Cheers
Ray

 

[74Sprint]

Thanks for the video Ray. Would this system work for a 9 cylinder up to 4000ish rpm?

Yes it would. On the LS engines it is a coil near plug setup so they use 8 coils. But I guess only one could be used. I'm going to use 8 coils for my car next summer with a cam & crank signal to keep the firing order correct. Sometime this winter I'll put it on my distributor test bench and see what the RPM is that the coil max's out at and let you know.

cheers
Ray

 

[74Sprint]

Ray:
FWIW I had a look at the specific transistor you are using. I didn't realize that the MOSFET you are using is an N-channel enhanced device. 3N06L22 .
Since you are using it as a high side driver (not recommended) it will never be properly turned on. The gate will need to go at least 1.7 volts above the source (Vgs ot Vth - See the spec sheet). Which means if the transistor is going to be fully ON the gate and source will be at almost the same voltage as the supply (12v) and the gate will need to be 12+1.7 - 13.7 volts minimum to keep it turned on. So this helps explain why the transistor is getting hot. It is not fully on (20milli-ohms resistance according to the spec sheet)
You should be able to measure this with a meter across drain and source when the transistor is on (no magnet on the hall sensor). The voltage from drain to source should be very low (like near zero).
A better solution would be to arrange for a low-side driver or use a P-channel device.
Even a regular bipolar transistor will suffer from this same problem. (i.e the base needs to be above the Emitter (and collector in this case when it's on) by .7 volts. An IGBT has another set of problems but suffers from the same issues.
Just FWIW.

Yup, as I mentioned above I just threw it together but, I will need to sort this out before the coming spring. I'm going to use a decade counter that will reset the count in/after the 8th count and use an 'AND gate' with the cam & crank signal as a lock-out to keep the sequence in time, I hope. LOL
I'm sorry if I don't get things moving along fast enough but, I have 2 projects that need to be updated and right now they earn me 10's of thousands in licensing fees. It took me a year to convince STMicro to release to me their RTOS which is the only one recognized by UL. Design and test = 1 year, UL testing = 1 year, Patent under small business = 1 year, cost for UL certification = $50,000, R&D = $?, patent = $2,500 it sucks it sucks it sucks.

Cheers
Ray

 

[74Sprint]

Ok more on this 'smart coil' thingy. So I finally biased and connected things up properly instead of just throwing it together. I went back to the 2n4403 PNP transistor tried different biasing to get things closer to where they should be, note this circuit is not optimized, that will come later. But it is useable the way I have it now.

So signal current is 0 with the magnet on the Hall and 4ma without the magnet. Signal voltage going into the coil is 0 and 7.65v magnet held on and off. In both cases the coil did not fire. The coil only fires when it sees the voltage rising above 4 volts. If the voltage is steady at 7.65v or 0 it will not fire. Temperature rise on the transistor was 1.5C and 1C on the Hall, which was nice for a change. I should mention I decided to try a SS441R Hall that I bought to try out, it was too sensitive. Hall on was at 3/4" and off at 1". The coil did get warm which is normal for these coils but, it was just warm to the touch and no way near hot. These coils are used in many different makes of cars under Delphi as manufacturer, I use RockAuto to find out which makes. Just search for part number "STANDARD MOTOR PRODUCTS UF262T" or "ULTRA-POWER UF262" or just "CHEVROLET > 2006 > SILVERADO 1500 > 5.3L V8 > Ignition > Ignition Coil" and search the part numbers. The coils were used from 1999 to even today I believe.

Because the 'smart coil' uses such low power for a signal it would be able to use something like an Arduino board to signal several coils for multi cylinder engine using 1 board for each coil for each cylinder. This means one can make a programmable ignition curves. But one would need both a cam & crank signal. Most Arduino boards can drive a 20ma load so it should work here.

Let me know if something is out of wack and I'll fix it up, otherwise I've ran out of time for this 'smart coil' thingy. But I will come back to it in the winter.

Cheers
Ray

 

[74Sprint]

A little more about the 'smart coil' circuitry. The coil fires when the LED goes out. This is because I'm using a PNP transistor which, kind of makes things look like they are working backwards. When the magnet is near the Hall-effect the Hall goes to ground or negative potential this turns on the 2N4403 which provides a positive going pulse which tells the coil to start charging. As soon as the signal goes level the coil fires. The dwell time is set by the coils internal circuitry and varies with RPM as the Hall's pulse width shrinks with higher RPM. GM uses Hall-effects pretty much like everybody else now a days for the crank & cam signals.

You don't need to use a 2N4403, you can use something with similar specs. I used it because I have about 500 of them. I went to the auto-wreckers today and got 8 coils, wiring harness' for them, and sparkplug wires for $110 CAD. They will sit in the garage until I get time to finish up an Arduino ignition system I'm working on for the racecar. I was going to do a CDI ignition but I think I'll do the distributorless TCI ignition first using the LS1 coils. Well that's about it for this puppy until later.

Cheers
Ray

 

[dsage]

The LED is backward. It's never going to light.
But, if you turn it around the transistor will always be on.
You need some current limiting for the base of the transistor. The Hall sensor and the transistor likely won't survive since it is trying to short out the supply through the transistor base emitter junction.
Using the points the transistor will be toast instantly.
But you should know all of this.

 

[74Sprint]

The LED is backward. It's never going to light.
But, if you turn it around the transistor will always be on.
You need some current limiting for the base of the transistor. The Hall sensor and the transistor likely won't survive since it is trying to short out the supply through the transistor base emitter junction.
Using the points the transistor will be toast instantly.
But you should know all of this.

Your right Dave I should know this better. I just recently lost my step-father and I guess my mind isn't fully focused.
Anyway, I switched the LED to the high-side so it comes on when the Hall goes to ground (magnet on) and coil fires when the LED goes out (magnet off). The LED is fine. You would be right about the transistor if it were driving the coil but it's just supplying only a signal. As for the transistor the base doesn't need any current limiting because of the 1K resistor on the collector and the 'smart coil' is only drawing 4ma. For points I forgot to say to remove the pull-up resistor R1 with points. I have the circuit running right now and it's been points closed for 30 minutes. The one problem is the points resistor is up to 28.3C so I recommend using a 1/4 watt instead o the 1/8 watt. The coil fires as soon as the points open and turning off the resistor and LED.

Thanks for keeping me on my toes Dave.

Cheers
Ray

 

[74Sprint]

This is the new circuit running like it was using points.

YouTube link below.
LS1 coil and points

Cheers
Ray

 

[dsage]

No. Actually you are wrong. Look at the transistor again. You have 12v on the emitter and with the points closed you have ground on the base (through the points). With nothing limiting the current through the B/E junction the transistor will fry. It's the same as putting a diode right across the supply. It doesn't matter what the collector is doing for a living.
I can't comment on how you are getting away with it. Perhaps your supply is current limiting.

 

[74Sprint]

Ok Dave, I see what your talking about, I thought you meant the LED. My bad, I had the drawing wrong but, the bread board is correct which is why it didn't blow up. K new schematic. Oh power supply is 22 amps, I should limit the current. I need to slow down and check myself.
Since I'm including the Gerber board files some people may not know about Gerber Viewer to read the board files and use to make boards.
Thanks again Dave.

Cheers
Ray

 

[dsage]

Your almost there. I think I've corrected at least three major errors in your circuits.
For anyone else that has been following along my advice would be Caveat emptor .
I'll stop commenting on your circuits now.