How to make a proper COP setup?

[game-hunter]

Let me first start off by saying I already have a standard coil on plug setup running on my 90 GSX. I went with a COP setup because I was tired of replacing plug wires and I'm running a cyclone intake so mounting the stock ignition system is a pain. The COP setup works but as most people who have COP setups know, it's not quite running the same as it was with the stock ignition system. It dies out at the top end and idles a little rougher. I can only imagine it does this because the spark is not as strong as stock. I'm hoping to get some technical feedback, perhaps schematics or other specs of the stock ignition system along with specs of the COP coils (I'm using Intrepid coils). I'd also like to hear some ideas on how to improve the COP setup.

My coils are wired up similar to this, the only difference being mine is for a 90 so I have the taco module wired up as well.

The big flaw that I see in the common COP setup is that the coils are in series. This means each coil can only get half the potential voltage they would get if they were wired up in parallel. This also increases the total resistance in the circuit thus lowering the maximum current. I believe the reason why we don't hook the coils up in parallel is because the stock ignitor/power transistor will not be able to properly handle the coils and burn out.

I also see a lot of talk about stock dwell time being wrong for a coil on plug setup. People seem to be saying the dwell time is too long. Maybe it is if the coils were hooked up in parallel but I don't see how shorting the dwell time will increase the charge when in series. If anything it seems like the dwell time should be longer for the coil to stabilize. Granted too long and you overheat/fry the coil. While I'm not running a aftermarket ECU I can burn my on chips and reprogram my stock ECU's dwell time if I need to.

I think the solution I'm looking for is a way to replace the stock power transistor with one that can handle the coils in parallel or drive each coil individually. Polarmoment seems be on the right track in this thread
http://www.dsmtuners.com/forums/bolt-tech/299790-why-cop-if-stock-better.html#post151540946
but I need more details to implement something.

Let me know your guy's thoughts and ideas are. I am capable of building my own circuits if I need to. But I would like to keep things as simple as possible so others could use it.

 

[->PrOjEcTGS<-]

Nice to see someone got some use out of my diagram

Anyway, I recently went back to the stock "wired" setup. Main reasons being that it is a pain in the ass to set timing when there is no wire to hook the inductor to. Secondly it's also a pain to change plugs also.

On to your question:

I have never opened up our CAS therefore I do not know exactly how it operates. But, on older cars you have a pickup coil that sends a pulse to the ignition module to fire. On a V8 that's 8 times of the camshaft rotation and 2 rotations of the crank. I assume our's sends 4 pulses per camshaft rotation. Firing each cylinder once. Now for a COP system to work like you want, we need to get away from the waste spark system. There needs to be something that differentiates Cylinder 1 from 4 and 3 from 2 tells what coil to fire. I don't believe our current CAS will do that. Therefore the whole system needs to be changed or there needs to be a new sensor introduced to work along with the stock CAS to have it fire the correct cylinder. Sort of like a distributor but using the signal before the voltage gets ramped up.

These are my thoughts.

 

[game-hunter]

Thanks for the diagram

Oddly enough I just checked my timing for the first time today with my COP system and it was quite easy. All I had to do was open the inductive timing connector and hold it on around the corner of coil 1. It picked up the pulses no problem.

As for the CAS thought. I don't think I necessarily need to get rid of the waste spark system. A common misconception of how the ignition system works is that people think the coils are energized for the whole duration of the stroke until they are needed. This may be how it worked back in the day when you only had 1 coil but today they're only charged for ~8ms before they are used. If you charge the coil for longer than that you tend to overheat them and they'll break.
Removing the waste spark system will lower the duty cycle on the power transistor and coils but unless they're actually overheating this should not be a problem.

 

[polarmoment]

you need better power transistors. the stock power transistors drop a lot of voltage. the correct way to do this would be to use several MOSFETS of proper D>S breakdown voltage (400V min) wired in parallel.

this essentially cuts the circuit resistance considerably, allowing for faster current rise time. most people don't consider the power transistor as a resistance source when calculating the coil circuit's total resistance.

the only issue is that the required MOSFETs are pretty spendy- on order of up to $15ea. the nice thing is you can run lower current MOSFETs since each one only carries partial current for the circuit. do NOT try to use bipolar transistors for this. MOSFETs have a positive junction temp/resistance coefficient, so if one transistor starts carrying more load, its resistance goes up and diverts some of the current through the others. bipolars have a negative coefficient. so if one starts to carry more current, it heats up and continues to decrease in resistance until it burns out.

there are also some ignition-specific IGBTs out there. most have built-in current limiting (the VB921 limits to 7.5A, the bosch BIP373 limits up to about 9A, and i forget what the fairchild ISL9V5036P3 limits to) that would likely work very well too. the advantage of the current limit is that if the dwell is too long, the current doesn't continue to ramp up.

if you decide on the parallel MOSFET method, the wiring is pretty simple. i forget if the DSM ECU is positive or negative edge trigger. if it's negative trigger, all you have to do is connect each respective coil pair's ECU output to the gate of each MOSFET through some arbitrary value resistor (100ohm would do- it's just there to even out the gate charge time across the transistor bank). connect the drains all to the coil negative terminal, and the sources to ground and you're good to go. you can connect the stock tacho interface to the coil negative side of the circuit too. oh, and use a heatsink!

the honda ignition modules would work well too, but would likely cost more depending on where you can find them. you'd only need two, but four would be ideal and would give the best performance.

 

[polarmoment]

Now that I think about it too, the first thing I would do is test the stock setup. This is pretty easy- Install a 0.1ohm resistor in series between the power transistor unit and its ground connection. You then measure the voltage drop across the resistor. Every 100mV of drop you see equals 1A of current flowing through the circuit.

There are a few ways to measure this:

I don't know how good the datalogging is on the stock ECUs, but if you have an extra input that you can feed voltage to, you can send it to that and log it. I don't know if the datalogging programs sample fast enough though. you'd need at least 50 samples/sec.

Find someone with an oscilloscope. Any auto shop that's worth your business should have one, even if it's a handheld unit (Fluke, etc). These take several millions of samples per second and can plot the voltage curve over time in a nice little graph.

the third way would be to use a good multimeter with a peak function to measure the voltage and record the highest voltage seen, representing the highest current value.

FWIW, the stock system charges to 6A, and the coils have about 3.5mH of inductance. COP coils will have lower inductance, but current has a greater effect on energy than inductance does.

 

[game-hunter]

Thanks for responding polarmoment.

It's going to take me a little while to get an oscilloscope. I would also like to measure what kind of delay the power transistor introduces and what the stock dwell time really is. Any idea what the dwell time is on a stock 300m or Intrepid coil?


In the mean time do you think the Honda igniters could handle two of the COP coils in parallel not serial like the pic above? If they can I may just go that route since they come in a nice proven package and I can probably get them fairly cheap at a junk yard. Also any idea if they have any strange protection in them? Like auto triggering if the dwell time is too long? After all they are called an igniter and not just a power transistor.

I also have some concerns on whether or not the stock ECU can drive the multiple MOSFETs/Honda igniters without burning it up. I suppose I could use the output of the stock power transistor unit to drive the new ignition system right? Anyone have a schematic of the stock power transistor unit?

 

[polarmoment]

Thanks for responding polarmoment.

It's going to take me a little while to get an oscilloscope. I would also like to measure what kind of delay the power transistor introduces and what the stock dwell time really is. Any idea what the dwell time is on a stock 300m or Intrepid coil?

no clue on the 300M dwell time. you could also look at the toyota camry V6 COP coils. the dumb coils have the same plugs as the intrepid coils, and i think their dwell is around 3-4mS. you can get a ballpark idea of the dwell if you know the inductance of the coil, but there really isn't a way to measure that without disassembling the coil itself.

as for an o-scope, there are some pretty easy DIY designs for the PC... one of them uses an ADC to change the voltage level to a digital signal that interfaces with the parallel port. software for it was free IIRC, and the hardware cost about $10.

In the mean time do you think the Honda igniters could handle two of the COP coils in parallel not serial like the pic above? If they can I may just go that route since they come in a nice proven package and I can probably get them fairly cheap at a junk yard. Also any idea if they have any strange protection in them? Like auto triggering if the dwell time is too long? After all they are called an igniter and not just a power transistor.

i'm not sure about putting them in parallel. when you put two equal resistances in parallel, you effectively halve the resistance. 0.3ohm might be enough to burn the igniter out. i don't think they have any current limiting circuitry in them, though i know they do have a soft shutoff so that when power to them is cut, they turn off slowly so they don't create a spark. i would try them in series first. remember, COP coils have lower inductance, so despite the higher series resistance, they'll charge faster than a stock DSM coil with the same resistance.

I also have some concerns on whether or not the stock ECU can drive the multiple MOSFETs/Honda igniters without burning it up. I suppose I could use the output of the stock power transistor unit to drive the new ignition system right? Anyone have a schematic of the stock power transistor unit?

the stock ECU should have zero problems with the MOSFETs. MOSFETs are voltage driven devices, not current driven. there is no current flowing through the gate (what the ECU is connected to) save for a very small amount to charge the gate capacitance (on the order of picoamps). when there's voltage present on the gate of the MOSFET, it conducts from drain to source. when there's no voltage present, it stops conducting. the stock ECU may need a pullup resistor if the output is an open collector.

it may be a good idea to use the stock power transistor to drive the honda igniters though, if just for the fact that the signal needs to be inverted anyway. they fire on the rising edge of the signal, which is what the signal would look like at the coil negative terminal from the power transistor. i don't doubt the ECU could drive them directly since it only needs to sink about 10mA to pull the trigger line low, but i'm pretty sure the signal needs to be inverted anyway.


again, your best bet before even making any changes is to find someone locally to you that has an ocilloscope and measure the dwell and the charge current of the COP units. from this, you can calculate inductance and current rise time for series/singular/parallel and make an intelligent decision on which way to go.

i don't check here as often as i'd like, so feel free to email me if you have any questions..
grue0038 at umn dot edu

 

[game-hunter]

as for an o-scope, there are some pretty easy DIY designs for the PC... one of them uses an ADC to change the voltage level to a digital signal that interfaces with the parallel port. software for it was free IIRC, and the hardware cost about $10.

Oddly enough I have played around with some simple DIY PC oscilloscopes that used the parallel port. In fact I even wrote a program to graph and display data for one. Having said that I can say they suck compared to the real thing. I know you can also use the audio in for a simple oscilloscope but figuring out the exact voltages and making sure you don't break your sound card/pc does worry me. Especially when we're dealing with large inductors that are bound to have some good kickback voltage. I'd rather just have the real thing. I've been putting off buying one for forever and I suppose this is a good as excuse as any.

the stock ECU should have zero problems with the MOSFETs. MOSFETs are voltage driven devices, not current driven. there is no current flowing through the gate (what the ECU is connected to) save for a very small amount to charge the gate capacitance (on the order of picoamps). when there's voltage present on the gate of the MOSFET, it conducts from drain to source. when there's no voltage present, it stops conducting. the stock ECU may need a pullup resistor if the output is an open collector.

I suppose I should have known that.. been out of working with electronics for too long I guess.

Thanks again for your feedback. I'll post the data on any measurements I take once I get ahold of an oscilloscope and get it all hooked up properly to test. This will probably be at least a couple of weeks.