Catching car's ignition signal

[cmcraeslo]

Hey guys.

I have 12V and 2 signal wires coming out of the ECU into the ignition coil pack. To my understanding, they are active low signals, which means they get activated by the ECU with low signal (ground). I'd like to catch that signal (individually) and I have problem reading them.

If I just connect the scope to a signal wire, it looks like this:

Spark is active for 3-ish milliseconds and then it bounces back to 12V and repeats the ignition again.

If I cut the wire and connect the wire coming out of the ECU to 12V using a pullup, I get 12V like I should (I get 12V without the pull up as well) and it does not change when I crank. I tried different resistor values, even pull down - no change. Just as if ECU is not triggering it. What could be the issue?

[Benta]

ECUs are very often smart enough to detect a wrong load, eg, your resistor.

[cmcraeslo]

Good point - but i guess ECU would try at least once?

[Benta]

Good point - but i guess ECU would try at least once?

Not necessarily. It'd normally do a system self-check when starting. If it finds something wrong, odds are that the ignition won't be released.

[David Hess]

If it is driving the ignition coil directly, it could be looking for the series resistance of the primary; use that for the value of your pull-up resistor.

[Circlotron]

If it is driving the ignition coil directly, it could be looking for the series resistance of the primary; use that for the value of your pull-up resistor.

I think you’re on the right track, but modern coils often have a primary resistance of something like 0.5 ohms but they get switched off for the spark at 7-10 amps or so. So a resistor of 1.5 to 2 ohms might be about right. Lots of watts though.

[cmcraeslo]

I tried with 3 ohm resistor (didn't have any lower at hand) and it started showing something so it's definitely the right way to go. Because I only want to get the signal out of the ECU and not waste power (unless needed), is there any other way to detect this signal? (if we ignore that ECU detects the error in the coils) ?

[Circlotron]

If the ecu is detecting a lack of current in the switched load then it must be switching it to ground to measure it. Perhaps the negative going pulse is only very short and you are missing it with the timebase the scope is set on. Speed it up a bit. Make sure you use some kind of pull-up resistor, say 100 ohms for this test.

Maybe it pulses just once then locks out.

[Labrat101]

Hi we help develop a car diagnostic system . about 5 years back.
What you are looking at is just part of the story .. Car ignition varies  from make / Model /year / engine size/ compression ratio / altitude /  O2 % .. etc .
 The ignition system is controlled by the central computer which measures about 8 sensors from the engine and fuel .
 The pulse you are seeing is about what you will see . from a GP scope . be careful your scope is mains grounded ..
 you have to use the diag plug with a loop back to tell the on board computer to output the correct signals .
 and you need the the manufactures data for you Car as every model differs .
 You need to use the correct Diag equipment .
  The signal you are seeing will not be of any use as this is only 1 fire pulse without the time from TDC  -8 -10 deg.

   Have fun
     

 BTW if you really like your scope Don't use it on a car ignition system without Isolated clap over probe .
Update .. 2 quick details ..
   1 . remove the plug from the engine connect it to the lead and rest the case of plug on chase on engine ground.
    if there is a spark its good.  Clean all plugs.
 2 60% of engine problems are caused by the immobilizer system .. check first

[mikerj]

If it is driving the ignition coil directly, it could be looking for the series resistance of the primary; use that for the value of your pull-up resistor.

I think you’re on the right track, but modern coils often have a primary resistance of something like 0.5 ohms but they get switched off for the spark at 7-10 amps or so. So a resistor of 1.5 to 2 ohms might be about right. Lots of watts though.

With an ignition coil the primary current will (relatively) slowly ramp up to the desired final value, with a dummy load resistor it will be immediate.  If the ECU has closed loop control of coil dwell time this could cause it problems.

[cmcraeslo]

Hello.
@Labrat101 - I'm not trying to diagnose the car's issues with this.

What I'm trying to do is to build ignition cut device, that is catching the signal, "processing it" and then output it. I have the device working on a car with ignition drivers built in the coils, but not when coils are directly driven.
Device uses very sophisticated cutting strategies, so it's not just "coil off" cutting so I need to have clean signal coming in constantly so I can process them correctly.

@mikerj - that is very advanced control. Do you know how many cars use this? This could be a deal breaker for me for sure.

[Labrat101]

Ok that you can do by using an induction pick up  .
There are clamp over that can be bought ..
But as you are building it . I would use ether a Hall effect transistor or a ferrite ring with a few turn .

 I don't fully understand why you are doing this .. you are supping up an old car
 if so you have to add the    N0²   

 You are getting into to some heaving stuff .  good luck

[cmcraeslo]

@Labrat101 - You don't understand. I need to cut the signal coming out of the ECU. I can do pickup reading on connected coil, but it's not connected in my case. I dont know how to explain it otherwise

[Labrat101]

@Labrat101 - You don't understand. I need to cut the signal coming out of the ECU. I can do pickup reading on connected coil, but it's not connected in my case. I dont know how to explain it otherwise

Ok .
I think I know what you are trying to do .
 But I have this feeling your are working on a industrial Project .By they way you are keeping 95% of what you are really
 doing  Not told . You have already invested some very expensive equipment in this update to a ignition system.

  This could be a deal breaker for me for sure.

 

  To break a signal is really not Hard but if this is for commercial gain and NOT a hobby etc.
 Sorry I am out  . TTFN

[HighVoltage]

Best to use is a current probe to watch the ECU signal.
Most of the time it is around 10 mA to 20 mA but some latest systems have less than 1 mA

Also, some systems use the amount of current from the ECU to the coils as a signal to get the
coil in to different mods. And others use the current as a feedback from the coils to tell the ECU how the last spark performed.

There are many different systems around these days.
What car is it and what is the coil being used?

[Circlotron]

What I'm trying to do is to build ignition cut device, that is catching the signal, "processing it" and then output it.

Are you making a rev limiter? If so, it might be easier to cut the drive to the fuel injectors instead. Easier on the catalytic converter too. No unburnt fuel going out the exhaust and no bangs and pops.

[joeqsmith]

Hey guys.

I have 12V and 2 signal wires coming out of the ECU into the ignition coil pack. To my understanding, they are active low signals, which means they get activated by the ECU with low signal (ground). I'd like to catch that signal (individually) and I have problem reading them.
...
If I cut the wire and connect the wire coming out of the ECU to 12V using a pullup, I get 12V like I should (I get 12V without the pull up as well) and it does not change when I crank. I tried different resistor values, even pull down - no change. Just as if ECU is not triggering it. What could be the issue?

What I'm trying to do is to build ignition cut device, that is catching the signal, "processing it" and then output it. I have the device working on a car with ignition drivers built in the coils, but not when coils are directly driven.
Device uses very sophisticated cutting strategies, so it's not just "coil off" cutting so I need to have clean signal coming in constantly so I can process them correctly.

I'm not too surprised by your first post.  I had made an engine simulator for my bike.  Getting the ECM not to fault out on the ignition coil took some effort. Attached a couple of pictures of the that load which was a custom made pot core that I mounted inside an aluminum housing and mounted to the case to draw the heat away.   The driver basically is a HV FET that can load the secondaries to cause the ECM to  fault under software control.   This ECM is fairly low tech and I wouldn't be surprised to find the load requirements for an auto to be more critical. 

Are you trying to make some sort of governor (2-step, studder box...)?  Is the car injected?  If so, are you handling the fuel as well?   

In most of my setups, I am using aftermarket electronics which provide a kill signal input.   In the case of this stock ECM, I am using some changes to the OEM software to perform the kill.  The simulator in this case allows me to test different software without the need of running an engine. 

[joeqsmith]

What I'm trying to do is to build ignition cut device, that is catching the signal, "processing it" and then output it.

Are you making a rev limiter? If so, it might be easier to cut the drive to the fuel injectors instead. Easier on the catalytic converter too. No unburnt fuel going out the exhaust and no bangs and pops.

Where's the fun in that??!!

[cmcraeslo]

Best to use is a current probe to watch the ECU signal.
Most of the time it is around 10 mA to 20 mA but some latest systems have less than 1 mA

Also, some systems use the amount of current from the ECU to the coils as a signal to get the
coil in to different mods. And others use the current as a feedback from the coils to tell the ECU how the last spark performed.

There are many different systems around these days.
What car is it and what is the coil being used?

If I understand correctly, you mean current probe on the wire from ECU to a 10k (as an example) pullup?
Test car I have is Mercedes Benz A160.
I have successfully tested this on a coilpack that has integrated drivers (VW).

[cmcraeslo]

Are you making a rev limiter? If so, it might be easier to cut the drive to the fuel injectors instead. Easier on the catalytic converter too. No unburnt fuel going out the exhaust and no bangs and pops.

No, I'm making flatshift (ignition cut) device. I have it working on a coilpack that has integrated drivers, but when ECU is driving the coils directly, it doesn't work.

[cmcraeslo]

Ok .
I think I know what you are trying to do .
 But I have this feeling your are working on a industrial Project .By they way you are keeping 95% of what you are really
 doing  Not told . You have already invested some very expensive equipment in this update to a ignition system.

  This could be a deal breaker for me for sure.

 

  To break a signal is really not Hard but if this is for commercial gain and NOT a hobby etc.
 Sorry I am out  . TTFN

I'm building an ignition cut device (that will allow me to shift while on full throttle on racing gearbox) and it has very sophisticated algorithms to cut it so I need to process the ignition signals very precisely and time it. I have it working fine on a coilpack with integrated drivers, but it doesn't detect the input signal on cars that ECU drivers the coils directly.

[Labrat101]

you have a few choices Hi voltage IGBT or  FET
I think IGBT maybe a better choice and have them oil mount cooled .
 There was a post a few up used Hi voltage Fet's  Both will give you the fast response time.

I have never tried this we only build the plug module for diag plug 5 yrs back for standard cars . Auto Test.
 if its for racing is it a 4 or 8 cylinder just for interest ..
Photo's word help as well   
Mr joeqsmith maybe more your Man ..

Have fun
 

[Circlotron]

I'm building an ignition cut device (that will allow me to shift while on full throttle on racing gearbox) and it has very sophisticated algorithms to cut it so I need to process the ignition signals very precisely and time it. I have it working fine on a coilpack with integrated drivers, but it doesn't detect the input signal on cars that ECU drivers the coils directly.

Yep, you definitely don't want to interrupt the coil while it is charging and consequently get a spark way to early! What about a single relay contact that cuts power to the fuel injectors, controlled by a microswitch on the clutch pedal? Why wouldn't that be good enough? Would the ECU throw an error with that?

[cmcraeslo]

Yep, you definitely don't want to interrupt the coil while it is charging and consequently get a spark way to early! What about a single relay contact that cuts  power to the fuel injectors, controlled by a microswitch on the clutch pedal? Why wouldn't that be good enough? Would the ECU throw an error with that?

It's for racing application where perfect engine cut is needed to unload the gearbox properly. It's for racing gearboxes so no clutch between the shifts. This has to be timed perfectly. What you're describing is a full cut which is something that hasn't been used in this industry for the last 10 years or so (at least the proper ones). Fuel cut on the other hand is very slow and not really useful in my case.
My device is capable of retarding timing of the original ignition pulse, then cutting the spark partially, depending on the configuration and load and making sure that the active spark is evenly distributed among the cylinders and then ramping the restore back to the original ignition timing. It's capable of doing this on up to 6 cylinders parallel, either wasted spark or sequential ignition.

Now, I have this working perfectly on low power/current signal, but when I tried this on a stock car, with regular ignition coil, there was this issue i'm having. So I need to somehow catch this signal's timing when it's not connected to the coil directly.

Driving the coil (output) is not a problem.

[joeqsmith]

Yep, you definitely don't want to interrupt the coil while it is charging and consequently get a spark way to early! What about a single relay contact that cuts  power to the fuel injectors, controlled by a microswitch on the clutch pedal? Why wouldn't that be good enough? Would the ECU throw an error with that?

It's for racing application where perfect engine cut is needed to unload the gearbox properly. It's for racing gearboxes so no clutch between the shifts. This has to be timed perfectly. What you're describing is a full cut which is something that hasn't been used in this industry for the last 10 years or so (at least the proper ones). Fuel cut on the other hand is very slow and not really useful in my case.
My device is capable of retarding timing of the original ignition pulse, then cutting the spark partially, depending on the configuration and load and making sure that the active spark is evenly distributed among the cylinders and then ramping the restore back to the original ignition timing. It's capable of doing this on up to 6 cylinders parallel, either wasted spark or sequential ignition.

Now, I have this working perfectly on low power/current signal, but when I tried this on a stock car, with regular ignition coil, there was this issue i'm having. So I need to somehow catch this signal's timing when it's not connected to the coil directly.

Driving the coil (output) is not a problem.

If we spun the motor at say 20,000RPM, or 333revs per second, or 3.0ms for one rotation.   Using a fixed kill time of 20ms, that's 7 crank rotations.   With your six cylinder we get 3 power strokes per rotation or 21 power strokes.   For this example, it seems long enough that there may be some gains.

My bike bike shifts under full load with no kill during the start (Robinson transmission).  After that, it shifts at 9800 with a 20ms kill.  So 20ms/6.12ms or 3 rotations.  It's an in-line 4, giving us only 6 power strokes to work with.  It's a 5 speed, so only 4 shifts on a pass and again, not all require the kill.    Actually, with this type of transmission, using kill could damage it.  Similar, if you chop the throttle, you can expect repairs. 

Your particular ECM may be looking for a certain combination of di/dt, peak currents and back EMF.   I would just measure the coil that is known to work and replicate it as a starting point.