How do it know?

[Fordum]

I wanted to revive this one.

I went out on a limb yesterday, and told dhcomp that his problem could be either the pickup coil or ignition module even though it was getting spark and fuel.

I know that active electronics can spontaneously output pulses under certain failure conditions.
And I was certain that the ignition module had those in it, but I was not sure how they worked.

So, I did some research. Found some block diagrams of the ignition modules.
They work about like I thought.

A sine wave input from the pickup coil goes into an AtoD converter which generates square wave pulses from the sine wave.
The pulses are fed to both the control module to trigger the fuel and tell it how fast its running, and also to some kind of summing amplifier which drives the coil.

If the DtoA converter starts free running, then you got troubles.

Anyway, in studying all that, I discovered how a DIS can detect which cylinder is misfiring.

Would not work with a distributor triggered system, but with the crankshaft sensor the computer is continuously monitoring the engine RPM on an instantaneous basis.

The article did not say if it actually sees a slight bump in RPM when each cylinder fires, but it does see a slight drop when one misfires.
Since it knows which one it just told to fire, it knows which one did not.

Then, if it sees a sudden brief rich condition at the O2 sensor, then that verifies that a misfire has occurred.

I work with high speed and high frequency electronics, and I know that an engine running at say, 4K is slow compared to even a 50MHZ system, and it would have no problem processing data fast enough to see something like that, but it blows my mind that an engine with that much rotating mass would slow down enough at one misfire for it to be detectable.

But, the article says that is how it knows which cylinder is misfiring.

Just wanted to put this one to bed.

 

[jekquistk5]

I wanted to revive this one.

I went out on a limb yesterday, and told dhcomp that his problem could be either the pickup coil or ignition module even though it was getting spark and fuel.

I know that active electronics can spontaneously output pulses under certain failure conditions.
And I was certain that the ignition module had those in it, but I was not sure how they worked.

So, I did some research. Found some block diagrams of the ignition modules.
They work about like I thought.

A sine wave input from the pickup coil goes into an AtoD converter which generates square wave pulses from the sine wave.
The pulses are fed to both the control module to trigger the fuel and tell it how fast its running, and also to some kind of summing amplifier which drives the coil.

If the DtoA converter starts free running, then you got troubles.

Anyway, in studying all that, I discovered how a DIS can detect which cylinder is misfiring.

Would not work with a distributor triggered system, but with the crankshaft sensor the computer is continuously monitoring the engine RPM on an instantaneous basis.

The article did not say if it actually sees a slight bump in RPM when each cylinder fires, but it does see a slight drop when one misfires.
Since it knows which one it just told to fire, it knows which one did not.

Then, if it sees a sudden brief rich condition at the O2 sensor, then that verifies that a misfire has occurred.

I work with high speed and high frequency electronics, and I know that an engine running at say, 4K is slow compared to even a 50MHZ system, and it would have no problem processing data fast enough to see something like that, but it blows my mind that an engine with that much rotating mass would slow down enough at one misfire for it to be detectable.

But, the article says that is how it knows which cylinder is misfiring.

Just wanted to put this one to bed.

Technology is amazing isn't it. When you compare most electronics to say the ever changing computer and cell phones of today. It is almost as if propaganda is beaten into us. Anything less than 2.5ghz is ancient and old. Not true, I am sure most people would be surprised on how many complex circuits run on a 10/12mhz clock.


Hell I bet even some people would be dumbfounded that most interfacing between boards and controllers is still using the SERIAL bus. I would say some folks wouldn't even know what that is. Sometimes, I think there is no way that will work, but after doing the math I realize I have plenty of room

 

[Fordum]

When I was coming along in computers, the fastest port was the parallel port. After all, the serial port could only send or receive one bit at a time, and the parallel port could handle an entire word ( 8 bits ) at a time.
So, when hard drives got going good, they used parallel cables for data transfer.

One day, I read an article that a new connection system was coming that was faster than the fastest IDE connection.

I thought that was stupid, how could serial be faster than parallel?
You could send 8 times the data per clock pulse.

Then I read up on it, and remembered the problem with parallel.
Its a matter of timing.
When the clock pulse hits, the device reads all 8 lines for data. But, when the speed gets higher, a slight difference between the length of the 8 data cables would mean that some data might not be there yet when the read pulse hits.

At MHZ data speeds, just having the inductance of one of the wires laying next to a piece of steel could cause the lines to get out of sync.

With serial, you only have one data line, so its not as critical at high speeds. This is one of the main reasons that SATA has taken over from PATA.

 

[imiceman44]

Are you familiar with Jacobs Electronics?
That is exactly what the jacobs omni pack does, it detects not only a misfire but irregularities and deals with it with a higher energy spark, so it will give each cylinder the amount of spark it needs to balance the engine firing.
That box with nothing else changed improves idle, response, and mileage.
It replaces the coil but still uses the distributor.

I wanted to revive this one.

I went out on a limb yesterday, and told dhcomp that his problem could be either the pickup coil or ignition module even though it was getting spark and fuel.

I know that active electronics can spontaneously output pulses under certain failure conditions.
And I was certain that the ignition module had those in it, but I was not sure how they worked.

So, I did some research. Found some block diagrams of the ignition modules.
They work about like I thought.

A sine wave input from the pickup coil goes into an AtoD converter which generates square wave pulses from the sine wave.
The pulses are fed to both the control module to trigger the fuel and tell it how fast its running, and also to some kind of summing amplifier which drives the coil.

If the DtoA converter starts free running, then you got troubles.

Anyway, in studying all that, I discovered how a DIS can detect which cylinder is misfiring.

Would not work with a distributor triggered system, but with the crankshaft sensor the computer is continuously monitoring the engine RPM on an instantaneous basis.

The article did not say if it actually sees a slight bump in RPM when each cylinder fires, but it does see a slight drop when one misfires.
Since it knows which one it just told to fire, it knows which one did not.

Then, if it sees a sudden brief rich condition at the O2 sensor, then that verifies that a misfire has occurred.

I work with high speed and high frequency electronics, and I know that an engine running at say, 4K is slow compared to even a 50MHZ system, and it would have no problem processing data fast enough to see something like that, but it blows my mind that an engine with that much rotating mass would slow down enough at one misfire for it to be detectable.

But, the article says that is how it knows which cylinder is misfiring.

Just wanted to put this one to bed.

 

[Fordum]

Are you familiar with Jacobs Electronics?
That is exactly what the jacobs omni pack does, it detects not only a misfire but irregularities and deals with it with a higher energy spark, so it will give each cylinder the amount of spark it needs to balance the engine firing.
That box with nothing else changed improves idle, response, and mileage.
It replaces the coil but still uses the distributor.

Yep, I always liked the ideas Dr. Jacobs had. But, I'm not sure his stuff is as good as it used to be since it got bought out.

 

[y5mgisi]

I thought misfires were identified by the crank/cam sensor "expecting" a signal at a certain time based on the given condtions, and not getting it right when its expected.

 

[imiceman44]

I thought misfires were identified by the crank/cam sensor "expecting" a signal at a certain time based on the given condtions, and not getting it right when its expected.

well Dr Jacobs reads it from the spark plug and the tach signal IIRC no other sensors required, that was the Genius of this guy, he found a way to get it done from what the engines already had even if carbed.

 

[Fordum]

I thought misfires were identified by the crank/cam sensor "expecting" a signal at a certain time based on the given condtions, and not getting it right when its expected.

Well, the crank/cam only looks at rotation as shown by pulses per second. So, it only sees the slight slowdown of rotation when a cylinder misfires.

And in fact, if you had multiport injection, the misfire could be caused by a faulty injector.
The system I mentioned would show that even if the plug fired when it was supposed to.

Dr. Jacob's system would in theory detect it also by the effect on the plug as it fired.

 

[paratrooper307]

If you've got a reader that can show you the datastream, you can watch the information the computer is receiving from all the various sensors. The computer can count misfires in each cylinder and will trigger a code if there are over a certain number within a period of time. Doesn't matter if its a bad injector, no spark, or a broken valve spring, it knows...quite fascinating. The modern CAN/BUS systems have a lot of information available.

 

[Blue85]

I thought misfires were identified by the crank/cam sensor "expecting" a signal at a certain time based on the given condtions, and not getting it right when its expected.

I think misfire detection comes primarily from the crank position sensor. The PCM will have a tolerance on when the next pulse should come, based on current RPM. If you are late enough, it implies a misfire. The dwell to the coil is determined before dwell starts, so you would need a consistent misfire to dial in a different dwell needed on that cylinder. Most OBD-II cars used waste spark systems, so even if something measured differently on one coil, you would still need crank position to know which of those two cylinders was firing at the time. Backing up even further to distributor based systems, most of those didn't even know which cylinder they were sparking.

Now with 1 coil per cylinder you could gain insight from reading the "back voltage" on the primary of each coil. If just knowing the voltage that should "kick back" under each load condition is enough, that's not too hard. Measuring actual waveform timing seems kind of complex, though. I don't know if the cost to benefit ratio is there.

 

[Blue85]

Thinking about this topic, I'm now remembering back when I worked for Delphi...

They made a system that monitors the current through the plug during the spark event. I'm thinking they superimposed some other signal in there that facilitates the measurement. It's called Ionsense or something like that.

 

[black dawg]

misfire counters have always amazed, but you do have to watch out cause they can be wrong.

the way I always understood it was that the computer watches crankshaft velocity between cylinders to find misfires.