Best ignition setup to run

[quickt]

Wish I had more details on the spark scatter issues you found. Is there a link to an old post?

We were working on a Mustang 347 SB ford. This car had the engine in and out several times with what appeared to be detonation. The car was just hi maintenance. He was making around 1200 RWHP revving the engine to 9500 RPM. He was using aftermarket engine management, MSD flying magnet, Distributor and a digital MSD box. Engine pulled clean made good power. Sometimes out of the clear blue sky would start to miss. Pull spark plugs and signs of detonating on different cylinders. Was never a uniform problem. The engine problem was never in the same cylinder. After being convinced that another engine management would fix this problem because they had the better deal he swapped. Problem did not go away. After installing many Electromotive systems on Cobras I was talking to them on the phone and they told me about a spark scatter problem with 3 tooth triggers and systems that would resync every 720 deg. We loaded up the stang on the dyno pulled out the timing light and pulled the engine to around 9000 while watching with a timing light and there it was. Timing was floating all over the place. Then it hit me The timing light only shows the number one cylinder. We put a scope on the car checking all 8 cylinders and thats when we really seen what was going on. The timing would vari on the other cylinders same as number one. Sometimes it would go off the scale 10 degrees or more. The engine never once sputtered. So after replacing the crank trigger and new box, distributor, wiring harness took every precaution to prevent noise nothing helped. Electromotive offered up a system to test. We installed it problem was gone. Not at any time with this system did we see a deg of variance. The guy bought the system. The car ran last to years with no more mystery issues. What we found is that when the system would miss a trigger event it would allow the timing to go out. It might happen with the entire firing order or just a single event. Never the less the problem was there. We had a total of 4 systems on the car including my personal favorite 3 of the systems had some degree of spark scatter issues. The funny part was you never would have known it. The car pulled clean made good power. They just thought they were hurting the engine because they were pushing it. They would pull timing lower boost but still had small issues. After the problem was corrected they could really lean on the car. I was sold to say the least. I was wondering how many engines have been damaged from this issue. Even more disturbing is some refuse to believe the issue exist. I am not a big fan of the TEC software. Its workable. I am sold on the ignition.

 

[lazaris]

We were working on a Mustang 347 SB ford. This car had the engine in and out several times with what appeared to be detonation. The car was just hi maintenance. He was making around 1200 RWHP revving the engine to 9500 RPM. He was using aftermarket engine management, MSD flying magnet, Distributor and a digital MSD box. Engine pulled clean made good power. Sometimes out of the clear blue sky would start to miss. Pull spark plugs and signs of detonating on different cylinders. Was never a uniform problem. The engine problem was never in the same cylinder. After being convinced that another engine management would fix this problem because they had the better deal he swapped. Problem did not go away. After installing many Electromotive systems on Cobras I was talking to them on the phone and they told me about a spark scatter problem with 3 tooth triggers and systems that would resync every 720 deg. We loaded up the stang on the dyno pulled out the timing light and pulled the engine to around 9000 while watching with a timing light and there it was. Timing was floating all over the place. Then it hit me The timing light only shows the number one cylinder. We put a scope on the car checking all 8 cylinders and thats when we really seen what was going on. The timing would vari on the other cylinders same as number one. Sometimes it would go off the scale 10 degrees or more. The engine never once sputtered. So after replacing the crank trigger and new box, distributor, wiring harness took every precaution to prevent noise nothing helped. Electromotive offered up a system to test. We installed it problem was gone. Not at any time with this system did we see a deg of variance. The guy bought the system. The car ran last to years with no more mystery issues. What we found is that when the system would miss a trigger event it would allow the timing to go out. It might happen with the entire firing order or just a single event. Never the less the problem was there. We had a total of 4 systems on the car including my personal favorite 3 of the systems had some degree of spark scatter issues. The funny part was you never would have known it. The car pulled clean made good power. They just thought they were hurting the engine because they were pushing it. They would pull timing lower boost but still had small issues. After the problem was corrected they could really lean on the car. I was sold to say the least. I was wondering how many engines have been damaged from this issue. Even more disturbing is some refuse to believe the issue exist. I am not a big fan of the TEC software. Its workable. I am sold on the ignition.

Lonnie,

Can the TEC ignition system be used on DFI, XFI , BS3 ECU's ?

 

[Alky V6]

You could run the Electromotive XDI system alongside another system that will provide fueling. It is a stand alone ignition only system. You have to mount a 60-2 trigger wheel to run the XDI or the TEC units. The 60-2 wheel is really the heart of these systems. If your fueling system can't use the 60-2 trigger wheel, you'll have to run 2 types of trigger wheels. One for ignition control and the other for fueling. Seems like a massive bother and expense to me to run two systems side by side, but it could be done.
Like Lonnie said, the WinTEC4 software interface is not the fanciest, but if you can get around that, it will do what the vast majority of us require. Actually, the more I work with it, the more I actually like it. I really like the datalog playback (flight recorder) with varying speed control. You can watch the gauge panel (dashboard) and it's as if you're making the run all over again. Run it in realtime or slow it down to catch more detail. Like the Datalogger deal.
The WinTEC4 software that Electromotive has come out with will be upgraded with new features as time goes on.

 

[TurboTR]

Thanks for the info quickt. The Tec sw, totally agree. For user's sake, hope they get that improved sometime.

Wanting to understand this better, and just some of my thoughts here. To make sparks, aftermarket ecu's do not need 360 or 720 deg sync. They just spark xx deg after a reference pulse comes in basically. And since the stang had a distribuor, neither the ecu nor an MSD box would need any kind of sync for spark. not sure aboutthe "digital" box that was on the car though. Again, it should just spark after each reference pulse and not need 360 or 720 deg sync. Assuming a delayed spark setup (leading reference angle)

That said, I have seen (with a scope) that the MSD sync pulses (analog VR pulses) from their distributor are very strong (high amplitude). Probably too strong IMO. Have not pursued that further, but was struck by how strong it was even at low speed (looking at a supercharged SBC truck at the time). An input circuit might have trouble with that signal level IMO.

And, MSD tends tohave less than stellar reliability, from what I've seen and heard. Maybe there was just a problem with that particular MSD setup? Is that possible here? Seems so, based on what you'd explained above.

Ok with your heads up information in hand, will be paying alot of attention to that now. Again, thanks. Planning some bench testing now. Will report back if/when we make any headway when work allows.

TurboTR

Edit- Sorry, forgot the forest for the trees here (lol). One big possibility that could explain the above issue is- reversed polarity from the VR sensor. I think this is not unusual to be honest. And it causes unstable reference pusles (and spark), like you describe above.

The Megasquirt page has a decent explanation of how a VR inupt works. For it to work reliably, it has to be correct polarity. A picture of the VR signal here is required to understand the following discusion; the green trace here makes a good example of one.

The quick reason is this: Assume a positive going pulse, followed by a fast high/low transition (the zero cross) on the middle of the sensor lug passing to a negative going pulse. The input circuit "arms" on the initial positive going portion (kindof where the arrow is pointing on the green trace in the picture), then "triggers" on the rapid (sharp) high/low transition that follows. The sharpness of tha transition is the key here to a stable puklse detection.

If you reverse the leads, then the input circuit doesn't "arm" until the low/high (now, because it's reversed) transition occurs. This is the edge that we're trying to detect, but we missied it. Now the circuit "triggers" instead on the slowly descending edge that follows, as the following positive going pulse decays away. Not good, not a stable, reliable trigger.

Is another possibility here at least, I'd say.

 

[Alky V6]

They just spark xx deg after a reference pulse comes in basically.
TurboTR

The spark occurs xx time after a reference pulse, since a low resolution type trigger wheel can't supply enough information to the controller to allow a spark event to occur at a more precise crankshaft degree angle.

 

[SS/GN]

Msd

What we found with carbed set ups,is the MSD 6a series is the most durable however we have a superstocker and every little bit of power we can find we keep.The msd 7 and 8 series are very easily damaged with voltage spikes from batter charging etc.An alternator can easily damage a box.We send the boxes back to msd and they say their is no problem.The box actually get's lazy.We have seen as much as 30 hp between boxes.The digital stuff in carbed set ups makes no more power than the 7 series boxes.This is one of the best threads in months.Thanks for the info,very interesting

 

[Alky V6]

What we found with carbed set ups,is the MSD 6a series is the most durable however we have a superstocker and every little bit of power we can find we keep.The msd 7 and 8 series are very easily damaged with voltage spikes from batter charging etc.An alternator can easily damage a box.We send the boxes back to msd and they say their is no problem.The box actually get's lazy.We have seen as much as 30 hp between boxes.The digital stuff in carbed set ups makes no more power than the 7 series boxes.This is one of the best threads in months.Thanks for the info,very interesting

If I were running a carb setup, there is no question, I would be running a EM XDI. No more msd box worries.

 

[SS/GN]

Yes indeed

We have to run a dist thats why I like my Buick (new challenge).I have seen the EM on a local dyno session and was very impressed how easy it was to use with a carb set up.

 

[quickt]

Thanks for the info quickt. The Tec sw, totally agree. For user's sake, hope they get that improved sometime.

Wanting to understand this better, and just some of my thoughts here. To make sparks, aftermarket ecu's do not need 360 or 720 deg sync. They just spark xx deg after a reference pulse comes in basically.

The Stang had a flying magnet Crank trigger. The dist was the cam sensor. Funny you mention that the ecu does not need to resync after start up. I found there is something funny about that. Unplug the input from the dist after start up if the car would go out of time it took a long time for it to get back on track. So if the ECU never looks at cam signal after start up why would that seem to make a difference? Plug the dist back in she would come right back. Unplug cam sync and start car still have same issue. I know that technically with the distributor you really dont need a cam sync because the dist mechanically aligns number one. But we found when the ECU misses a trigger event it goes nuts even if for only that one event. So the question I have asked with absolutely no reasonable answer is. What happens when the ecu misses a triggering event? How does it get back to square one? With the 60-2 wheel it counts 1-60 hits the -2 tooth (number 1 TDC) and starts the count over again in essence re-syncing every 360 deg.

Now the real problem I have with this whole deal is what happens when you take out the mechanical alignment if the firing order the Dist provides and go DIS? Then the ECU makes all the decisions.

 

[TurboTR]

Well all we can say is, the FAST I have runs with no cam signal at all. It doesn't need it. That answers the question definitively.

The C3I needs it though to maintain 360 deg sync, so it can fire the right coil.

The FAST in my case just tells the C3I when to fire (and when to start the dwell as well, the Achilles heel of the stock Buick C3I at high rpm). It does NOT tell it what coil to fire.

If you miss a ref pulse your'e in trouble. In my case there is no ecu re-syncing involved, but the C3I would be out to lunch then. Until it could sync up again. The distributor would not need to re-sync of course.

But the missing pulse could totally screw up the ecu rpm and position calculation (to be able to time the spark), and the ecu would likely not know what's going on. Don't know what kind of fault tolerance could be built in for that.

The flying magnet VR setup, IF the polarity was backwards as described above would have lots of spark issues. Again, IF, not saying it was. If I saw spark moving around like that, VR polarity is the first thing I'd check though.

And having wrong polarity ALSO makes the overall timing move in a general direction as well when rpm changes. This happens because the amplitude of the VR signal also grows and shrinks radically with changing rpm. And so, if the signal polarity is wrong the point that the ecu triggers on (on the wrong part of the signal rather than the zero crossing) moves around with rpm as well. That could also totally explain why the engine suffered unexpected damage, even when set conservatively. Just saying..

TurboTR

 

[S2V6RACER]

It seems like this study done on the 9500 RPM Mustang would be magazine article worthy data if true. You would think there would be more press about these issues. Lonnie, on the 24x (or 58x) BS3 COP test you did on this Mustang, did you see 10 degree spark scatter or was that from another system you tested? Kudos to the guys involved in this test, it must have taken forever and a day to get the tune up right every time.

 

[quickt]

It seems like this study done on the 9500 RPM Mustang would be magazine article worthy data if true. You would think there would be more press about these issues. Lonnie, on the 24x (or 58x) BS3 COP test you did on this Mustang, did you see 10 degree spark scatter or was that from another system you tested? Kudos to the guys involved in this test, it must have taken forever and a day to get the tune up right every time.

Ted why ask questions? You cant answer any when you are asked. Its obvious what your motive is here. The ignition system used was listed. Once again help with your reading comprehension. COP or 24X was never installed. There were 2 systems on the car during the course of the summer. First was the XFI and then the BS3 Chuck installed himself. I installed the Thruster and then the EM. The EM being the only DIS on the car. We spent about 7 days on the car with the last two systems. I have told many about it. I had conversation with Cal while all this was going on telling him about it. He was aware of it as were many others.
Why share info? There is always a guy like you to argue with the findings if you dont like what you hear. The info is out there. Let people draw their own conclusions. You like anyone else can choose to think its FOS. I have my opinion as to what was going on. Right or wrong its my opinion there cheif. Why dont you get off my A$$ and find someone else to follow around the boards.

 

[Alky V6]

It seems like this study done on the 9500 RPM Mustang would be magazine article worthy data if true. You would think there would be more press about these issues.

What is the primary job of magazines? It's to sell and promote product. That is how they attract more advertisers to help support the magazine. Do you really think you would find an article in a magazine explaining the shortcomings of popular products? That would just put all their advertisers on notice, and would be the beginning of the end for that publication. Do you agree?

 

[quickt]

But the missing pulse could totally screw up the ecu rpm and position calculation (to be able to time the spark), and the ecu would likely not know what's going on. Don't know what kind of fault tolerance could be built in for that.

This is the question that for what ever reason cant be answered. If the wheel is 3, 4, 24 or 58 or any number of teeth with equal spacing 360 degrees. How does the ECU get back on target when the ECU mis-reads a trigger event?

I am sure the trigger was the problem. The chances of 3 ECUs of different brands having similar issues would be slim to none. It was my thought that more resolution would have helped.

 

[RmvBfrFlght]

This is the question that for what ever reason cant be answered. If the wheel is 3, 4, 24 or 58 or any number of teeth with equal spacing 360 degrees. How does the ECU get back on target when the ECU mis-reads a trigger event?

I am sure the trigger was the problem. The chances of 3 ECUs of different brands having similar issues would be slim to none. It was my thought that more resolution would have helped.

In the event of a missing pulse to the ECU, it will drop an ignition event (one cylinder won't fire). This missing pulse will also drop an injector firing.

So until the cam sensor comes around again the injector firing and spark plug firing will be one cylinder behind.

The C3I is out of sync until the cam sensor comes around again.

GM went to lengths to provide a decent reference pulse signal from the C3I to the ECM. There is both a REF- signal (ground at the C3I), and then the REF+ signal. Inside the ECM there is a differential amp that uses the two signals together.

The REF- is not grounded inside the ECM. It goes to one leg of the diff amp, with REF+ going to the other leg. This is done to reject common mode noise.

Another aspect of ignition timing on the TR is the acceleration of the engine and crank snout flex/twist. The ECM calculates the RPM via the time between reference pulses. With the RPM value it will then convert the desired timing event from degrees of crank position to a time value after the previous reference pulse.

{edit: the ECM actually uses the time between reference pulses for the conversion from crank degrees to time after the reference pulse. Note that the ECM also uses the time between reference pulses to calculate the RPM value.}

So the actual signal to fire the C3I is based on the previous cylinder reference pulse (from the 3 window wheel). It is easy to see how a rapidly accelerating engine will have this calculation off by some amount.

Along with any twist at the crank snout affecting the time the reference pulse is generated.

By adding teeth (or windows) to the crank sensor along with the proper electronics can provide a better reference signal to the ECM. This is most likely why the EM stuff works so well.

{edit: In this case the ECM is the EM ignition controller.}

It can 'see' the rapid acceleration in RPM along with crank twist. By averaging over many teeth the crank twist can be nearly eliminated.

RemoveBeforeFlight.

 

[Alky V6]

With the explanation offered as to how the spark event is calculated between large gap trigger pickup points, and the amount of spark scatter that's been observed, there is a massive flaw either in the programming that's responsible for the spark event calculation or a mechanical flaw with the pickup points (crank and camshaft). How much do you think timing chain slack, crank twist, camshaft sensor endplay, cam sensor gear backlash, etc. really plays into this? Does the same amount of scatter occur on a fresh engine with a tight chain, tightened endplay on the camshaft sensor? If there isn't much difference, it would have to be a simple flaw, or much more likely, a simple shortcoming in the programming. Or, the engineers just figured the accuracy was good enough for a 250hp production V6?

 

[RmvBfrFlght]

With the explanation offered as to how the spark event is calculated between large gap trigger pickup points, and the amount of spark scatter that's been observed, there is a massive flaw either in the programming that's responsible for the spark event calculation or a mechanical flaw with the pickup points (crank and camshaft). How much do you think timing chain slack, crank twist, camshaft sensor endplay, cam sensor gear backlash, etc. really plays into this? Does the same amount of scatter occur on a fresh engine with a tight chain, tightened endplay on the camshaft sensor? If there isn't much difference, it would have to be a simple flaw, or much more likely, a simple shortcoming in the programming. Or, the engineers just figured the accuracy was good enough for a 250hp production V6?

I browsed through this thread and from a quick look only found one reference to a large amount of spark scatter. It was Lonnie working on a non-TR: "We were working on a Mustang 347 SB ford." Although I may have missed a post where this was found on a TR, just let me know the post #.

{edit: OK, I found it, Lonnie states in post #25: "When the C3I is removed and dist replaces it. It also seems spark scatter will get much better until you put the timing light on at the chassis dyno at hi RPM. Then you see the scatter again maybe only 3-5 deg but its there. The C3I is much worse."

That is interesting. Figure for an example that "high RPM" is 6,700. That is a crank revolution (360 degrees) every 9 milli-sec (rounded). A 5 degree timing error is 125 usec. (9 ms / 360 * 5 = .125), not a lot of time. But more then enough to be important. End of edit}

The Buick '7148 ECM does the timing calculations correctly. Any scatter would be related to the Hall Affect pickup (crank) and/or crankshaft twist.

With a properly adjusted cam sensor the timing chain and cam sensor end play & gear backlash won't have any affect on the spark timing. The cam sensor only sets up the cylinder firing sequence.

RemoveBeforeFlight

 

[Alky V6]

I browsed through this thread and from a quick look only found one reference to a large amount of spark scatter. It was Lonnie working on a non-TR: "We were working on a Mustang 347 SB ford." Although I may have missed a post where this was found on a TR, just let me know the post #.

{edit: OK, I found it, Lonnie states in post #25: "When the C3I is removed and dist replaces it. It also seems spark scatter will get much better until you put the timing light on at the chassis dyno at hi RPM. Then you see the scatter again maybe only 3-5 deg but its there. The C3I is much worse."

That is interesting. Figure for an example that "high RPM" is 6,700. That is a crank revolution (360 degrees) every 9 milli-sec (rounded). A 5 degree timing error is 125 usec. (9 ms / 360 * 5 = .125), not a lot of time. But more then enough to be important. End of edit}

The Buick '7148 ECM does the timing calculations correctly. Any scatter would be related to the Hall Affect pickup (crank) and/or crankshaft twist.

With a properly adjusted cam sensor the timing chain and cam sensor end play & gear backlash won't have any affect on the spark timing. The cam sensor only sets up the cylinder firing sequence.

RemoveBeforeFlight

Understood. So the cam sensor signal is not 'expected' a certain time period after the last crank ref signal? But it does setup timing of the injector firing in relation to crankshaft angle. Correct?

I wish someone had the exact parameters of the OEM ignition control programming to share. That would be very interesting.

 

[TurboTR]

Thanks Remove

Since the EM setup has 1 crank position marking pulse every 6 deg, it should be able to position a spark more accurately in general than the 1 pulse per cylinder setup. As long as it has an algorithm to match.

So that leaves the real question- assuming it likely can't quite match the 60-2 tooth approach in overall accuracy, is the 1 pulse per rev setup still "good enough"? How close is "good enough"?

quickt, DonW, et al have generated enough interest here- now we need to get the 1st gen FAST onto the test bench and dig in. Hopefully this week, assuming work allows a free evening. My own FAST still uses the stock Buick approach of 1 ref pulse per cylinder @ 10 deg BTDC (predictive spark), so it should be working from the most stale crank reference information possible and be worst case, you'd think. Hopefully we'll soon see.

We have high end DSO scopes with extensive Math function capability, so hopefully we can test this in a straightforward way with the scope.

For testing, this really becomes a phase detection problem. Just looking at the EST pulses in time as you slew the frequency of the CRS (rev the engine) won't work for us here except in CRS frequency = steady state.

quickt we should also be able to make the crank input test signal drop a ref pulse too and see what happens.

TurboTR

 

[Alky V6]

Sounds good. Nothing beats good ol' testing.