The PFC and dwell
 

Did some work with this today and figured I'd post my initial findings. I want to go back and play with the settings to confirm their operation but what I have already may interest some. I hooked up my Innovate LMA-3 to the trailing coil trigger feed. I typically use this to log RPM. This time I also setup the LMA-3 to log dwell on this channel. You end up with a percentage value that corresponds to the percentage of on-time versus available cycle time. Think of it as duty cycle if you will. The log is shown below. This was done with no load. Since the PFC shows no adjustment for dwell v. load I have to assume at this point the dwell will remain constant under load. It will be easy to verify this later on.


http://www.ludwigmotorsports.com/log.bmp




Doing a calculation to find actual charge time is simple enough. Below I have arranged the PFC settings to what I have actually logged. Unfortunately I don't see a correlation between the PFC numbers and actual dwell or charge time. I did not rev the car out to 8000 rpm to get a reading there. You can get an idea of what the PFC would interpolate at 7000 and compare it to the actual numbers.


http://www.ludwigmotorsports.com/results.bmp




Thoughts?

The charge time looks about right for the coils I'm using until the revs pass 5500-6000. I'm going to play with raising the numbers in the dwell settings on the PFC to see if that really will raise the charge time at the coil.

Thanks for the interesting data. I think the PFC dwell settings have always been somewhat of a mystery, even to the experts (Chuck). I'd be interested to see what happens when you increase the PFC values.

I'd also be interested to see someone do the same test with an HKS twinpower installed because there have been claims that all it does is increase the dwell time.

Posed the question "what do the numbers in the datalogit represent on the dwell settings" on EFI101 and someone responded that they represent degrees of crankshaft rotation. Makes perfect sense if this is the case and sounds entirely plausible to me. The highest number the datalogit will accept is 255.

Here's where things got funky for me. Obviously, with a wastespark leading configuration you can not charge the coil(s) for 255* of crank rotation on a rotary application. You need to charge it, discharge it, and allow a slight amount of recovery time in 180* of crank rotation. At 7500 RPM (the highest revs this particular engine should see) we have 8.57 mS available during one 360* revolution for the crankshaft. So we would have about 4.2 mS to complete each 180* spark event on the leading coil(s). Every system I have worked with until now would allow you to program in whatever charge time you wanted and would automatically discharge the coil at the appropriate time even if the charge time was greater than the available cycle time at high RPM. It's look like the PFC might not do this.

When setting the dwell to 255 at 8000 RPM (obviously a greater charge time than is available on the leading coil(s) on a rotary) the engine hits what can best be described as a very smooth rev limiter around 7000 RPM. In this case the actual rev limiter was set at 8400 so we were not on that. I can only assume that the PFC was holding the true charge time and not discharging the leading coils for every event. Lowering the charge time back below 200 allowed revs to 7500 RPM (the highest I wanted to rev the engine anyway).

So, it seems that there is a practical limit to the upper end of the dwell setting, at least in a rotary application. I'll go over my data some more and make another post on my final settings and how they relate to logged charge time.

I don't know much about dwell. How could adjusting dwell times improve performance or driveability?

Dwell (in points, distributor speak) or charge time (in distributorless speak) is the amount of time the primary ignition coil is charged before the spark event. Any given coil will have it's need for peak charge time. Some quite short, some quite long. Anywhere from 3 to 10 mS usually. Charge times shorter than this leaves the coil less than saturated and produces a weaker spark than a fully saturated coil. Any charge time over the amount needed to saturate the coil simply produces heat and will eventually ruin the coil.

That makes sense. Using the RPM & PFC dwell values you posted you can calculate a charge time and it looks to be pretty constant over the RPM range. However, it doesn't seem to line up with the actual charge times you recorded in your log.

Interesting stuff none the less. I wonder what the ideal saturation time is for an FD coil.

Ok this is weird. So I'm looking through my Banzai Racing PFC basemap (for a series 4 Turbo II) and all the dwell settings are set to 30, both leading and trailing. I wonder why that is. I know that two trailing coils are essentially spliced together due to the differences in FC and FD ignitors, but really the two generations aren't too different in the way they do ignition from what I understand. I also wonder if playing with those settings could improve spark performance. I have been getting intermittent ignition blowout despite having an MSD 6AL box on the leading (FC) coil and B10EIX plugs gapped to .28 .

I wonder what the relationship is between optimum dwell time and plug gap, if any.

True. The higher RPM values do match up relatively close. I need to put a scope on it instead of relying on the LMA-3. I think there is certainly some loss of accuracy there. It was just the handiest way at the time to take a quick look at what was going on.

Revisited this with the scope today. Below are the results. The scope measures actual charge time in mS instead of trying to convert a measurement from the Innovate logger. I'm confident these measurements are pretty accurate. The scope would fluctuate slightly by about .25mS up or down at each measurement but no wild swings at all. The "Theoretical Charge Time" is calculated by assuming that the Datalogit dwell number is indeed a measurement of degrees of crankshaft rotation. Such that for 1k RPM for example 12* would represent 360*/12* equals 30. At 1k RPM one revolution takes 60mS. Dividing 60mS by 30 nets us 2.0mS which is exactly what the measured charge time is. The numbers are very close. I can't explain why they begin to separate at 7k. Part of this may be that there is not an input for 7k and I extrapolated the 174* measurement assuming the calculation is linear between 4k and 8k. Also, I didn't want to hold the engine at 7k for an extended period of time so the measurement may not be the most accurate.

In the end I feel safe in assuming that the Datalogit numbers do indeed represent degrees of crankshaft rotation and will program accordingly from now on.

http://www.ludwigmotorsports.com/dwell_2.JPG

This is pretty interesting... people reading this should remember that adjusting the dwell time has no effect on charge time if using "intelligent" igniter's to include any brand or type of CDI, such as the twin power, MSD etc.

True. FWIW I have been measuring this on a customers car equipped with the 034 inductive coils.

^^ that explains why changing dwell did jack shit then.

I just found this thread out of my own interest in dwell times. Ludwig, I've usually found that when the powerfc has a 0-255 range, it usually represents 0-100%. Would that make more sense in your crank calculations?

thewird

what PFC number in the 0-255 range represent 3.0, 3.5, 4.0 and 4.5ms dwell???

Not based on what I measured with the scope. When I used the numbers as degrees of crank rotation, the charge time jived with the available time in each rev range.

That depends on the RPM range.

Old post, but what could you of meant by this?

I am assuming arghx is using an inteligent igniter and that is why changing the dwell times in the pfc had no effect on his ignition system.

Old topic, I know. I have a customer with an S5 with AP Engineering PFC and like you say both leading and trailing dwell is set to 30 everywhere. for leading this doesnt interest me as the car is running a Dynatek ARC-2 with stock leading coil, but for the trailing side its worrying me a bit since I have no clue if this number is meant to be there for some reason because of the funny trailing coils setup on the FC or if its just running way to less dwell at high rpm.

back from the past. This primary currents I measured on a FC using the banzai 30 settings at all rpm. Its 850 rpm leading, 6000 rpm leading, 850 rpm trailing and 5000 rpm trailing.

It appears the FC igniters have current limiting to 8 amps inside. I wil redo the measurement on a FD to verifiy if these show similar behavior. Somehow the 30 value seems for the leading already on the limit at 7000 rpm where charging happens again when the spark is still going from the previous cycle. Whats even more strange is why the FD basemaps run such a big numbers for dwell. The voltage correction is same at around 1 at 14v

https://cimg6.ibsrv.net/gimg/www.rx7...9049eaea4.jpeg
https://cimg7.ibsrv.net/gimg/www.rx7...35ccc6068.jpeg
https://cimg8.ibsrv.net/gimg/www.rx7...198d67ac9.jpeg
https://cimg9.ibsrv.net/gimg/www.rx7...35edf76c4.jpeg

some more testing. Stock S5 turbo traces and a FDS8 with PFC with the default FD basemap dwell settings.

FC stock ecu trailing 750 RPM
https://cimg1.ibsrv.net/gimg/www.rx7...a41bb3e68a.jpg

FC stock ecu leading 800 RPM
https://cimg2.ibsrv.net/gimg/www.rx7...cf12ab479d.jpg


FC stock ecu leading 6000 RPM
https://cimg3.ibsrv.net/gimg/www.rx7...f4a4f35463.jpg


FC stock ecu trailing 6000 RPM
https://cimg4.ibsrv.net/gimg/www.rx7...a050bc8e14.jpg


FDS8 PFC basemap settings leading 850 RPM
https://cimg5.ibsrv.net/gimg/www.rx7...c6a53dff7e.jpg

FDS8 PFC basemap settings trailing 850 RPM
https://cimg6.ibsrv.net/gimg/www.rx7...d0fb922b5f.jpg

FDS8 PFC basemap settings leading 6000 RPM
https://cimg7.ibsrv.net/gimg/www.rx7...7a376c3382.jpg

FDS8 PFC basemap settings trailing 6000 RPM
https://cimg8.ibsrv.net/gimg/www.rx7...6394b4b14b.jpg

FDS8 PFC basemap settings leading 6000 RPM with R-Magic aftermarket ignitor (forgot to add the cursors :/)
https://cimg9.ibsrv.net/gimg/www.rx7...b3d02c1537.jpg

Still cannot really understand why the FD PFC in the FC with the banzai/AP settings at 30 troughout has such a long dwell at high rpm. doesnt make sense

After measuring the signal wire it becomes clearer why the FC ignition setup works just fine with the ver short dwell number present in the banzai/AP base tunes.

It seem these ignitors are a mixed breed between the smart J109 units found on pre FC era cars, and the dumb ignitors as found on FD's. In fact they seem to calculate their own dwell and rely on the falling edge of the trigger to fire. there is a minimum criteria so if one feeds them a very long trigger signal they will also dwell very long but limit the current. pretty damn awesome ignitors if you ask me :D

here is the proof; 6000 rpm leading with the 30 setting as per AP/banzai tune. Its clear that it automatically starts charging way ahead of the trigger signal going to 5v and only relies on the falling edge to fire
https://cimg7.ibsrv.net/gimg/www.rx7...5b3241b123.jpg

here is the opposite where I feed it with a very long trigger pulse at idle rpm and it just governs the coil current waiting for the falling edge to blow its load
https://cimg8.ibsrv.net/gimg/www.rx7...8a08306343.jpg

awesome tech

doublepost