N5X Ignition coil testing, part 1

V8bait

Lieutenant
Nov 2, 2016
500
773
0
Texas
As some of you know, the last couple months I have been helping reverse-engineer the BMW ignition system in order to help develop a path for upgrading it. The stock system isn't horrible, but it has many shortcomings as everybody who owns a high powered N54 is well aware of. That being said, since others seem to also be developing upgrades, I think I'll start releasing some information on what I've found for the benefit of the community. Unlike some other threads, this is all new information, and will be expanded and refined the next 1-2 months as testing continues.

1) What are the hardware limitations of the DME? We'll use the older MSD80 to expand on this question. If the hardware is good, it opens up different options than if the hardware is lacking to begin with.

Nearly all BMW's use TCI systems for spark (transistorized coil ignition), which means they have a transistor IGBT (also known as an ignitor) to drive an inductive coil. The transitor is located inside the DME, as opposed to some cars which have a module. The transitor used is a Fairchild ISL9V5036S3ST class IGBT, the datasheet can be downloaded here. It's a very nice transistor, capable of over 30 amps per channel and 250W heat dissipation, among other key features.

IGBT.jpg


The way the circuit functions, is you have a 5v logic signal to the IGBT in the DME. That grounds an output to the coil, energizing the coil (coil has constant 12v and a secondary ground in addition to the coil - signal). When the 5v logic signal ends, the ground path ends, and the coil fires. This is a "dumb" system since the transistor (aka ignitor/IGBT) is separate from the coil. Basically, the transistors function as the points in an old muscle car. Smart coils (LS, Audi, Honda) have the 5v logic going straight to the coilpack, which has the IGBT incorporated.

2) What are the current ignition coils capable of? Bosch/Delphi OEM coils were used to establish a baseline, to see just how powerful the factory system is, and what the limits are for dwell time.

The best way to test an ignition coil is with an oscilloscope. There are a few key things to be concerned with including primary coil amperage draw vs dwell time, primary coil saturation point, secondary coil peak voltage potential, and secondary coil output. Some of these measurements are more difficult to obtain that others. Here are some of my results that might be relevant to everybody on the platform for OEM coils (as always I encourage others to independently verify this but I'm confident in these results):

a: Bench testing OEM coils. Top trace (green) is function generator signal (dwell), yellow is current draw, lower trace is high tension (secondary voltage, spark starts when primary current collapses and ends at the squiggly, better seen in lower image with upgraded coils)
OEM Bench dwell.png

b: Testing OEM coils on a running car. Current draw can vary based on operating conditions but is a quick and easy way to see if the tables in the XDF are functioning as expected. In a nut shell, at the beginning of this project, they were not. Now they are, and similar to the bench testing the coils saturate around 3mS at ~14V. Note- current clamp was being finicky today and no good secondary coil data due to technical difficulties with interference (no spark plug wires makes it difficult to obtain):
Composite on car OEM coil.png


If you want to know the mJ output of the Bosch, it's low. The DME can handle more amperage than this, the coils were simply saturating. More will be released on this later.

Just for fun, here's what an upgraded coil looks like on the test bench at 1, 2 and 3mS with significantly stronger secondary output:
comp MM.png


3) Is the DME hardware limited by software? Just because the hardware is good, doesn't mean BMW won't stop your fun.

Some of us have tested the Dwell table before and found that the car did not react as expected. Some of these people have voiced concerned that despite the new tables we still do not have proper dwell control with the DME. I feel my results here are compelling. Regardless. Working with Jake and MHD, the Dwell table appears fully functional and at least one cause of current limiting has been resolved.
Lol
4) Which coils should be considered for upgrades?

The answer is always LS swap, for everything car related. Except LS are smart coils and tend to discharge early if over driven. Price, reliability, power are the players. I would love to hear opinions on what would constitute the best coil for these cars, as it seems there's a few drastically different versions coming out in the next 1-2 months.
 
Last edited:

AD-ENG

Corporal
Oct 22, 2016
122
140
0
39
I think a good choice would be the GM IGBT 2wire coils with short plug wires or CBR coil (if they fit in the hole).

4g63 guys have good success with Chrysler 300 coil (2wire dumb), but I don't know much about them expect they have a long thin boot that may fit the narrow n54 bore.
 
  • Like
Reactions: V8bait
Feb 1, 2017
42
47
0
Nice to see someone testing things rather than "just throwing random coils from rock auto at it".

I did however get some different results from testing stock coils.
3A53CDE7-B434-43AA-AA5C-29609E8DB151_zpsuug8uwy0.jpg

About 7amps with stock dwell at 14v, which coincides with other factory coils I've tested in the past. I also displayed the voltage and amps per division as this is useful when determining dwell time.

The R8 coils tested produced significantly higher amperage readings of 15.6 amps at only 2.5ms of dwell while still not fully saturated.
240AE94A-1A6D-43AF-A37D-2D15A625DC0A_zpsttyfv3np.jpg
Knowing the coils you're using put out more MJ, I'm interested in seeing the amp draw numbers from them.
 

V8bait

Lieutenant
Nov 2, 2016
500
773
0
Texas
If this guy is helping, then this is worth paying attention to...

Jake is great at the DME stuff for sure!

Nice to see someone testing things rather than "just throwing random coils from rock auto at it".

I did however get some different results from testing stock coils.
3A53CDE7-B434-43AA-AA5C-29609E8DB151_zpsuug8uwy0.jpg

About 7amps with stock dwell at 14v, which coincides with other factory coils I've tested in the past. I also displayed the voltage and amps per division as this is useful when determining dwell time.

The R8 coils tested produced significantly higher amperage readings of 15.6 amps at only 2.5ms of dwell while still not fully saturated.
240AE94A-1A6D-43AF-A37D-2D15A625DC0A_zpsttyfv3np.jpg
Knowing the coils you're using put out more MJ, I'm interested in seeing the amp draw numbers from them.

Could be accurate, the current clamp was acting up all week with interference. It's been consistent though so I'll be using it for comparison for now. Thanks for the data! For current measurements I really hate using hall effect and other virtual measurement tools but they're fast and easy :)

Was this on the car or on a bench? If bench, what are you using to drive them? On car measurements are difficult so I was mostly checking dwell based on the current draw interval. But you're not kidding, the Audi draw some amperage. On our tests we were seeing similar amperage on the bench from the Audi, but I'll leave those results to you sir.
 
Last edited:
Feb 1, 2017
42
47
0
These results were on the vehicle.

How did you set up your bench testing? The square wave (dwell signal) without any primary voltage fluctuation seems odd to me unless it was used as a signal vs actual current being drawn through it. Being 0 to 14v(?) vs 14-0v on a actual car seems odd as well.

Glad we can share data :)
 

V8bait

Lieutenant
Nov 2, 2016
500
773
0
Texas
These results were on the vehicle.

How did you set up your bench testing? The square wave (dwell signal) without any primary voltage fluctuation seems odd to me unless it was used as a signal vs actual current being drawn through it. Being 0 to 14v(?) vs 14-0v on a actual car seems odd as well.

Glad we can share data :)

Bench setup is using a common TCI module with the function generator to emulate the cars system closely, so no voltage drop. Yeah on-car testing I was getting amperage like that if the capture window was wide, I saw upwards of like 7-30A on the OEM coils if the scope window caught too many events, so I zoomed in until I found a consistent window as I was more interested in dwell control. As for voltages you can flip the sensor over and get the other direction, doesn't really matter.

Big fan of data and sharing here, much as I'm allowed lol. The more scopes we get the better the data will be. Part of the fun for me is the pursuit of information!
 

R.G.

Lieutenant
Nov 17, 2016
668
326
0
Henderson, NV
Ride
E92 335, F10 M5
Look at the brain on this guy.

Glad Motiv is helping. I remember Chris BBQin more then a handful of plugs tinkering with this years ago
 

V8bait

Lieutenant
Nov 2, 2016
500
773
0
Texas
Have you tested any plasma coils or plasma ground yet? Coil ground tuning?

Are you referring to all the Okada things? Our ignition system is transistor based with a small high voltage capacitor in line on the supply side, it's redundant to add any of those devices as the voltage at the primary is already being increased. I haven't tested any of them because I have major doubts they'd do anything good enough to justify that price tag.
 

Erichale77

Lieutenant
Nov 14, 2016
827
447
50
Clearwater, FL
Ride
07 E92 coupe
I see. Well I have the plasma coils and ground installed on my car. Perhaps i can help with supplying you some data once motor is up and running again. I'll be the guinea pig for you if you let me know what you need and to get it.
 
  • Like
Reactions: V8bait

V8bait

Lieutenant
Nov 2, 2016
500
773
0
Texas
I see. Well I have the plasma coils and ground installed on my car. Perhaps i can help with supplying you some data once motor is up and running again. I'll be the guinea pig for you if you let me know what you need and to get it.

That would be great, i tried sending you a message but the system isn't letting me. PM me or shoot me an email we'll figure something out!