Technical Ignition advance on E85

rac

Sergeant
Nov 14, 2016
341
204
0
Australia
Ride
135i ST
I understood from comments elsewhere that our heads are actually rather inefficient, if we are viewing efficiency by CFM?

Filippo

optimum ignition timing is driven by the time it takes for combustion to take place in relation to piston position.
you can have a high flow head that results in slow combustion and therefore needs more timing and vice versa. actually this is why poorly ported 4 valve heads can often flow higher bench numbers but make less power/become less efficient, when things like swirl and tumble that reduce combustion time have become less effective.
 

V8bait

Lieutenant
Nov 2, 2016
500
773
0
Texas
Someone get a timing light and verify, problem solved. Either bmw has an offset or Syvecs missed one, back in the day retrofitting an edis you had to mount the CPS sensor right to get the timing right, bmw sensors don't pick up like ford and gm etc etc.

Regardless, MBT is just a number. My old Chevy runs way different advance than a C7 Corvette with DI. Ultimately you run the lowest curve that gets the power you want.
 
  • Agree
Reactions: langsbr

bahn

Sergeant
Platinum Vendor
Nov 5, 2016
250
414
0
Iowa
Someone get a timing light and verify, problem solved. Either bmw has an offset or Syvecs missed one, back in the day retrofitting an edis you had to mount the CPS sensor right to get the timing right, bmw sensors don't pick up like ford and gm etc etc.

Regardless, MBT is just a number. My old Chevy runs way different advance than a C7 Corvette with DI. Ultimately you run the lowest curve that gets the power you want.

Agreed, I'll put a timing light on it tonight to verify ADE's findings (I'll probably have to mark the crank pulley and block from the underside as there's not a lot of room from the top with stock outlets). For fun I found some good information of the CPS reluctor ring as well. The DME triggers on the falling edge of the CPS square wave, the falling edge of the next tooth after the reference teeth (2 missing teeth on the reluctor ring) is tooth #1, the falling edge of tooth #10 should be TDC of cylinder 1. I'll see if I can confirm that tonight just for tits and pickles.
 
Last edited:
  • Like
  • Funny
Reactions: MattGu40 and V8bait

bahn

Sergeant
Platinum Vendor
Nov 5, 2016
250
414
0
Iowa
EDIT: The findings in this post have been proven inaccurate in a further test I performed, please see the latest test here: https://bmw.spoolstreet.com/threads/ignition-advance-on-e85.3729/page-4#post-52831

Alright, some good news and bad news. The bad news: Must have bumped the timing light around too much when I last used it an broke the bulb. The good news, didn't matter anyways as I was able to get a probably more accurate measurement with the oscilloscope. The CPS is pretty hard to get to so I used a spare harness I have to identify what pins on the DME harness are for the CPS. Pin 29 (Yellow) on the left DME connector is the CPS signal wire, I back probed this and ignition coil #1.

As I said before the reluctor has 60 teeth (2 are missing, this is the reference tooth). The missing reference tooth is 60 degrees out from TDC of cyl #1 (NC-OFS_TDC0_REF_CRK) and with some math we can calculate out that each falling edge of the CPS square wave is 6 degrees of crank rotation. With that we find tooth #10 to be TDC for cyl #1. For this test I had my ignition timing at 0 degrees as seen by the reported MHD timing for cyl #1 below, if this was the actual ignition timing we should see the spark discharge right at the falling edge of tooth #10 since the DME knows the coil dwell time. Instead we see it just past tooth #11 (actually around 11.25), with some more basic math we find our actual timing at -7.5 degrees from cyl #1 TDC.

N54_Timing_Advance.png


Screenshot_20180802_183636.jpg
 
Last edited:

rac

Sergeant
Nov 14, 2016
341
204
0
Australia
Ride
135i ST
Alright, some good news and bad news. The bad news: Must have bumped the timing light around too much when I last used it an broke the bulb. The good news, didn't matter anyways as I was able to get a probably more accurate measurement with the oscilloscope. The CPS is pretty hard to get to so I used a spare harness I have to identify what pins on the DME harness are for the CPS. Pin 29 (Yellow) on the left DME connector is the CPS signal wire, I back probed this and ignition coil #1.

As I said before the reluctor has 60 teeth (2 are missing, this is the reference tooth). The missing reference tooth is 60 degrees out from TDC of cyl #1 (NC-OFS_TDC0_REF_CRK) and with some math we can calculate out that each falling edge of the CPS square wave is 6 degrees of crank rotation. With that we find tooth #10 to be TDC for cyl #1. For this test I had my ignition timing at 0 degrees as seen by the reported MHD timing for cyl #1 below, if this was the actual ignition timing we should see the spark discharge right at the falling edge of tooth #10 since the DME knows the coil dwell time. Instead we see it just past tooth #11 (actually around 11.25), with some more basic math we find our actual timing at -7.5 degrees from cyl #1 TDC.

View attachment 14190

View attachment 14191

World order restored.... Great post.
 

Jake@MHD

Major
Platinum Vendor
Nov 7, 2016
1,593
2,058
0
Philly
@bahn are you able to check some other loads / rpm, preferably outside of idle since that is just DME torque reserve strategy? Possibly set your timing map to all 30* or something 0 - 100 load, 0 - 3500rpm (so request from table is constant) and lightly press the throttle to get it there and outside of idle control.
 

V8bait

Lieutenant
Nov 2, 2016
500
773
0
Texas
Yeah idle control uses timing to keep it smooth, much faster than trying to adjust an iac or throttle. But that should disable over a certain rpm, which would be easier to capture on a scope than moving rpm. Cool to see some testing though nice work. I have a few scope shots of the cps and vanos etc from a while back but never thought to check advance with it, clever.
 

bahn

Sergeant
Platinum Vendor
Nov 5, 2016
250
414
0
Iowa
@jyamona I can try to grab that tomorrow but it's going to be a little difficult, hopefully my leads can reach with the scope in the passenger seat. As for under load, without having the car on a Dyno that would be very difficult. I'll do some steady state tests at different rpms though but I expect about same results. Thanks for the feedback guys.
 
  • Like
Reactions: Jake@MHD

NoQuarter

Major
Nov 24, 2017
1,662
1,066
0
Indiana, USA
Ride
Z4 35is, 535xi, X5 35i
Don't know how much it matters, but I think what you are calling a tooth is actually a space.

The tooth sends the signal to ground so the falling edge is the beginning of a tooth and the rising edge is the end.
 

bahn

Sergeant
Platinum Vendor
Nov 5, 2016
250
414
0
Iowa
Don't know how much it matters, but I think what you are calling a tooth is actually a space.

The tooth sends the signal to ground so the falling edge is the beginning of a tooth and the rising edge is the end.
It is a space but it's referred to as a reference tooth.
 

Jake@MHD

Major
Platinum Vendor
Nov 7, 2016
1,593
2,058
0
Philly
@jyamona I can try to grab that tomorrow but it's going to be a little difficult, hopefully my leads can reach with the scope in the passenger seat. As for under load, without having the car on a Dyno that would be very difficult. I'll do some steady state tests at different rpms though but I expect about same results. Thanks for the feedback guys.

No problem, thanks man! I'm assuming that the instant you scoped, timing was actually perhaps 0.5* or similar, giving an 8* offset (some whole number). It would be easier to see outside the controlled idle where you can set ign table to a constant advance :) Also would just like to confirm it is indeed a constant offset for peace of mind, although it most likely is.
 

bahn

Sergeant
Platinum Vendor
Nov 5, 2016
250
414
0
Iowa
No problem, thanks man! I'm assuming that the instant you scoped, timing was actually perhaps 0.5* or similar, giving an 8* offset (some whole number). It would be easier to see outside the controlled idle where you can set ign table to a constant advance :) Also would just like to confirm it is indeed a constant offset for peace of mind, although it most likely is.

Multispark got in the way of today's test, couldn't get consistent results. I'll try again tomorrow with multispark disabled.
 

bahn

Sergeant
Platinum Vendor
Nov 5, 2016
250
414
0
Iowa
@jyamona - Took me longer than expected to respond due to some inconsistencies I was seeing that I wanted to test further. I'll have to start off by eating some crow as my previous test at idle was inaccurate due to how heavily the DME uses ignition timing to stabilize idle speed. I let the car idle for a while and watched my scope in rolling mode and the inconsistencies were plain to see. The previous finding is also made moot by the MSD81 documentation suggesting the CPS gets pulled to reference voltage (5V) when a tooth is present in front of the hall effect sensor however that doesn't appear to be true. Instead it is pulled to ground when a tooth is present which is common for automotive gear speed sensors. I noticed this after capturing the waveform so I had to invert the signal later causing the blue channel 2 trace (IGN #1) to be inverted so please ignore that.

I started off by wanting to verify that reluctor tooth #10 is indeed TDC for cylinder #1 so I performed the following test:

* Scope in roll mode
* Coils unplugged & #1 spark plug removed
* Rod in cylinder #1 spark plug sitting on top of the piston
* Key on to provide reference voltage to the CPS
* Turned engine over by hand until I found the missing reference teeth and counted teeth from the reference until cylinder #1 stopped rising.


This test confirmed tooth #10 is TDC for cylinder #1.

Next I hooked the scope back up and flashed the car with a tune that locked timing at 1000-3000RPM @ 6.0 degrees and 3000-4000RPM @ 12.0 degrees. I also disabled multispark although I saw a very long almost double discharge at 4000RPM but it didn't cause a problem with the results.

First test @ 1500 RPM:
CPS_1500.png


Timing was reporting a constant 6.0 degrees, scope captures were very consistent. With the capture inverted correctly you can see the discharge was on the falling edge of tooth #9. If we follow along with the math again then we know 1 tooth = 6 degrees of rotation so our spark is indeed being delivered at 6 degrees before TDC.

Next test @ 4000 RPM:
CPS_4000.png


Timing was reporting a constant 12.0 degrees, scope captures were also very consistent. With the capture inverted correctly you can see the discharge was on the falling edge of tooth #8, two teeth before TDC or 12.0 degrees.

From my testing, the MSD8x is delivering the exact ignition timing it is reporting. I now believe Syvec is either triggering on the incorrect tooth or they are looking for reference voltage in the presence of a gear tooth instead of a ground which would put their delivered timing about 3 degrees more retarded.

I'm out of time tonight to edit my original post tonight but I will update it tomorrow morning for consistency.
 

veer90

Lieutenant
Nov 16, 2016
1,000
774
0
West Nyack, NY
Ride
e90 335i 6MT
Now that we've verified MSD8x has correct reported vs commanded ignition timing, can anyone comment on the tweaks I made to the standard BMS BEF timing table? Just initial impressions during normal driving the engine is noticeably smoother, however there are still improvements to be made.

I only smoothed it and added 1 degree at WOT (logging 6* at peak torque and 9* at redline), didn't make any major changes or timing increases in the low load cells.
 

Attachments

  • timing 3d map bms bef vs smoothed.PNG
    timing 3d map bms bef vs smoothed.PNG
    559.2 KB · Views: 202