Megasquirt Ignition Options (falling edge, stock distributor, etc)

[Ondonti]

After looking at some files and also datalogs, I am finding that Joe, just like Ed, has high timing error.
Both are using Rising edge ignition input capture and using some trigger angle offset like 75 degrees. Nathan Wilbur was using 64 degrees, "rising edge" but he posted no datalogs.

Joe had -12.7% to 12.7% error during an actual pull and his error is less during cruising (about half). The error constantly goes up and down extreme to extreme almost like each ignition event is 12-24% different from the one before it as far as actual timing....

My timing error is about 3% to -2.9% at most during pulls. My error also spends most of its time at less then those peak errors and during shift or decell is slightly more erratic (not really important).

I was using falling edge and much lower timing error but my car is unhappy after sitting so I have been looking at others files including Joe's very recent stuff to make sure I didn't mess more things up in my tune when it got corrupted.

I don't understand the reasoning behind rising edge use or how people are deciding to set it up. I just set falling edge and did nothing else per DIY forever ago. Never thought about it.

I did try checking my timing with a light just using the starter to bump it over. it bounced around a bit but I got it down to around 12 degrees with manual adjustment. I did not disconnect any sensors. I don't know what that represents with MSII Extra when the engine is not running!!!!

[Sundance 6g72]

megasquirt might have a cranking timing setting. best to set it with the motor running

i need to fix my cam seal so i can investigate this.

[Shadow24]

What code revision are you using or basing your theory on? There has been a lot of work done on the MSII-Extra code in the last couple years. I am updating my MS when it gets re-installed (hopefully soon) as I am over a FULL revision out of date (2.x.x to 3.2.5 or latest).

[Ondonti]

What code revision are you using or basing your theory on? There has been a lot of work done on the MSII-Extra code in the last couple years. I am updating my MS when it gets re-installed (hopefully soon) as I am over a FULL revision out of date (2.x.x to 3.2.5 or latest).

3.2.1

I am just looking at datalogs and the timing error is something we have talked about in the past. I am more concerned with rising edge vs falling edge (per DIY) but I noticed this while investigating differences in peoples setups.

[Shadow24]

hmmm, it looks like the DIYautotune article defaults to the 6-"tooth" input. Has anyone tried using the MSII-extra wheel decoder and the high data rate 360-10 input?

Was Ed's issue on the 12v, not his new 24v 3.5?

[Ondonti]

hmmm, it looks like the DIYautotune article defaults to the 6-"tooth" input. Has anyone tried using the MSII-extra wheel decoder and the high data rate 360-10 input?

Was Ed's issue on the 12v, not his new 24v 3.5?

Not sure if his problem went away or if this was a distributor problem or if he had this problem while also trying a trigger wheel on the crank.
I don't know much about the decoder but that does sound cool.

[Shadow24]

It probably is worth some investigation as my thought is that the 6 tooth isn't as accurate due to its lack of resolution. If your able to run a 360-10 wheel and still run the distributor, I think that would increase the accuracy of the ignition timing where rising or falling edge means +-1* variance at most (guestimate).

[Sundance 6g72]

i dont even know what this 'problem' is or how it woll effect me

[Shadow24]

it is an ignition accuracy issue that we are referring to. Ondonti is bringing up the fact that multiple installations using the 6-"tooth" low data rate input from the 12v distributor tends to have a fair variance in asked for ignition timing vs. actual observed timing. The initial thought was that the capture mode (rising edge/front of the tooth vs. falling edge/back of the tooth) could be one possible issue.

Possibly switching to the high data rate input via the wheel decoder in the MSII-extra firmware may provide better timing accuracy all around as it goes from 6 "teeth" to 350 "teeth" (360 with a gap of 10 teeth, similar to a 36-1 EDIS trigger wheel).

If you are running OK as is, it probably isn't that big of an issue, but more of an area that can be improved upon.

[c2xejk]

I had the "problem" when position was with the original 3L 12v distributor. It was still there when I switched to the 36-1 wheel. And I just checked, it is there with my current 3.5L 6g74 distributor setup.

So far I have yet to see reason to believe that is any kind of "real" issue. I don't know what timing err % is or how it is calculated...

[shackwrrr]

Howabout modding the distributor and making a 72-2 wheel for it (36-1 equivalent at crank speed) Then using a VR or Hall sensor. No need to keep the distributor function since wasted spark would be a simple addition. I have no 3.0 experience other than seeing them in the junkyard so dont hate on my idea too much.


Also, since the Dizzy drive is off a cam, a sloppy timing belt could cause some error. Just like a neon with a bad hydro tensioner setting a cam sensor performance code.

[Ondonti]

I had the "problem" when position was with the original 3L 12v distributor. It was still there when I switched to the 36-1 wheel. And I just checked, it is there with my current 3.5L 6g74 distributor setup.

So far I have yet to see reason to believe that is any kind of "real" issue. I don't know what timing err % is or how it is calculated...

So are you running rising edge with that as well? I feel there must be something going on there for our values to be so different and its possible that no matter how good your wheel is, that MS is not up to the task of using that information properly...
I think 6% variance in my datalogs is a lot better then 24% variance though I agree that the meaning behind that statistic is unknown to me beyond what I assume (what Shadow 24 says, timing demanded vs timing observed).
Joes timing is up 12+% up and down seemingly each cycle. That is equivilent to seeing 22.5 degrees and then 17.5 degrees of timing when you were demanding 20 degrees. My timing error is equal to 20.6 up and 19.4 degrees down when demanding 20 degrees. Even that amount of error would be a problem if I was pushing the limits. I have friends who talk about huge gains from less then 1 degree of timing (getting timing "right" vs close).

I know there are more accurate ways to measure crankshaft position but it would help if someone (obviously not me) was more familiar with the limitations of megasquirt to use those measurements. It is possible that our low data rate used properly functions better then other setups in real life.

BTW I am doing some reading and others seem to be capping off at 12.7% but I need to read more. They are reporting problems associated with error and ways they think they can solve it. Some using 36-1 wheels. Old thread I was looking at (2010).

[Shadow24]

My personal experience with a 36-1 and EDIS was a pretty stable and accurate ignition timing... Which thread are you looking at Ondonti?

Also, are you and Ed running the original B&G MSII code or are you both running the MSII-Extra? Everything I have read supports the MSII hardware being more than sufficient for accurate timing, provided you have a good clean input. For instance, the base code runs fast enough for .1* accuracy in the ignition table, so it would seem that there is something external of the MS that probably is causing the timing error...

[c2xejk]

MS2-Extra. I should note that most of the time the value is much less, there are just a few short times were the error is larger. So far I am still not sure that my timing is ever off by any significant amount...

When I last looked there was talk that Timing Err % really didn't mean much the way it was calculated. Some talk of changing it but I have not heard anything about a change.

[Sundance 6g72]

What if you looked that the timing advance in a datalog and then looked to see what it says in the timing map at that exact moment. IE. You are doing a pull and are at 200kpa x 4000rpm and the log says you are running 25*... if your timing table is demanding 25*, then what is the problem?


Im curious in the difference between rising edge and falling edge. Every once and a while i get this slight 'ignition miss' or something that I always thought was a lean condition but im not sure.

[c2xejk]

Does MS show the "actual advance" or the "desired advance"?

[Ondonti]

Does MS show the "actual advance" or the "desired advance"?

Another bit of information that would help in understanding datalogs.

Joe's error is extreme and constant.

BTW One person was talking about the clearances on their VR setup and how that was creating error but that trying to improve error caused other problems.

This is just what I was reading last night.
http://www.msextra.com/forums/viewto...?f=101&t=36666

[Sundance 6g72]

"joe's is extreme and constant" Thats fantastic :/ I should tell all the people I win against that My ecu has no idea how much timing it's really running and that its a mini van motor with an ebay turbo VROOOOOOOOOOOOOOOM

[Shadow24]

Does MS show the "actual advance" or the "desired advance"?

I believe that the MS logs the "requested" advance, i.e. whatever advance is needed for the given conditions after adding or subtracting for temperature, RPM, MAP etc (see the timing equation in the megamanual for more specifics). If that value differs from the "observed" advance, as recorded visually with a timing light, then there may be an issue. Again, if there is a confirmed error %, it may be due to the low data rate input from the 6-"tooth" part of the distributor.

Also, that thread on msextra.com is from '10, and per Ken, the % off value was borderline useless due to no smoothing in the code that derived the % off value. With 2+ years additional work on the MSII-Extra code since that thread/post, I would suspect that if you did a fixed timing and checked it with a light, you wouldn't get much variation.

[Sundance 6g72]

when I set the timing, its spot on.