Getting a new trigger wheel added to the code

[jsmcortina]

If you have an engine setup that is not presently support by Megasquirt, you might not be out of luck.

One of the key design objectives we have with the MS2/Extra code is to support plenty of engines natively. i.e. original wheels and sensors.

Any of these "special" engines requires a custom wheel decoder to be written.

Before that process can start, the following need to be satisfied.
- compatible sensors (hall, opto or standard VR)
- accurate degree timing diagram of crank and cam wheels (see attached template)
- ability, time and commitment to test the code

So, if you pass the test, start a new topic with all those details and get me motivated to write the code!

James

[jsmcortina]

Here's an example timing chart - this one for Subaru 6/7. We don't need all this detail to get the code written though and this chart doesn't have the full 0-720 degress drawn out.

subaru_6_7_timing_chart.gif

James

[UnaClocker]

Loving the Chrysler 2.2/2.5 turbo wheel decoder you added. And there are at least 3 other guys that I've spoken with that are also running it, probably a handful more that I haven't spoken with, and tons more out there in the wings that'll run it someday.. Thank you so much for adding it.

[quark2501]

You are gold for the community.. ;-)

[FSG_PB_Patrick]

Guess I was stressing his nerves lately with my Nissan CAS mod. Sorry James

[OB6D]

Hmm, well if he's ready for a new challenge, i've put my request here: http://www.msextra.com/viewtopic.php?f=91&t=30423

[jsmcortina]

I've split off the Jeep and Nissan 25 wheel discussions into their own topics.

James

[molak]

For the future releases of msextra 3.0 it might be usefull to add the dizzy patern of "Magneti Marelli 1avp / 1avb". This ecu uses a single coil and a dizzy with a single hall efect sensor. The nice feature about it is that is capable of doing fully sequential injetion thanks to an un even gap patern.

The dizzy has 3 gaps corresponding to cylinder 2, 3 and 4. Those gaps begin at 72 degrees BTDC and end at 6 degrees BTDC . A fourth Bigger gap corresponds to cylinder number one, this one also begins at 72degree BTDC but ends later, at 0 degrees BTDC.

When the engine first starts the injectors fire simulstaniusly. A soon as the ecu detects cylinder number one it switch to fully sequential.

I best thing about this setup is that almost any hall or optic dizzy can be transformed into a fully sequential setup by making the gap corresponding to cyl 1 some degrees bigger..

Here´s a diagram of the dizzy in question..



APMS mean BTDC
PMS means TDC

[h22]

K here's another one. No fancy graphics or scope patterns yet, but they will follow later this week if time permits.

KTM 990 superduke 2cyl 75degree V twin, also super moto and RC8 sport bike, possibly quite a few others.......

24-2 tooth crank wheel, front cyl is at TDC at aprox tooth 19, 285degr. teeth after the skipped teeth. ( will have to verify this, going off memory...). I will have to check the rear cyl., if it is actually #1 then its TDC would be aprox 210degees or 14 teeth after skipped teeth ( i think ). there is no cam sensor, and I know that the current code probably has no way of supporting this motor, but it would be nice!

One oddball thing i did notice with this motor is that somehow it DOES NOT seem to fire in wasted spark! I have had it running on both microsquirt and MS2E-2.0.1 (fuel only and only firing one coil for the MS2E), and while both MS codes fire the coil(s) in waisted spark, the factory ECU somehow is able to determine the compression stroke ,possibly its an input to the CPU tied to a sensing resistor on the coil ground, could see a higher flyback voltage spike on the compression stroke compared to the exhaust stroke, or possibly it logs and stores the time differential between tach ( toothed wheel triggers ) during cranking or idling conditions to determine what stroke? I had it hooked to a 4 channel lab scope, and logging both stock coil signals and MS coil signals, the stock signals were 1 per cycle while the MS was 2/cycle (720degr). it ran the same regardless of witch ECU was actually controlling the spark, as it should, but I was a bit perplexed when i saw the factory ECU was only firing at 1/2 the rate of the MS.

I don't know for sure, but I do know that the stock ECU is able to do it. I bring this up as it may be a useful feature if we (by "we" i mean the smart guys that can actually write code and do math and stuff, not me) could implement something similar in the MS2 or MS2E code. this could potentially lead to no need for a cam sync, as well as ( if it is time/trigger activated ) to fun stuff like cyl contribution and misfire features.

With the current A-B-G ( alpha beta gamma.... i think i have that right) prediction already in the MS code, it may be as simple as looking at one particular error factor compared to the next,on a given tooth spread, then taking the greater of the two and setting up a pseudo "sync".
Ok, that got a bit off topic, but it is food for thought.
Joe

[Matt Cramer]

The crank wheel itself is fairly normal, besides it being an odd fire engine. The not running in wasted spark is a pretty neat trick, but I have a guess how it's done: Is there a MAP sensor installed in a single runner on this engine? A lot of motorcycle engines use this for cylinder identification by reading the pressure dip on the intake stroke.

[h22]

yea, that could be it, it has one MAP per cyl. defiantly could be how its done. has the MAP widowing been ported over to MS2E? that might be a neat but seldom used addition.

[FSG_PB_Patrick]

That would be quite complicated with the current concept. You would need to start in wasted spark mode, then look at idle where there occurs pressure drop in one cylinder (even without a second sensor) and then switch to sequential ignition.

BTW: You can run it wasted spark with the oddfire wheel code.

[dontz125]

What about a comparator circuit, joining the dedicated MAP to the 2nd trigger? Treat it as a dual wheel, but instead of a cam signal, 'sudden drop in MAP in this tract' = next TDC is #1 power. No new code required, just an adaptor / filter circuit?

[racingmini_mtl]

The problem is that the "just" part might not be that simple. Since the MAP signal varies a lot, this needs a dynamic threshold.

Jean

[dontz125]

Heh - my post was 'just' too simple, obviously, missing thoughts and idea. I thought about it some more, and I had meant to note that there would be 2 MAP sensors. For a 1243 arrangment, (just to pull an example out of the air), the main MAP signal 'M' could be on cyl #3; with the trigger sensor 'T' on #2, so the next firing cyl would be #1. If a filtered T value suddenly drops below a filtered M value by a threshold amount amount, trigger.

I know OpAmps can be used for straight Boolean operation (if V(T) < V(M) then ON); can they be used for offset comparisons - (if V(T)+0.5 < V(M) then ON)? Would it be a simple offset, or - as you said - a dynamic offset? How can that offset be provided? Could something like an LM2907 F-to-V converter (Vout = Vcc x Fin x R1 x C1) be used to provide that dynamic offset?

Could it be done with an analog circuit, or are we talking a significant code mod?


1p2c4i3x 1x2p43i 1i2x4p3c 1c2i4x3p

[MS2tester]

would it be an idea to have a user defineble wheel in the code
maybe something like 0 to 360 drg with a setting high or low per drg
in this way new wheels would not have to be programmed just user defined

[jsmcortina]

would it be an idea to have a user defineble wheel in the code
maybe something like 0 to 360 drg with a setting high or low per drg
in this way new wheels would not have to be programmed just user defined

Defining the wheel like that is ok and the code presently does it like that internally. However, how do you synchronise to the wheel? That is the difficult part.

James

[MS2tester]

i assume you mean where to start reading for top dead cilinder 1
assuming a missing tooth wheel a gap would be present somewhere in the 360 sequence
or even a single drg tooth as a reference then start counting in teeth or crank dgr
could you explain how the code now reads
a fixed amount of dgr after reference point being either a missing tooth or a cam signal
or is it counting teeth

[FSG_PB_Patrick]

i assume you mean where to start reading for top dead cilinder 1
assuming a missing tooth wheel a gap would be present somewhere in the 360 sequence
or even a single drg tooth as a reference then start counting in teeth or crank dgr
could you explain how the code now reads
a fixed amount of dgr after reference point being either a missing tooth or a cam signal
or is it counting teeth

The code is counting flanks or signal edges, be it either rising or falling edge. Every edge is defined a tooth and every tooth knows how much degree it is until the next tooth shows up. Now's the trick, every wheel definition must present it's own sync algorithm while cranking and a control loop if it's still synced. And that's non-trivial, I did it myself with the Nissan.

[BaldTurboFreak]

The file for the trigger format is no longer availible. I have a trigger I would like to see supported.
It's identical to this trigger : http://www.msextra.com/doc/ms2extra/MS2 ... -2-2-2.htm
aside from the fact that the tooth count is changed to reflect the H6's needs.
Instead of 11 and 14 teeth in between the larger slots we have 8 and 16 as per this photo here:
http://www.romraider.com/forum/viewtopi ... =25&t=5746.

Would that be a difficut code swap? I have paypal beer money in hand!