fluke

Hi,
I'm almost done building my MS-I V3, I would like to use the stock crank and cam angle sensors but I'm a bit lost.

Should I assemble the MS as per diyautotune instructions for the DSM install? :
http://www.diyautotune.com/tech_arti...alon_laser.htm

kingofl337

Yes wire it up the same as you would for a 90-97.

fluke

Ok, I've seen the signal plots for the '99 crank and cam sensors:
http://pages.sbcglobal.net/jcuadra/c...ing/timing.gif

how does this compare to the signal plots for the <97 setup? I can't seem to find a plot of these signals.

kingofl337

might end up both being a 5v pull up based on that.

They work the same thought a "high" voltage dragged to ground
when the "window" passes by.

Reverant

iTrader: (10)

Here is what I got when I connected my oscilloscope to a 90-93, optical-type CAS:

fluke

Excuse my stupidity but what is the 5v pull up for? Does this just add 5 volts to the signal?

Doesn't MS expect to trigger on the high voltage? These signals from the '99 look like they're triggering on the ground state (or falling edge).

What about the reset signal from the cam sensor? On the <'97 it's a long single high voltage pulse while on the '99 it's a pair of ground state short pulses, will MS be able to decode that?

How hard would it be to build a simple circuit that inverts these signals so MS is receiving high voltage triggers? Maybe I can set MS to trigger on the falling edge instead of the leading edge and avoid inverting the signal.

Reverant, where is ground on that signal plot? What is the value of the high voltage? Where is the #1 TDC?

Thanks for your help everyone!

Reverant

iTrader: (10)

As the above plot would suggest, the cam signal on a <='97 is TWO pulses, one short an one long. See the trigger wheel here:



The plot is taken with +12V on pin 2, GND on pin 1, CKP is the pin 3 and CMP is pin 4 on the sensor. The output signal voltage depends on the input signal voltage but IIRC it's about 10V. The plot was taken with the CAS out of the engine so I don't know which peak is #1 TDC fire.

fluke

Yes, but on the '99 the CMP signal is 2 short pulses instead of the one long pulse; that's what bothers me.

Reverant

iTrader: (10)

If you look at the wheel (the inner track), you'll see that it's also 2 pulses, one short and one long - not just one.

fluke

Right, but if you look at the '99 plot it's 1 short @ 0 deg. then 2 short @180 as opposed to the <'97 that's 1 short at 0 deg. then one long @180.

I just don't know enough about megasquirt at this point to know whether it will be able to decypher the signal, if not I guess I'll have to come up w/ a conditioning circuit.

Has anyone gotten MS to decode a signal like this one?

Reverant

iTrader: (10)

I don't think so, which is why I am fitting the CAS from a NA to my NB.

Jim

Reverant

iTrader: (10)

Just to be sure, I scoped my own engine:



Top channel is CMP, bottom channel is CKP, CMP is 20mS/div,5V/div and CKP is 1mS/div,5V/div. Any less time resolution on the CKP and the pulses wouldn't show up - I wonder why Mazda recommends 10mS/div in the service manual!

fluke

Reverant,
the CMP signal coincides w/ the plot I posted, but your CKP looks different; it looks like an inversion; strange. That makes things more complicated...

You said that you are using the CAS from a NA; is this mounted to the space in the back of the cam cover where the stock CAS in the early NBs went? I've been reading about ppl using the CAS from the <'97 NB, which is "better", the NA or the NB CAS?

I was concerned with losing some accuracy w/ this method because:
1- Belt stretch
2- Play between the CAS drive shaft and the cam (does the "key" fit tight in that slot on the cam?)

Am I getting too carried away w/ precision? Did the NA use a crank sensor for timing in addition to the cam sensor?

Reverant

iTrader: (10)

Yes, I was alienated by the inverted signal as well.

Exactly, on the exhaust side cam to be exact.

There are two sensor variants:
1) The optical type, which was fitted on the intake side cam of the 1.6 90-93 NA Miatas
2) The magnetic (Hall Effect) type, which was fitted on the exhaust side cam of the 1.8 94-97 NA Miatas.

The magnetic is supposed to be a bit more accurate, but I wouldn't bet on it. However, you should know that the magnetic type fits on NB (98+) cars like a glove, while the optical will fit but it will not bolt using two studs, only one will fit (different housing).

1) Belt stretch is insignificant, keep in mind that Mazda has the cam sensor on the cam on NB cars as well!
2) The key fits perfectly in the slot, no play whatsoever.

Yes, you are getting carried away as far as precision is concerned. The NA did not use a "crank" sensor, as this was integrated in the CAS: Both the CKP and CMP sensors were fitted in one single housing.

fluke

Reverant:
that's awesome, thanks so much for the info; you sold me on the cam mounted sensor.

So before '99 all timing was off of the cam? Interesting.

fluke

Sorry if this is a repeat question, but let me make sure that I'm getting this all straight...

MS sees the short and long pulses from the cmp and distinguishes between the two?

What are your settings on msns-extra? The onese that setup the cmp and ckp that is.

Thanks again for the help, i'm pretty confused.

Ok, I just found this on another post:
https://www.miataturbo.net/forums/at...6&d=1153057055

But how does MS know which spark event goes with which cmp signal (short or long) I guess I was looking for a setting in msnse like "long pulse = fire cylinders 1 and 4, short pulse =fire cylinders 2 and 3"

Reverant

iTrader: (10)

I haven't digged through the source code of the MSnSe firmware to understand exactly how it interprets the settings in the wheel decoder config page, so I'm not really sure I can help. So far all of the NA guys seem to be OK with these settings so for now I trust them without questioning why.

fluke

lol, cool.

kingofl337

The length of the signal is not used by the megasquirt. The megasquirt uses the 0 and 180 signals as resets for the crank.
It fires like this

Fire Coil A
Fire Coil B
Reset
Fire Coil A
Fire Coil B
Reset

fluke

Ok, I understand how this would work if there was only 1 signal per 2 crank rotations since you could tell MS: "OK, when you receive the cmp signal reset because that's cylinders 1,4 TDC." As I understand it, MS has to know not only when the cylinders reach TDC but whether it's cylinders 1,4 or 2,3 so that it doesn't spark at BDC.

I don't see how MS could do this with 2 rising voltage signals per crank rotation as is the case in the '99+ sensor, unless you set it to distinguish between the single and double pulses (but how?) Since the <'97 has 2 signals per 2 crank rotations (1 cam rotation) each signal will reset for cylinders 1,4 or 2,3 (I don't know which) w/ wasted spark.

As a matter of fact, I don't see how you could distinguish between TDC w/ a compressed charge and TDC scavanged even w/ a crank mounted missing tooth setup. Guess this is why we use wasted spark. It would be possible to get away from wasted spark w/ a single signal from the cmp though, but why bother.

Basically, unless you can distinguish between the double pulse and the single pulse I don't think you can use the '99+ cmp.

I would be interested in inverting the '99+ ckp signal and using that since there are twice as many signals per rotation and therefore double the precision. But, since it looks like I have to use the <'97 CAS for the cmp signal I might as well make it simple and use the ckp as well...