WestfieldMX5

Yes, but doesn't Jason's circuit above modify the output of the NB sensor to a NA signal ?

Joe Perez

iTrader: (8)

The circuit which Jason just posted converts the NB's cam pulses into an NA-style output. Problem is, I'm not convinced that this actually needs to be done. CMP is nothing but a reset line for the crank counter, it's not particularly time-sensitive. Look at the link which Jason posted concerning "funkiness". Forget about the rising-edge latency on the crank pulses- I'm referring to the fact that the single CMP pulse and both of the double CMP pulses all fall entirely within the same pair of adjoining CKP pulses. Thus, I think you could probably just feed the CMP pulses into the MS1 as-is.


The part that really matters is building a circuit that electronically eliminates the two TDC pulses from the crank wheel, leaving the other two pulses (which are, IIRC, either 60 or 70 degrees BTDC) alone. It is not necessary to lengthen them into an accurate emulation of the NA CAS, you simply need to eliminate the two extraneous pulses. One could, I suppose, use the four pulses in a set/reset configuration for a flip-flop, as the MS1 cares only about falling edges on the primay input. You'd just need to make sure that the device initialized itself properly with the CMP pulse before it started outputiing CKP pulses.

JasonC SBB

It's easier to convert the NB CKP pulses into NA style pulses than to eliminate 2 of them.

Joe Perez

iTrader: (8)

By which, I presume, you mean to use them to toggle a flip/flip, yes?

Either way you look at it, CKP is the important one. I believe you could feed NB CMP into the MS1 as-is and it will work. But CKP needs to be two equally-spaced pulses, with falling edges somewhere in the 60° BTDC neighborhood.

Matt Cramer

Well, you could move the falling edges or send it only one pulse per cam revolution instead of two - this can be changed with MegaTune settings. But the pulses as-is do not work with the code.

Joe Perez

iTrader: (8)

To clarify- the NB cam pulses (single-double-single-double) do not work with the MS1 code? Or the crank pulses?

Matt Cramer

Cam. The crank pulses will probably work.

JasonC SBB

You have to use the CKP RISING edge (NEVER falling) to trigger the divide by 2, and you reset it whenever the CMP signal is low.

JasonC SBB

So you guys decide which circuit you need and I'll post a schematic.

Zaphod

Well - for me a working one would do... (as I am totaly lost on all the specific electronic things)

Joe Perez

iTrader: (8)

I thought somebody tried this recently, and we determined that because the four pulses on the crank were not evenly spaced, the MS1 would not read them appropriately.

Why, precisely, will the stock NB CMP signals not work?

Matt Cramer

It seems they built a bit of "wiggle room" into the MS1/Extra code for unequally spaced teeth on the primary trigger. Really the only way to find out if it's enough room is to try it.

The cam signal though resets the tooth count each time it gets a pulse. Two closely spaced teeth will create an RPM spike.

Joe Perez

iTrader: (8)

That's fine. All i was saying is that the MS1 CPU needs to see a falling edge once every 70° (or so) BTDC and BBDC on pin 14, IRQ1. The rising edges can happen any time- they don't even have to be consistant.

Joe Perez

iTrader: (8)

I was all ready to say "It has been tried" and post a link, and then something hit me- the one person who has tried (mikeflys1) cut a single tooth off and tried to run it as 4-1 wheel. Whole different animal there as the MS is now timing every single tooth. In 4/2 mode, if my understanding of the code is right, the timer only counts between "significant" teeth, and the ones in between just increment the counter.


Are we sure about that?

I figure that if it's merely resetting a counter, then you could have twenty CMP pulses in a row, so long as they all fell in the same CKP-CKP gap. The thing that would cause a spike would be if a CMP pulse occurs both before and after a CKP pulse- logically causing three sequential #1 CKP pulses (the real one, the false one, and then the next real one.)

I need to get myself a spare MS so I can experiment with conditions such as this. You guys have any crippled ones lying around?

Matt Cramer

It would depend on the exact phasing, but I've been working on a dual trigger setup that had noise in the secondary trigger and each time it pulsed at the wrong point would give a huge spike in RPM. It had more teeth on its "crank" trigger, however. If it had a pulse on the crank input between the two pulses from the cam wheel, you'd definitely have it act crazy. If there wasn't any pulse, it's harder for me to say.

Joe Perez

iTrader: (8)

Yeah. I understand noise pulses in the wrong spots causing grief- that's the whole reason I abandoned CAS in favor of a 36-1 wheel on my car. But I honestly believe that if two CMP pulses occur in the same gap between CKP pulses, it will have no adverse effect.

Does anybody out there have an MS1 they'd be willing to loan me for a couple weeks so I can simulate this on the bench and come up with a definitive answer?

Edit: even just a naked MS1 CPU chip will do, if somebody has one lying around. (it'd be nice to also get the RS232 transceiver chip, so I don't have to place a digikey order)

andriusmx

i`m allso trying to ms my 99 euro 1,6 miata, and i think i have the same problem. I haven1t found a 90 - 97 CAS yet, but is it really impossible to use it on NB 1,6 ? so that would create a lot of problems...
It seems i allso will need a circuit for converting the NB CKP pulses into NA style pulses.

P.S. i have MS 1 3

arga

I tried something like this a few years ago to get MSII (no extra) to work. The circuit would turn the trigger wheel into a tach signal and another would distribute the single ignition signal to the two coils. I used this for about 10 months eventually running both spark and fuel but I was never really happy with it. I had the same missed pulse problems Abe and Pat had when they ran parallel. I eventually warned other to not follow.

https://www.miataturbo.net/forum/showthread.php?t=5407

https://www.miataturbo.net/forum/showthread.php?t=6730

https://www.miataturbo.net/forum/showthread.php?t=7375

JasonC SBB

I need to know exactly what Zaphod's MS1 needs. I can make a layout, mill a PCB, and build a circuit for Zaphod real quick. A friend is flying to Germany and he can then mail it to Zaphod. This discussion is taking longer than it takes me to design and build the circuit.

Zaphod, you are gonna owe me a tour of the best whorehouses in Prague!!! (You live near the Czech border, right?)

JasonC SBB

And noise pulses won't affect a 36-1 setup?

While I don't disagree a 36-1 wheel will be superior to a 2 or 4 tooth setup wrt timing accuracy, it seems to me the right noise filtering AND shielded cable setup would solve the problem just as well.