Zaphod

Hi guys,

I just finished my MS1 3.57 parallel install in my 99 (1,6 european).

I tried to start the car but it won't fire up. I get no RPM reading in the Megatune, CLT, AFR, MAP, MAT are fine. How can I make sure my 90.97 CAS is working and sending a signal.

What else could be the problem?

Here are a short log (where I can see nothing at all) and my actual MSQ.

MS firmware was flashed at the stim, MSQ as well. So I don't think I fried my coils...

Greets

Joe Perez

iTrader: (8)

First, can we assume that MegaTune showed RPM when you ran on the JimStim?

Here is what the wiring at the CAS should look like. This is from the perspective of you, looking at the end of the connector into which the wires go:


W/R needs +12V, and B/LG needs ground. W is CKP and should connect to pin 24 of the DB37, and Y/L is CMP and should go to pin 25.

To properly verify the operation of the CKP, in my opinion, requires a two channel oscilloscope.

Zaphod

Nope!

paul

iTrader: (21)

thats why you shoulda installed the MS to monitor first to make sure everything was working. leave fuel and spark to the stock ecu till you confirm it is.

Joe Perez

iTrader: (8)

**** that, but at least bench-test the thing. That's what the stim is for. It's much easier to troubleshoot on the workbench than inside the car.

Zaphod

Well I read a lot that the RPM wouldn't work with the stim.

Was I wrong?

No problem to take the MS out again and test it on the bench...

P.S. Yes I was thinking about let MS monitor first - but I was also thinking... **** that!

Maybe I was wrong.

Edit: CAS is wired correctly.
If I can assume that "port b" in MLV is showing one of the cas inputs, then there is a little activity.

Greets

Joe Perez

iTrader: (8)

The old-style stim can only produce one trigger output, so the MS will not run in that mode once it has been configured for a Miata (which has two triggers.)

The JimStim, however, has two trigger outputs and can properly simulate a Miata, thus, the MS will run fine on it.

Not quite.

porta, portb, portc, and portd are hardware registers inside the CPU. Each one of them is an 8 bit number, where each bit represents the state of a single I/O pin on the CPU. If you look at page 1 of the schematic, you will see that most of the general-purpose I/O pins have a symbol such as "PTC0" or "PTD3". The first means that this pin (pin 7, which drives LED D14) is represented by Bit 0 (the least signifigant bit) of portc, and the second means that this pin (18, which is JS11) is represented by bit 3 (the fourth bit) of portd.

These are of relatively little use to the average user, being intended for developers to have a window of sorts, to see what is going on inside the chip at any given time.

Joe Perez

iTrader: (8)

On further reflection, I should be clearer on how to read the ports.

First, if you look at page 1 of the MegaSquirt schematic, you'll see that portB covers pins 23-30, which is all of the analog inputs, plus JS4 and JS5. So we'll ignore that one for the moment.

Instead, let's look at portC. Assuming that you followed the instructions at DIYAutoTune's writeup on how to modify the board (you did do the internal mods, right?) then you've got pin 25 of the DB37 (pad IAC1A) connected (through a 1k resistor) to pad JS8 which goes to pin 11 of the CPU. That pin is assigned the designation PTC4, meaning that its state is represented in bit #4 (the fifth bit) of the portC register.

So pretend we're looking at the portC value in MLV. Let's say that it has a value of 17. Convert that to binary: 00010001. I don't know how much you know about binary, but in this form it's read from right to left, and the bits in this case are numbered from 7 to 0. So a decimal value of 17 means that bits 4 and 0 are true, which means that PTC0 and PTC4 are both high. PTC0 is the output which drives the "Squirt-1" LED, so we know that LED is on. PTC4 is our old friend JS8, which, assuming you did everything right, is your CMP signal- so that signal is currently true.

Now, imagine that the very next reading of portC is 3. Decoded to binary, that is 00000011. So now, PTC4 is off, and PTC0 and PTC1 are both true.

Etc...

In your log, portC is stuck at 3 the whole time. This indicates that PTC0 and PTC1 are on,- those are D14 and D16, which are used to drive SparkA and SparkB. This suggests to me that you are getting no spark output. Also, PTC4 never toggles, so we are likely not seeing your CMP signal coming in. And that's very likely your problem. No incoming CMP signal.

Check, and then re-check, the internal modification to the MS PCA between the IAC1A pad and the JS8 pad.

Zaphod

Hi Joe,

I think I understood most of the things you said - even if the binaries thing is quite a while away - sometime in my schooltime, like 20 years ago...

I bought the board pre-assembled at diyautotune with all the mods already done:

MegaSquirt-1 PCB v3.57 (SMT) Engine
Management System

MOD_13571... Modifications for Miata / DSM ignition
input and output on a V3.57 board. The
CKP signal comes in on pin 24 and the
CMP signal comes in on pin 25. Spark
output A comes out on pin 36, and spark
output B comes out on pin 31.

MOD_1357-... Add jumper for KnockSenseMS to
Megasquirt-I V3.57. Run a jumper from
JS10 to DB15 pin 3.

MOD_1357-... Launch control input mod on pin 29

MOD_1357-... Add boost control mod-kit to V3.57 board
and bring output out on pin 27 on the
DB37.

JimStim-C JimStim v1.4 Assembled Unit
'JimStim' Stimulator with Wheel Simulator
version 1.4

SM-WireBund 23" Wire Bundle -- Same wire used in
harnesses, color coded and printed
labeling. Great for building a harness or
wiring up connector to the factory
connector.

That is what I bought.

I will recheck my wiring first. Maybe there is a disturbance in the CMP line.

How would I set the Jim stim up to acknowledge my RPM signal?

Greets

Zaphod

O.K. it did all the test procedures to the jim stim 1.4 - the tach signals seems to work. Still - I don't get a rpm reading with the stim.

Jim stim set up with switches 2+3 at on, 2 tach signal to pin 25 (CMP), no pull-up jumpers.

Thanks for your help.

Edit!
I just jumpered my jimstim to hall sensor (was set to VR mode) and I get rpm reading!

Greets

Zaphod

Well - I think I solved the problem of the missing CMP signal - the 90-97 CAS can be installed in the 99 1,6 NB head, BUT there seems to be no "holder" for the CAS "rotator/rotor(?)" at the cam. I felt with my finger and also did a pic - I think there is simple the extension for the CAS missing.

So I have to check if the NA (90-93) intake cam works on the 99 1,6, or I have to use the sensor circuit Jason mentioned.

How is this circuit working? (Are there a lot of things to change at the MS?)

Greetings from the German Problemfall

Sven

AbeFM

iTrader: (3)

I don't know which circuit Jason was talking about - but I do have a sense that any input mods you'd want to make of that level would have to go either outside of the MS box, or inside of the case, but on a protoboard. Assuming it's my copy of the OEM circuit, you'd have to get a lot of resistors in there.

Of course, I'm a fan of a stand alone on stock sensors set up - CAS has inaccuracies and few advantages.

Um, you keep saying intake cam? On my (1.8L) '00, the driver for the CAS is on the exhaust cam. Perhaps a picture of the motor could help. :-)

I have heard that, on some motors, an "exintake" - putting the longer duration exhaust cam into the intake cam's location - can yeild significantly more power - but this may or may not require modifying the valve cover, etc, to accomadate it, and you'll likely have to find a way to index the cam. It's fine, but a big job.


Don't worry about the binary stuff to much - it's not that hard, but also not that necessary.

If you want to try skipping the CAS I'd be happy to help - though I don't know a lot about the 3.57 board. I wonder what the launch control looks like...

Zaphod

Yes, at the 99 1,6 the CAS is at the intake cam, just like at the 90-93 models. (But the cams between 90-93 and 99 are not interchangeable)

I always hated me for not buying a 1,9 first hand - but too late now.

So it seems I need a circuit to read the factory sensors. I would be really glad about any help. Would they work with the MS1?

Please feel free to discuss this problem further - I will not be on here for about two weeks.. I am going to holidays with my family and need some time with my two little kids (and of course my wife).

After the holiday I will try to solve the problem asap - I want to get the car running again...

Greets

Joe Perez

iTrader: (8)

If the Euro 1,6 '99 uses the same crankshaft wheel as the US '99 (and I'm almost 100% certain that it does) then no, it will not work with MS1. The reason is that the wheel has four teeth which are not evenly spaced. Two are at approximately 70° BTDC, and two are at exactly TDC. You would need to cut off the two which are at TDC, in which case the stock ECU would no longer work.

AbeFM

iTrader: (3)

Picking up a spare wheel and doing that mod might not be a bad way out of the predicament, but then there goes the whole thing of doing it in parallel.

I guess I would start down the path of modding your MS to run off stock sensors - it's all there, and the timing will be more accurate.

if that's what you're going to do, I'd suggest ordering an MS-II chip now, so it's there by the time you return from your vacation

JasonC SBB

I can design a circuit that takes the NB signals and "removes" 2 of the teeth.
I can possibly make the circuit also modify the CMP signal to something the MS1 can understand.

What shape does the CMP signal have to be so the MS1 can understand it, if the CKP signal has 2 of the teeth "removed"?

IIRC I made a circuit that also created a CMP signal (aside from CKP) that resembled the NA signals. I have to go dig it out and make a timing diagram.

JasonC SBB

OK I studied what a simple circuit can do. The resulting CKP signal will be like an NA. Rise at the first NB CKP tooth rising edge (10° BTDC), fall at the 2nd NB CKP tooth rising edge (70° BTDC).

CMP signal will resemble the NA signal. However its edges will not be at the same exact point in the engine cycle as the NA CAS, but its edges will be between the correct NA CKP edges and not "cross over" any of said CKP edges. Is this good enough?

Lastly, IIRC, the synthesized CMP edges will be 360° crank out of phase with an NA - i.e. you'll need to swap cyls 1&4 and 2&3 injector and plug wires.

Joe Perez

iTrader: (8)

The exact spacing of the CMP does not matter so long as it falls in the right place between CKPs. You're doing a flip-flop on the CKPs I assume, using CMP for sync?

Zaphod

Hi,

I don't care to swap the wires, as long as I don't have to do a lot more to get the MS running. (And as long as the circuit gives me the right signals to get everything running correctly)

Would the circuit fit inside the MS case? I got a 3.57 so no proto area - but I would fit it inside the box anyway - if possible.

Greets
Sven