AbeFM

iTrader: (3)

Yeah - everyone starts off the same way, afraid to jump in with both feet, and keep the stock ECU around "just in case". But it really doesn't do anything but get in the way of your clutch foot.

BUT, if you're going to get rid of it, I'll again strongly recommend my OEM circuit for the inputs, I had SO many problems with noise. If you want to spend months wondering why your car bucks, misfires, kicks, belches fire, blows your cat right out your tail pipe, stumbles, pops and bangs - let alone the wear on your tach from slamming into the 0-rpm stop twice a minute... by all means, don't do it.

-Abe.

kingofl337

What issues did you have with your MS-II sharing the oem sensors and ecu? Did you use the drawings I posted as an input or something else?

Abe what are the final schematics you are using for standalone operation?

Are either of you running the 2.0.1 release code?

AbeFM

iTrader: (3)

I can't answer what problems he had, but recently helping a friend work on his car, there were issues with the altenator not running, the TPS not reading right, coolant not reading right, etc. Basically, if it's at all important to anything, you have to cut it so the ECU doesn't see it. If it's not important, what does the OEM need to see it for.

Once you take that out, you can replace the carbon canister mess with a check valve. :-P

As far as the schematics, the only real issue is the one I had posted names impossible pins for the inputs, so I need to figure out which are the right ones to use, but it's just the same dual op-amp circuit the OEM uses, nothing to it.

I haven't run the 2.0.1 release code yet. I was waiting for it to get stable - since the betas were more stable than the RC's and release versions (strange scheme they use there). Now my only concern is though I've asked several times, they haven't explained how they run the motor without having sync. I'm not sure I'm comfortable having the motor fire a tooth off.

On a 36-1 wheel, ok, you're off by 10 degrees. Maybe. On our 4 tooth wheel, you could fire 80 degrees or 110 off? Weird.

So, I'm probably going to try it this week, but it makes me uneasy.

Joe Perez

iTrader: (8)

Here's my rendition of the corrected schematic for Abe's dual input circuit. Cap values are in uf.


The IC is spec'd as an LM393, though if you live near a Fry's, you can use an NTE943M instead. Abe suggests experimenting with smaller values of C1 and C2. With my POS CAS, I had to use MUCH larger caps, which caused lag on the order of 350us- much more than can be compensated for in software.

Netnames are relative to a R3.0 PCA.

AbeFM

iTrader: (3)

Ah! Much nicer. Only, I'm not really a fan of the LM393 - I've been using the much cleaner/faster/lower-current dual bi-fets from radioshack.

I didn't try the LM on my car, but at work I've had things suddenly work when I switched to the not-designed-in-the-70's version of the dual opamp. :-)

But yes, that's the circuit, and very well cleaned up!

Joe Perez

iTrader: (8)

The TL082 you mean? Same pinout and supply rating, so it should drop in just fine.

AbeFM

iTrader: (3)

Indeed, that's the one.

Too bad I didn't think of it earlier, it would have been darned interesting to see if your circuit got any better swapping out that one component. I sort of don't see how it would have, but then again, I don't see why mine worked better here either. By the specs they should be the same.
-Abe.

kingofl337

Joe, why JS8 instead of JS10? How do you change the second input to JS-8 or do you have to edit and recompile the code?

Abe, did you remove the bias resistors R4/R7 for the temp sensor inputs? I can't pass ODB-II checks if I remove the evap system. Unfortunately NH has checks, I'll definitely keep an eye on what the MS and OEM computer are saying. If I can't make it work I will have to work on a more standalone solution.

I appreciate your advise this is all very good info that may help us to nailing down a 99+ Miata MS-2 manual.

kingofl337

Oh another question for Abe, weren't you having issues triggering on the circuit above during cranking?

Joe Perez

iTrader: (8)

Because as my signature and many posts indicate, I'm using an MS1, not an MS2.

This is one of the relatively few necessary wiring differences between the two CPUs. On the MS1, pin 11 is coded as the second trigger. On MS2, that's one of the CANbus lines.

Hell, your input circuit was the fourth design I tried. The problem wasn't the circuit, it was my POS CAS. Anyway, I'm quite happy with the crankwheel in place now.

AbeFM

iTrader: (3)

I'd love to see a good manual for an MS-II install for the miata. Wouldn't take much. And if it could be done with schematics instead of pictures, people would actually learn something. :-)

Yes, pulling out the pull up resistors should help. I ran in parallel for like a few days. Basically, I kept fixing problems until there was no connections left to the OEM ecu.

If it were me, I'd just swap computers for the tests. Put in the OEM, drive for a day, take your test. Then swap back. I made all my changes (diodes, wiring the fans together, banking the injectors/coils/etc) all into the boomslang, so aside from the extra temp sesnor, and boost controller) I just unplug the boomslang and put in the OEM ecu.

If I am willing to leave the boomslang hanging loose, then it's literally moving the plug over.

And taking out the 550cc injectors. :-P

AbeFM

iTrader: (3)

Oh, all my starting issues stemmed from two things:
1) Wrong polarity, the signals put out by the OEM circuit (which are conditioned VR sensors) are unreliable on their leading edges. Simply choosing the other side fixed things up
2) The code sucked (still sucks) for the motors. But someone (arga?) caught that it ALWAYS thought it was on cycle 1, so it would take a while to sync up. Now it starts anywhere from 1/4 turn to 1 and a 1/4. But better than the more-than-three it would sometimes take before.

It still starts too slow in my opinion, but there's a lot to tweak in the code to fix that. I'm going to work on PID idle (and boost control?) first, then maybe more I/O. After all that the slow starting is my biggest boogaboo.


Joe - Yep, I do like the crankwheel. Might do it myself at some point. Got to get on this FreeEMS thing. Sequential spark and fuel! Idle that means something. Etc.

Joe Perez

iTrader: (8)

I don't know why everyone keeps complaining about idle. Mine is damn near perfect when cold, damn near perfect when hot, and only slightly less than perfect when in mid-warmup.

AbeFM

iTrader: (3)

The idle code on the MS-II is not the idle code on the MS-I, I believe. It could be, but I don't think so. I know that it was written specifically for the MS-II, though it might be a direct copy.

The MS-II version has two states: Oscillatey, which makes it stall, and too slow, which makes it stall. And, it's been shown to be mathematically wrong!

Now, just setting the motor to idle at 1k instead of 800 rpm works flawlessly. Go figure.

M-Tuned

iTrader: (6)

What headaches did you have.. I'm about to do the parellel install. Just need to get a weekend without some weddings. Atleast they are not mine.

cjernigan

iTrader: (13)

He was running MS2 in parallel that was his main problem.

Joe Perez

iTrader: (8)

Indeed. On page 1, kingof1eet asked why the stock ECU could be causing any harm when the IAC was connected to and being controlled by the MS. Of course, connecting the IAC to the MS means that it must be disconnected from the stock ECU, which will cause a code to be thrown. This defeats the purpose of doing a parallel install for many.

I firmly believe that a parallel MS install can be made to function correctly, even an MS2 on an NB. However I suspect that it will be more work, not less, to get it running properly as compared to a standalone install. The only valid reason for attempting this in my opinion is where OBD-II authentication is necessary for the purpose of passing a smog check, and even at that I would be tempted to simply de-megasquirt the car for a week once a year when it's time for the test.

AbeFM

iTrader: (3)

+1



No kidding. There should be a fee for running in parallel. Well, there is, but you pay it after you start, not up front.

Unequivocally, parallel installs are harder than stand alone installs. Ask again, and I'll slap you.

cjernigan

iTrader: (13)

I don't know guys my parallel install works pretty great. Even though the harness takes a little more time I find it incredibly convenient to be able to not have to mess with the idle valve, A/C circuit, or the alternator control. All of which are easy to deal with but I have never had to. Not something i'm necessarily proud of because I now know nearly nothing about controlling those three except what I read but isn't there a saying "Ignorance is bliss."
Poor MS2 guys and their bad parallel install experiences.

Joe Perez

iTrader: (8)

And I keep begging you to write one. It will become a sticky.

Really, just having specific details on your alternator controller would be a huge help. Part number, picture, wiring diagram.