Sclippy96

I have a MSPNP2 that I just installed in a stock '95 miata with no A/C. Drivetrain is mostly stock right now with the exception of a Monsterflow intake, stock headers with highflow cat with a supertrapp dumping before the axle. I have an MTX-L wideband which is currently wired to send a narrowband signal thru the oem harness. This will change to wideband once I have a chance to take the dash tombstone off again to get to the gauge mounted in the radio delete plate.

I had usb-to-serial driver issues for my windows 7 laptop initially. My workspace doesn't have wifi and I didn't connect the ECU to the laptop before I started the car and test drove it. It's starts right up first try, purely on the base MSPNP2 map loaded in the box from the factory I am idling rich at 10:1 and down a few hundred rpm, any light throttle tip-in will peg the gauge at 20:1 and the car will shudder and hesitate as it goes super lean. If I jump on the throttle moderately to full the AFR numbers are decently close to 14.7 though. I am assuming I need to lower the fuel numbers in the fuel VE idle cell till it is closer to 13:1-14.7:1 AFR, then look at giving the tip-in/low throttle cells more fuel to solve the really lean spots. I drove it about 5 miles home from my garage workspace and the MTX-L numbers are decent when I hop on the throttle harder than 25%, its at low throttle situations where it pegs the gauge full lean. Makes for a funny drive when I have to hop on the gas hard and coast then rinse, lather and repeat.

I think I got the serial adapter drivers figured out, I am able to get loop-thru in the Tunerstudio mini-terminal now when I jump serial pin #2 & 3 on the adapter. I'll try getting connection later today.

Right now I have my MTX-L gauge wired into the dash using the switched hot that used to feed the now deleted radio and the big negative wire that used to physically screw into the back of the radio. With the OEM ECU and this wiring setup gives I get rock solid 14.7:1 AFR under idle and cruise unless I hop on the throttle and it will go as high as 11:1 at certain RPM's which is normal for the OEM ECU. I read everywhere that the wideband should share the ECU ground with a Megasquirt. Does this mean I should run a dedicated ground wire for the MTX-L back thru the firewall around to under the throttle body to where the ECU harness grounds to the block? Obviously in any case I will be switching back to the wideband signal wire from the MTX-L and changing the settings in Tunerstudio to expect a LC-1 wideband signal before I do any tweaking to solve my drivability issues at idle and low throttle.

[edit] Off throttle downhill decelerating will hit perfect 14.7:1 AFR on the gauge right now with the MSPNP2, which is a little odd since thats the time that the OEM ECU will go full lean as it cuts the injectors under downhill deceleration. It might be a placibo effect but the car doesn't seem to slow down as quickly off throttle when the injectors are still firing fuel into the cylinders to achieve a 14.7:1 ratio under decel.

[edit rev2] Yes, I did physically set the fixed timing to 10 degrees the traditional way with the OEM ECU before installing the MSPNP2.

Braineack

iTrader: (62)

I didn't read the rest, but you still need to sync the MS with the CAS.

The MS has no idea if the timing is correct, just what it wants to control it at. It's your job to make sure that when the MS is demanding 10° the MS is acutally outputting 10°.

you do this but telling the MS to lock the timing at 10°, verifying it with a timing light, and if it's not at 10°, you adjust the trigger angle in the software until the timing matches. Then you unlock the timing and go back to running of your spark table.

Sclippy96

Update: I have been able to get the computer to talk to the MSPNP2, but the connection isn't always repeatable. I'll have the serial to usb adapter connected to a COM in the device manage and set in Tunerstudio MS and I spent an hour on every COM port and usb port plus wiggling all the connections but nothing, then almost magically it will start working on the same exact settings I've tried before.

I think I may have finally got it to connect when I changed the baud rate in the device manager to 9600 then back to 115200. When I get frustrated I usually pull the serial and jump pins #2 & 3 of the serial adapter and connect to the current com port in the mini-terminal of tunerstudio MS and see if I'm getting loop-thru connection and I know that the serial adapter drivers are indeed working properly.

Once tunerstudio connects it has been easy-peasy so far, when I finally got the ECU to connect to the computer I noticed why the car was running so crappy, the tune loaded in the ECU I got from my buddy was in Alpha-N mode, I switched it back to the base default speed density settings it started running a million times better.

I got the fixed timing all squared away, needed -2.20 degrees of offset trigger angle to get the timing light to hit dead on 10 degrees with fixed timing selected.

The MTX-L wideband sensor/gauge was reading about 0.5 richer than the Tunerstudio was showing. I was wondering if it was a bad ground on my wideband so I went ahead and ran a ground directly from the black and white wires of the MTX-L to the spot where ECU ground wires connect on the block just under the throttle body on my '95. The numbers were slightly closer, but still showing about 0.3-4 richer on the gauge. I initially calibrated the AFR table in tunerstudio for the wideband using the MTX-L's default output which is 0v=7.35AFR & 5V=22.39AFR, but I messed around and set the AFR table to 0v=7.35AFR & 5v=21.6AFR and the both gauges seem to match up within 0.1AFR now.

After I got the wideband squared away I started VE Analyze Live and watched the idle AFR get down to 14.0ish AFR in a couple of minutes. My idle was a little low so I adjusted it up to 900RPM with the idle screw. Drove it home about 10 miles from my workspace and hit as many cells as I could for VE Live. My laptop will lose sync while driving sometimes, but might be because I don't have a solid state drive in it yet and it's wigging out. So far so good, I'm going to try installing the 650cc EV14 injectors next weekend, changing the req_fuel accordingly and let VE Analyze Live work it's magic again. If that goes well it might be on to the turbo the weekend after that.

Sclippy96

I have the car running awesome now on the MSPNP2 with the ev14 650cc injectors. Some hurdles I had to overcome were 2 bad usb to serial adapters, I finally bought the usb to serial adapter sold on the EFI analytics website and it works great. I also saw that my old laptop had an expresscard slot and I bought a expresscard to serial adapter and that works perfect too. For all newbs trying to save a dollar on your usb to serial adapters, DON'T do it. If you feel you have to buy something cheap at least make sure it has a FIFO chip, the Prolific chip adapters are not good for our use.

With the 650cc injectors I set the required fuel to 4.8 and messed around with the dead time finding that 1.220(ms) seemed to be the sweet spot for these injectors. My VE table number at idle is ~54 with a target of 14.0AFR, I usually see an average pulsewidth of 1.750 at idle. The car idled great, the engine stumbled badly when any sort of quick throttle was used, I saw the AFR gauges go super rich as it stumbled, never lean. I was perplexed, but then thanks to MT.net is saw that when you install larger injectors you have to shorten the pulsewidth in the Accel Enrich TPS Based AE table by same ratio you reduced the required fuel when you went from stock injectors to bigger injectors. Once I ramped down the Accel Enrichment it runs like a pearl. My numbers are 2.5pw adder/100TPSdot to 4.8pw adder/1000TPSdot with a 95TPSdot Threshhold. I turned the threshold up to 600TPSdot when I did the VEAL and tried to be as steady as I could with the throttle to not engage the AE as it autotunes. I am happy with my rough tune and am confident to take the plunge and go ahead and install the turbo in the coming weeks. The de-newbing continues...