TeamSwift

Ok, I have delved into the land of Megesquirt2 and I have a couple of bugs that are plaguing me.
It does start but runs horrible and will not idle
Bugs
Erratic tach signal (steady when just cranking)
Erratic timing advance when using table (stable when in locked timing)
Erratic afr at all rpms

The setup is as follows
G10 motor with 75mm vitarra pistons
1990 chevy sprint turbo top end and turbo
Stock 220cc injectors
Deleted mav
Gm coolant temp sensor and IAT sensor
Wahlbro 255 fuel pump
Single coil oil bath old school no igniter run off of pin 30 with transistor db373 on circuit d14
Stock distributor with vr sensor
Aem wideband O2
Stock static timing


Megasquirt setup

MS2 V3.0 board
Firmware 3.4.2
Basic trigger
0 trigger offset


I have been at a loss at what this is being caused by and have tried everything that I can think of there is just not a lot of info for the g10 motor and MS2. I know that t3ragtop and a couple of others have run a similar setup throughout the years and any advice would be greatly appreciated. I know that you all might need more information to give advice if there is anything else that you need to know
I will answer to the best of my abilities. I think that I have a setting wrong with the ignition setup but for the life of me IU cannot figure it out.

Well i hope that i have figured out some of the problem. iSc valve was malfunctioning and i believe in hat was one of the reasons for hat idle was so terrible also hopfully explain why the timing was all over the place due to the incosistant tach input.

Also changed the crank trigger angle to 13 this seemed to help smooth everything out.

i never ran the single coil setup on megasquirt. i started with the 36-1 trigger wheel and a v/r sensor. that proved to be a bad deal in that the v/r sensor acts like a generator. when i got the timing dialed in the signal was in the mud at idle and at 6000 rpm the v/r was making 70 volts which interacted badly with the on board signal conditioning circuitry. that caused timing resets and and misfires at high rpms and larger loads.

in retrospect i believe that it was that exact condition that caused the demise of another member's 'squirted turbo3 while he was at high boost, high revs, and high load.

i switched from v/r timing sensing to hall effect devices and the voltage generation problem, the timing resets, and misfires disappeared.

the problems i ran into trying to use the megasquirt configured to run the 3 banger came from the fact that you actually have to run a "spoofed" configuration for a 6 cylinder engine which automatically defaults to a lost spark set up. using a sequential configuration and 3 channel direct fire ignition gives you way more control of both injection and ignition and it was that configuration that really allowed me to use the tuner studio app to dial in the engine management.

i'm not sure what injection mode you are running but i believe that trying to use throttle body injection with a g10 intake manifold and throttle body is in itself problematic. it didn't take me long to shift to multiport fuel injection using a turbo3 intake set.

_________________
1991 Blue Geo Metro Convertible highly modified 1.0L Turbo3 5 spd. - 1991 Red Geo Metro Convertible customized with a Twincam 5 spd.

My Turbo3 Project
My Cardomain Page -Ol' Blue
My YouTube Channel
My Photo Gallery
SAAB Sonett II

Thank you for the reply. I am running the g10t mpfi head and manifold the Canadian one. I am running the stock dizzy with 3 lobe vr sensor

using the distributor for engine timing might be your problem as there are only 3 data points in 2 crank revolutions. that's pretty low resolution for the timing count. the cam only goes around once for every 2 times the crankshaft turns. that means the distributor only goes around once for every turn of the cam while the crank turns twice, 3 timing pulses in 720 degrees of crank movement.

my 36-1 trigger wheel gives the megasquirt one tick every 10 degrees of crank revolution. that would be 72 timing pulses to your 3.

so what i'd suggest is moving your engine timing count to the crankshaft. you can still use the distributor to direct the spark to the correct plug but the data supplied the controller would have a much higher resolution. that timing count is very important for the injection process.

_________________
1991 Blue Geo Metro Convertible highly modified 1.0L Turbo3 5 spd. - 1991 Red Geo Metro Convertible customized with a Twincam 5 spd.

My Turbo3 Project
My Cardomain Page -Ol' Blue
My YouTube Channel
My Photo Gallery
SAAB Sonett II

I was afraid that you were going to say that about the triggers. I had been hopeful mainly due to the fact that my trigger logger was so clean. The problem is that we have a race in 3 weeks and I am afraid that I do not have time to re engineer a 36 tooth setup. I have read your build thread and I know it was a couple years ago now since then has there been any pnp solutions created or manufactured that would allow a 36 wheel.

Like this maybe

http://www.ebay.co.uk/itm/Swift-GTI-MK2 ... 7675.l2557

On a separate note any one would not happen to know the hertz that the ISC should be operating at I found that the g13b isc is at 20 hz but I can not find anything on the g10t ISC.

i found that the hard part is bolting the trigger wheel onto the crank accessory drive pulley so that it is concentric. you want to have as little run out as possible so that the signal doesn't fade in and out.

when i set up the isc i just let the megasquirt treat it as an on/ off sort of thing and protected the isc valve's coil with a diode. i don't try to pulse it or anything as it is only in operation until the coolant temp comes up and while fuel enrichment is in play. i let the cold start air valve located on the bottom part of the throttle body take care of the idle until the coolant temp comes up and that doesn't take long.

also, i added the big transistor high current driver to the cold start coil circuit. it will run a full amp or a little more to the isc coil without blowing up.

i also don't let my turbo3 idle all that much. the 450cc/min injectors are very fat at idle and the engine is running pig rich at 1100 rpm, no load. i let the engine idle just long enough to let the oil pressure stabilize and then i drive using low throttle angles until the coolant temp normalizes. then it's on like donkey kong.

_________________
1991 Blue Geo Metro Convertible highly modified 1.0L Turbo3 5 spd. - 1991 Red Geo Metro Convertible customized with a Twincam 5 spd.

My Turbo3 Project
My Cardomain Page -Ol' Blue
My YouTube Channel
My Photo Gallery
SAAB Sonett II

What version board are you running? Just asking so that i can know how to reference. I did get it to idle well and now my only problem is that if i give it gas to quickly off of idle it stumbles and after i rev it up and release it goes lean for a sec and then stabilizes. Also if i rev it up to say 2000 it will osilate about 150 rpms.

i'm using the ms2 ver 3.0 controller modified with a jbperf 4 channel injection/ ignition driver board populated with 3 channels.

i de-populated the main board of many components that weren't needed with the jbperf mod including the crystal oscillator. the processor uses the crank trigger wheel count instead of the oscillator to clock data through the register which makes the processor run with great stability. it also allows the controller to operate the predictive mode with better accuracy.

i developed a hall effect device sensor set for the crank and cam angle counts using little chinese hall effect device signal processor boards that output a zero to 4.56 vdc amplitude square wave at any frequency, tested to be rock solid up to 10,200 rpm, and then depopulated the signal conditioning components from the megasquirt's main board. i connected the hed signal conditioning boards' outputs directly to the processor on the daughter board landing the wires onto little solder pads.

i also removed the injection and ignition drivers from the hear sink on the controller board to cut down on heat inside the enclosure.

i'm running the latest msextra firmware and run the latest version of a licensed version of tuner studio.

i use the mosfet ignition drivers on the jbperf board to switch 3 hitachi ignitors from a gt dohc and run 3 cop coils from a ford v8 on each spark plug. the injector drivers on the jbperf board drive the 3 lo-z 450cc/min injectors directly.

it took a little work to set the dwell on the ignition circuit to keep the cop coils from heating up but that configuration renders super throttle response. there's nothing but go, go, go when i stomp the gas pedal. as a matter of fact, i have to use the throttle judiciously or the drive tires just roast, even in third gear.

_________________
1991 Blue Geo Metro Convertible highly modified 1.0L Turbo3 5 spd. - 1991 Red Geo Metro Convertible customized with a Twincam 5 spd.

My Turbo3 Project
My Cardomain Page -Ol' Blue
My YouTube Channel
My Photo Gallery
SAAB Sonett II

I also am running a ms2 v3.0 with registered tuner studio, I built but not nearly as modded as yours. After tuning the vr pots a bit I am getting a great trigger signal and clean comp logger. It starts right up and idles great during warm up but after that it starts to oscillate RPMS. It also does this if I try to hold an rpm steady

I think it is from my twitchy WB signal that fluctuates about 3 points. It is a AEM WB but I have is mounted in the stock o2 bung hole with about a 1ft down pipe (need to get it to a muffler shop to connect to the old 2" exhaust) could this be screwing with the WB and in turn screwing with the AFR which the MS is trying to correct and giving me a oscillating rpm

as per your suggestion, i am running the isc valve in an on/off fashion but have a problem now not getting enough air after it is warm to allow for a steady idle. tried to adjust idle air intake with idle speed screw on top of throttle body but am unable to get a steady idle without the car dying. while driving, boost seems to be intermittent and have great power through one acceleration pole but the next will be gutless in comparison. i believe to be a tuning issue and not mechanical due to its itermittent happiness. thank you for your help- also, if there's anyone in the denver area that would like to help me tune, i will pay you.

Would anyone know where to get a g10t intake manifold? It's extremely hard to find here so far. Trying to go boost but the tbi setup is the worst

I can supply one, off of Canadian g10t. Pm if interested.

_________________
1995 Swift w/16V 4.39s, 3tech cam, Esteem t-body, Header, needs more.
1995 Gt Mustang "Boss Shinoda" package.
1999 F150 4x4 Supercharged
1967 Mustang 428 auto, never ending expensive project
1993 Civic si h22a, fell in my lap, couldn't resist!