TeamSwift

I want to run individual injectors (MPFI) to eliminate the wet manifold setup.

Can the stock ecm drive 3 injectors?
Is the pulse width going to be way off?
Is there a way to change the constants?

Batch fire is fine, sequential is not important here.
I understand the turbo models ran 3 injectors, but those are hard to source, and I dont plan to run a turbo.
I know megasquirt (and other ECMs) can do this, I just want to see if the stock ecm can handle it.
I know the fuel pump will have to be changed, and a propper (43.5psi w/ vac) regulator added.
I have crunched the numbers for 60hp, BSFC of .55, 80% duty cycle for 13.75lb injectors. (14lb injectors from a GM 2.8l 60deg v6 should be fine, I can tune with the regulator)

Thanks folks,

Jon

the stock g10 runs a big resistor along with the single tb injector.

although i have not done it, i would think that you could bore the 3 bosses on the na g10 head for the injectors (just like the turbo3) and use a turbo3 fuel rail with it's fuel regulator and probably it's injectors. the trick would be coming up with a suitable intake manifold and throttle body. again, the turbo3 parts would simplify things.

i started out with that in mind but when i considered it, it was so close to just adopting the turbo3 system in it's entirety that i just gave it the full monte. theoretically, you should get a good improvement in low range torque. just be sure to consider port runner length for the rpm range you want to design for. a lot of people do multiple throttle bodies on short tubes and then find out the the engine has a really peaky rev range and wants to run at 4500 or 5000 rpm. off idle, that short runner mod has got no balls.

i've never run into anybody who really does ecu modifications on 3 bangers. there are no available chip swaps, no code dumps, or any aftermarket suppliers for engine management. that why we go to mega squirt.

_________________
1991 Blue Geo Metro Convertible highly modified 1.0L Turbo3 5 spd. - 1991 Red Geo Metro Convertible customized 1.0L 5 spd. - 1992 White Suzuki Swift GT

My Turbo3 Project

My Cardomain Page -Ol' Blue

My YouTube Channel

My Photo Gallery

SAAB Sonett II

Interesting....

Do you know the impedance (ohms) of the injector and the resistor? Is the resistor run in series or parallel with the injector?

If I can run all 3 injectors in series, that would be great!
It looks like the resistor might be there for impedance matching the injector to the driver in the ecm. Which leads to more assumptions that there is the possibility that the ecm is already set up for multiple injectors, and it has a cheap patch (the resistor) to run one injector.
As it stands, from tests I have found online, the stock injector fires every 360* of crank rotation, or twice a cycle. This is essentially batch fire when applied to multiple injectors.

Intake manifolds are my forte, I am wanting to apply a design I have for "progressive diameter and length runners" with will accompany another experiment with variable valve timing. I know what you are referring to when you talk about peaky power... it’s really hard to tune ITB for more than one rpm range effectively. If you dig into the Helmholtz theory you can easily see what’s going on and why things happen the way they do.

My goal is to have a good platform to use for testing, IE cheap, easy to acquire, and well documented in all aspects. I just need to stray away from the TBI to study my theory.

It’s a shame no one has gotten into the g10 ecm. I bet there is a lot of potential in there that would be easy to take advantage of.

Does anyone have more input?


Thanks,
Jon

i never measured the injector impedance or the resistor's value. the resistor is the aluminum package bolted to the firewall on the left side behind the strut tower. it has white high temperature leads and a cylindrical yellow 2 wire connector. the schematic shows it as a series connected resistor.

connecting multiple injectors in series isn't a good idea. each additional coil in series sees the lag of current from the previous coil, each successive coil affects the current flow, pintle reaction time, etc. the single 210cc/min turbo3 injectors measure 14 ohms each, i can offer that much further info. i'd wire them in parallel and make up the current regulation with a single appropriately sized resistance.

_________________
1991 Blue Geo Metro Convertible highly modified 1.0L Turbo3 5 spd. - 1991 Red Geo Metro Convertible customized 1.0L 5 spd. - 1992 White Suzuki Swift GT

My Turbo3 Project

My Cardomain Page -Ol' Blue

My YouTube Channel

My Photo Gallery

SAAB Sonett II

hmmm, if I wire them in series, place the last injector of the series on the last cyl to fire...then I could make "sequiential" injection.

Kidding.

Realistically, it being batch fire, and the injector firing twice per rpm, indicates that there is fuel injected constantly, no matter the realitive location of the cycle of a particular cylinder. It would seem that a delay caused by coils in series would be a moot point.
In other words, it would be no different than how the TBI works now, other than the fact that the injectors are closer to the valve, and there are 3 times as many.

Have you seen this?
viewtopic.php?f=11&t=25103

_________________
1994 Metro - MPH project (getting a DOHC G13B)
1994 Metro - MPG project (getting an XFi G10)
1992 Swift - Parts car (gone)
1991 Swift - Parts car (gone)
1990 Swift - Parts car
1997 Metro - Parts car (gone)
1993 Metro - Parts car
1989 Swift GTi - Parts car
1998 Metro - Parts car

haha, good memory mate, I was thinking of the same thread, been a long time since Don was here

_________________

I've done a code dump on a 2001 Swift with G10, I'll have to find it and post it up. I haven't seen an earlier model yet.

rhinoman, when you post the code, it will be the first time ever that i will have seen it.

until you make it public and someone puts it to use to modify a g10 ecu in a way that improves performance - it doesn't exist.

you're the code jockey. i've only played with digital code since about 1976 and i've only been pissing around with suzuki 3 bangers since 2003. i might have missed something.

nobody does it because there's no money to be made by doing it.

_________________
1991 Blue Geo Metro Convertible highly modified 1.0L Turbo3 5 spd. - 1991 Red Geo Metro Convertible customized 1.0L 5 spd. - 1992 White Suzuki Swift GT

My Turbo3 Project

My Cardomain Page -Ol' Blue

My YouTube Channel

My Photo Gallery

SAAB Sonett II

Here you are:

http://rhinopower.activeboard.com/index ... D=37583298

I haven't delved very deeply into the code, I was only really interested in the serial port stuff, the bootloader and the OBD2 and OBD1 ports. I should have some schematics somewhere, scrawled on bits of paper. I was supposed to move home a month ago but it didn't happen and most of my stuff is still boxed up.

ok folks, ive got the code for the g10, i read that other post where it seems the guy has the same idea i do and now im wondering if he ever got it finished...it looks like he was real close.

on a side note, from my fiero and ecm learning days, i recall the 1227730 ecm being a heavy hitter for home brew setups, running things from 2.0l 4cyl to 32v northstars and everything inbetween including boost. I am going to dig into the 7730 bin and see if i cant change some things around for DIS and batch fire, individual injectors.

sorry for the briefness, i am at work and have been (and will be) busy for the next few days.

rhino,

there it is! seriously, that's the first dump of 3 banger code i've ever seen.

that's from a mk5 ecu, obd11 like stuff? the ecu is different from the mk2/3 obd1 type. is the uP the same as in the older ecus? maybe not and the sensor and control set is considerably different on the older cars.

it make sense that you would be able to dump the code and disassemble it on the newer car, though, and i'm impressed that you actually took the time to do it.

_________________
1991 Blue Geo Metro Convertible highly modified 1.0L Turbo3 5 spd. - 1991 Red Geo Metro Convertible customized 1.0L 5 spd. - 1992 White Suzuki Swift GT

My Turbo3 Project

My Cardomain Page -Ol' Blue

My YouTube Channel

My Photo Gallery

SAAB Sonett II

I haven't seen any of the earlier 1ltr ECUs, Denso used a variety of processors but started with the 68HC16 in around 1995/96, there are also some HC11 based ECUs around the same period. Before that its a bit random with various 6801 based processors, some of which are easily read, some are a bit odd.
The 2001 code looks like it was ported from a much earlier model, it doesn't make any real use of the TPU which may not even use Denso's EFI microcode.

i know that nopi.com offers a ecu reflash from jet

JET Computer Chip Upgrade Box Program - consists of box and paperwork for you to ship your EC to Jet to have chip soldered in - you pay freight to Jet, they ship it back for free
Mfg# 65001, NOPI# 514+0035

Price $264.66

It's been a LONG time (years) since I have posted here, but I have experiance swapping from a TBI setup to MPFI while still using the stock TBI computer, but it was on a different car. But, I figured I would post up and tell you guys what I did and how it worked.

My car was a 1991 Chevy Cavalier 2.2L TBI. I first did some research and found the factory injector was about 2.0 ohms. I then found an MPFI 2.2L later model in the junk yard and I measured the injector impedence which was 3.6 ohms. I found that the 1992 and 1993 2.2L Cavaliers had 4 injectors, but they fired all 4 injectors at the same time, and fired them once per revolution just like the stock TBI setup. 1994 and 1995 models ran batch fire injection, and 1996 and newer ran true sequential injection for emmissions purposes only.

I had it easy and was able to swap the entire intake manifold which simplified things. So anyway, all the sensors were the same so I had no check engine lights and I was able to use the factory wire harness, this is something that you have to take into consideration. The only splicing I did was on the injector wire harness, which suprisingly the TBI injector harness and the MPFI harness used the same color codes which made it easy.

After it was all put together, suprisingly it ran pretty good, but it did run a bit rich at idle. Low end power was SLIGHTLY lower, mid range power was about the same, but high rpm power (4,000 rpm and above) was GREATLY improved!! Pre swap the car dynoed at 82 whp, post swap it dynoed at 107 whp!!!

I put 5,000 miles on the car post swap and had ZERO problems out of the car. I raced the car at autocross events a LOT so that 5,000 miles were not easy highway miles.


btw, I did this back in 1999.

_________________
2010 Nissan Frontier SE Quad Cab 4.0L V-6

Someday I will get another Metro, but this time it will be all out built 1.3L!!!

I didn't want to start a new topic especially since I'm planning on doing the exact same thing. so I have already modified my cylinder head for the 3 injectors and built a stainless steel fuel rail. My only annoyance is that the fuel rail is a bit misaligned but nothing that should cause a serous problem.

I just have to start to figure out the ecu

Keep the updates coming interested to see know your project turns out