Twiss' supercharger+EFI

[Rhinoman]

The 8V Vitara post 1996 are all tunable and a few earlier JDM models, the Jimny G13BB is an excellent ECU but it would be a lot of work to get it to run on boost. For an easy turbo you would want a MAF based system or an ECU from something that's already boosted or has support for boost. You'd probably have to look at a later (32-bit) Suzuki ECU or switch to another make, the old GM and DSM ECUS had a lot of support as do the Subaru ECUs. MX5 ECUs are quite cheap and are very similar to Suzuki ECUs.
There are a lot of definitions available with the KWP flasher but I haven't played with them, they are mainly Euro cars.

[Edweird]

Rhinoman - I've read that many stock ECU's are reflashable - do you know if any of the Vit ones are? If you can, maybe one of the generic ECU's would be suitable for this project? That might make the engine harness simpler, for one thing...

Putting aside the freescale, what ECU would you go for? I'll freely admit that I know *nothing* about ECU's, other than occasionally wondering if a cheap ARM CPU like the one in the Raspberry Pi would make a good CPU for an ECU (at about $5 a pop, a 600mhz 32bit CPU can't really be had any cheaper, now can it?)...

I'm pretty sure a Raz-Pi could function as an ECU if you really wanted it to. Maybe have two, one handling the mathematics and the other one interfacing with the car.

[twiss]

i reckon the Pi could handle it, but you'd have to code it from scratch, which I can't really be bothered to do...

Rhinoman - With these stock EFI systems, is it difficult to change the kPa and RPM values easily in the fuel tables?
I was lead to believe that it was more a case of adjusting the mixture at each point in a pre-set table.

Correct me if I'm wrong as I am a newbie to injection

[Rhinoman]

There are over sixty tables in an OEM ECU but yes if its remappable then you adjust the values within those tables. Many ECUs have non-linear axis for the fuel and ignition tables and use a look up table to convert real RPM to 'table position' so its not too hard to rescale them.

The problem with a MAP based ECU from a N/A motor is that the table scaling starts at atmospheric and falls to zero, there is no provision for boost. If you fit a 2-bar sensor then all the parameters that rely on MAP need to be rescaled, this includes functions like EGR and Canister Purge as well as many diagnostic checks. The ECU also uses manifold pressure for baro and that is maxed by diagnostics at around 1.1 bar, over that the ECU flags a fault and drops to fail-safe operation using TPSvRPM.

On some ECUs its possible to disable the baro update and move the MAP sensor test to a higher pressure but then the MAP sensor runs out at about 1.3bar, on a good day (bad day?) you can get to about 4psi of boost OK.

What you need on any ECU you want to tune is a good definition, ie a file that tells you where the tables are in the memory. Which EFI sytem do you have at present, the Denso or the Mitsubishi? the Denso is very similar to the early Mazda MX5 ECU, I modified one of those to be reprogrammable and there is a guy in Finland getting 170bhp from the 1.6. It wouldn't be too difficult to modify the Suzuki ECU in the same way. Unlike the MX5 the Vit has a good diagnostic link and I found most of the tables in the Vit ECU and then looked for the equivalent in the Mazda ECU.

[twiss]

I don't actually have any fuel injection at the minute as I'm on an SU carb...

Which is partly why I thought about MS in the first place. If I went for the suzuki setup, on top on the intake manifold, I'd have to use the zook throttle body, ECU, loom, keys, etc.
... and I would still end up with a distributor

See I definitely want to get the injection working before going near a supercharger, as setting it up with an SU is a completely different matter!

I'd like something where I can change the kpa values above atmospheric pressure easily.
I anticipate a lot of tuning for the supercharger, and would like to keep one table NA and one charged, for example while i was tuning it... then if I needed to drive I could change the table, take the belt off and move the air filter.

[Rhinoman]

I'd like something where I can change the kpa values above atmospheric pressure easily.
I anticipate a lot of tuning for the supercharger, and would like to keep one table NA and one charged, for example while i was tuning it... then if I needed to drive I could change the table, take the belt off and move the air filter.

You don't need to do that. On a MAP based system the main tables/maps have Kpa and RPM axis, if you're not on boost then you won't be using the cells covered by that range of kpA values. A MAF based ECU uses tables that have AirFlow and RPM axis, off boost you will be hitting the cells that have less airflow.

[Rhinoman]

It would be fairly simple to modify a dizzy to drive a distributorless setup. Fit four hall effects and one vane and gate the signal from the ECU. I've considered that myself, I even bought four COP modules, no idea where they are now as I've got stuff in storage in three different locations for three years. I don't know if there is that much advantage, one day I'll do some tests. I looked at the Megasquirt coil analyser but it doesn't work with low impedance coils.

[twiss]

I've checked the MAP sensor I would use and it works up to 20psi, which is more than enough for what I want
Do you think it would hurt to move the MAP sensor into the engine bay? The MS design has the MAP sensor on the board (stupid) and you are supposed to run a vacuum hose to it... problem if the ECU is going to be in the glove box!
I was looking at running it out of the (stupid) DB-37 connector so I could have a short hose from a relay board under the bonnet to the engine
I don't know if it will cause problems having it at the end of a 3-4' cable

One cable im looking at saying resistance of 25ohms per 1000ft

i found my GFs 3-cyl skoda has logic level COPs that fit in my engine
i would quite like to not have HT leads!
for the timing signal I was probably going to put a timing ring on the cam pulley and just blank off the dizzy

[Rhinoman]

The MAP sensor can be mounted in the engine bay - it is on a Suzuki system anyway. The 2-bar MAP sensor gives an output of (approx.) 0->5V, a one-bar sensor gives an output of (approx.) 0->5V, the ECU doesn't know the difference. With a two-bar sensor it will read 2-bar as 1-bar which in the fuel and ignition tables isn't an issue but for other stuff it is; for example, the baro reading will be half what the ECU is expecting and you'll drop to fail-safe mode - you may not get a CEL for that.

[twiss]

ahh ok it works on voltage level not resistance... so the cable length won't matter
Thats at least 1 design flaw overcome lol

I don't think MS even supports a CEL haha

Just looked into it and its a 2 bar sensor that works up to 250kPa

I was going to say as well I'd power up the ECU with the sensors out of the car to make sure its all correct before using it... I think half barometric pressure in open air would be a bit obvious