LocostBuilders - powered by XMB

locost MS

02GF74 - 20/6/06 at 08:47 AM

well I got m'self some bike TBs and planning ahead for fitting them.

Thing of using PIC (yep back to that again) or two to run them as follows.

The timing will be run separately from the fuelling, maybe by having its own controller. As there will be a crank position sensor, this will act as a dizzy with fixed timing map plus a vacuum sensor will apply and advance, much like a distributor vacuum advance unit.

I haven't done enough reading, but the fuelling will be dependent on the output from an oxygen sensor in the exhaust - petrol will be supplied so this is in the correct range and there will be adjusted to be richer (?) when the engine is cold i.e. input from a water temp. sensor.

Is this a workable system or not?

martyn_16v - 20/6/06 at 09:13 AM

quote:
Originally posted by 02GF74I haven't done enough reading, but the fuelling will be dependent on the output from an oxygen sensor in the exhaust - petrol will be supplied so this is in the correct range and there will be adjusted to be richer (?) when the engine is cold

So are you planning on using closed-loop feedback from the O2 sensor to run the engine in all conditions? This is a bad idea - a sensor failure will kill the engine pretty quickly at high load, and O2 sensors do like to fail. To be able to control the fuelling with any kind of accuracy it would also need to be a wideband sensor, which needs some fairly tricky control of the pump cell current in order to make sense of it's output.

There are several other operating conditions that will require enriching (or leaning) of the mixture, such as acceleration and deceleration enrichments, and cranking needs fuel dumping in.

Personally i'd just go for Megasquirt. Everyone may be doing it these days but it is a competent system and you'll struggle to build an effective working system from scratch for much less. If you have PCB making facilities you could make it even cheaper than normal, secondhand HC08 processors loaded with the MS firmware should be available for pennies if you ask around now that people have been upgrading them to MS2, and the components themselves are cheap as chips.

[Edited on 20/6/06 by martyn_16v]

paulf - 20/6/06 at 12:18 PM

I did the whole installation for Ł200 , I built the megasquirt and simulator from scratch but had to pay Ł50 for the PCB and processor, i could have programmed my own HC08 but had to buy it with the board and map sensor, the other parts I sourced from RS and Farnells.
I dont think it would be very feasible to build from scratch without lots of development time etc, when you look into what is involved with the megasquirt firmware and programming and logging software it would take a lot of time and effort to get even a basic system working.
Paul

quote:
Originally posted by martyn_16v
[
Personally i'd just go for Megasquirt. Everyone may be doing it these days but it is a competent system and you'll struggle to build an effective working system from scratch for much less. If you have PCB making facilities you could make it even cheaper than normal, secondhand HC08 processors loaded with the MS firmware should be available for pennies if you ask around now that people have been upgrading them to MS2, and the components themselves are cheap as chips.

[Edited on 20/6/06 by martyn_16v]

tks - 20/6/06 at 12:53 PM

You should use a basic map.

i agree with the thing that you could use the wide band sensor to give you tuning advice!

but to rely on them and to adjust the mixture on the fly?

i also thingk that the sensor need time to react to changes in the mixture..

and thats where the basic mapping is for.

also i wouldn't use 2 seperate systems for this because the things are close to each other.... and to communicate there isn'tmuch time for that.

Tks

02GF74 - 20/6/06 at 01:02 PM

quote:
Originally posted by tks
also i wouldn't use 2 seperate systems for this because the things are close to each other.... and to communicate there isn'tmuch time for that.

communicating? I envidage a sikmplistic system that replaced distributor, 1st PIC and other the carb, 2nd PIC.

gong back on what has been said, then the fuel uses the throttle position, vacuum and tempeerature (like carb) - I was thinkingt he the oxygen sensor would be simpler.

Oh well. need to read up more about this stuff.

MikeRJ - 21/6/06 at 01:55 PM

The number of inputs and the calculations required to get the correct injector pulsewidth are surprisingly complex (they are all listed on the MS website). The hardware to drive injectors is also not trivial, especialy if you are using low impedance injectors.

A simple igntion controller is quite doable in a PIC, but I think you are letting yourself in for a whole world of pain making a fuel computer from scratch.