Injector timing

[icedvolvo]

Hi People,

I need to build my own injector module for a motorbike (why is complicated ...) ...

I am going to start with a piggyback module that just intercepts the pulses form the existing ECU and modifies the pulse width according to some basic inputs RPM and TPS.

However in the latter stages I want to extend that to intercept the other engine data and basically create my own ECU ...

So the question is:

1: does the start timing of the injector pule matter e.g. should it always start say 2ms before the intake cycle?

2: the theory says that injector pulses should be less than 80% of the cycle time but in reality can you go to 90% ... 100% ...

3: one of the problems we want to solve is the way the standard ECU cuts the fuel COMPLETELY during engine braking .. my simplistic strategy is to program our piggy back ECU to do something like:

if ECU_PULSE > 0 then OUR_PULSE = ECU_PULSE * modification factor (engine in power mode)
else ECU_PULSE = 0 then OUR_PULSE = MIN_ECU_PULSE (engine in braking mode but don't allow fuel to go to 0)

Is this a reasonable strategy ???

thanks for any assistance

[mk e]

1) Yes the timing of the pulse matters. Most (as in ALL) ECUs calculate based on angle and the chips designed for ECU use generally have angle clocks built in to easily calculate the correct time at any rpm ...like a co-processor to help unload the core.

2) you can set the injector pulse above 80-85% but you often don't get much if any more fuel...it depends on rpm and pressure and the injector itself. The reason is that it takes time for the magnetic field holding the injector open to decay and allow it to close, then it takes time for the valve to move, then the same on the opening and this time is already included in the ECU pulse width calculations as deadtime or whatever.....so by 80-85% the pulse width or the effective pulse width is already at or near 100%.

3) Building/programming ecu is not to easy and the solution you're proposing will probably not work very well. There are many aftermarket ECUs that have already solved all the problems, many use megasquirt, i'm found of enginelab.net because it allows (forces) me to program my own control logic but lots of option. If you're going to make something made have a look at one of the open source projects like rusefi.com

good luck.

[icedvolvo]

Many thanks for your reply

Timing: yes I wondered about that .. sensor knows when TDC is but I don't think bikes have an angle/cam sensor so I'm not sure they know whether its intake or power stroke but I guess in theory at least you want to inject as the air pulse/wave moves past the injector and of course the timing will change with RPM so there is a algorithm based on TDC and RPM when to open injector .. wonder how much difference it makes particularly at higher RPM when the cycle time starts to get closer to the injector time .. i.e. in a bike engine at 8000 rpm the cycle is only 7.5ms so the injector is probably open most of the cycle ...

OK got the reason about the 80% stuff. I forgot about the the LC constant of the electromagnet leading to open/close lag

I have to do this myself cos there is no commercial solution for this bike. I am planning just to use a cheap Arduino and a MOSFET with some simple programming in three stages.

- Stage 1 Initially it will just mirror the existing ECU pulses
- Stage 2 extend pules with RPM and TPS to fix the lean running at low TPS
- Stage 3 monitor pulses and RPM to prevent ECU cutting fuel on engine braking

Thanks again for you info, every bit helps

[mk e]

Have a look at speeduino

Careful with mosfets....I'm not an electronics person but remember being warned about fly back voltage. Look and see what the rusfi or speeduino guys are doing.

Timing stops mattering as the pulse length gets longer and approaches full on. It matters most at low rpm/power.

Step 1 is always be certain about what signals you have....I can't believe it's a simple tdc sensor on the crank, more likely is a pattern wheel that gives the ecu both rpm and the ability to calculate tdc. With no cam sensor you with be running waste spark and probably semi-sequential injection.

What 8w the issue with the OEM fuel cut? Is this a roadrace application?

[icedvolvo]

Thanks for your reply.

Thanks for the speeduino hint, I was hoping someone was doing this

Yes will have to protect the tranny (whatever it is) from inductive kickback ...

As to timing I thought that would be the case, as the RPMs go up I understand that at high RPMs the "sound waves" created from the intake pulses start to blend over each other and hence so can the injector pulses, I suppose that's why direct injection is a better tech

This is a motorbike so its pretty simple! As far as I can see physically and wiring diagram the only position indicator is TDC sensor from the crank ... Im not sure what sequential pulsing means but I will put a scope on the existing one soon and have a look at what happens ...

This is a bike which has been nobbled by compliance to Euro V emissions rules .. which might be fine for a multi cylinder engine but for a big single it just trashes it, makes it almost un rideable. There are two problems to solve:

1: the bike runs so so lean to keep the catalytics hot at low throttle that it is gutless and jerky and dangerous especially in the wet

2: on engine braking the ECU cuts the fuel completely but if you "blip" the throttle as you do on down change on a bike then it stalls cos there is absolutely NO fuel at the instant u crack the throttle again.

This is a adventure bike ... its an obscure brand but becoming very popular .... and I like playing with bikes and electronics so I thought I would try to make a piggy back unit to fix the problems ...

Thanks for your reply and info .. appreciated ...

[Circlotron]

2: on engine braking the ECU cuts the fuel completely but if you "blip" the throttle as you do on down change on a bike then it stalls cos there is absolutely NO fuel at the instant u crack the throttle again.

Maybe include in your build a microswitch on the clutch lever to tell the ECU to start injecting again before the clutch is fully disengaged.

[englertracing]

Microsquirt and an adapter to the stock harness, develop the tune, sell the kit.

[icedvolvo]

2: on engine braking the ECU cuts the fuel completely but if you "blip" the throttle as you do on down change on a bike then it stalls cos there is absolutely NO fuel at the instant u crack the throttle again.

Maybe include in your build a microswitch on the clutch lever to tell the ECU to start injecting again before the clutch is fully disengaged.

Thats clever, maybe even tweak the tps so the ECU doesn't think its been shut down ... hmmm .. thanks

[icedvolvo]

Microsquirt and an adapter to the stock harness, develop the tune, sell the kit.

Thanks .. I wil l have a look at microsquirt ... I was sure someone would have something similar ...

[mk e]

It's a bike so be sure whatever you do is watertight....I think MS3pro is the only ms product that would meet that need, last I looked anyway. Certainly you could repackage it but just something to keep in mind

[Rick!]

Dynojet Power Commander is also an alternative. They have a unit for KTM singles so that might be an option to keep one from reinventing the wheel. The interface connector may be the tricky part.

http://www.powercommander.com/powercomm ... 32&yr=3633

[adam728]

This is a adventure bike ... its an obscure brand but becoming very popular .... and I like playing with bikes and electronics so I thought I would try to make a piggy back unit to fix the problems ...

CSC?

[Walter R. Malik]

At low engine speeds the O.E.M. companies time their pulse shot of fuel to happen after the exhaust valve closes from overlap so, it all goes into the cylinder and none is wasted.
At higher engine speeds the higher duty cycle time with the injector does not allow this.

[icedvolvo]

This is a adventure bike ... its an obscure brand but becoming very popular .... and I like playing with bikes and electronics so I thought I would try to make a piggy back unit to fix the problems ...

CSC?

SWM

[icedvolvo]

Dynojet Power Commander is also an alternative. They have a unit for KTM singles so that might be an option to keep one from reinventing the wheel. The interface connector may be the tricky part.

http://www.powercommander.com/powercomm ... 32&yr=3633

Yes I have been speaking with Dynojet technical people. They have a unit for a closely related engine (Husqvarna TE630) but the plugs and sensors are not compatible and the unit is now obsolete ... I did wonder about adapting this unit but without proper access to the programming it would be very difficult and probably easier to do my own ...

If Dynojet would allow me access to the full programming and code I would be happy to develop this for this SWM bike but they are just not interested .. shame SWM is becoming a very popular brand ...

All these Euro V compliant bikes have the same crap problem with running too lean!!! The ECU (an Athena GET) is encrypted and locked and Athena and SWM have both said they will not do a race version of the ECU firmware!!

So I have to do my own .. The above info was really good re timing but if I knew what sort of circuits the ECUs use I could do a really quick one now with an Arduino board ... so what do they use for the injector switch: MOSFET, BIPOLAR, SCR ??? How do they protect from inductive kickback diode/resistor/coil .. in other words what do the injector switching circuits look like ??? So far I have not been able to find out ...