10 replies to this topic

[fixitsan]

The need to meet emissions regulations has forced engine manufacturers to do their best to reduce fuel consumption. During an overrun condition, where you're traveling at perhaps 5000rpm and decide to slow down by closing the throttle, in most cases the ECU stops pulsing the fuel injectors. This is probably true for all fuel injected bikes, cars and vans made in the last 15 years or so.

 

The transition between pulsing and not pulsing can sometimes feel harsh when riding smoothly. My negative experience has been when riding around a couple of roundabouts, downhill, I back off the throttle a little to control the speed, but the fuel-cut process begins and I then I slow down more than I expected, unsettling the bike a little.

 

When riding using all the beans it's not so much of a problem, but if you're trying to take it easy with a pillion on board it makes the bike feel more nervous. Different manufacturers try to deal with it in different ways but ultimately a hard cut-off in fuel delivery is reached quickly, and I think some reports of modern bikes feeling more 'jerky', than carburetted bikes, can be due to the implementation of this system.

 

The system I'm working on aims to create phantom injection pulses during this fuel-cut period.in an attempt to reduce the negative effects and harshness of fuel-cut, by trying to provide a more carburetted feel !

 

Interfacing to the injectors is easy, one side of the injector has a constant 12V supply, and to force it to inject fuel the ECU pulls the other side of the injector coil to ground briefly

 

A master timing signal is required. Initially I wanted to use one of the phases from the alternator stator, which has 6 AC sinewaves per crank revolution, but that was electrically too noisy to be reliable. The crank position hall effect sensor was tested, which  a very clean AC pulsed signal once per crank revolution, but requires rectification and filtering. So I moved to using the Cam sensor, which is also called the 'cylinder identification' sensor by Yamaha. The neat thing here is the output from this sensor works on 5V logic level and outputs 1 small pulse per cam revolution, which equates to 2 crank revolutions

 

The two signals from the injectors are taken from the ECU switched side, and all 3 signals connect to a PIC microcontroller, through fusible resistors and optical isolators.

 

 

 

 

 

 

 

Attached Files

•  InjectCircuit.jpg   196.78KB   0 downloads

[fixitsan]

This short video shows the ECU injector pulses on the top yellow trace, and the blue pulses on the lower trace are my  computed ones.

Because the gearbox is in neutral there is no fuel cutting (yellow pulses are always present), because fuel cutting only happens when in gear with the wheels turning.

 

It shows I managed to get good alignment of when a pulse should happen, through computation , during deceleration.

 

The trickiest part here is that as engine revs rise the injectors start firing sooner and earlier in the crank rotation cycle

 

https://www.youtube....rts/KrCO2ASf2GQ

Edited by fixitsan, Today, 12:02 pm.

[fixitsan]

This is the spreadsheet I produced to help me derive the formulas needed to calculate when an injector should be fired.

 

I use one of the microcontroller's inbuilt timers to create timed pulses every few hundred microseconds. The cam pulse is the reference mark when the timer is started and stopped, the number of pulses is read, and the timer is restarted again from zero to capture the next series of timer pulses through one cam rotation.

As the cam increases RPMs the number of timer pulses are fewer. And conversely at lower revs it takes longer for a cam rotation so there are a greater number of timer pulse cycles per RPM of the engine.

 

InjA refers to injector A, and InjB is injector B

 

For each injector threre are two columns, one is for the measured delay (number of timer pulses elapsed) from the start of a cam pulse to the start of the injector firing. The second column for each injector is the calculated time period which is used by the software to delay before firing the phantom pulses

 

The numbers in both columns need to be within about 50 timer cycles of one another to get good accuracy and therefore smooth running. There is some scope in the software to tighten this up further but for now it seems to be good enough to go ahead and do some real world testing on the road.

 

 

EDIT, Column D, "InjB", should read InjA.

 

 

 InjectSpread.jpg   134.23KB   0 downloads

Edited by fixitsan, Today, 01:49 pm.

[LAURENCEAUX]

Wow -you're talking way over my head.

Edited by LAURENCEAUX, Today, 02:18 pm.

[fixitsan]

Wow -you're talking way over my head.

 

Ha, sorry.

 

I know a few other forum members have experience with microcontrollers so I threw it all in there.

 

Basically, it's a 'carb emulator'

[Boosh]

This sounds a fascinating project!

 

Could you achieve the same end result by shorting the neutral switch wire to ground (after it has been detached from the neutral switch itself, on the end of the gear selector drum, but that doesn't matter where).

 

Then run a new neutral switch-connected-wire; attached to the neutral light bulb on the dash (the bulb being previously disconnected from the 'factory' circuit) and run down the bike to the now-disconnected gearbox switch?

 

My suggestion is not an elegant 'technical' solution like yours, it's a bit 'ghetto' by comparison! But I wonder if it would work?

 

On my bike, I did once mention on here, I have fitted an extra brass contact in the gear selector drum, bridged to the existing neutral contact, so that 6th gear causes the neutral light to operate as a sixth gear indicator too. I never knew that I might be inhibiting the fuel cut off in that gear. It works well as an indicator, but a bit too bright at night for comfort.

 

I think in the car world, the advent of 'fuel cutoff on the overrun' was claimed to achieve overall fuel savings of around 5%, FWIW. On a bike? I don't know. I would trade that for the smoother ride!

 

The low speed jerkiness is the worst feature of my bike. I'm interested in how this turns out!

 

P.S. I appreciate my bike, as it is, could be accidently started with the gearbox in sixth, as opposed to only when in neutral. That's never come close to being a problem so far.

Edited by Boosh, Today, 05:24 pm.

[Bjørge]

You're on a high level, Fixitsan! I'm working on a solution to trick my Toyota diesel into thinking the dpf is sound and no need to cut power😎
A solution made up from 1 silisium diode, 1 germanium diode and 3 resistors...
Good luck!

[dapleb]

You are trying to emulate carbs.....cous they is better than injection.

This is why no one tries to emulate injection. ◉⁠‿⁠◉

[Bjørge]

You are trying to emulate carbs.....cous they is better than injection.

This is why no one tries to emulate injection. ◉⁠‿⁠◉

My Mk1 was out of business for 2 years because of carbs...I'll take injection surge any day😎

[fixitsan]

This sounds a fascinating project!

 

Could you achieve the same end result by shorting the neutral switch wire to ground (after it has been detached from the neutral switch itself, on the end of the gear selector drum, but that doesn't matter where).

 

Then run a new neutral switch-connected-wire; attached to the neutral light bulb on the dash (the bulb being previously disconnected from the 'factory' circuit) and run down the bike to the now-disconnected gearbox switch?

 

My suggestion is not an elegant 'technical' solution like yours, it's a bit 'ghetto' by comparison! But I wonder if it would work?

 

On my bike, I did once mention on here, I have fitted an extra brass contact in the gear selector drum, bridged to the existing neutral contact, so that 6th gear causes the neutral light to operate as a sixth gear indicator too. I never knew that I might be inhibiting the fuel cut off in that gear. It works well as an indicator, but a bit too bright at night for comfort.

 

I think in the car world, the advent of 'fuel cutoff on the overrun' was claimed to achieve overall fuel savings of around 5%, FWIW. On a bike? I don't know. I would trade that for the smoother ride!

 

The low speed jerkiness is the worst feature of my bike. I'm interested in how this turns out!

 

P.S. I appreciate my bike, as it is, could be accidently started with the gearbox in sixth, as opposed to only when in neutral. That's never come close to being a problem so far.

 

 

Wow, that's an elegantly simple approach, I might have thought along those lines myself, given another year or two. !

 

It could be taken in a different direction, by letting the neutral switch work as intended while stationary, but as soon as output shaft pulses/wheel pulses are detected , switch a relay over to ground it out, with the downside of it appearing to be in neutral when moving, but a two pole relay could also turn off the lamp.

 

I'm definitely going to check this one out, , it would be great if fuel cut was controlled just by the neutral switch !

You're on a high level, Fixitsan! I'm working on a solution to trick my Toyota diesel into thinking the dpf is sound and no need to cut power
A solution made up from 1 silisium diode, 1 germanium diode and 3 resistors...
Good luck!

 

I had DPF problem's with my Alfa 159. Instead of coring it out I realised you only needed to attack the rear of the filter to scrape it away a couple of millimetres because it's a wall-flow type of filter.

If it regenerated it would throw an error and make endless smoke, so I made an OBD dongle which counted the miles and when it reached 480 miles (20 miles short of the first regen) the dongle sent a code which made the ECU reset the DPF regen counter.

 

 

The lengths manufacturers have to go to just  to give hobbyists something to do is amazing !

[fixitsan]

This sounds a fascinating project!

 

Could you achieve the same end result by shorting the neutral switch wire to ground (after it has been detached from the neutral switch itself, on the end of the gear selector drum, but that doesn't matter where).

 

 

Three beers at least owed to you.

 

It works !

 

Fuel cut is active at all times, but, when the neutral switch wire is grounded it doesn't happen at all.

 

The neutral light is always on though

 

I won't get a chance to ride the bike with a grounded neutral wire till i come back at the weekend, but it's obviously the better solution.

 

Just need to add a switch on the bars to activate it !

 

 

Thats saved me hours of work !

 

 

Looking at the wiring diagram, in order to get a functional neutral lamp when stationary just needs a resistor and a relay to be added behind the dashboard to bypass the lamp, and some sort of intelligence to check for zero sprocket rpm.. Although the downside is that finding neutral while coasting to a stop would be impossible, but I dont think I ever do that anyway.

 

I'll try this out after the weekend and see if it feels much different when riding, when fuel cut is defeated

Edited by fixitsan, Today, 08:27 pm.