[electroguy]

I have a Royal Enfield Classic 500 EFI Engine. I ruined its ECU once while doing some testing. I have a bit of experience with building electronics and thus I decided to design my own custom ECU for that, would be a good learning project. So I started reading some books, Including the service manual before getting started.

I got stuck while figuring out a very basic fundamental. The engine is equipped with a Crank Position Sensor, with 30-1 teeth, the missing tooth denoting the TDC. But surprisingly, there is no camshaft position sensor .

There are two cases for a TDC, first, just before the intake stroke, and second, just after the compression stroke. Both of these cases should be differentiated from each other, and the camshaft position sensor provides this reference, which is not present in this engine.

Now my question is, how come the ECU know that which TDC indicates what stroke? How does it decide whether to turn on the injector or fire the spark plug?

[im_srini]

Originally Posted by electroguy

I have a Royal Enfield Classic 500 EFI Engine.
The engine is equipped with a Crank Position Sensor, with 30-1 teeth, the missing tooth denoting the TDC.
But surprisingly, there is no camshaft position sensor .
How does it decide whether to turn on the injector or fire the spark plug ?

Hello,
I'm no engine expert, but could it be that the spark-plug is fired at TDC ( factoring in ignition advance ) irregardless of which stroke the engine is in ?
For example, the F10D engine in my Alto uses a 'wasted spark' design.
The spark plug is fired at, or before TDC, in the exhaust stroke as well - only thing is that it has no effect.
With respect to fuel-injection, I remember reading in the Mega-Squirt documentation relating to sequential & batch injection that timing precision of the fuel spray is not very important as the fuel droplets sit behind the intake valve until it opens.
It's merely enough if you get one activation of the injector prior to the intake valves opening.
Anyway, will leave it to the engine & Bullet experts to explain things better.

[electroguy]

Originally Posted by im_srini

Hello,
With respect to fuel-injection, I remember reading in the Mega-Squirt documentation relating to sequential & batch injection that timing precision of the fuel spray is not very important as the fuel droplets sit behind the intake valve until it opens.
It's merely enough if you get one activation of the injector prior to the intake valves opening.

Yes, the firing of spark plug is not a problem, and the timing precision of the fuel spray is not very important, but the fuel should not be sprayed twice in a cycle, precisely, it is of utmost importance for fuel injection system to know when the intake valve opens.

[Sutripta]

Originally Posted by electroguy

but the fuel should not be sprayed twice in a cycle,

Why will it be sprayed twice in a cycle?

Regards
Sutripta

[im_srini]

Originally Posted by Sutripta

Why will it be sprayed twice in a cycle ?

Being a tad forward here, but I'm guessing the reasoning of 'ElectroGuy' is that since there's no camshaft-position-sensor & only a crankshaft-position-sensor in the engine in question, it's going to detect TDC for each rotation of the crank, hence if the ECU decides to 'fire' the fuel injector based on TDC, there would be 2 of such activations.

Originally Posted by im_srini

With respect to fuel-injection, I remember reading in the Mega-Squirt documentation relating to sequential & batch injection that timing precision of the fuel spray is not very important as the fuel droplets sit behind the intake valve until it opens.

Finally found it, here's the URL for anybody's who's interested.

Quoting the section:

...
However, sequential injection does not necessarily mean you are injecting into an open intake valve all the time.
The intake valve is only open less than 30% of the time in a typical 4 stroke engine.
Once you are trying to produce more than about 25% of maximum HP your injectors are firing for longer than the intake valves are open.
If your maximum HP is correctly calibrated to a safe 80% duty cycle, your injectors are injecting well over 50% of the time on closed valves.

At higher rpms, it becomes increasingly difficult to inject while the valve is open.
For example, if your Req_Fuel = 15.0 ms, & your maximum duty cycle is 85% then the interval between injections cannot be closer than 17.6 ms.
The time available to inject during the entire 4 stroke cycle is:

timecycle = 120/RPM

The intake is typically open for less than 240° of 720° in an engine cycle in a hot street engine, about 1/3 of the time for a complete cycle.

So : timeinject = 40/RPM

&

RPMmax = 40/timeinject

Using the numbers above, RPMmax = 40/0.0176 = ~2300 rpm

Above this RPM, it is not possible to inject the entire amount fuel through the open valve at 100% VE & 100kPa.
At idle, however, VE may be 30%, & kPa might be 35 kPa, yielding a pulse width of ~1.8 milliseconds, which certainly could be injected during the intake valve opening.
So sequential injection is primarily effective at idle, & not much different from batch injection as pulse widths get larger at higher engine speeds & loads.

In any case, when fuel is injected while the valve is closed, it will simply stay in the port until the valve opens.
In some cases, this period of time may allow enough heating of the mixture to better vaporize the liquid fuel, improving efficiency & emissions.
...

.

[electroguy]

Thank you srini, this was very helpful, and I think I can proceed now with the programming, the doubt seems to have cleared. What I will do is, I will fire the spark plug on every TDC (keeping the advance in mind), and will inject the fuel every second TDC. What valv is open should not matter. Right?

[Sutripta]

Originally Posted by im_srini

Being a tad forward here, but I'm guessing the reasoning of 'ElectroGuy' is that since there's no camshaft-position-sensor & only a crankshaft-position-sensor in the engine in question, it's going to detect TDC for each rotation of the crank, hence if the ECU decides to 'fire' the fuel injector based on TDC, there would be 2 of such activations.

ECUs do what they are programmed to do, so

Originally Posted by electroguy

and will inject the fuel every second TDC.

is on the right track.

Anyway of distinguishing between say inlet and power strokes? (Esp. in a single cylinder.)

Regards
Sutripta

[Whiplash7]

Originally Posted by electroguy

What I will do is, I will fire the spark plug on every TDC (keeping the advance in mind), and will inject the fuel every second TDC. What valv is open should not matter. Right?

Originally Posted by Sutripta

Anyway of distinguishing between say inlet and power strokes? (Esp. in a single cylinder.)

Offcourse what you said should work but if you have enough resources in your control unit you could do the following to detect which TDC corresponds to the compression stroke. Normally during the power stroke I would expect an increase in the rate of change of the crankshaft rotation which the control unit should be able to identify. Using this I think you should be able to differentiate the TDC thus injecting and sparking only when required.

[im_srini]

Originally Posted by Sutripta

Anyway of distinguishing between say inlet & power strokes ? ( Esp. in a single cylinder )

Originally Posted by Whiplash7

Normally during the power stroke I would expect an increase in the rate of change of the crankshaft rotation which the control unit should be able to identify.

Great idea !
And I suppose you've answered Sutripta's question too

ElectroGuy, it would be great if you could detail here the design of your home-built ECU as you go along, particularly the software architecture & design.
.

[Sutripta]

^^^
I wonder if this method is going to be reliable. Afterall, a flywheel is there to specifically mask this.

Any other strategies?

Regards
Sutripta

PS - The RE ECU syncs in one cycle. Necessary so that the kickstarter works. RDs ECU needs two cycles, so needs the electric starter to start. Just a thought.

[Whiplash7]

I know its hard to imagine. But I'm sure even with the flywheel there will be some variation also due to the fact that there is only one powerful cylinder in the engine.

He could inject and spark at both TDC's until sync is achieved.

[Sutripta]

^^^
I don't think imagination (or lack thereof) is the issue.
I feel some other method is used to achieve sync.
Race Dynamics sometimes participates on this forum. Maybe he can shed some light.

Regards
Sutripta

OT. - Before starting how does one decide how much fuel to inject per squirt?

[electroguy]

Originally Posted by Sutripta

^^^
OT. - Before starting how does one decide how much fuel to inject per squirt?

The fuel maps serve this purpose. I am searching for RE fuel maps but have been unlucky. So what I will do now is I will calculate the pulse width theoretically, start injecting the minimum fuel, and check if the engine starts. I just want to start the engine with my ECU, until I find the fuel tables.

Originally Posted by im_srini

ElectroGuy, it would be great if you could detail here the design of your home-built ECU as you go along, particularly the software architecture & design.
.

For now, I just have an ATmega16A, 2 breadboards, a few IRF550 MOSFETs, an L293D as gate driver. I am quite busy with my academics, therefore I am finding it a bit hard to work on it, but I will surely complete it someday,

[Whiplash7]

Originally Posted by Sutripta

^^^
I don't think imagination (or lack thereof) is the issue.
OT. - Before starting how does one decide how much fuel to inject per squirt?

Then why else do you think it won't work

Originally Posted by electroguy

The fuel maps serve this purpose. I am searching for RE fuel maps but have been unlucky. So what I will do now is I will calculate the pulse width theoretically, start injecting the minimum fuel, and check if the engine starts. I just want to start the engine with my ECU, until I find the fuel tables.

I would check the amount of air available (air mass sensor in intake manifold) and inject fuel in the ratio of 14:1. If required you should provide a rich mixture for starting based on temperature for a cold start.

[electroguy]

Originally Posted by Whiplash7

I would check the amount of air available (air mass sensor in intake manifold) and inject fuel in the ratio of 14:1. If required you should provide a rich mixture for starting based on temperature for a cold start.

Yes, that's also an option in case I don't get fuel maps. Any idea btw where I can get the maps for Classic 500 EFI?