Ecotrons Engine Management

[TZ350]

Because someone asked.

There are three physical injectors but there are actually only two "logical" injectors, one big one small.

The two "physical" injectors in the B ports are there for symmetry of fuel distribution and are fired together as one Logical injector.

I use three physical 124g/min injectors but the EFI CPU sees only two logical injectors, a 124g/min slow speed injector and one 248g/min high speed injector.

In broad terms, because there is much the same amount of air consumed per revolution. Both 124 and 248 Logical injectors each deliver much the same amount of fuel per revolution.

Attachment 332298

Both 124 and 248 injectors deliver much the same amount of fuel per revolution.

Because the 124g/min injector at lower speed has 10ms to do the job.

And the 248g/min injector at higher speed only has 5ms to do the job so needs to be twice as big.

Time available dictates the relative injector sizes.

[Drew]

Because someone asked.

There are three physical injectors but there are actually only two "logical" injectors, one big one small.

The two "physical" injectors in the B ports are there for symmetry of fuel distribution and are fired together as one Logical injector.

I use three physical 124g/min injectors but the EFI CPU sees only two logical injectors, a 124g/min slow speed injector and one 248g/min high speed injector.

In broad terms, because there is much the same amount of air consumed per revolution. Both 124 and 248 Logical injectors each deliver much the same amount of fuel per revolution.

Attachment 332298

Both 124 and 248 injectors deliver much the same amount of fuel per revolution.

Because the 124g/min injector at lower speed has 10ms to do the job.

And the 248g/min injector at higher speed only has 5ms to do the job so needs to be twice as big.

Time available dictates the relative injector sizes.

Does it work properly yet?

The ecotrons seemed to be promising early on, but from what I read here the system is shit and a complete balls up as far as software goes.

Looks like there's a checklist as long as your arm that needs to be checked thoroughly every time there's a miss.

[TZ350]

Does it work properly yet?

Narr ... not yet.

Here is a run of torque curves from my Suzuki GP125 taken at WOT - 90% - 80% - down to 20% throttle.



Below 20% (Red line) is where I am having my problems. Not sure why but probably more a 2T with EFI thing that I don't understand yet than the Ecotrons system itself.

[TZ350]

I did some further testing on the ion sensing system, choking the inlet to induce a misfire... I expected the measured ion current to drop to zero like it does with a four stroke... of course it doesn't! I think because the exhaust gas isn't purged from the cylinder every cycle like in the four stroke, the ions from the combustion are still there for the next cycle, whether there's fire or not, which seems obvious now. Nothing is ever clear cut with the two stroke cycle.

So next step will be to make a circuit to capture the pressure peak each cycle, then feed it into the ECU for datalogging, I should be able to detect misfire with that method.

Very interesting ......

[Drew]

Two stroke is always gonna read rich from measuring anything the exhaust I would have thought. Diagnosing a miss from ecu data would be a bloody nightmare.

That whole paragraph should be read as a question. I have no idea about it all, it's just ramblings.

[TZ350]

Two stroke is always gonna read rich from measuring anything the exhaust I would have thought.

Yes, always a bit rich but when it goes lean, like you say, it is hard to tell if it is really lean or so rich that its not firing and the O2 sensor is just seeing the wasted oxygen. I am still learning about this beast.

[TZ350]

Attachment 332511

The Ecotrons EFI Alpha-N map load increases in 16 steps. So some time ago I used a flow meter to find the 16 TPS positions that equated to 16 even increases in airflow.

The curve followed this throttle plate curve I found on the net. I guess the offset is due to my curve starting at a slightly open position for the idle air.

Attachment 332512

Interesting that 50% of airflow happens at 20% TP throttle position. The green shaded area is the area I am having most trouble with getting the engine response with the EFI Alpha-N map right.

Attachment 332513

A series of dyno runs from 100% throttle down to the 20% green line. Also interesting that 50% air did not make 50% power. I don't understand that, I would have expected 50% air would equal 50% power.

The areas below 20% throttle represent a very small area on this collection of graphs but I have found it to be a huge problem on the track with throttle response. They also represent half the steps on the Alpha-N map.

When everyone knows big power is at WOT who would have thought that so much importance was in that first bit of barely opening the throttle.

There are always losses in friction and heat transportation so the curve of powergain is seldom equal to air usage.

Back when programmable EFI was in its infancy they used to save room and data and more importantly time, by using the multiple smaller steps only in problem areas if that makes sense.
ie they did a basic map much like a 4 circuit carb pilot, cutaway, needle, main.
then used the rest of the 16 steps to full in the areas where it stumbled
smaller engines and single and twins and rotaries were far less forgiving then 4 and 8's.
I remember Riley Will saying something on Pitlane about someone (Aprilia i think) basically plotting the flow of a carb and then recreating it with the efi, to get rid of the twitchiness.

............... EFI

[FastFred]

Attachment 332512

Interesting that 50% of airflow happens at 20% TP throttle position.

Seeing it mapped out like that sure makes it much easier to visualize.

[TZ350]

Attachment 332640Attachment 332639

Below is an extract from DirtBike that talks about the EFI KTM250. Interesting it mentions exactly the same problems with throttle response that I am having and that it mentions a crankcase pressure sensor.

https://dirtbiketest.com/fresh-dirt/two-stroke-fuel-injection-ktms-path-to-tpi/#ote9PlJBGtOBGJI6.97

Now using accepted and understood parts it all came down to processing the information and getting the mixtures right. Quickly. The window of time for the injector to be open and getting the large amount of fuel a two-stroke needs is the challenge, especially in even smaller displacements we are told. Plus when a two-stroke has a “flame-out” inside the combustion chamber, it can take up to 50 strokes for the engine to recover. This was a big issue in high RPM, low throttle times. For instance when you are going down a hill with the engine revving and then you give it a little gas. Making the throttle response as good as that antiquated carburetor was very difficult. There is not a lot of time for the computer to decide how much fuel and when to shoot it from one stroke to the next since the conditions inside the two-stroke engine can change drastically and very suddenly. Even small things like having the engine’s crankcase pressure sensor and fuel injectors mounted so that vibration was not affecting them were lessons learned in testing. Luckily the new 2017 engine was designed from the beginning to eventually have a non-carbureted fuel system so it would vibrate less and also be able to produce as much electrical power as would be needed.
Read more at https://dirtbiketest.com/fresh-dirt/...1B4YRrV2OU2.99

Attachment 332641Attachment 332642

The crankcase pressure sensor seems to be taken off the back of the cylinder and go up to a sensor in a plenum behind the motor.

Very interesting .....

[Drew]

From skimming that article, I didn't see the even bigger improvement KTM made to their new 2 stroke.

Look up why they put a counter shaft in the new bike.

It's absolutely fucking insane the issues (including snapping rear shock shafts) From the vibration a high revving two stroke creates.

You're a clever cunt TZ (TZ got auto corrected to RD and I feel compelled to make old age jokes), so start considering balancing that shit up.

[TZ350]

You're a clever cunt TZ (TZ got auto corrected to RD and I feel compelled to make old age jokes).

Thanks, and when I was a little RD I always wanted to be a TZ when I grew up ...

[Drew]

Thanks, and when I was a little RD I always wanted to be a TZ when I grew up ...

I bring it up because KTM found that the rebound port on the shock had to be made perpendicular to the crank rotation, or the vibration would stall the fluid in the port.

One can see the potential for fuel to do the same thing on the way I to an injector.

[TZ350]

That is a very interesting point. Thanks.

[TZ350]

Attachment 332723

Red line is throttle, Blue is rpm and Yellow LamWO2 sensor.

Same rpm and load/fuel injection amount at both cursors.

Nothing happens until the engine slows down to about 7k rpm where the pipe starts to work against the engine. Then it fires and Lam goes from lean to near right. Then there is a period of rough running pulling out of the pipes reversion hole. Then takes of at Lam 0.88 as it gets on the pipe.

The puzzle is, why does it fire when the pipe works against the engine and not while the pipe is in tune with the rpm as the motor is slowing down? ... same throttle and fueling both sides.

TZ does your ecu change fuelling using feedback from the Wideband at all?

No

I'd like to see injector pulse width in that graph as well

I should be able to do that.

Are you saying the first "lean" section after throttle opening is only lean because it didn't fire and therefore could actually be too rich fuel/air to fire?

Yes I think you are right. It looks lean but it is too rich to fire. But why????

Same rpm, same throttle position, different airflow.

Above 7k RPM the on-pipe and off-pipe air flow is too different, the on-pipe fuelling (what you've tuned for) will not support combustion off-pipe. Way too rich, true lambda probably around 0.5 or worse.

As the engine speed falls below 7k rpm, the on-pipe and off-pipe air flow is not so different, true lambda is approaching the rich limit until it supports combustion.



Once combustion begins the small increase of airflow from the pipe effect will bring the lambda above the rich limit, 'clearing' the cylinder.

Yes Nath I think you and Brett are right, what you say makes sense. Thanks.

[TZ350]

I'd like to see injector pulsewidth in that graph as well

Attachment 332767

Pink line is injector pulse width. Blue RPM and Brown is wide band O2 sensor.

WOT, same RPM, same injector pulse width, different Lambda.

I think Nath88 is onto it.

Same rpm, same throttle position, different airflow.

Attachment 332725