ESE's works engine tuner

[husaberg]

Initially I thought these were rich areas and to get it running nicely all I had to do was carefully adjust the corresponding parts of the map.

But I now think the problem is with the injection on time being much less than the transfer port window open time and consequently poor fuel homogenisation.



Attachment 321109Attachment 321108

I have a butterfly type throttle body, it will break my heart if I have to discard that ball valve throttle. But I guess you are right, and in the interests of science I will have to try it.

Those Cagivas were throttle body injected, so I don't think that's the full answer.
I think massive amounts of injector pressure is couple with small injectors with their light and super quick responding pintles are though.

You could also do a base line fuelling with something like a old cold start injector which were not pulsed at all.
These crude injectors were used as a form of a choke on early 80's cars.This down and dirty system was used on modded turbos cars up to the nineties pre aftermarket programmable ecu's.
This could allow for you to use the smaller injectors while still getting enough fuel and time.

Failing that you could always try to set up a skip a stroke system with the bike set up as a 180 twin with the ECU controlling alternate strokes. this would by some time and allow some additional fuelling overlap.

[TZ350]

you could always try to set up a skip a stroke system with the bike set up as a 180 twin with the ECU controlling alternate strokes. this would by some time and allow some additional fuelling overlap.

Clever idea, for slow speed, use two injectors half the size so as to cover twice as much time.

The irony is that we thought that the injectors would not be quick enough, so short on injection time at high speed and power but that is all good and the real problem turns out that at low speed we have to much spare time. LOL

It might be possible for the digital ignition at low speed to turn of the EFI's slow speed injector and to turn on a very small injector that flows 100% of the time like the starting injector you posted. Ecotrons has very small injectors suitable for model aero engines, one of those might do.

...... I thought I was beaten but people have come up with lots of good ideas.

We might get this thing working yet, thanks.

[breezy]

Yes...

In the Ecotrons EFI software there is a way of mapping (seperate to the fuel map) the injection end point.

Attachment 321118Attachment 321117Attachment 321116

As best as I can tell, in the manual the picture is a 4T map and the Injection end point is so many degrees before TDC on the next firing stroke.

For me it would be something like 120 degrees if I wanted to finish the injection squirt as the transfers close.

I have tried 90 - 180 - 120 - 270 and 360 before TDC, and 180 ( or BDC) worked the best.

wouldnt you think that as the fuel held in the transfer tunnels, is what is actually pulled into the cylinder when the transfers open, to charge the cylinder , that this would be the place to inject the fuel? some of the transfers would supply air only and injected transfers a mixture of both or even closing a couple of transfers off from the crankcase and using them to supply fuel only.

[wobbly]

The problem with that idea is that you will end up with differing A/F ratios within the scavenging streams.
This will then in effect produce almost a stratified charge scenario within the compressing mixture.
Unless you can do a heap of CFD or try an endless combination of air, or fuel, or both in different places, the chances of fluking
the best ( or even a good ) charge distribution is near zero.

[ken seeber]

Didn’t Harold Bartol ages ago have a hybrid style system, essentially a carb covering all the low load/speed operation and then an injector to do all the high speed stuff? All as a result of not being able to emulate the metering and fuel distribution/mixing characteristics of a carb at the lower settings. Obviously when he was doing this the computing power, software and injectors available would have been crude to what is available now.

As Wobbly says, there would be all sorts of stratification going on with just injecting into only one pair of transfers. In saying this though, Fletto seemed go get his going and the Athena was the same I think.

Maybe 5 injectors is the go.

[ken seeber]

The 5 injectors concept was a bit of a joke, However, what about drilling a hole thru the material between the A & B passages, the hole being directly in line with the injector axis? This would allow some of the spray to enter the A passage, possible minimising the stratification.

[husaberg]

Didn’t Harold Bartol ages ago have a hybrid style system, essentially a carb covering all the low load/speed operation and then an injector to do all the high speed stuff? All as a result of not being able to emulate the metering and fuel distribution/mixing characteristics of a carb at the lower settings. Obviously when he was doing this the computing power, software and injectors available would have been crude to what is available now.

As Wobbly says, there would be all sorts of stratification going on with just injecting into only one pair of transfers. In saying this though, Fletto seemed go get his going and the Athena was the same I think.

Maybe 5 injectors is the go.

I am pretty sure the system Harold employed on the KTM's was a carb covering all the fuelling other than an injector augmenting the overrun to stop it seizing.
It was on either pitlane or Frits or Wob spilled. Riley Will said pretty much the same thing as well.
http://www.kiwibiker.co.nz/forums/at...4&d=1425089570
http://www.kiwibiker.co.nz/forums/at...9&d=1425089727
http://www.kiwibiker.co.nz/forums/al...p?albumid=4857

KTM's 125 FRR features a unique and exclusive fuel injection system that is not designed to replace, but instead to work in conjunction with the carburettor. It allows fine tuning of the fuel supply and lubrication under extreme circumstances -- such as over-revving whilst back-shifting with closed throttle -- and it does it with unparalleled precision. The engine can be set up to run lean at any rev range and under any load, without the dangers of piston seizures.

With the right set-up, the KTM engine is not only very fast, but also strong in acceleration, with smooth, easily controllable power delivery. "Riders aren't robots. And if the power kicks in too aggressively, they will feel a risk factor at the exit of each turn. It's like shying at an abyss. They will open the throttle later and more cautiously, and inevitably lose time,"� explains the Red Bull KTM 125 Team. "Therefore, we won't stop working on our carburetion until our engines run like electric motors, with a very linear and predictable power curve."�
http://www.roadracingworld.com/news/...prix-of-china/

Another snippet from Riley Will was.
Riley later says he timed the injector to squirt 10deg before inlet open and finish 10deg after inlet close (this was for his own mule engine not the KTM)

[TZ350]

wouldn't you think that as the fuel held in the transfer tunnels, is what is actually pulled into the cylinder when the transfers open, to charge the cylinder , that this would be the place to inject the fuel?

Yes I certainly did think so too and I thought the injection cycle should be timed to match the transfer window, but I was wrong. The engine seems to run better the longer the injector fires for. Which can be for way more than the transfer period.

I have drawn myself a picture so I can visualise it.

With transfer port injection it looks like timing is critical and the length of the injection cycle has to match the period that there is bulk outward air flow in the transfer duct and I guess this will be significantly shorter than the transfer open period.

At 12,000 rpm the transfer is open for 1.8ms and looking at the Ecotrons run data the injectors at 12,000 rpm are open for 1.9ms so we are pretty stuffed for much more progress.

Maybe I think its time to change tack and start thinking about Injection into the inlet.



Transfer injection on an engine that revs to 8-9k would work very well (Flettner has done it on a big Kawasaki single) and with careful timing you could get some fuel efficiency's but the much reduced time available at 12k rpm to get things done. Lack of time makes it hard to do on my little 125.

The Injector is timed to fire when the transfer is open. But the areas on the map that are giving me trouble are areas that the injector is firing for less than the transfer open time.



The engine can make good power and in this test, at 12,000 rpm the injectors are firing for 100% of the total time available ie 360 deg of crank revolution.

[TZ350]

what about drilling a hole thru the material between the A & B passages, the hole being directly in line with the injector axis? This would allow some of the spray to enter the A passage, possible minimising the stratification.

My side injectors are aimed across the transfer ports to hopefully get some fuel into both the A and B ports.



Hard to do on an air cooled motor. But I would have prefered to have my injectors vertical and spraying against the air flow as I think this is the best way to get good mixing.

I think this is Flettners YZ250 cylinder. It only has injectors in the B ports but it runs Ok, in fact very well. The difference between a YZ250 and my 125 is that the YZ is running E85 and less than 9,000 rpm. Above 9K seems to take 2T EFI onto another level of complexity.

[Grumph]

E85 is very misleading as far as setting up injectors goes. Any Methanol or Ethanol base fuel can be run very rich without adverse effect on response.
Look at how long the Hilborn constant flow system has lasted for methanol use - crude metering but it works.
As an aside i'd remind you that Irving suggested constant flow via one mainshaft and the big end....

[ken seeber]

OK, I was looking at this pic.

[TZ350]

OK, I was looking at this pic.

Yes that setup was pretty crude.



The current setup has the side injectors splayed out a bit and the middle one aimed up under the piston.

[Moooools]

Yes that setup was pretty crude.



The current setup has the side injectors splayed out a bit and the middle one aimed up under the piston.

I have a question in the midst of all this talk of injector positions:

Why not start with the injector where the carb would normally be situated before trying all of the crazy positioning?

I am sure there are benefits to the transfer port injection, but surely getting a handle on the injector sizing and mapping before delving into that area will at least remove a few unknowns from this largely uncharted territory.

[TZ350]

I have a question in the midst of all this talk of injector positions: Why not start with the injector where the carb would normally be situated before trying all of the crazy positioning?

Probably because we are not trying to emulate a carburetor. I particularly want to avoid the fuel laden blow back associated with a carburetor and see if I can take advantage of the possibilities offered by EFI. Its the learning that is rather challenging and very interesting.

I am sure there are benefits to the transfer port injection, but surely getting a handle on the injector sizing and mapping before delving into that area will at least remove a few unknowns from this largely uncharted territory.

You are right, it's pretty much uncharted territory, for me anyway. But I am beginning to feel confident that I have started to get a handle on what size injector to use and some idea of what shape the map should be. Now that I am getting on top of those things I am starting to see other issues that need addressing.

There is not much helpful information about 2T EFI on the net. So I have had to figure out most of this from first principles for myself, I have posted a lot of original work, and constructive help on this journey is gratefully received.

Riley says he timed the injector to squirt 10deg before inlet open and finish 10deg after inlet close (this was for his own mule engine not the KTM)

The injector turns on for as long as it needs to to deliver the fuel required. You cant just turn the injector on 10 deg before the inlet opens and off 10 deg after closing. But the timing of the inlet gives a window of opportunity of about 6ms at 6,000 rpm and 3ms at 12,000.



From the screen shot it can be seen that at 6,000 rpm the small injector fires for 3.8ms and at 12,000 rpm the big injector fires for 2.7ms.

So the injection events can fit inside the inlet window.

But the fly in the ointment is that air does not continuously flow inwards but there are periods in the rpm range where air in the inlet tract blows both ways. So its just as likely to carry the fuel out of the engine as in.



I have tried external injection, both low and low + high injectors together timed to the inlet event but they did not run as nearly as well as the internal injectors did.

The big high power high rpm transfer injectors are working passably, its the low speed, trailing throttle less than 20% load injector that is causing me grief.



The next move is to try the slow speed injector in a throttle body and as its only on for 60% or so of the inlet event I will try varying its timing. Hopefully there will be a sweat spot .

[husaberg]

The injector turns on for as long as it needs to to deliver the fuel required. You cant just turn the injector on 10 deg before the inlet opens and off 10 deg after closing. But the timing of the inlet gives a window of opportunity of about 6ms at 6,000 rpm and 3ms at 12,000.



From the screen shot it can be seen that at 6,000 rpm the small injector fires for 3.8ms and at 12,000 rpm the big injector fires for 2.7ms.

So the injection events can fit inside the inlet window.

But the fly in the ointment is that air does not continuously flow inwards but there are periods in the rpm range where air in the inlet tract blows both ways. So its just as likely to carry the fuel out of the engine as in.



I have tried external injection, both low and low + high injectors together timed to the inlet event but they did not run as nearly as well as the internal injectors did.

The big high power high rpm transfer injectors are working passably, its the low speed, trailing throttle less than 20% load injector that is causing me grief.



The next move is to try the slow speed injector in a throttle body and as its only on for 50% of the inlet event I will try varying its timing. Hopefully there will be a sweat spot .

I think Riley was referring to the max flow duration used rather than all revs
Funny enough the rest sounds like 3 times past the main Jet..........