Ecotrons Engine Management

[TZ350]

Have any of the ones you know of on the site done an RS125R Honda with the kit ?

No, not that I know of. I am still trying to find a way to get good drivability with EFI on a highly strung 2T like a RS125.

[TZ350]

My first 2T EFI attempts were with an air cooled Suzuki GP125 engine and they don't have power valves. Made 31 RWHP @ 12250 RPM but still had the throttling problem below 25%.

Attachment 337065Attachment 337063Attachment 337064Attachment 337062Attachment 337066

The first air cooled cylinder had direct through the B port and under piston injection. The second air cooled cylinder had under piston injection and injection across the A and B port entrance at the crankcase. I liked that arrangement, it gave promising results.

Flettner has talked about something similar, two primary injectors firing against the air flow in the B ports and secondary's in the A ports but with them all firing when the trapping efficiency is peaking and the motor is making real power. This arrangement would keep fuel away from the exhaust port when the trapping efficiency was not that great and air was spilling out the exhaust.

The water cooled cylinder has B port injection but squirting against the air flow. All the cylinders have three physical injectors but only two logical injectors as the B port injectors are fired together. The fourth picture shows the underside with the piston at TDC.

I have tried a lot of different combinations, primary injector in the inlet, in the crankcase, in the B ports with the secondary in the alternative injection ports. And all sorts of other combinations like firing from one side of the crankcase across the top of the flywheels and directly into the incoming inlet air.

The least successful was direct inlet injection from the outside (the 5th picture), my favorite was under the piston and horizontal across the transfer ports but that is difficult to arrange with the water cooled cylinder. But under piston for the slow injector and injecting in the B ports against the air flow makes good power and looks to be more economic than a carburetor too. With EFI there is much less of a fuel puddle on the dyno deck beneath the inlet. With a carb there is a much bigger puddle.

Along the way I have learnt a lot and figured out most of the puzzle but I am just missing that last piece of the jigsaw, how to account for randomly changing air flow through the motor below 25% TP. It seems that the more highly strung the motor the more unstably changeable the air flow is below 25% throttle position.

Its like trying to sail an unstable "P" class learner yacht in gusty changeable wind conditions, its much easier when the wind is steady and consistently from one direction.

Attachment 337067

I am hoping that a MAP sensor in the crankcase and an Arduino Nano interpreting what it sees might be able to make sense of the air flow below that troublesome 25% throttle position.

[TZ350]

Item 17 is a combined pressure (vacuum?) & temperature intake sensor.
The exploded parts pics don't indicate where it is fitted.

Interesting that they are only using a MAP manifold absolute pressure sensor and a inlet air temperature sensor in the inlet. MAP and inlet air temperature, that is the same as my Ecotrons system so there is some hope.

Single point injection, after the throttle body and straight into the reeds.

This is where it gets interesting, 300cc cylinder with a single injector spraying into the reed block.

For myself I have several rules of thumb for 2T EFI. (a lot of this I have learnt from Flettner and Speedpro).

The longer the squirt duration the easier it is to get it right, so use the smallest injectors possible.
A long squirt mimics a carb.
A 360 deg squirt would be ideal especially if it coincides with the engines point of maximum charging efficiency.
Inlet injection maybe Ok for reed valve inlets but it is not so good for rotary valve engines. I have tried this.
There is a sweet spot for timing the finishing of the squirt. Usually somewhere around BDC to TPC.
Cylinders 250cc or greater are easier to get right.
MAP sensors have a settling time 2 - 6ms and they get confused if things get to busy.
Less than 9,000 rpm and you only need 1 injector. Because a small injector up to 9K rpm still has enough time to get the job done.
Above 9,000 rpm and your going to have to quickly start thinking about two stage injection using a small and big injector.
Injectors in the "B" ports are aimed downwards against the air flow.
Or better yet, if its physically possible, from the back aim the injectors horizontally across the bottom of the "B" and "A" ports.
For two stage injection, the single small injector is aimed under the piston for homogeneous mixing.
Alcohol based fuels are much more forgiving of changeable air flow, because they can be run real rich during moments of crappy air flow.
The more industrial the engine and/or application the easier it is to tune properly.
The more the engine depends on symbiotic reinforced wave action for making its power the harder it is to tune below 25% TP.

A EFI Honda RS125 engine would be a challenge to get throttling properly for negotiating corners.

Way back in the Olden Days, when we used to race with carburettors, we'd tune them to idle well, then lower the slides so they wouldn't/couldn't.

That is good info, tells me something about the minimum fuel required on closed throttle over run.

[TZ350]

.
I have enjoyed a bit of success today.

Used EngMod2T to simulate a series of runs so I could look at the variations in crankcase pressure just to see if there is a consistent change that can be measured and reflect changes in air flow through the motor.

These are pressure ratios, absolute crankcase pressure divided by atmospheric pressure.



This series of runs were taken at 8,000 rpm and is where the engine is really starting to get on the pipe.

The objective today was to see if the Arduino could read the MAP sensor and output a sensible result.



MAP sensor plugged into a spare injection port in the crankcase and the laptop and volt meter setup to monitor the Arduinos (true) analog voltage output.

The Arduino program looks for the highest and lowest pressure and the average for each cycle.

Then it outputs the sum of the average plus the difference.



Flashed the bike up and gave it a couple of blips. The analog voltages coming out of the Arduino look to be in the ball park and changed in a way that reflected real pressure changes (read, air flow) in the crankcase.

Its looking good for tomorrow. When I will connect the Arduino to the EFI CPU's MAP input.

With the new simulated MAP sensor input I expect the fueling to get messy but as long as I can use ECOCAL to data log some of the EFI CPU's output including the MAP trace I will be happy, ecstatic if it looks like it is the answer to seeing the un predictable air flow through the motor at less than 25% throttle position.

Fingers crossed.

[jato]

Top effort TZ...I'm sure many people are following your work with great interest and admiration!

[TZ350]

.
Ok, another little step along the way.



Gave the Arduino concept a try out today.

The EFI CPU was still running on only the Alpha-N map so I expected it to run fairly badly (to rich) but only really wanted to see what the crankcase pressure readings looked like and whether it looked like it might be useful for seeing changes in air flow at low throttle settings.

Fired the Beast up and gave it a few blips.



Map sensor = Green line, TPS = Blue and RPM = Brown.

The MAP sure looks responsive enough and the numbers make sense.

It looks like we can now see real time changes in air flow at throttle settings below 25%.

The next move is to re do the mapping to include a VE table (may take a week) and see what the drivability looks like.

I am starting to feel hopeful.

[speedpro]

I would still have a bottle of wine on standby

[Ulstermanone]

Thanks for the reply
I have race many many two strokes and mostly GP machines
The RS125R witch ran from 1995 to 2004 most upgraded to an Air box system that gave a few psi of booster pressure and maybe would make below the 25 percent a bit more stable ?
()
I been building a Moto3 machine but its a lot different and soon to be taken for its first track day

[crbbt]

So I was mentally designing an engine I don’t have the money to build.

Thinking about the issue of when the fuel injectors will delivery fuel when not require.

Which got me curious. When the fuel supply is cut off but the air supply isn’t. How do the rpms respond?

I’m guessing a bit of a lag due to their being mixture in the crank case.

[TZ350]

When the fuel supply is cut off but the air supply isn’t. How do the rpms respond?

I’m guessing a bit of a lag due to their being mixture in the crank case.

That is a good question. How many rotations before the wet fuel on the walls in the crankcase and inlet tract is consumed at WOT.

It must be several rotations at least. Previous comments by those who have studied it are, that it takes 5-6 revolutions to completely change all the air/fuel mixture in the crankcase with fresh. So with diminishing fuel available from the crankcase walls you would think at least 2-3 revolutions at full load and WOT after the fuel supply is cut dead.

If you find an authoritative answer someplace I would be interested in what they have to say.

[TZ350]

More Dyno adventures .....

Had a chance to try the Arduino MAP simulator today.

RPM brown line, TPS blue and Arduino MAP simulation is the Yellow line.



The Arduino uses a MAP sensor to measure crankcase pressure and looks for the maximum and minimum pressure and outputs the difference as a simulated MAP value to the EFI's CPU.

Bike sounded good and was even starting to show signs of improved drivability.

I was rely starting to enjoy myself laying consistent runs and making little improvements with ignition/fuel and PV adjustments.

Things were going real well, on the home run I thought, then I heard the dreaded dyno death rattle......

Quick as a flash, in with the clutch and let the beast coast to a stop.

(&^%%$#$#^+_!!! could be the problem this time.

Turns out that it had drunk all its cooling water during 20-30 runs. .... wasn't expecting that ....

Have not taken it apart yet but I suspect that small pressure drilling I put in a while back may be leaking or the head "O" ring. Not many other possibilities.

...... ......

I would still have a bottle of wine on standby

......

Well I need that bottle of wine. Just as things started to show glimpse's of the promised land I got the dyno death rattle.

Found the water had all disappeared. Expecting the cylinder and piston to be toast I took the top off. Turns out the cylinder is ok but it had run the bigend. Looks like the water washed the oil out and ruined the brg. Piston is toast from banging the head, I guess that was the rattle I heard but the head is Ok.

[TZ350]

Blue line = RPM. Brown line = LamWO2. Pink line = Injection pulse width. TPS = 100%

Cursers are both at the same TPS, RPM, Alpha-N load, and Injection PW.

On the left the engine is coasting down with the throttle held open and steady, LamWO2 a little lean, engine picks up at about 7,000rpm when air flow matches Alpha-N fueling and accelerates back to full power and rpm.

A little lean in reality can be an indication it is too rich to fire.

Left and Right sides both pass through the same Alpha-N map cells. Nath88 clued me into this. The air flow is different, there is much more air flow on the right than the left. After the exhaust wave has collapsed it does not re-establish until the motor has been firing again for a bit. The motor coasts down until the airflow matches the fueling at that point on the Alpha-N map (TPS/RPM) and the motor starts firing again and re establishes the wave action in the pipe.



For the same RPM and TP throttle position you can have very different air flows through the motor. It all depends on the strength of the wave action in the pipe. The wave action collapses when you throttle off for a corner and takes some time to re establish itself after throttling on again. This is where those pesky 4T's get the jump on you at the track.

Alpha-N is a bit ridged as it only works with RPM and TPS so it cant flexible adjust the fueling because it cant see changes in airflow.

VE volumetric efficiency works with RPM and MAP or MAF and so it can flexibly adjust fueling because it can see changes in air flow.

But the problem is that all the EFI systems I have seen are aimed at 4T's and all of them don't have an easy way to identify changes in air flow through a high performance two stroke motor. Ie a 2T motor that is heavily dependent on its pipe for its performance.

A work around is required to marry a 4T EFI system to the needs of a 2T. Traditional 4T ways of measuring air flow are not appropriate for 2T's so some other way of doing it is required.

Ecotrons 2T tuning software only has the Alpha-N map available and other EFI systems like link have both but for any system some sort of work around is required to make them useful for two strokes.

[speedpro]

We need to be careful that we don't criticize Ecotrons for not being able to do something they never intended to do. Their focus, the one that attracted us, was the supply of cheap engine management systems. Being cheap there are limitations or restrictions.

[TZ350]

Their focus, the one that attracted us, was the supply of cheap engine management systems.

That is true. I love that they are making EFI systems affordable for budding experimenters. I could easily afford a Link but I would rather be able to add to affordable EFI 2T experimentation by finding a way for the Ecotrons 2T system to be used with high performance 2T's.

We need to be careful that we don't criticize Ecotrons for not being able to do something they never intended to do.

I am pretty sure Ecotrons intend to sell 2T EFI systems.

The Ecotrons 2T software is basically their 4T version with the VE table disabled because according to the Ecotrons tuning manual the MAP sensor (and I guess any MAP sensor for that matter) becomes confused when used in a 2T inlet tract.

They say that Alpha-N is sufficient for 2T's, well it might be on some bigger or more pedestrian 2T's but as we now know, it is not sufficient for a 2T that relies heavily on its pipe to make power.

I have asked for the VE table on my copy of Ecotrons 2T tuning software to be enabled so I can try my own approach to measuring airflow in a 2T.

I am not asking for Ecotrons to solve my tuning problems, I don't think they understand 2T's well enough to be able too anyway and that is Ok.

I only want the tools to work with for myself, which is an active Alpha-N and VE table.

I doubt it is difficult to re activate the VE map in their 2T tuning software.

Maybe they are just not interested.

[speedpro]

I've been thinking about a carburetor lately. I need one for my scooter. I thought about a conversion to EFI but it's already a mission and I have an EFI project already.

Anyway, thinking about the various circuits and where in the load/throttle range they have an effect it struck me that possibly an EFI system could be setup very quickly to mimic a carburetor. The huge fuel tables with all the load cells are really only a benefit if you want to very precisely control fuel for emissions. If you only want to race then emissions aren't the concern.

What I thought was for the load ranges, say TPS, simply go with 0% 10% 25% 50% and 100% and say 4-5 rev values, say idle to redline with 4 intermediate values. You could let the software take care of calculating the amount of fuel between actual specified values in the cells. The whole fuel map might only have 20 cells. I'm going to try it on my FZR once I get back to playing with it. If it doesn't work I'll just reload the last good config and carry on.