ESE's works engine tuner

[TZ350]

Connected up a signal generator and oscilloscope to the Beast and had a look at how the EFI injector pulses behaved at a simulated 2,000 to 12,500+ rpm. And then with the motor running.

Five things I learned about my EFI setup today.

1, I don't have to worry about how fast the injectors open/close, and probably all good (modern) injectors are plenty fast enough with a latency < 1.75ms and good for > 16,000+ 2T rpm.

2, That with a simulated engine rpm signal from 2,000 rpm to 12,500+ rpm the oscilloscope showed perfect injector traces and cross over with no problems.

3, That when the bike was run on the dyno the oscilloscope clearly showed the EFI system swapping to the second injector and it running for a bit then as the rpm tried to climb the ECU started dropping pulses and the engine running became increasingly unstable until it was violently surging from injector 1 to injector 2 and then back again.

4, At the swap over point the second injector starts working but the injector driving signal becomes unstable. This only happens when the rpm is changing dynamical and not when the rpm is being changed relatively slowly by hand.

5, It makes me even more sure that its nothing to do with the injectors or whether the Alpha-N map is lean or not, the 2nd injector is coming on. So I think its in the software setup there is a box that should be ticked or un-ticked or some variable with a wrong value. For some reason the system is stable when bench tested with a simulator but with exactly the same settings, unstable under dynamic running conditions.





The oscilloscope trace is taken from the active or -ve side of resistors pushed into the injector plugs. The second injector is actually two injectors in parallel. As there will be a spare second injector plug when I run the motor up it will be easy to get a signal from the second injector for the oscilloscope so I can see what happens at cross over from the smaller slow running injector 1 to the bigger power and rpm injector 2.



Simulated 12,500+ rpm with the signal generator.



Both injectors at a PWM of 3.1ms



Perfect matching traces at 12,500+rpm.

I am measuring on the active (-ve) side of a resistor inserted in the injector plugs. Now it has to be remembered, this is only a trace of the ECU's signal to the injector and it tells us nothing about the shape of the curve of the fuel flowing from the injector during each pulse.

How the EFI system works, is that the ECU supply's 12V to the injector and when it wants it to fire, the ECU closes a switch and grounds the -ve side of the injector to 0V so current flows through the injector coil. This voltage swing to earth or oV is what I am tracing and is the flat line you can see at the bottom. The flat line is 3.1ms long.

The curved upper line is the ECU opening the switch again and the -ve side of the injector rising back up to 12V. The vertical drop is the -ve side being switched to 0V again to turn the injector on once more. The multiple trace lines are not my shaky hand but at 3ms the trace gets re drawn very quickly and the camera catches several images.



The spiky line is the trace from a real injector. The high voltage spike is from the inductance of the injectors coil wanting to keep the current flowing when the ECU opens the switch to turn the injector off.

When the motor was running I could clearly see the trace swap over to the second injector, the second injector would pick up and start to run and the rpm would start to increase just before the ECU began dropping pulses and the motor rpm started swinging ever more wildly between injector 1 and 2.

I did not have enough hands to get pictures or video of this, but it was good to see, as it confirms the problem is not in the EFI units hardware or injectors.

[mr bucketracer]

i know when it comes to fuel injection a real good battery is needed , my mate glen that race's f3 has a 1000 dollar battery and is only good for 1 race . thats why i have left the charging system on my aprilia

[TZ350]

i know when it comes to fuel injection a real good battery is needed.

I have a battery charger connected to the battery all the time I am working with the EFI system on the dyno. When I get it going properly I will have to junk the old total loss system and upgrade the bikes electrics to include a decent charging system.

[F5 Dave]

I was just about to say, have you put a meter on the battery when it is running?

[Moooools]

Attachment 291448

Thanks, but I am a glutton for punishment and want to try and conquer this thing.

If you do end up switching to the link I am more than happy to do the loom for you if you pay for materials. And also help you set it up.

Connected up a signal generator and oscilloscope to the Beast and had a look at how the EFI injector pulses behaved at a simulated 2,000 to 12,500+ rpm. And then with the motor running.

Five things I learned about my EFI setup today.

1, I don't have to worry about how fast the injectors open/close, and probably all good (modern) injectors are plenty fast enough with a latency < 1.75ms and good for > 16,000+ 2T rpm.

2, That with a simulated engine rpm signal from 2,000 rpm to 12,500+ rpm the oscilloscope showed perfect injector traces and cross over with no problems.

3, That when the bike was run on the dyno the oscilloscope clearly showed the EFI system swapping to the second injector and it running for a bit then as the rpm tried to climb the ECU started dropping pulses and the engine running became increasingly unstable until it was violently surging from injector 1 to injector 2 and then back again.

4, At the swap over point the second injector starts working but the injector driving signal becomes unstable. This only happens when the rpm is changing dynamical and not when the rpm is being changed relatively slowly by hand.

5, It makes me even more sure that its nothing to do with the injectors or whether the Alpha-N map is lean or not, the 2nd injector is coming on. So I think its in the software setup there is a box that should be ticked or un-ticked or some variable with a wrong value. For some reason the system is stable when bench tested with a simulator but with exactly the same settings, unstable under dynamic running conditions.





The oscilloscope trace is taken from the active or -ve side of resistors pushed into the injector plugs. The second injector is actually two injectors in parallel. As there will be a spare second injector plug when I run the motor up it will be easy to get a signal from the second injector for the oscilloscope so I can see what happens at cross over from the smaller slow running injector 1 to the bigger power and rpm injector 2.



Simulated 12,500+ rpm with the signal generator.



Both injectors at a PWM of 3.1ms



Perfect matching traces at 12,500+rpm.

I am measuring on the active (-ve) side of a resistor inserted in the injector plugs. Now it has to be remembered, this is only a trace of the ECU's signal to the injector and it tells us nothing about the shape of the curve of the fuel flowing from the injector during each pulse.

How the EFI system works, is that the ECU supply's 12V to the injector and when it wants it to fire, the ECU closes a switch and grounds the -ve side of the injector to 0V so current flows through the injector coil. This voltage swing to earth or oV is what I am tracing and is the flat line you can see at the bottom. The flat line is 3.1ms long.

The curved upper line is the ECU opening the switch again and the -ve side of the injector rising back up to 12V. The vertical drop is the -ve side being switched to 0V again to turn the injector on once more. The multiple trace lines are not my shaky hand but at 3ms the trace gets re drawn very quickly and the camera catches several images.



The spiky line is the trace from a real injector. The high voltage spike is from the inductance of the injectors coil wanting to keep the current flowing when the ECU opens the switch to turn the injector off.

When the motor was running I could clearly see the trace swap over to the second injector, the second injector would pick up and start to run and the rpm would start to increase just before the ECU began dropping pulses and the motor rpm started swinging ever more wildly between injector 1 and 2.

I did not have enough hands to get pictures or video of this, but it was good to see, as it confirms the problem is not in the EFI units hardware or injectors.

Are you sure you aren't seeing interference from unshielded injector leads? I can't quite gather if you hooked the signal gen up and ran it up to 12500 with all the injectors in. But even if you did a potentially weaker signal from the crank pickup may be having a bad time in amongst some noise.

What sort of crank pickup are you using? Reluctor? And a single tooth timing wheel or multi tooth with one missing? Could you switch to a Honeywell GT101DC hall effect? They are mad bulky but produce an awesome clean signal and are very hardy. They are the general standard for motorsport and only set you back about $60 from Motorsport Electronics.

Is your pickup sensor lead shielded? Probs do that if it isn't. Interference is a bitch. As are a lot of things with FI but as soon as you have it right you will love it.

As for the batteries don't even worry as long as the charging system produces enough power to keep up. I think we pull 9A on the racecar to run the engine, a DAQ system and an electric water pump and radiator fan. We have a meaty charging system (160W) which keeps us well in the green. I wouldn't even dream of running total loss on FI.
Actually the standard WR450F is FI and is set up to have the battery removed! It just has a capacitor that charges enough from a kickstart to prime the fuel pump and get the thing to go. Incredibly impressive.

[TZ350]

If you do end up switching to the link I am more than happy to do the loom for you if you pay for materials. And also help you set it up.

Thank you for the offer, I think if I had of borrowed Flettners Link and with your help I would have been up and running long ago.

but I am keen to get the Ecotrons working properly if I can, then we can afford to fuel inject some of our other bikes too.

I have been able to run the Ecotrons on a single bigger injector to 12,000 rpm which suggests its not lean map settings and I have tried other tricks like under reporting the size of the second injector so the system will run rich after cross over. With the settings rich on one injector the engine will continue to run (badly) but not hunt violently like it still does on two however rich I make the second one.

The problem seems to be that the Ecotrons EFI system with 2T split injection runs and crosses over to injector 2 OK all the way to 13,000+ rpm when its driven by a signal generator.

But when its running live with the motor it becomes unstable after the cross over point at about 8,000 rpm.

The only difference I can see between the signal generator and real running is that with live running the motor forms a feed back loop.

And some how after the cross over point to the larger injector 2 the EFI control becomes unstable and that re enforced instability leads to violent hunting until the average engine rpm drops back below the cross over point and the motor recovers.

The Ecotrons EFI system was originally developed for GY6 CVT scooter engines and in the Ecotrons advanced calibration menu there are any number of hysteresis settings maybe something here is responding to what it thinks is runaway (but normal 2T) acceleration and tries to hard to damp things down.

Anyway that is where my current thinking is.

[TZ350]

Are you sure you aren't seeing interference from unshielded injector leads? I can't quite gather if you hooked the signal gen up and ran it up to 12500 with all the injectors in. But even if you did a potentially weaker signal from the crank pickup may be having a bad time in amongst some noise.

Actually I have not looked at that, ....

[Moooools]

But when its running live with the motor it becomes unstable after the cross over point at about 8,000 rpm.

The only difference I can see between the signal generator and real running is that with live running the motor forms a feed back loop.

And some how after the cross over point to the larger injector 2 the EFI control becomes unstable and that re enforced instability leads to violent hunting until the average engine rpm drops back below the cross over point and the motor recovers.

The Ecotrons EFI system was originally developed for GY6 CVT scooter engines and in the Ecotrons advanced calibration menu there are any number of hysteresis settings maybe something here is responding to what it thinks is runaway (but normal 2T) acceleration and tries to hard to damp things down.

Anyway that is where my current thinking is.

So if I have this right there is closed loop feedback that tries to determine if the engine is doing something it is not supposed to be and corrects it, and you don't have control over that feedback? Or have you just not found the window that controls it yet? I take it you mean a hysteresis setting such as 'when acceleration gets over X remove Y much fuel until it is below Z acceleration'?

The more I have seen of the Ecotrons software the more it looks to be way over complicated. The fact that you have to input your injector size alone seems like a hugely redundant feature. Just so you have an idea here is the process we run on the Link:

We have a master fuel control that has a the pulse width control and a master fuel multiplier.

There is a setup sheet that deals with the base timings and setup of pickups and firing order.

A table that is RPM vs TPS for fuelling and ignition timing.

A table for start enrichment.

And that is everything to get the engine to run.
If you can I would highly recommend turning every damn fun feature of the Ecotrons off until you have just the fuel, ignition, and a start correction running. Get rid of any closed loop feedback to eliminate sensor issues. By this stage you should be down to 2 variables that control every point, rather than 10 and then some more that you don't know. Try to do all of your changes just on your fuel and ignition tables, avoiding making changes to the master fuel enrichment and the injector settings. Even if there are still some spots that run rough don't worry about it until you can get the thing to rev to whatever you want it to rev to. Then put in a rev limiter. Then put in acceleration and overrun enrichment. Then do the rest of the closed loop control as you see it needs it. Usability first, then complexity.

If all fails then use the link.

[pete376403]

Years ago I read an article in Cycle Magazine about development of the XR750 (Harley! Fourstroke!!) which mentioned a problem they had with eddy currents surrounding the dyno which would completely upset the Fairbanks Morse magnetos, so to successfully dyno the engines they had to use a battery and coil ignition, even though they raced with magnetos.
( I appreciate there is a world of difference between a slow-revving, magneto-ignited, carburetted four stroke and what you are doing, but you seem to have eliminated all the obvious things...)

Could you place a faraday cage around the ECU while doing the dyno runs?

[TZ350]

So if I have this right there is closed loop feedback that tries to determine if the engine is doing something it is not supposed to be and corrects it, and you don't have control over that feedback?

I am guessing, but to account for the behavior I think that there must be a closed loop and no I can't find the correct variable to change, there are lots of them, and some of them are hidden.

Ecotrons does not seem to be able to grasp the issue and yes their setup adapted from a road going 4T that has to meet emission standards is overly complicated for a racing 2T.

Their conversion is a bit of a bodge. Like they just turned off the VE table although I could still see it and make changes to it but nothing changed in reality, it took me two weeks to work that cunning little trap out.

I am not getting happier ....

[husaberg]

Moto guzzi was using fuel injection with compressed air in the 50's
if i trip over it i will post it.

i found it interwrestling.

Rob with the Mirco squirt there is a software program that can self trim the mixtures does the econotrons have a silimilar setup?
http://www.megamanual.com/megatune.htm#autotune
The big potential advantage i can see with an electronic fuel injection set up is you could set it up to be "fuel cooled" to lessen power fade rather than any hp gains......

i had a gander at this the other day re the injector size..
http://www.rbracing-rsr.com/calcinjpulse.html
http://www.rbracing-rsr.com/fuelinjectors.htm
http://www.rbracing-rsr.com/efiharley.htm

[husaberg]

Barker headless the other side.

[wobbly]

Just a random thought - 2 injectors in parallel, is this causing an impedance mismatch that the injector driver circuit cannot handle.
Most manufacturers I have seen specify the correct impedance,and or have a different connection for high/low impedance injectors.

[Moooools]

I am not getting happier ....

This will cheer you right up. Pretty good price on the new Link Atoms: http://www.trademe.co.nz/motors/car-...-677326274.htm

Did that help?

Just a random thought - 2 injectors in parallel, is this causing an impedance mismatch that the injector driver circuit cannot handle.
Most manufacturers I have seen specify the correct impedance,and or have a different connection for high/low impedance injectors.

Seems like that could be an issue. But the fact that they go fine on the signal gen would seem to rule that out.

[TZ350]

This will cheer you right up. Pretty good price on the new Link Atoms: http://www.trademe.co.nz/motors/car-...-677326274.htm

Thankyou for the ecouragement, I have had a look at them and it will be where I go next if I have to.

Just a random thought - 2 injectors in parallel, is this causing an impedance mismatch that the injector driver circuit cannot handle.

Ecotrons confirm that adding another injector plug wired in parallel on a channel for two injectors is OK, but as its turned out, I have only been using 1 injector on each channel so the extra plug has been a handy measuring point for the scope.

The main problem is that I have seen on the scope screen the ECU dropping pulses just after the cross over point to the bigger second injector. This does not happen when the ECU is driven by the signal generator.

And if I set the system up to run only one injector it will run well past the troublesome cross over point for two injectors so I think the map and pickup are not the problem.