You are correct, it is interpolated.

The reason for adding this table was make the injector table reflect more the fuel quantity. Before the Baro compensation overcompensated a little because it also worked on this dead time.

The values in this table are only an initial guess by me.
It wont hurt to set them all the same. The effect should be minor.

Note if you use the ECUController to change these, it will automatic adjust the injector duration to produce no net change. I don't recall which voltage cell it used for that at the moment, but you could set them all to zero this way

Why don't you start another thread with your actual problem or send me an email/pm.

Probably it has nothing to do with the ECU. Hard to describe the sympthoms. In short, the bike changes temper too much and too often, for no obvious reason. Unless it's because of fluctuating battery voltage.

I have a charging problem, I only have net charge above 3-4000 rpm so my voltage fluctuates between 12 and 14 V all the time. New regulator is ordered. I also seem to have problems with my LC-1, or rather the Bosch sensor. The dreaded error 8. Maybe time to buy a new sensor. So maybe all problems will go away as soon as these problems are fixed.

Anyway this made me think: there is no reason my battery voltage should affect how the engine runs as longs as it's over 12 volts. If this charging problem is the root cause of my 'issue', there is a possibility the InjDur table Cliff supplies have a little too much - or too little - variation. But that is just a theory and I'll try to investigate it further.

Baselining injectors is doable at home, but the fire fighters might not like it 

For what it's worth, here is the table I used until Cliff "published" one. I got it from somewhere on the 'net, it's NOT confirmed correct at all for our injectors so I started using Cliff's instead:


The difference from 12-14 volts, between this table and the other one, is not that much. It's in the order of 5% at cruising.

YaBB Administrator wrote on 06/25/09 at 19:27:51:

Yes. And of course, as long as closed loop is in working order the dialled in dead times will be of very little, if any, importance but that did not occur to me until after I wrote the previous post.

I rode home 50 kms open loop when the sensor went tits up and that part of the trip was absolutely free from mood changes. But it did also not involve city driving and low voltages.

I've now made hundreds of kms closed loop and everything is still OK. I notice a much more stable O2 reading - once tuned it stays close to 0% between rides. That was not the case before, probably because of varying voltage combined with an incorrect dead time table.

I'm thinking I could spot the injector dead time at 14.1 volts (my now stable regulated voltage) like this:
1. set dead time to 1000 across the table
2. tune a chosen cruising map area to a certain AFR
3. manually make that cell 10% richer
4. cruise again in the same map cells and notice how closed loop will lean it back towards the target. Verify what battery voltage we have.

If the closed loop correction is -10%, the dead time of 1000 ms is correct for 14.1 volts (or whatever voltage we had). If not, we can calculate the dead time knowing the pulse width that was used. I'll probably get some new grey hair while figuring out the exact maths even though it must be pretty simple really. Maybe doing all this with a dead time of zero is easier.

Very clever. I think that will work.

I was trying to think of a way to find those values. I was thinking of special debug code to double pulse the injector.

I'll task you to get some proper values for everyone.

The OEM software has this option since years. They call it "Diagnostics". Once you could have this option coming with a no longer available shareware even.
With this you can trigger the nozzles, the coils and maybe even the rev-counter. This option and, right, a proper voltage supply was what you'd need for this dead time thing.
Fiddling with Lambda values probably means wasted time only. Just think of how accurate some units are working resp. how accurate the rest of the sensors is working resp. how accurate the interpolation between single map cells is working and so on.

Hubert

Quote:


Maybe close enough, this from Epicycle:
"Ducati typically use 16 ohm units which have a fairly long dead time, typically 1 - 1.1 mS (milliseconds). 1 millisecond being one thousandth of 1 second."
http://www.epicycle.com.au/fuelhard.htm
Since examples are 888/916 these have 3 bar pumps. Voltage I dunno, maybe around 14?

Dead time seems like a small but important parameter:
http://www.injectordynamics.com/dynamic-characterization.html
http://www.yawpower.com/injectordeadtimesarticle.html

Seems strange that not even MM specs does not include dead time.
http://www.magneti-marelli.com/download_zone/4_fuel_parts/PICO_01.pDF
At least one standard value of f e at 14.0V & X fuel pressure or similar should be supplied. Or not?