V12 Unhappy when leaning

[Philip Lochner]

I find it impossible to get mixtures leaner than about 14 without the engine beginning to run rough. This posting to hopefully obtain opinions as to why that might be and what I could do.

Today I got the V12 Jag starting with both ignition (this has been working for some time while original ECU was doing the FI) and FI (this is what I got working today).

Trouble is that these injectors are LoZ = 2.5Ohm each - with 6 of them per bank.

Running them without series resistors is not an option (exceeds 14A MS current limit during cranking) so I used the car's original resistor pack (easiest from wiring perspective) with four 5Ohm resistors. However, what Jaguar did was to group the injectors into 4 groups of 3 injectors each and running each of these 4 groups into a single 5Ohm resistor.

So you have (R = Right L = Left):
Gr 1: 1R, 3R, 5R
Gr 2: 2R, 4R, 6R
Gr 3: 1L, 3L, 5L
Gr 4: 2L, 4L, 6L

Having only two outputs with MS, I used Gr1 and 4 on Inj_1 and Gr 2 and 3 on Inj_4. I am using 2 squirts / alternate per engine cycle. This seemed closest to what the OEM ECU did. If my understanding is right this means that only 6 injectors fire at any one time, ensuring minimal pressure drop in the fuel rail. This arrangement should therefore spread the open injectors evenly around the entire rail.

I also found that the opening times of the injectors seem to vary between 2.8 and 3.2ms. For this reason I have entered 3.2ms as the opening time in MT. I presume this long opening time is due to the series resistors restricting current and hence flux buildup through them. I measured over these series resistors and one could clearly see the voltage rising exponentially as the pulse starts with 3 little dips soon after cresting which I presume was when the pintle hit the open limit.

At idle, I have an idle pw of 5ms giving and AFR of 14. The OEM ECU was idling at 16.5 with a pw of about 3.5ms.

My opinion as to why I can't seem to lean it is that with almost 0.5ms of difference between the slowest and fastest injector, some injectors are not opening much longer than others resulting in some cylinders running much leaner than others which are likely running quite rich with the net result of an average 14AFR.

With me trying to lean it out, some cyls become extremely lean causing misfires and roughness. Am I on the right track here? If so, what to do? Install a 2Om res in series with each injector trying to come as close as possible to the 14A limit in order to reduce the opening times to as little as possible? (Ouch, this would imply new FI harness to be made...)

All comment much appreciated.

[jsmcortina]

Another (equally awkward) possibility is to duplicate the Megasquirt injector driver circuits (right back to the FET driver) so you have four channels and therefore twice the currect capability.

Reducing the resistor values to give the maximum allowable current appears the best option.

James

[racingmini_mtl]

Another thing you could do is use peak and hold drivers. I have made a small board which as 4 drivers. You would need 3 boards for your 12 injectors. Have a look at http://www.msefi.com/viewtopic.php?t=19661

Of course this is not a cheap solution once the boards are populated but it should resolve any issue regarding injector opening and current handling. You should be able to fit the three boards in the top half of the standard MS case or you could put them in another box.

Regards,
Jean

[Philip Lochner]

Another thing you could do is use peak and hold drivers.

Thanks, I'll try lowering the resistor values first (although I am somewhat sceptical about that option but it is the easiest and quickest) and will then knock on your door for some of these drivers. I would prefer to buy them built and tested as I am more interested in getting control of the car than building circuits at this time.

[Philip Lochner]

Today, I installed a 2Ohm resistor in parallel with one of the four 5Ohm resistors in my setup (see previous post) and VOILA, opening time came down to 1.5ms from around 3ms. The engine started and I could see it running rich on the one bank as the 3.2ms opening time was now clearly wrong for that group of injectors.

Thinking I was close to a solution I promptly installed 2Ohm resistors on the remaining three 5Ohm Rs and turned the key. Instead of starting I heard the familiar "crrrr" as the injetors osscilate due to, I think, the 14A current limit.

I don't understand why the current limit should kick in though (if that is what is happening). If I have three 2.5Ohm injectors in parallel in a group, that has an effective resistance of 0.83 Ohm. This in series with (a 5Ohm // 2Ohm) = 1.43Ohm resistor gives a total of 2.26Ohm. I = V/R even assuming 12V during cranking, this gives 5.3 A. In my setup I've got two such groups on one injector output which gives a total of 10.6A which should still be well clear of the current limit.

Will try with a 5Ohm res on Monday...

[Philip Lochner]

When cranking, the number of injector pulses double and their duration “halves” (more or less). What could cause this?

This results in my “starting procedure” having to be as follows:
1) crank for 1 sec;
2) floor the loud pedal (this activates “flooding control” by switching injectors OFF );
3) now the engine starts on the residual fuel injected during the first second.

This seems to have something to do with the injector impedance because if I increase the impedance, all is fine, but lowering it brings out this behaviour. I have disabled the current limit circuit by shorting out the legs of the 0.05Ohm current limiting resistors.

[jsmcortina]

During cranking the behaviour has always been to fire on every tach pulse.

Does the actual pulse width match the reported one and how does this compare to your cranking pulse width settings?

James

[Philip Lochner]

During cranking the behaviour has always been to fire on every tach pulse.

Thanks James, that explains that one. At least I now don't have to continue searching for issues there. (I'm at the stage where I think I know it all but now realise I still have so much to learn...)

Does the actual pulse width match the reported one and how does this compare to your cranking pulse width settings?

This is the thing, the actual pw (measured by scope) does not correspond with either that reported by the pw gauge or the cranking pw's set in my WUE page

[Philip Lochner]

During cranking the behaviour has always been to fire on every tach pulse.
James

See, I was trying to be too clever. I measured the starting PWs with the OEM ECU when cold and when warm and plugged these figures into MS. However now that I know it squirts on ever tach pulse, I realised that I had to drastically reduce my starting PWs.

Now it starts without having to use "flood clear". Starts very quickly when cold but it still does not start quickly when hot - but it starts.

Thanks again James. May such help never seem trivial to you or any of the experienced guys.