Strange values in the Fuel table

[jsmcortina]

As you are running high impedance injectors, you could remove Q9 and Q12 to disable the "active flyback", because there's a fault somewhere, also remove (or lift one leg of) R30 and R34. Then re-measure the voltage at the banded end of D5 and D7.

James

[PierreL]

James,
We just made the changes.
Now, we have 12v on D5 and -40mv on D7

Is that what you did expect ?

Many thanks again for your help,
Pierre

[jsmcortina]

Not what is expected.
If you look at the schematics for the injector drivers http://www.msextra.com/doc/general/ms2v3schems.html#p4 I don't see how 12V is getting to D5. This is with injectors disconnected - right?

James

[PierreL]

We did measure with the engine running and injector squirting.
Did we do it in the wrong way ?

[jsmcortina]

Re-measure without the engine running.

James

[PierreL]

Without the engine running i do measure12v on D5 and +-0mv on D7.
Have you an idea ?

As we have a mono cylinder engine, only 1 injector, could a workaround be to command that injector using another circuit than the one we use ?
It's just a question, maybe is it stupid ...

[jsmcortina]

The 12V is likely coming through the injector and if so is normal - unplug the injector and be sure.

James

[PierreL]

Injector unpluged i have 300mv on D5 and +-60mv on D7.
On D5 the value starts from +- 80 to 300 mv and then my multimeter crashes.
On D7 it is more stable
Does it help ?

Many thanks again

[jsmcortina]

Sounds like it is working ok now.

James

[PierreL]

Unfortunately not.
I can do whatever i want with the values in the VE, the engine seem to receive always too much fuel.
I go down to values of VE = 3% (!!!), the engine should not even start. But the (brand new) lambda wide band goes hardly under 9. I checked that lambda on a carburator engine, and values are more realistic, so it seem that the lambda works ok.

I rechecked that the PW is affected by my changes, and it is on the tuner studio gauge. But the engine turns like if i didn't do anything.

I don't know where to look at now...

[PierreL]

I made another test : i did once again let the injector sqwirt in a test tube while i was changing the VE values.

Here is what i did obtain:
VE=100%, injected qty =4,85 ml
VE=70%, injected qty =4,35 ml
VE=40%, injected qty =3 ml
VE=10%, injected qty =2.75 ml
VE=3%, injected qty =1,95ml

If VE=100% is the reference, based on the fact that approximatively for VE=10% we should injetc 10 x - , we have the following differences:

VE=100%, injected qty =4,85 ml, should be 4,85ml => correct at 100%
VE=70%, injected qty =4,35 ml, should be 3,395 => correct at 128 %
VE=40%, injected qty =3 ml, should be 1,94 => correct at 157 %
VE=10%, injected qty =2.75 ml, should be 0,485 => correct at 567 % (!!!)
VE=3%, injected qty =1,95ml should be 0,145 => correct at 1340 % (!!!)

The lowest the VE values, the biggest difference between what i really inject and what i should inject.

Thanks for your help...

[jsmcortina]

It does sound like an electrical fault causing the injector deadtime to be negative. Unless you have the tools and skills to debug the circuit further I'd suggest you send off the Megasquirt (and injector) for analysis and repair.

James

[PierreL]

I would like to...
But we have to be in Rotterdam on next monday, 14 may, for the competition.
And i suppose that nobody can repair the MS quickly enough ? We are here in Belgium...

[PierreL]

Hi James,
If you have an idea where to repair our board...
Or buy a new one quickly...

I made other calculations this moring, based on the results i had yesterday:
For all VE values, it seems that the system is ALWAYS injecting +- 6 ms longer than expected (varies between 5 and 7, variation probably caused by the relative poor accuracy of my measures...)

Thank you for you time & support,

Pierre

[PierreL]

i have interesting news:
- we did borrow a board from another tem, and it gives exactly the same result as our, so it seems that the board is ok
- we double checked that : we did put the oscillo on the injector, and the signal do respond exactly to the PW computed by the board.

So, we have try another, brand new injector.
It is always the same: to much fuel injected, confirmed by the fact that the (quantity @ 10% VE) is far to be (quantity @ 100% VE) /10 !!!

Does anybody have any idea ?

[HidRo]

I'm curious about that engine you are running. What calculations did you use to get to the Req_Fuel?
Engine size, injectors, etc?

[PierreL]

Hi!
Our engine is 4 stroke, 25cc.
We have an injector 14cc/min
We did let tunerstudio compute the required fuel for us, and then checked it: 12ms.
Do you have an idea?
Pierre

[HidRo]

What is your fuel pressure? Is it the same as the injector is rated to?

[PierreL]

in addition to my previous posts:
the shape of the signal sent to the injector is right.
But what about its power ?
I mean: do we have enough volts & amps?

on the oscillo, v is rising down from 12v to 10. Is that enough ?

[jsmcortina]

Two thoughts
- likely your entered deadtime is too large
- at 30% VE you are probably in the non-linear region of the injector

James