Can someone check my logs for my sanity

[ellisd1984]

Hi, I'll appologise in advance as I've been messaging quite a bit recently on this topic but I just couldn't seem to get my head around whats been going on. I'm hoping someone can have a quick look at my logs and make sure I'm on the right path.

I was told when checking deadtime its a good thing to change the injection mode to multi point and have it fire twice a cycle.  This didn't seem to work at all I assumed it was because the pulswidth was essentially halved and going further into the non linear range of pulsewidth so to try and get things into at least the same ballpark have tried using the target Lambda method.

I've been chasing my tail trying to get a stable idle using the deadtimes supplied from the seller my injectors.  Using their deadtimes when changing target lambda from 1.0 to 0.9 I would get an actual change in the region of 0.75- 0.8 so was told my deadtimes are out.  I've changed them and have got lambda to track roughly,  pressuamably the logs confirm that these deadtimes ar at very least close?  I realise there's other factors such as charge temp correction and fuel charge temp but for now I need to concentrate on deadtimes.  I would assume that if my deadtimes are out my Short pulse width adder tables I likely to be out too.  Am I better setting this to 0 and will altering this table affect deadtimes? 

 

Few things I would like confirming.

Does it make sense that change from 1.0 lam to 0.9 happens almost instantly, but going from 1.0 to 1.1 takes more time?  My assumption was it would be the fuel film reducing casuing the slight delay.

What could cause the lambda to track fine going richer (1.0 to 0.9) but when going from 1.0 to 1.1 lambda goes really lean?

It seems very small chages in VE are causing quite large changes in Lambda, would this make sense?  I've never used modelled fueling before so I'm doubting everthing I do at the minute.

 

Thanks in advance

 

https://www.dropbox.com/s/5jahxlwo5d0p9tc/PC Datalog - 2021-01-6 10%3B15%3B28 am.llgx?dl=0

https://www.dropbox.com/s/g9jjlx6k0dsfiq4/PC Datalog - 2021-01-6 9%3B53%3B07 am.llgx?dl=0

[Adamw]

1 hour ago, ellisd1984 said:

I was told when checking deadtime its a good thing to change the injection mode to multi point and have it fire twice a cycle.  This didn't seem to work at all I assumed it was because the pulswidth was essentially halved and going further into the non linear range of pulsewidth

Correct, that method is no good for large injectors.  

 

1 hour ago, ellisd1984 said:

I would assume that if my deadtimes are out my Short pulse width adder tables I likely to be out too.  Am I better setting this to 0 and will altering this table affect deadtimes?

Thats a hard one to answer because the injector drive circuit that was used to test the injector makes a big difference to the injector response.  But I would probably leave the SPWA table with the suppliers numbers and just adjust the deatimes using a lambda target step change as you have done.  

 

1 hour ago, ellisd1984 said:

Does it make sense that change from 1.0 lam to 0.9 happens almost instantly, but going from 1.0 to 1.1 takes more time?  My assumption was it would be the fuel film reducing casuing the slight delay.

Many factors make it quite variable, but it is normal. 

 

1 hour ago, ellisd1984 said:

What could cause the lambda to track fine going richer (1.0 to 0.9) but when going from 1.0 to 1.1 lambda goes really lean?

Most likely you are getting unstable combustion or a faint misfire at 1.1 lambda, not many engines will run happily at that which is why I suggested stepping to 1.05 on the lean side.

 

1 hour ago, ellisd1984 said:

It seems very small chages in VE are causing quite large changes in Lambda, would this make sense?  I've never used modelled fueling before so I'm doubting everything I do at the minute.

That is the problem with running large injectors in the non-linear region.  If you tell the ecu that airflow has changed by 2% then the ecu will only adjust the inj PW by 2%.  So if that gives you more than 2% change in lambda then that means the injected fuel volume isnt proportional to the PW.

[ellisd1984]

Thanks Adam, It is starting to stick in, I promise

Ok So If I managed to get things running well in say the 2000 rpm area (hopefully closer to being in the more linear range of the injectors) it would indicate the deadtimes are ok and then drop down to say 1000rpm and the swings get larger would it make sense to add/subtract a little from the SPAT to try and bring it into line?

[Ducie54]

Can you link ur current map. Try these dead times and zero out your SPAT

 

6.000	7.000	8.000	9.000	10.000	11.000	12.000	13.000	14.000	15.000
4.108	3.354	2.599	2.145	1.691	1.468	1.245	1.111	0.977	0.881

[ellisd1984]

My map file attached is what I've spent the morning messing with,  I'm finally at a point where I can work with with things now.  Your dead times are similar to what i was given with the injectors but they just didn't seem to work.  I think half the problem was what I was considering quite small changes in map where needing the adjacent fuel cells increasing /decrease to keep the lambda right,  I was initally just putting the same number in all adjacent cells and that was causing my lambda to swing about all over.

deadtimes set for 13 to 14.5v.pclx