Closed loop becomes crazy

[olivenoire]

Hi there
I figure an important problem,

Sometimes my cars becomes totally lean at idle or something 10-12% TPS
The problem disappeared totally at wot but at partial throttle it is not good at all
I can see o2 sensor voltage cycling from 0 to 1 volt.
I can see my large band AFR alterning from 14 to 17...

TPS and INJ are near constant
CMD.CAM is constant too
And AFR is random

Sometimes edu stays in open loop instead of closed loop. In open loop, AFR are not bad
SO I wonder what can

And other problem : I try to fix it with deadtime, but in deceleration, sometimes (often) there is a little surge (injectors goes 1.4ms during a fraction a time) with 0% TPS...

Any advise welcome!

slow stable strim-variation o2.kdl

slow monté strim-variation o2.kdl

slow monté strim-low o2.kdl

slow intéressant 2.kdl

map_immo_final drive_deadtime.kal

[olivenoire]

With pictures, maybe it is easier to understand...
english is not my mother tongue...

clip_6 by olivier gx, sur Flickr


clip_5 by olivier gx, sur Flickr

[olivenoire]

Nobody here?

I will try to put some better log

[olivenoire]

Nobody to helpme?

here the full log...

I really wish someone could help me

[olivenoire]

Nobody wants or understand my problem

Discussing with some mate, I^m beginning to understand my deadtime are probably totally wrong.
I've send my kprov3 to CPl to upgrade to a v4
After while I will try to tune DT carefully at idle etc

Obviously latencies are the key...

[Hondata]

Your deadtimes are far too low, even for stock injectors, let alone 850cc injectors. What injectors are they?

[olivenoire]

Hi thank you for your reply

I'm actually quite sure my problem comes from injectors too

These are my strange and very noisy injectors
http://www.siemensdeka.com/product/80lb ... -fi114992/

About dead time, I've put what I've found here

injectors DT.png

[Hondata]

I do not believe the values are measured in the same way. You'll need to find your dead time values by trial and error. For a starting point they should be around 1.2ms @ 14V and 1.5ms @ 12V.

[olivenoire]

Hi

Thanks again

Ok I understand.
I will give a try as far and will give you a feed back

Kinds regards

[olivenoire]

Hi

just received my brand new kprov4 yesterday!
Sound great stuff. Need to find how to plug ae30-4100 in digital....

But I've tried tuning DT.

Actually 1.2ms @ 14V and 1.5ms @ 12V is a bit too much. AFR stays at 10,5 at idle. Maybe I have to try longer?
IAT are too much too because it is 30°C outside actually so it is not easy to tune, probably because of compensation. As a result fan turns on and BAT is changing, because of piping and intercooler become hot

At first step I discovered these values are not bad at idle (furthermore with these values, surge at deceleration desepeared, and that's a great thing ):
13.00 13.96 15.02
1.00 0.88 0.83

As soo I add only 1% more Strim goes -5 or 7% without been able to stay at 0

But sometimes once again when cruising at 9-10% TPS, AFR goes lean. So I will try again increasing velues

I will try again and share log etc

Obvisouly it will be a total empiric and long way to sort this. But I'm pretty confident.

regards

[olivenoire]

Today was a new day to test my AFR gauge

It is actually impossible to tune injectors DT better that I did.
I really wonder if the problem do not comes from injectors themselves.

Will try again tomorrow, but this problem is the most difficult to resorb I ever had with this car!

But today I clearly confirm the problems occurs when injection time is near 1,5ms. I wonder why it opens so little.
Will try to reduce the injector size kmanager
Probably it will increase inject time??

I put some maps and logs if someone wants to have a check