Latency and the effects on Decel fuel cut

[roastin300]

Z31 88N/A 16 Bit ecu. Running 660cc high impedance top feed injectors, but 37psi, making them 612cc, and on E85 so essentially they are around 460-490cc.

Only way to get it to run OK, is with a K=360 and Latency > 1200

When I let off the gas, I do not achieve fuel cut when the TPS is OFF on deceleration and initial throttle release if my Latency is above 1100-1200. If I drop it down to around 850-950, decel works as it should, wideband goes completely lean to 21.8. However within about 30 seconds of lowering the Latency, the car starts running very lean at all times and stalls. I HAVE to increase Latency back up to 1200 or above, however, the fuel cut on throttle release stops working again, sometimes even goes rich to 12-13 when the throttle is released.

What is the relation between Latency and the decel operation?

Also 1200-1400 seems like a high latency value???? Compared to others I have seen on the boards.

[Matt]

check this for you over the weekend in regards to fuel cut. I didnt think that latency should affect it

[Matt]

Okay as promised I've looked into this

Latency will add a static amount of extra time to your overall injection time

During fuel cut the injection time during cut is 0.60ms but with latency increased from 750uS to 1300uS (ie extra 650uS) the injection time during cut is 1.88ms (about double the latency value)

On my bench performed a cut 3000rpm by putting TPS idle switch on. It went from 5.3ms to 1.88ms

This means that its still performing an adjustment to injection latency there is indeed a fuel cut. If you take a data log of a run whilst performing this you should see your 'injection time' drop below 2ms during the cut period until fuel cut is achieved (about 1500rpm is when the cut lower limit is reached)

You will need to post your tune and some log files showing the cut so I can see what is happening here

[roastin300]

I will try and get some logs for you. I also noticed it cuts fuel completely as it should when the engine is cold. As it warms up, the wideband starts revealing much richer conditions when the throttle is released. Very confused at this point.

[roastin300]

I lowered my TTPmin WAY down to a value of 1 from 1600rpm up, and decel fuel cut seems to be working much better now. But I still notice the relationship when I decrease latency down it works much better. However, these injectors want large Latency values. I am trying to get away with around 1100 now. Have you heard of any injector having a latency value upwards of 1100-1400? These are some sort of OEM top feed injector that is modified and then sold as 660cc's. You can tell because it has a stock part number on it and the last few numbers are burned off by a soldering iron to hide it. Must be their little secret. Thanks for the help

[Matt]

Resized injectors can cause all sorts of hassle. We dont recommend them because of the troubles they can cause. Nissan ECUs can control the injection quite well, but latency is normally only that high when running say 1000CC but you are running E85 and resized injectors may require a lot of fuel added at idle to keep it going well

TTPmin could be too high given the latency and still putting to much more fuel in.

You can try increasing TTPmin around idle to add more fuel (its a floor on minimum injection time so you can artificially add more fuel by increasing this)

[roastin300]

The thing is, when I lower latency, it runs lean in all areas of part throttle while trying to drive. Not just TTPMin. I tried to raise this value at idle to acheive exactly what you are saying. I could raise the areas higher in the rpm which would probably cure it, however, I would not have fuel cut on decel anymore like before.

I think it just needs that large of a latency value. What baffles me is the injectors were flow tested at 43psi and they flowed 660cc on pump gas. When running 19psi of boost (Approx 400hp) with E85 my duty cycle is only 30%. I think I saw it peak 50% once. I would expect the duty cycle to be more....Maybe I should try new injectors. The car runs decent, however, I feel like it changes daily and acts different quite a bit.

[Croustibat]

I lowered my TTPmin WAY down to a value of 1 from 1600rpm up, and decel fuel cut seems to be working much better now. But I still notice the relationship when I decrease latency down it works much better. However, these injectors want large Latency values. I am trying to get away with around 1100 now. Have you heard of any injector having a latency value upwards of 1100-1400? These are some sort of OEM top feed injector that is modified and then sold as 660cc's. You can tell because it has a stock part number on it and the last few numbers are burned off by a soldering iron to hide it. Must be their little secret. Thanks for the help

Let me guess ... venom injectors, bought from T**performance ?

I bought some. Crappiest thing ever. Inconsistent spray and flow, 1000+ latency. I had 2 flow tested, one was flowing around 560cc, the other would just lock itself open from time to time. I binned them. A quick check over the internet showed many people with dead engines because of this.


Another important thing ... You say duty is around 50% max. The CA18DET ECU injects two times per cycles when a threshold is reached. If your ecu works the same, it will show 50%, but will actually be 100% !

Just to let you know, i am running e85 too, with base fuel pressure set at 45psi, and boost at 15-18psi max. I am hitting the "50% duty" at 15psi ... note that the duty includes opening and closing time, meaning you can go higher than 50%. I did hit 64% with the oem injectors at 12psi, which was quite bad actually.I swapped injectors as soon as i saw that.

[roastin300]

Not venom injectors. I bought them from "Osidetiger" and they were flow tested within 1% of each other. They appear to be some sort of modified stock injector. With a soldering iron, they burned out the last number or 2 on every injector so that the customer cannot see the entire number and know what injectors they are starting with.

Hmmm, so when I am seeing 50% duty cycle, that doesnt directly mean the injector is open 50% of the time? Can anyone(Matt) confirm this is or is not how the Z31 ECU works.

I cannot get this car running great, mostly when cold, when throttle is varied (rolling into the throttle at different rpms as well as decel) and wanting to idle. I have tuned many different control systems (Haltech, SDS, Megasquirt) and had better results. Im determined to figure it out, just not the greatest results so far.

Sometimes will idle fine, but typically idles for about 5-10 seconds then just goes lean and stalls out every time. It did that when I first installed nistune, then went away and now its doing it again.

[Croustibat]

as far as i can tell, nistune allows you to modify your ECU, it is not a standalone ECU. The "half duty" is the result of a feature on the CA18DET ECU. I dont know if your ecu does the same, but if it does it really is easy to spot : make a log with duty cycle or time, and check if its value suddently halves on acceleration, then double when decelerating.

the CA18DET ECU starts injecting twice at some point, and the nistune has no way to see that. The duty cycle shown is the time of 1 injector opening. If there are 2 openings in the cycle, it will show 50% max.

check it yourself, and you will see. As the 2 ECUs use the same board, maybe they work the same ?

i had the same problem as you had when restarting my car after a rebuild. Dont forget there is a cold/warm start enrichment table that is used for some seconds when starting. Maybe it idles nice and then go lean because the car started to warm and enrichment tables are not used anymore.