Fuel Dynamics: How can we check fuel mass?

[swingtan]

I'm not sure how many people have actually played with dynamics, it seems to be a little discussed topic on here. I'll probably get a round to a more general discussion of dynamics in a different thread, so this one can remain specific to the topic.

I'm trying to work out how to measure the dynamic fuel mass, or at least log the ECM's calculated fuel mass. Having a PID for this would shine a massive light on dynamic tuning, as you would be able to see the calculated fuel mass and compare it with the WB O2 data to see what is going on. It would also help to give a clear indication on what the settings for dynamics are doing.

My beliefs are that dynamics become more important...

• The closer you get to having perfect steady state fueling.
• If you have a manual vehicle.
• If your driving style has a lot of significant throttle movement.
• During the warm up cycle of the engine.

It is worth looking at a PID for dynamic fuel load or is there a different approach that may help here?

Open for comments.

Simon

[WeathermanShawn]

I think it is a great idea. I have never been able to totally wrap my head around the concept yet. Any ideas on how to do it.I.E. Injector data, WO2 data, etc..

I am supportive..just a little over my head right now...

[swingtan]

yes, it's a bit tricky at the moment.

I understand the concepts of wall wetting and dynamics fairly well. I just need to clarify how the tables affect the fueling in the tune and see how much effect they are having. If I can get a better understanding, I'll do a dynamics guide but I don't want to do it till I work a few things out.

Simon

[5.7ute]

Simon, I have done a fair bit on this which is covered on this forum.
Basically from using the PCM's calculated airmass (DMA pid) IFR & AFR you can calculate what portion of the pulsewith is the pure fuelling section. Leaving only the offsets to remove to give you the transient fuelling contribution. If you use B3601 in the search function you should be able to find what I have done.
I did ask Paul if this was a pid we could get added but as you know they already have enough on their plate.
If you want to dig into this deeper let me know.
Cheers Mick

[swingtan]

I can see where you are going, but there are some difficulties in this...

• It assums that the base fueling is spot on.
• It assumes that the delay in the WB reading is constant or non existent. At the very best it assumes the delay is predictable.
• It assumes that the injector flows and corrections are spot on.

I can see the idea though. Given enough log data you should be able to pick a trend for how the dynamics are working, but I was hoping for someing a little more precise. Something we can log easily and tie back into the tune file.

Can you give more of a hint on that thread Mick?

Simon.

[5.7ute]

I can see where you are going, but there are some difficulties in this...

• It assums that the base fueling is spot on.
• It assumes that the delay in the WB reading is constant or non existent. At the very best it assumes the delay is predictable.
• It assumes that the injector flows and corrections are spot on.

I can see the idea though. Given enough log data you should be able to pick a trend for how the dynamics are working, but I was hoping for someing a little more precise. Something we can log easily and tie back into the tune file.

Can you give more of a hint on that thread Mick?

Simon.

I will find the threads & post the links. There are bits here & there & you may have to read between the lines, Its been a while since I done them.
The main thing I was using it for was to remove any cells showing transient additions/subtractions from the data set to get VE dialled in more accurately. Since the dma pids are what the pcm is using in the calculations, you can get the main fuelling portion of the pulsewidth logged with a reasonable accuracy. Obviously you would ensure the pids used in the calculation are contained in the one packet to minimise errors from data frame offsets.
With the wideband lag I found it to be one frame at high airflow & two at low to mid airflow. This can be allowed for with a calc pid to bring them in line.
I agree that there are a few assumptions that need to be made. But in the end there a ton of variables that will keep us chasing our tail if we get too OCD.(fuel pressure fluctuations for one)