Page 1 of 1

Tuning returnless fuel pump

Unread postPosted: Thu Feb 13, 2020 8:56 pm
by VJordan
How would I go about tuning the fuel pump , I have a hesitation shifting wot the fuel pressure spikes when off the gas then drops and goes lean when back on the gas. I realize that’s gonna happen with returnless because the fuel has nowhere to go and most people solve it with deleting the pprv but I would rather try to tune it out is it possible with the pids?

Also on cvaf1 the adx dashboard fuel pump input voltage doesn’t change so can’t really log it and the o2 sensor voltage doesn’t change either

Re: Tuning returnless fuel pump

Unread postPosted: Thu May 21, 2020 1:23 pm
by p4ndiamond
I also need to figure this out as well .

Re: Tuning returnless fuel pump

Unread postPosted: Thu May 21, 2020 3:24 pm
by decipha
I need to add those two parameters then to log them. What date adx do u have?

Re: Tuning returnless fuel pump

Unread postPosted: Sun Jun 14, 2020 7:04 pm
by cajun cobra
Hey Michael I am also very interested in this as well not that I have this problem on the ole cobra but I just happen to be searching through the terminator table of contents over on modularfords and came across some interesting info on this subject of which I copied and pasted a small portion below. There is a ton more info on this subject over there as well. Anyway it got me trying to search for these tables just out of curiosity to see how they compare. I found the fp voltage table but nothing related to PID controller they speak of. Am I just not finding it or is it something that hasn't been added in yet?

Let me go into some more explanation about what causes the problem, why it’s such an issue, and various solutions.
The crux of the problem is a combination of the fuel pump voltage table and a software problem from Ford. When you are heavy into the throttle and accelerating the fuel pump voltage table needs to be 100% accurate, no exceptions. Why? As long as the fuel system is full of fuel, meaning everything from the fuel pump to the injectors, changes in fuel pump duty cycle will result in fuel pressure changes almost immediately. This is the benefits of using an incompressible fluid and having the system full of fuel.

Have you ever seen a stock Cobra have this problem? The answer is no, you haven’t. Why? The fuel pump voltage table is very, very accurate. What happens if your fuel pump table is NOT accurate?

If the fuel pump voltage table is not accurate you have to rely on the PID controller to supplement the fuel pump voltage table because it’s not getting the correct pressure from the table alone. Now, when the fuel flow requirements of the engine change dramatically, meaning you have pulled off the air pedal some or all the way(it doesn’t control gas anymore, it controls air) what happens to the PID controller? It doesn’t freeze or something like that, it keeps adding the same voltage it was adding before, so the fuel pressure spikes dramatically. The PCM says that there is a huge error between actual pressure and desired pressure, so the integral term of the PID controller “winds up” and starts chasing the error. I will assure that if the PID controller is not adding any, or only very little voltage, the pressure will NOT spike. If you don’t believe me, find a totally stock Cobra and see if it spikes. Ford has an internal requirement of +/- 5 psi error EVER on a stock vehicle. So when you build 10’s of thousands of them, probably a max error +/- 10 psi.

Now here is some of the problem with the software. Every PID controller in the PCM, idle spark, idle air, rear O2, etc, has a max clip for each of the individual terms in the PID controller. You can see them if you look in the software (I don’t think I have them all in there, but most all of them are). But guess what? There are no clips for the PID controller for the fuel pump. So it can “chase” the error and add/subtract to the voltage delivered to the fuel pump all the way from zero fuel pump volts to full battery voltage. This is the heart of the problem, there not being a clip on the PID terms.

So what are the potential solutions?

Well the first and foremost thing you have to do is have an accurate fuel pump voltage table. And when I mean accurate I mean accurate from WOT all the way back down to idle. If you are using a higher HP engine, you need to ratio the fuel injectors and what not so that peak fuel flow is only 3.99, so you don’t go beyond the table. One day I’ll work on finding the clip for this in EEC cars.

Two, if you change the injector values by .55, or 55%, then you need to multiply the Slope_of_MAP_vs_aircharge, in the speed density section by 1/.55 or 1.82. This will make the inferred MAP calculation correct, which is important.

Code:
Note, Jerry is referring to ratioing in general,
and means if you ratio the injector values any.
He was using the above given values as an example.
Then, if your voltage table is accurate enough, lower the integral gain in the PID controller to a value that’s very low, something around .10. If the fuel pump table is accurate enough, you shouldn’t need much in the way of the PID controller to keep voltage correct.

Now if the hesitation is still there, what do you do? At this point you need to work on bringing the “error” down, so you don’t get a lot of PID windup. There are two ways to do this. Command a higher fuel pressure at WOT, assuming that you have enough fuel pump for it. A higher fuel pressure will result in less “error” between the actual and desired fuel pressure. So commanded a delta P of 50 or 60 psi so when the pressure spikes, the difference in pressure is less so there is less error, and hence less PID windup.

The same happens if you change the transfer function. If you make it so that no matter what, it thinks the max delta P is 60 minus 40, or 20 (by clipping the transfer function to a max of 60 psi) then it won’t see as high of an error, and hence it won’t add as much voltage.

If you want, any of you can share any of this with the outside world.

Re: Tuning returnless fuel pump

Unread postPosted: Sun Jun 14, 2020 7:40 pm
by decipha
hey tony thats the wrong strategy, they are using cvaf1 and not fbgi0. I've had the fuel pump histogram going for fbgi0 a couple years now I thought I had it in cvaf1 as well but apparently not.

Thats all pretty common knowledge but a few of those things arent necessary.

Its easy enough to dial in the pump voltage though, no reason not to.