Fuel pump PWM strategy

[Hugh]

I wanted to design a fuel system up so that I could PWM the primary pump to hold a relative fuel pressure of 3000mbar, hoping to reduce the fuel temperature and also to assist with holding the fuel pressure steady to help with mapping accuracy on fuel control.

Once the system is fully functioning, the hope is that when pump 2 and pump 3 kick in, the PWM strategy will keep the fuel pressure steady at 3000mbar.

The following refers to a S6 plus plug and play ECU in my R35 GTR

There is a fore innovations triple pump set-up in my car with 3 TI Auto F90000274 pumps that are suitable for E85 and PWM.

To get the system running I used one solid state relay and two traditional relays triggered at higher load.

If the system is not flexible enough to switch in the traditional relays and hold fuel pressure, there would be the simpler option of PWM all three pumps together rather than switching pumps in.

Here is how I wired the Solid state relay – please note the Fly-Back diode to safely manage the current generated when the pump is PWM'd off.

I ended up using a 1N5408 diode rather than the 1N4007 on the schematic for added safety.



Here is the solid state relay I used https://www.schneider-electric.com/en/p ... SP1D440BDT

Picture of the SSR installed - I later changed the power on wire (bottom left) to be an ignition on switched wire rather than permanently powering the SSR.



Initially I struggled to get the pump to run with PWM using the PWM3 / FUEL11 pin which has a pullup resistor.

I switched to PWM6 / Fuel 14 which has no pullup and initially the circuit still it did not work, so went back through the settings to ensure there was not something set wrong.

Bingo, here are the settings I used that got the PWM working.

PWM pump control enable – Enabled
Output Drive Type – LOW-SIDE
PWM frequency – 100Hz
Off Duty – 0%
Priming Duty – 35%
Run on Duty – 40%
Pump PWM Y Axis breakpoints - 200 850 1500 2150 2800 3000 3200 3400 3600 3800 4000
Base Duty all points set to 33%

Referring to Ryan’s help video.
https://www.youtube.com/watch?v=IVIdYESOuOQ

I created a new table with Relative fuel pressure on the X axis and PWM duty multiplier on the Y axis.



Closed Loop Fuel Pump Pressure Enable – Enabled

Fuel Pump Pressure target table values set at 2000mbar above MAP pressure.



Fuel Pump Pressure Error Breakpoints -1000 to 1000
Fuel Pump Pressure Proportional Duty Adder – not in use
Fuel Pump Pressure Integral Duty Adder – not in use
Fuel Pump Pressure Integral Duty Adder Minimum – -50%
Fuel Pump Pressure Integral Duty Adder Maximum – 50%
Fuel Duty Minimum – 0%
Fuel Duty Maximum – 100%

The car starts and runs, I did try and make the system work with the following controls without luck - please let me know if you made your PWM with the following controls.

Fuel Pump Pressure Proportional Duty Adder
Fuel Pump Pressure Integral Duty Adder

When I used them the numbers I used made the PWM climb rather than seek 3000mbar, so may just be a poor choice of numbers.

I have still to drive the car hard to see what happens with the fuel pressure.

Any thoughts?

Thanks

Hugh

[stevieturbo]

I tried closed loop briefly....oddly it wanted to do the opposite of what I expected.

Figured not worth the hassle vs a simple open loop table

I use a Crydom 60A relay ( pricey ) and drive 2x450's at present. At the low end around 38% duty, ramping to 100% duty up top. I see a little pressure drop, but nothing I'm concerned about.
I tested mine static at various frequencies to see if there would be one where the pump might seem to work better and I reckoned 200Hz was a little better.
Some cheap SSR's may not be happy there though. 60A Crydom says it can do up to 850Hz, or 100A up to 650Hz, but they seem like high numbers compared to others I looked at.

If and when I ever do need more fuel I'll connect up the 3rd pump and switch to a 100A Crydom and PWM all 3 of them in the same manner. Just oversizing the relay really so it'd never be overloaded. The 2 x pumps at present around idle pull about 30A peak ( avg with PWM would be lower though but there are still 30A peaks ) so with 3x pumps I'd expect 45A peaks, rising with load/pressure.

I wired the SSR up as per the Haltech drawing listed elsewhere on here, with the resistor across the terminals on the SSR input. No clue what it does though. I could stage the pumps...but I see it as a waste of outputs and just adds more relays and wiring for no need, with perhaps more points of failure.

In my case if the SSR fails...engine stops. Simple...replace SSR and all should be good again, but so far there are no indications that's likely to happen

[Hugh]

Thanks Stevie,

Agree it is a waste of outputs to separately drive each fuel pump.

I wasn't really sure how well the SSR would work, but wanted to see how controlled the fuel pressure would be.

I tried using the Fuel Pump Pressure Integral Duty Adder map today, with the other maps set to zero to see what the response is.

I video'd my screen on a quick pull, you will see the relative fuel pressure dip at around 2640 RPM where I have a 5% adjustment in the MAP that is not adding enough duty, when it hits 2765PRM there is a 600 mbar error and the 10% adjustment sorts the fuel pressure pretty quickly.

Fuel pump 2 kicks in at 3141 RPM, so effectively ends the test.

Think I will up the correction values in the Fuel Pump Pressure Integral Duty Adder map and run again.

My SSR does not have enough current capacity to drive all 3 pumps, but if it would not hurt them to drive all three of them at very low flow rates, it would simplify the system to have one SSR, or three smaller SSR's hooked up to the one PWM control point.

Youtube video link below.

https://youtu.be/a6yArj_ZSHo

[stevieturbo]

If going to the trouble of multiple relays, then you might want to stage them and there could be many good reasons for taking that approach.

I just figured it was easier to use one relay, one point of failure etc ( or success ) which is why I opted for the higher capacity unit from the outset, also bolted to a decent sized heat sink.
So even despite running at low duty cycles with a lot of current through it....basically running the SSR at less than 50% duty ( in terms of load/current ) all the time shouldnt pose it any trouble with heat or anything.

But if FP is dropping somewhere ( and 0.6bar is a lot ) on a system with a conventional reg, then I'd say the base duty table is wrong in the first place and should be higher in those areas so only very small corrections are made.
If it was a returnless PWM closed loop setup with no regulator...it'd be a different story, but not sure I'd be ready to test that out yet.

At the minute at low load, mine sits around 3200mbar but does drop to around 3000 at high loads. If I did add a 3rd pump to that...it will be interesting to see how that affects things and if I can run a low enough duty at the low end that it doesnt blow the FPR open and then cause issue up top.

I havent tried any lower than 38% duty to the pumps though. But if with 3 pumps I did try say 33% and things were happy...that almost be same a single 450 for low load use.

[Hugh]

I have had the second pump triggered at 30% and the third pump triggered at 55% for the last couple of years, so this is not a new part of the fuel system.

The numbers in the Fuel Pump Pressure Integral Duty Adder map were a first guess and the video was the first pull, so was pretty close.

Will increase the map offset numbers and should be pretty close to working reliably.

Will need to compare to earlier run logs to see how steady the relative fuel pressure numbers are before making any decisions, but am pleased to have something to start working with.

I did try different PWM frequencies, the solid state relay was much noisier at 200Hz compared to the 100Hz that I currently have it set at.

Here is a video of the car starting, it needs about 24% duty cycle to hold 3000mbar relative fuel pressure when idling.

https://youtu.be/LoVHAMqJWXA

[stevieturbo]

In many respects, a much higher PWM freq should be better...but that'd be with a proper controller. We're not really doing that with a SSR.

Although the FPR sort of mucks the test up a bit, I simply ran the pump as a test ( engine off ) at various frequencies and noted which yielded the highest fuel pressure. The mechanical FPR is obviously fighting that test a little, but there were still small changes, and for me 200Hz seemed to see the highest pressure, so I will assume pump etc was happiest there.

I did same test for my water injection setup with a little Hella SSR but through a fixed nozzle and was very surprised at the different range of pressure and flow seen right from as low as 50Hz to 600Hz or so was the highest I tried. 50Hz was a lot better, at higher frequencies the pump would barely run at low duties.

[MReilly]

How are you guys actually setting the regulator itself up in this scenario then?

Idle with guessed lowish duty, set reg at target as normal? Do you try to find the lowest duty possible + tightest reg combo that you can use at idle, so that the pressure comes up faster with increased duty (tight reg)? Other way around targeting a slacker reg?

Looking at your tables Hugh, and without any experience of this at all so probably completely wrong, I would probably tighten the control range up on that multiplier table. Depending on your fuel pressure reg the overpressure for large flowrates can vary a bit, but it never seems to be massive in my experience, maybe .25bar, (switching 2nd or 3rd pumps in early doesn't massively jump pressure for example), so I would keep the control range tight, targeting slightly above your regulator setpoint, and have a massive jump in duty if pressure gets more than .25-.5 bar down from target. I'd aim slightly above regulator setpoint because as soon as you don't deliver enough fuel the pressure drops off rapidly due to the reg closing, once that happens that is that. So your actual real world range of control is against the spring in the reg with some return flow, except for a probably very very tiny range of fully closed reg against injector flow and line bulge (I'd work up to that level of control I think!).

This test by Andy Wyatt of Adaptronic shows the variances with different regulators to fuel flow rate. For this type of control obviously you would actually want to be choosing the reg the other way around to the purpose of this test, as you want to see the flow rate effect on the reg to provide the control feedback, but the data is still relevant enough; https://adaptronicecu.com/blogs/article ... difference

Fuel lab make a regulator with a built in return flow turbine that you can take an output from, would probably be quite an interesting metric for your tinkering here.

Keep us posted on this buddy, interested to see how you get on with it, thanks for what you have shared so far.

[Hugh]

MReilly,

Some interesting points you make about the way that the FPR will deadhead, then open, I guess it would be possible to log this action by reducing the PWM below the FPR cracking pressure and see how it reacts as the pump duty and pressure increases to verify what that looks like, might be a task for Saturday morning!

The FPR was set quite some time ago to give a relfp1 value of just over 3000 with the previous fuel pump system, this was set at idle. The GTR has a venturi system in the fuel tank that takes a fair bit of pressure to function, so the backpressure from the different components is not easily broken down.

I am also not sure how quantifiable any of the outcomes would be with the PWM of the solid state relay modifications versus running the pump flat out, just seems to me to be the right thing to do.

I have a large Fuel Labs FPR, so was not expecting there to be an excessive pressure build up when staging in pumps 2 and 3, however with pump 1 running flat out at idle, the base fuel pressure was around 3500mbar, so the FPR was not capable of easily passing the excess and also there was the theoretical unnecessary heating up of the fuel that was not desirable.

For my drive to work this morning, I tightened the pump pressure break points from +/- 1000 mbar to +/- 600mbar and also increased the response numbers for the error.

There did not seem to be any surprises with the level of control, so will up the response to bigger numbers again for the journey home.

I may delay the point that pump 2 kicks in to see what happens to the fuel pressure on a short blast.

The hope is that pump 1 can be left to sort out the pressure without significant under or overshoots.

I like the idea of tying all 3 pumps together, the way that Stevieturbo has, just not sure I am ready to do that yet.

[stevieturbo]

How are you guys actually setting the regulator itself up in this scenario then?

Idle with guessed lowish duty, set reg at target as normal? Do you try to find the lowest duty possible + tightest reg combo that you can use at idle, so that the pressure comes up faster with increased duty (tight reg)? Other way around targeting a slacker reg?

I see it as keep it simple...and most regs are designed to operate with certain flow ranges of fuel

I set my base pressure at idle, around 38% pump duty, then just let both it and the PWM ( open loop ) do their thing, ramping from that 38% to 100% at higher loads. And adjusting anywhere in between that may not have been keeping up, but generally the mechanical FPR is still doing it's job, but overall it will be dealing with lower and probably more consistent levels of fuel throughout its operation.

So you more or less use a reg for say up to 1000hp...because you're never really feeding it with more than 1000hp of fuel that it has to deal with out the return

[DaveRodda]

The one you listed requires an isolated ground on the fuel pump. Possible I guess, but never been there.
I’m currently building one for Marvin from a brake controller using a rotary switch to achieve approx. .5-.7 volt steps.
I like the steps, always the same, every time, but considerably more work to build.
The slider on the brake controller is the simplest. This voltmeter would take the guess work out of a reostat, or PWM slider system.

seo consultant