Fuel pump: internal or external...

[Cement Legs]

Hey guys, we are debating on the issue of where to mount the fuel pump. We are aware of pros and cons for both. Stock pump off of a motorcycle: harder to maintain and service but does not require additional plumbing and mounting brackets therefor it saves weight and space (depending on design). External pump such as one from a Ford Ranger: easier to maintain/service due to location but requires addional fittings and space, adds to weight and cost report. I don't believe there is a cut and dry 'better solution' in this situation so I would love to hear people's arguments for either set-up, based on more experience than we have. It's fine and dandy to have the greatest design in the world on paper but if you can't build it...

Cheers

[Paul V.]

As someone who has been bitten by the inaccesible internal pump I would recommend an external. We had a pump failure at detroit in 2001 and we had to run to a local store and buy this huge external pump.
long story short,in my humble opinion, if your tank is not super easy to access, get one in line. The servicability is worth the few ounces of sprung weight. It may be worth some points in the design tent in that respective category.

Paul Vaughan
UAH FSAE

[Dan Deussen @ Weber Motor]

Here at SVSU we have been running an in-tank pump as long as we have been using fuel injection. The pump-flange assembly (PFA) we use has an integrated pressure regulator and is therefore a returnless system. I can honestly say that the in-tank systems have never failed us, so accessibility for maintenance was never an issue for us. Guess who makes the PFA we are using http://fsae.com/groupee_common/emoticons/icon_wink.gif !!!

[Cement Legs]

I would tend to lean towards making a system with an in-tank fuel pump as long as we design it for quick maintenance. Everything right down to the engine would be removable and serviceable to some degree so making a fuel tank that can be disassembled shouldnt be a problem in my mind. Having a spare pump on hand is always a good idea for either solution. But I'm interested in what you are talking about, Dan, when you say "The pump-flange assembly (PFA) we use has an integrated pressure regulator and is therefore a returnless system". Could you elaborate on this a little for me. It sounds to me like you only have a line leaving the tank, is this correct? What is the procedure for setting up a system that way? Pics would be nice if you are so obliged.

Cheers

ps I checked out that website but I must have skipped all of my Italian courses... wait... I never took any... http://fsae.com/groupee_common/emoticons/icon_confused.gif

[DanO]

we have run an in-tank pump for the past 4 years and have had great results. Having the internal regulator also is nice for a returnless system. If it was designed well you generally dont have many problems with fuel pumps. Just make sure you can acess it relatively well. We use a large snap ring to hold it in and it makes for easy changes. The judges did like our system too.

[B Lewis @ PE Engine Management]

I think that either approach can be used with success. The one problem that I see over and over is teams discarding the fuel filter or using the wrong type of filter in the wrong location (ie filter designed for carb application in between the pump and the fuel rails). Also, if the car will sit for more than a couple of months, use some fuel stabilizer. Cleaning a gummed fuel pump and injectors is no fun.

[mario_rbr]

Something anyone has not mention is the fact that a external fuel pump may produce a fuel swirl due to the returned lines. Has anyone had any trouble regarding this issue?
What about the heat that the fuel pump produces? when it`s internal the heat is disipated easily by fuel sorrounding, but has it created a problem for external fuel pumps?
Regards

[Ian M]

We used a Walbro external pump last year and it worked quite well. For us, the external pump getting hot wasn't a problem. We used the F4i's rail and regulator and then ran a return line to the tank.
Our problem now is that the RR rail we are running doesn't have that built in regulator. So now we are trying to figure out how we are going to regulate the pressure. How do the internal non-return pumps regulate pressure? Do they kick on and off as needed? Is there a way to run an external pump without a return line?
I am all about ease of maintenance so I tend to lean towards the external type, but either system designed the right way would be fine in my opinion. What is the return lined fuel swirling you are referring to?

[mangel83]

We used the stock pump of the bike for our engine this year (2003 Suzuki GSX-R600, internal, returnless). As Ian said, it kicked on and off as needed.

It worked great for us, until the day before dynamic events when we were downloading the car from the truck, one of the guys kind of slipped and the pump was slammed against the edge of the truck, a fitting welded to the pump broke and we had to disassemble the damn thing, reweld the fitting and reassemble the pump.

This little issue along some trouble passing noise test, costed us skid-pad, and only had time for 1 of the 4 acceleration runs. Next year we're running an external fuel pump and pressure regulator, some extra pounds in change for a little peace of mind.

[Wesley]

I know parts houses sell deadhead regulators. (return-less style)

Some say its hard on the pump because the it limits flow, which most external electric fuel pumps depend on to keep cool. They are marketed and run in race apps enough though that I don't think that should be a problem.

[Dan Deussen @ Weber Motor]

Wesley,

Most return systems still circulate the fuel. The pressure regulator is built into the tank and that is where the excess pressure is relieved to. So the pump still gets anough cooling.

[Wesley]

That makes sense.

But I was also referring to running an external pump with an external returnless regulator. But looking around they only offer the returnless external style for carbed applications. So nevermind... http://fsae.com/groupee_common/emoticons/icon_rolleyes.gif

[Kurt Bilinski]

For an EFI engine, I'd expect fuel pressure would reference MAP, so does this in-tank pump with the regulator have a line running to the intake manifold?

[mario_rbr]

Originally posted by Ian M:
What is the return lined fuel swirling you are referring to?

We are using a mallory external fuel pump which has an air bleed system at the bottom of it and is plumbed to the fuel return line coming from the regulator, a dragster mechanic warned us about this mixture may cause a swirl into the fuel tank. has anyone experience this issue in the past?
As Manuel mention last year we didn't have to even think about it because an internal fuel pump was used, so please any advise based on previous experience on external fuel pump is apreciated.

[Wesley]

Yeah, fuel delivery is controlled by the pulsewidth on the injectors, not fuel pressure. Though fuel delivery is a function of fuel pressure, it's typically a constant anyways.

[Ian M]

Ok this is good stuff, lets keep it going.

Question #1: How does the stock bike pump regulate the fuel? You say they read off of the injectors and when the injector pulsewidth increases they pump more fuel? This doesn't make much sense considering you can hear the pump kick on and off before a vehicle is started.

Question #2: Would there be a way to run an external, returnless, system and tell the pump to kick "on and off"
I am going to guess the answer to that is no because it would be kicking on and off all of the time and that is assuming you could create the "smarts" to control it from the injectors/line pressure.

Question #3: Can the bike pumps adjust how much they push or do they truly kick on and off? I know on my bike and most cars it kicks on until it builds enough pressure and then kicks off before startup. This leads me to believe it is controlled by an internal sensor and not the injectors at all.

Question #4: In a return type system is there a type of in-line regulator available you could put in after the rail to keep the pressure at 50 psi?

Sorry if some of the questions are kinda dumb. Never had to worry about regulating because our f4i had a factory regulator at the end of the rail. We used a Walbro pump and I would highly recommend them. We even ran ours out of gas once after being told that will destroy the pump because it will spin too fast and it held up just fine.

[Ian M]

I somewhat answered one of my questions by looking at an RR repair manual. The fuel is brought through the strainer, pulled through the in tank pump, at which point it hits an in tank regulator. This regulator "dumps" the fuel back into the tank like a return line. I believe this regulator is also tied to the ECU which shuts off the pump when the regulator valve opens. At least that is my explanation for the pump kicking back off after the initial turn of the ignition switch and it reaching 50 psi. So probably once running the pump wouldn't kick off very often because the regulator probably doesn't open up all the way. On the new RR, which is a returnless system Honda calls this their "Absolute fuel pressure system because it is supposed to keep 50psi at the rail at all times. Now I may look for an in-line regulator to see if we can reuse the Walbro.

[Wesley]

This is how I understand it:

When you turn the key to the run position, the pump does as you say, and turns on to pressurize the fuel system in preparation for an engine start.

When the key is in the run position and the engine is running, the fuel pump is running continuously, regardless of engine speed or load.

So, I would assume the fuel pump relay is indeed controlled by the ECU, but only so much as it is instructed to prime the lines when the key is moved to "on," and to stay on while the engine is running.

The fuel pump does not build up pressure and shut off while the engine is running, it runs continuously. (Which is how carbed ECU-less engines can run an electric fuel pump, albeit at a lower fuel pressure)

However, the speed or voltage of the pump is not modified in any way by the ECU. It can only turn the pump on and off, as a function of key position and engine status.

Because fuel delivery can be modified simply by increasing or decreasing the pulsewidth, there is no reason to kick off or vary the speed of the fuel pump, it would require an extra layer of control circuits.

Of course, I could also be completely wrong since I am applying knowledge of cars instead of bikes. (As a side note, do bikes have the inertial switches for fuel cutoff like cars do?)

I know they make in line returnless regulators for carbuerators at 6-7psi. I don't know if they do for FI pressures.

[Chris Allbee]

Wesley has the priming pulse right. After the engine is running the pump stays on. As far as inertial sensors...not to my knowledge. they do have a bank angle sensor to detect if the bike has been laid over. Hey Ian, if you wanted to use your old fueling setup you can always use an F4i rail and regulator on the RR. We had that setup on the dyno for a little while and it worked just fine. And I do believe you can buy aftermarket fuel pressure regulators, we had an AEM one that replaced the F4i's stock one a few years back.

[Charlie]

Lots of incorrect assumptions here. Why don't you guys just take a look at your pumps and discover how they work?

There is only one 'true' returnless pump and it is very uncommon. It is a pump controlled by pulsewidth modulation through an ECU. This requires a fuel pressure sensor. I call this a 'smart' pump. I am quite sure that no-one here is talking about this style of pump, so forget about it!

On all other systems, 'returnless' is a total misnomer. All high pressure 'dumb' fuel systems return fuel. It's just a question of where they do it! A regulator does two things: restrict fuel rate, and return the rest. On a style where the regulator is on the engine, it restricts after the fuel injectors, and returns the rest to the tank. On a style where there is no engine mounted regulator, the regulator is in the tank. It restricts fuel at the tank, and a line is teed off from between the pump and the regulator, and goes off to the engine.

The fuel pump runs constantly in both situations. The systems are basically identical, just have parts in different places.

[Ian M]

Yeah Chris, I was actually thinking of getting an F4i rail, regulator, and injectors, which is probably a good solution. I was hoping the RR injectors would fit in a F4i and Rail and Regulator but no such luck due to the RR's having a different taper on the nozzle end. The AEM aftermarket that you mentioned intrigues me though! Thanks!

[ad]

Originally posted by Charlie:
Lots of incorrect assumptions here. Why don't you guys just take a look at your pumps and discover how they work?

There is only one 'true' returnless pump and it is very uncommon. It is a pump controlled by pulsewidth modulation through an ECU. This requires a fuel pressure sensor. I call this a 'smart' pump. I am quite sure that no-one here is talking about this style of pump, so forget about it!

On all other systems, 'returnless' is a total misnomer. All high pressure 'dumb' fuel systems return fuel. It's just a question of where they do it! A regulator does two things: restrict fuel rate, and return the rest. On a style where the regulator is on the engine, it restricts after the fuel injectors, and returns the rest to the tank. On a style where there is no engine mounted regulator, the regulator is in the tank. It restricts fuel at the tank, and a line is teed off from between the pump and the regulator, and goes off to the engine.

The fuel pump runs constantly in both situations. The systems are basically identical, just have parts in different places.

Thanks Charlie,

I always love your comments, informative yet firm :P

Coming down AUS this year?

[bigwig]

One things a lot of guys have said in here is fuel pressure is constant. That is often not the case. Many fuel pressure regulators are 1:1 ratio. If you measure fuel pressure at idle vs taking the vaccum line off the FPR the pressure should vary about 10psi.

[Wesley]

By constant I meant the force imparted to the gasoline by the pump is constant - the impeller spins at a constant speed. Pressure generated depends on how much fuel is being allowed out of the line.

And I have seen regulators for regular non-modulated electric fuel pumps that have no return.

[Charlie]

Originally posted by Wesley:
And I have seen regulators for regular non-modulated electric fuel pumps that have no return.

Can't say I've ever heard of this, and I can't imagine how it'd work in a high pressure application. Works with carbs, but IMO that is irelevant to this discussion. Not trying to sound like I know it all, but I'd have to see it to believe it. Perhaps you can provide an example.


One things a lot of guys have said in here is fuel pressure is constant. That is often not the case. Many fuel pressure regulators are 1:1 ratio. If you measure fuel pressure at idle vs taking the vaccum line off the FPR the pressure should vary about 10psi.

Well if the regulator is in the tank line pressure is always constant (with reference to atmosphere) since there is no vacuum line to the reg. In a rail mounted regulator that is referenced, pressure changes in reference to atmosphere, but is constant in reference to manifold pressure (and constant across the injector). So I guess it depends on how you define constant pressure http://fsae.com/groupee_common/emoticons/icon_wink.gif


Coming down AUS this year?

Afraid not, the new dates don't work wiht my schedule. Maybe 2007.

[Wesley]

Originally posted by Charlie:
Can't say I've ever heard of this, and I can't imagine how it'd work in a high pressure application.


You're probably right, I have been unable to find this style of regulator for a fuel injected system.



Well if the regulator is in the tank line pressure is always constant (with reference to atmosphere) since there is no vacuum line to the reg. In a rail mounted regulator that is referenced, pressure changes in reference to atmosphere, but is constant in reference to manifold pressure (and constant across the injector). So I guess it depends on how you define constant pressure http://fsae.com/groupee_common/emoticons/icon_wink.gif


Rail pressure is constant in relation to the MAP, I'll agree with you there. But before the regulator, the pump will always be imparting the same force to the fluid, but the pressure will vary slightly as the regulator allows more and less pressure into the rail.

Essentially, the pump pushes with the same force regardless, and the pressure varies as the flow out varies from the regulator.

But yeah, theres no such thing as a constant in the real world. Blast you Heisenberg. http://fsae.com/groupee_common/emoticons/icon_wink.gif

[Cody the Genius]

I just wanted to mention that from a tuning standpoint that having a MAP referenced fuel pressure is crucial to maintianing effective fueling at varied load sites. The injectors are merely nozzles that open and close for a set amount of time and the actual fuel flow though the injector is dependent on the time that the injector is open (pulse width) and the pressure difference across the nozzle ( rail pres. - MAP) during that time.


This means that the amount of fuel sprayed at 0 inches of vaccuum is not the same amount of fuel sprayed at 10" of vaccuum for the same amount of pulsewidth. This can play hell with a tune up especially if your load is based off of throttle position. Having a manifold refenced fuel pressure means that the pressure gradient from rail to manifold will always be the same pressure as the rail pressure is regulated lower as there is a decrease in manifold pressure. It can make tuning a million times easier and better when the fuel pressure gradient can be a constant then the pulsewidth can be much more proportionate to the amount of fuel flow.


IMHO there is no reason not to run manifold referenced fuel pressure when all it take is the use of a vaccuum line running to the regulator.

[Ian M]

So you're saying that when the injector fuel flow is changed drastically (i.e. full throttle/closed throttle) the rail mounted regulator cannot adjust fast enough and therefore causes a period of fuel pressure inconsistencies. In order to alleviate this problem you place a vacuum line from your intake to your regulator, which allows your regulator to react to this drastic change and therefore keep your fuel pressure more constant?

I guess this is where the in-tank regulators are more consistent, although I would think that they would even experience this a little bit because their regulators are so far away from the injectors, but I guess being in front of the injectors would be better.

I am slowly learning and educating myself on engine management and all of you guys's information/knowledge is greatly appreciated.
Oh, if my understanding of this process is way off than don't be afraid to tell me I'm retarded!
I enjoy this constant fuel pressure information because I noticed a considerable amount of fluxuation in our stoichiometric readings from our Lambda sensor last season, so I am trying to eliminate the variables, so I can pinpoint the problem to my fuel table setup and adjust from there through more trial and error. Kind of off topic, but are there any good ways of figuring out basic fuel (MAP sensor based) and ignition tables besides trial and error.

[Wesley]

I'm learning a lot too Ian. I didn't know they have vaccuum referenced regulators.

So basically, the fuel pressure has the vaccuum line in order to boost or lower fuel pressure in the rail as MAP varies. This is just because if you have say 50psi of gague fuel pressure in reference to the atmospheric pressure, it may be 53psi in reference to the MAP, which is under negative pressure(relative to atm.) So the vaccum line serves to keep the fuel pressure/MAP pressure as close to its set value as it can.

If you just referenced outside atmospheric pressure, you would decrease the amount of fuel injected into the runners as MAP increased.

So basically, you would have a rich condition at WOT, and it would run leaner at part throttle and idle. You could theoretically adjust the pulsewidth to compensate for that (and the ECU will automatically, if the duty cycles of the injectors aren't being exceeded.) But why make that much more computation when you can use a simple vaccuum reference to keep fuel pressure/MAP constant.

Right or wrong? Like Ian said, I'm a new guy too, so please correct me if I am wrong. I'm here to learn not pretend I know everything. (which I do sometimes... http://fsae.com/groupee_common/emoticons/icon_rolleyes.gif)

[Ian M]

After checking it out, I see that the Honda F4i regulator has the vacuum line built in. We used an F4i last year and the regulator vacuum line must have been capped the year before because I didn't even notice, or think about it being a vacuum line. Now I am preparing an RR and planning on using an F4i rail, injectors, and regulator, so I'll do it right this time. In the past our cars have been kind of thrown together and just been made to work well enough to race. This is the 3rd year car and we are finally starting to get some knowledge and experience built up.
This is also kind of off subject, but what if we cut off the crankcase breathers? Isn't really a necessity for a race engine is it?

[BeaverGuy]

I believe what he is saying is not that the regulator can't react fast enough but that the effective fuel pressure is vastly different. The fuel pressure is controlled by a valve in the regulator that has a spring and a diaphram. On one side of the diaphram is a reference pressure, P3 and the other side is the Fuel at P1. P2 is the pressure in the runner. When the throttle is suddenly opened or closed P2 changes. If P1 stays constant which it will if P3 stays constant the pressure across the injector, P1-P2 will change. This now means that for a given injector pulsewidth the fuel flow will be different at different throttle positions. This is means that fuel flow is now not only dependent on pulse width but also throttle position.

Wesley you are correct. Now with the ability to completely control the pulsewidth at all throttle positions in the ECU I don't see this as terribly important. There is also the issue that the pressure in the runner at the injector can be vastly different than where the pressure in the manifold. These are the opinions that we took when we tuned the '05 car and as a result we didn't run a reference line to the manifold.

One thing I didn't think about before, but came to mind after reading Cody's post, is that your refference won't be constant if it is atmospheric. Most times atmospheric pressure doesn't vary much but occasionally it does. Late in our testing in '05 and we had a weekend where barometric pressure dropped by around 3%. We were running a TPS scheme and were now noticeably richer than previously. Had our regulator been referenced to manifold pressure it wouldn't have been as bad. Also, a good reason to run a MAP based system or at least get your barometric correction maps done as soon as possible.

[murpia]

Originally posted by BeaverGuy:
...This now means that for a given injector pulsewidth the fuel flow will be different at different throttle positions. This is means that fuel flow is now not only dependent on pulse width but also throttle position...

Almost: the fuel mass-flow for an atmospheric-referenced regulator is dependent on manifold pressure, which depends on both throttle position and engine rpm.

Here's how I see it, there are some elements here of the speed-density (MAP) vs. alpha-N (TPS) 'debate' from a few years ago...

Many setups will work, but some are easier to calibrate than others and are more 'robust (http://en.wikipedia.org/wiki/Robust_control)'. Don't forget OEMs (both bike and car) don't always choose the technically superior solution, they want to save weight and cost and meet emmissions standards (as always).

Consider system 1, an in-tank regulator referenced to atmospheric with a TPS-based engine control strategy:
Possible sources of fuelling errors include: changes in atmospheric pressure (more or less fuel for the same pulsewidth), changes in throttle flow (dirt buildup, linkage slop, sensor calibration), impossible mapping (you need to visit every throttle / rpm / atmospheric pressure point on the dyno), I could list many more.

Consider system 2, a plenum-referenced fuel-rail regulator with a MAP-based engine control strategy:
Atmospheric changes are compensated by the plenum-referenced regulator and the MAP-based fuelling (OK back-pressure still changes, but is less significant), changes in throttle flow are irrelevent - you are fuelling based on your MAP sensor and if you drive to a mapping point you can't reach on the dyno your fuel-calc can still be based on the MAP-based calculation of air density.

System 2 is therefore more robust - of course it's not perfect, but it has some compelling advantages. There are always issues of course, some engines don't have stable plenum pressures which foul up both the fuel pressure regulator and the MAP sensor (stick a 'filter' in the sensor hose) and transients will always present problems. Also, fuel temperature can be less stable with a fuel-rail mounted regulator. As always, do the analysis, make your choice, and present accordingly to the judges.

Regards, Ian

[Poe]

If you just referenced outside atmospheric pressure, you would decrease the amount of fuel injected into the runners as MAP increased.

So basically, you would have a rich condition at WOT, and it would run leaner at part throttle and idle


If the pressure regulator vacuum line was not attached (referenced to atmosphere), fueling would be richer at vacuum conditions (light load, decel), but would be the same at WOT, naturally aspirated. The vacuum decreases fuel pressure in the rail, boost would serve to increase fuel pressure. The vacuum is used to keep a constant (as close as it can be at least) pressure differential between fuel pressure and manifold pressure.

[Wesley]

Oops, right, forgot the MAP was negative pressure usually.

So you'd be lean in both cases if you were running an engine with power adders?

[Poe]

both cases being what? It would run rich under manifold vacuum, and lean under manifold pressure if the vacuum line was attached during tuning and was one day just disconnected.