Fuel Pressure Regulator

[emry]

Airflow unfortunatly does not like to operate in freeze frames. While our engine has a fixed cylinder volume, the air that makes it into it is a function of many different factors. Load is one that will affect the flow of air through the engine. At any given moment our airflow may be different than a previous moment, even if our TP angle and RPM are constant. Since manifold vacuum is directly proprtional to load this make for a effecient method of computing pulse width. Clearly a alpha-n system is capable of working very well, a well set up speed-density system is just more accurate, just as a mass air system is still more accurate. More accuracy means better economy, less emissions, more power, hopefully less pinging on accel and less popping on decel.

[dlaing]

Again, sorry to all who find reading through all of this minutiae boring or a waste of time. Ryland is bringing up valid points and I feel like giving him valid answers is what a technical discussion board is all about.

I do after all have some qualifications even if Dave doesn't believe me.

What would my beliefs have to do with your qualifications?

Until you present proof, I can't be sure you are correct.

There have been countless pages of posts on this forum making the assumption that there is a relationship between vacuum and fuel pressure, and you are the first to assert otherwise.

FWIW I believe you enough to recommend a change to Ryland's FAQ

http://www.v11lemans.com/forums/index.php?...view=getnewpost

that says

"This port allows the regulator to vary its output pressure with the intake manifold vacuum when connected to the throttle body intake manifold ports. At wide open throttle, when intake manifold pressure is nearly atmospheric, it has no appreciable effect. At part throttle, it lowers fuel pressure to lean the mixture."

But I would still like to see proof.

I guess you ain't a "question authority" type of guy...

[dlaing]

So unless you manage to explain to me why load should affect instantaneous pulse-width, I'm still not following. How much air mass is being sucked (pushed) into the engine is built into each map cell. The ECU must compensate the map value for air pressure and temperature but it does not have to compensate it for load. Load will always be handled by the map cell used for the next revolution...

I'd agree that each map cell can accurately ENOUGH determine the proper pulse width to match the ESTIMATED air mass, but it is only ESTIMATED and it will vary with load.

How much it varies I don't know.

All I know is that if are going down hill and you whack the throttle to 100.00% and if you are going up hill and whack the throttle to 100.00% you will get an instantaneously different demand for timing and fuel mixture despite the fact that you are at the same map cell.

Before the engine position sensor can determine that the RPMs have changed, the piston on the power stroke is facing more resistance going up hill than down.

What happens on the next couple of engine rotations, I am not sure exactly.

The ECU receives a different time measurement between engine rotations depending on load, but at the first and second engine rotation after the throttle is opened,the RPMs have not changed enough to change the map cell position significantly.

Maybe we can add a sensor to the polymer cush drive to measure load?

As more rotations pass I would guess that the map gets closer to sending the ideal signal for the ideal pulse rate, but the load always creates a different demand for ideal pulse width.

I suspect it is better to tune for the high load than the low load.

Maybe someone else can explain better or show me where I am wrong...

[Dan M]

Yes, it needs to know airflow to get it right. Not load. What I am trying to say is this: Provided we have a perfect map value for each cell, according to Rylands post below the pulse-width of each cell will be right at high load AND low load at that cell. So the answer to your question is; it will know the previous pulse-width was right because it looked it up, and compensated it correct. Furthermore, it does not need to know whether it was wrong or right because it doesn't have to use any data from the previous revolution.

 

It can't compensate unless it knows all the data. Systems that are like ours that have a set map and no input for airflow or vacuum do not compensate they just apply the fuel for a given throttle angle and rpm. They do not compensate. They would need to know load to compensate. The reason for air flow meters and map sensors is to determine load so the computer can compensate for changing conditions LIKE LOAD!

 

 

I'm thinking just one four-stroke cycle at a time, freeze framed. On high load, yes rpm will raise quickly but that just means that for the next very cycle we use another map cell that is accurate for that exact rpm and throttle. Likewise for low load.

 

What your saying is the computer is able to predict what is going to happen. They use these sensors to have accurate fuel delivery, that means quick reactions to conditions. The conditions monitered are RPM, throttle angle, and load. You can call it airflow, you can call it vacuum but the calculation that is made from the data these sensors provide is to determine LOAD. I don't think all these manufacturers are throwing things like map sensors and mass airflow sensors on vehicles because they are not needed.

 

This however assumes a perfect interpolation between map cells. It also assumes that the ECU is fast enough to not lag in it's perception of rpm, and fast enough to perform said interpolation in almost no time. But hey, this is a theoretical discussion. It also assumes that Ryland is correct in saying:

 

 

So unless you manage to explain to me why load should affect instantaneous pulse-width, I'm still not following. How much air mass is being sucked (pushed) into the engine is built into each map cell. The ECU must compensate the map value for air pressure and temperature but it does not have to compensate it for load. Load will always be handled by the map cell used for the next revolution...

 

 

Thank you Emry!

 

Dave, I actually did the testing and offered the data. As you said there were many assumptions in prior posts. I do this sort of testing every day. The reason I jumped in to this discussion was to offer solid data collected by actual testing not assumptions or conjecture. I don't know what more proof is needed. I have no authority here, I'm just presenting facts as I've found them. I just wonder why you accept all of the assumptions as gospel but then try as you will to poke holes in results from actual testing.

 

 

EDIT: Raz, check this out (read the first sentence please) http://www.kemparts.com/TechTalk/tt06.asp

 

more stuff here: http://autorepair.about.com/cs/generalinfo/l/bldef_495.htm

 

 

Why is everybody measuring load???

[dlaing]

Dave, I actually did the testing and offered the data. As you said there were many assumptions in prior posts. I do this sort of testing every day. The reason I jumped in to this discussion was to offer solid data collected by actual testing not assumptions or conjecture. I don't know what more proof is needed. I have no authority here, I'm just presenting facts as I've found them. I just wonder why you accept all of the assumptions as gospel but then try as you will to poke holes in results from actual testing.

We are thankful for the data you offered.

I poke holes in the testing when it is questionable.

I have not tried to poke holes in anything that you showed test results for.

I have simply said that "I am not sure I believe you" in what you did not show test results for!!!!

And even then, based on no test results and just your word that I am not sure I believe, I suggest Ryland change the FAQ that says "This port allows the regulator to vary its output pressure with the intake manifold vacuum when connected to the throttle body intake manifold ports. At wide open throttle, when intake manifold pressure is nearly atmospheric, it has no appreciable effect. At part throttle, it lowers fuel pressure to lean the mixture."

 

How is that accepting assumptions as gospel?!?! I am leaning more towards your assumptions, than the previous assumptions, but you think because I don't buy them lock stock barrel, I am accepting the other assumptions as gospel?????

You have shown no EVIDENCE!

 

If your testing showed more than simply, "I'm guessing the spring is overcome at about 10"Hg."

or that you simply applied 15"Hg of vaccum and read the gas meters, then I would not question that vacuum/atmosphere side of the regulator acts as a cut off switch at a specific vacuum rather than a progressive modifier to the fuel pressure regulation.

If you find it unreasonable for me to question that in the face of NO evidence, than tough.

If you had said that you applied vacuum progressively, or in stages of 5"Hg, until you saw the switching point then I would believe you.

Or maybe specs from the regulator's maker???

 

Ratchet's testing showed nothing happens, and I poked holes in it, and thankfully your evidence showed his testing as essentially useless and possibly misleading. And not one Mea Culpa from ol' hatchet.

 

And both Ratchet and Raz suggested that Load won't effect fuel demands at a given fuel cell and I did my best to poke holes in that.

[Dan M]

If you had said that you applied vacuum progressively, or in stages of 5"Hg, until you saw the switching point then I would believe you.

 

That is exactly what I did say. Try reading post #43

[dlaing]

I am not accepting the following as gospel, but it does support our earlier assumption:

"The fuel arrives at the regulator under high pressure (>3 bar), and unused fuel returns to the fuel tank from the regulator, thus forming a continuous flow of fuel around the system. The pressure regulator sits at the end of the fuel rail, and guarantee's that the pressure in the fuel rail is constant. The regulator varies the fuel pressure proportional to the vacuum applied to the vacuum pipe feed pipe on top of the regulator. When the throttle is depressed, and the vacuum in the inlet manifold reduces, the fuel pressure regulator increases the fuel pressure, to cope with the extra demand from the injectors."

http://www.zetecinside.com/xr2/injection.htm

Looks pretty similar to our FPR

[dlaing]

That is exactly what I did say. Try reading post #43

Thanks!

Your evidence is fine. I completely believe you.

Sorry, I don't like reading red text. Too many horror films???

Ryland, will you please fix the FAQ, already?

[Dan M]

I am not accepting the following as gospel, but it does support our earlier assumption:

"The fuel arrives at the regulator under high pressure (>3 bar), and unused fuel returns to the fuel tank from the regulator, thus forming a continuous flow of fuel around the system. The pressure regulator sits at the end of the fuel rail, and guarantee's that the pressure in the fuel rail is constant. The regulator varies the fuel pressure proportional to the vacuum applied to the vacuum pipe feed pipe on top of the regulator. When the throttle is depressed, and the vacuum in the inlet manifold reduces, the fuel pressure regulator increases the fuel pressure, to cope with the extra demand from the injectors."

http://www.zetecinside.com/xr2/injection.htm

Looks pretty similar to our FPR

 

This is correct. Except the fuel arrives but it is not under any pressure until the regulator restricts it's flow to create pressure. Otherwise it would just flow back to the tank unrestricted under very low pressure. It does vary pressure with vacuum as so much has been written about in above posts (please read the text). It just has 2 values that it varies between. It is not constantly variable. Everyone that I have ever tested since they have been common on cars (20+ years) has had 2 pressure values. With vacuum applied and without.

[dlaing]

This is correct. Except the fuel arrives but it is not under any pressure until the regulator restricts it's flow to create pressure. Otherwise it would just flow back to the tank unrestricted under very low pressure. It does vary pressure with vacuum as so much has been written about in above posts (please read the text). It just has 2 values that it varies between. It is not constantly variable. Everyone that I have ever tested since they have been common on cars (20+ years) has had 2 pressure values. With vacuum applied and without.

I read the word proportional to indicate infinite values, constantly variable, no just the 2 values that you measured.

Given your measurements I think they were incorrect when they said, "The regulator varies the fuel pressure proportional to the vacuum applied"

Sorry again, for not reading your reply to Ryland.

[Ryland3210]

Airflow unfortunatly does not like to operate in freeze frames. While our engine has a fixed cylinder volume, the air that makes it into it is a function of many different factors. Load is one that will affect the flow of air through the engine. At any given moment our airflow may be different than a previous moment, even if our TP angle and RPM are constant. Since manifold vacuum is directly proprtional to load this make for a effecient method of computing pulse width. Clearly a alpha-n system is capable of working very well, a well set up speed-density system is just more accurate, just as a mass air system is still more accurate. More accuracy means better economy, less emissions, more power, hopefully less pinging on accel and less popping on decel.

 

All right, I'm going to take this up a notch!

 

I'd like for the moment to narrow the discussion to the effect of load on airflow.

 

Let's suppose that TP angle and RPM are constant, but load changes as you suggest. For RPM to stay constant, something else would have to change. For example, fuel delivery. In this case, a map sensor would have no way to directly know load had changed in order to maintain constant RPM, so we have an imaginary situation.

 

Another way airflow could change would be for atmospheric pressure, temperature, humidity etc. to change, all of which affect density and viscosity.

 

Manifold vacuum is not directly proportional to load, it is inversely proportional to load. I think I know what you mean, but let's not confuse cause and effect. It's not that somehow a change in load causes a change in manifold vacuum (by what mechanism I cannot conceive), it's that to maintain constant RPM, a change in manifold vacuum caused by a change in throttle position/fuel delivery is required.

 

Here's the bottom line for me: All other things being equal except for throttle position and RPM, the only way load can instantaneously affect airflow is if the rate of change of RPM is so fast that the change in time between stokes is significant compared to the time it takes sound to travel from the intake valve to the butterfly. Since I believe that is not the case, I still maintain that once again, a smart and fast ECU with appropriate model can do without a MAP.

 

So why to they exist? Because it can better compensate for variations in atmospheric pressure, filter pressure drop, and throttle restriction versus position than systems which only sense one or two of these variables. Moreover, it can do the job of all three with a single sensor.

 

My understanding is that MAP sensors can more accurately control mixture not because they somehow sense load, but due to reasons above.

[emry]

It's not that somehow a change in load causes a change in manifold vacuum (by what mechanism I cannot conceive),

 

Constant throttle, constant rpm, going up hill = High load, the engine is overcoming the mechanical losses in addition to gravity, etc.

 

Same throttle, constant rpm, going down same hill = Low load, the engine is being driven by the rear wheel which will help over come the mechinical losses, etc.

 

Yes going up the hill we would end up opening the throttle to maintain a constant rpm and going downhill we would close the throttle to maintain a constant rpm.

 

We often forget that our engine to run at constant RPM the crankshaft really us undergoing vast changes in accel and decel. During just after TDC compression our expanding fuel mix is exerting a great amount of pressure down on our piston, our crankshaft is in a state of acceleration. After approx 3/4 of the power stroke our crankshaft now is transitioning to decel. It continues to undergo a reduction of rpm with the greatest amount of decel during the next compression stroke.

 

Yes our second piston will affect this, but lets keep it simple.

 

Up hill our mechincal efficiency is decreased, downhill our mechanical efficiency is increased. The volumetric efficiency of an engine is a function of its mechanical efficiency. Volumetic efficiency is a measurement of how well an engine fills a cylinder to its capacity.

 

I still maintain that once again, a smart and fast ECU with appropriate model can do without a MAP.

Yes it can. Our bike are an example of this. But it will never be better that its orginal estimate of the predicted airflow through an engine. Do our engines run differently with tight or lose valves, why cant a alpha-n system compensate for that?? It has no way to measure airflow. It merely knows what should be going through. Which may or may not be the case. Valve setting affect our valve timing and airflow right?? A MAP sensor would relay that info to be used for fueling calculations.

[Guest ratchethack]

All other things being equal except for throttle position and RPM, the only way load can instantaneously affect airflow is if the rate of change of RPM is so fast that the change in time between stokes is significant compared to the time it takes sound to travel from the intake valve to the butterfly. Since I believe that is not the case, I still maintain that once again, a smart and fast ECU with appropriate model can do without a MAP.

Good point, well made.

[Dan M]

Here's the bottom line for me: All other things being equal except for throttle position and RPM, the only way load can instantaneously affect airflow is if the rate of change of RPM is so fast that the change in time between stokes is significant compared to the time it takes sound to travel from the intake valve to the butterfly. Since I believe that is not the case, I still maintain that once again, a smart and fast ECU with appropriate model can do without a MAP.

 

So why to they exist? Because it can better compensate for variations in atmospheric pressure, filter pressure drop, and throttle restriction versus position than systems which only sense one or two of these variables. Moreover, it can do the job of all three with a single sensor.

 

My understanding is that MAP sensors can more accurately control mixture not because they somehow sense load, but due to reasons above.

 

You said it. The model cannot compensate for the variables. Air density changes power output for a specific throttle opening. Hence changes load. The system that does not sense vacuum or air flow cannot compensate, that is why is is not accurate!

Here's a test for you. Run an engine at a set throttle opening on a hot, dry day with a vacuum gauge hooked up. Mist cold atomized water in front of the air intake. Watch vacuum gauge. The change is small but the vacuum rises as does RPM because the power output of the engine also rises. What happens when power goes up? Load goes down. Same throttle opening. A fixed fuel model cannot know this or compensate.

 

Funny how these two different sites refer to each of these as load sensors. Don't you think?

http://www.kemparts.com/TechTalk/tt06.asp

http://autorepair.about.com/cs/generalinfo/l/bldef_495.htm

 

Bottom line is I think we all know how this stuff is measured and works. I think we are at odds on the terms we are using.

Hint: Never say dampen around here.

[dlaing]

Here's the bottom line for me: All other things being equal except for throttle position and RPM, the only way load can instantaneously affect airflow is if the rate of change of RPM is so fast that the change in time between stokes is significant compared to the time it takes sound to travel from the intake valve to the butterfly. Since I believe that is not the case, I still maintain that once again, a smart and fast ECU with appropriate model can do without a MAP.

The bottom line for me is to start with a good map and twist the throttle as needed.

Our slow dumb ECU could be improved by adding on bells and whistles, but a better solution would be to simply make a better map.

The difference between high and low loads at the same map position could be improved on, but unless we are chasing an unknown small increase in MPG, there is no worthy benefit.

Simply making it rich enough with the right timing for the high load situation is adequate.