Speed controller, circuit change please help?

[55pilot]

My assumption from your description is that you are planning to dynamically measure the requirement for additional fuel by monitoring an O2 sensor in the exhaust. When it senses a lean mixture you will speed up the pump to increase fuel pressure and thus the amount of fuel delivered per unit time the injector is open.

Is this right?

Actual, that is not what he is trying to do, based on what he has posted. He is looking at the absolute manifold pressure, as reported by the MAP (Manifold Absolute Pressure) sensor and increasing the fuel flow to the engine in some undisclosed way.

I am much more familiar with aircraft engines than automotive engines, so my knowledge may be a bit sketchy. The mass flow of air in the engine is proportional to the RPM and the absolute pressure of the air. The ECU looks at the two, computes the mass of air and decides how much fuel to deliver to achieve a slightly lean mixture.

If you boost the mixture past the upper limit for the MAP sensor, the engine is going to underestimate the air mass and will deliver less fuel. In the mean time the exhaust oxygen sensor will see too much oxygen, suggesting to the ECU that the fuel flow should be increased. The ECU is going to increase the mixture up to a point, but it has limits to prevent it from going past isometric. It is unknown how it will all play out.

If one starts adding fuel willy nilly, there is a risk that the mixture can get too rich, which will kill the catalytic converter (assuming the vehicle has one).

What else is unclear how he intends to add the extra fuel to the engine. Considering that his story keeps changing to conveniently rebut concerns, I am not sure if we will ever see a straight answer.

---55p

 

[pxr5]

Ok.
Please understand that i'm 100% into buildintg high performance bikes & cars, so when i say this project is for a car & a bike both are correct statments.
As to fueling a turbo engine, common ways are remap/ chip the stock ECU, run an after market ECU, alter the input signals from sensors,
All the above are expensive & or difficult.
I'm looking into using a rising rate fuel regulator on the stock system, which will flow more fuel in step with boost.This is a very common & safe way to run up to 6psi boost. After that point the stock injectors won't flow any more fuel & will malfuction.
You can use bigger injectors, but then the low RPM running is disrupted.
I will add extra injectors in the inlet manifold & run additional fuel pump, MAP sensor etc these will only operate when on boost,the fuel pressure ( fuel flow ) will change in step with the boost. ( pump speed being modulated via the added MAP or even from a pot' hooked up to the throttle cable).
I have built a square wave 20hz to 20khz 14 Amp injector driver which also adjusts via a trim pot, so I'm sure I'll obtain the fueling as required.
I can also use a signal/s from the stock fuel injectors when seeing boost & then modulate the fuel pressure to get the correct A/F ratio.
I have a Wide Band A/F gauge & sensor which allows you to dial in a correct A/F ratio whilst driving..
My proposed system apart from raising the fuel pressure ( when on boost ) will not have any affect on the stock system.

Also, as I have said this speed controler will be used to operate a water injection system.
Instead of a fuel injector, by using a water solenoid valve, this will stop any detonation & allow loads of boost : 0 ) .
So the is a lot to be gained by working on this, I make you a promise- I'll get this system up & running, you will get some pictures of the results in due course.It will be the back end of the summer, I have many other things to do!
I forgot to say, high boost is only possible for a very few seconds at a time so as long as you don't go lean, it's not hard to get in the ball park & have a good working system.

 

[(*steve*)]

<moderator hat:ON>

I understand your frustration, but perhaps we should let him try to clarify what he's trying to do.

Whilst I agree that some of his statements seem contradictory, it is not evidence of any malfeasance on his part.

<moderator hat:OFF>

 

[(*steve*)]

OK, so you will be using either a voltage (from the MAP sensor) or a pot (on the throttle) to control the pump speed.

Lets assume you are using a MAP sensor that gives you a 0 to 5v signal. Do you want the pump to start running slowly at 0V and be running flat out by 5V, or start running at some other level (say 3.5V) and be running flat out by (say) 4.5V?

 

[55pilot]

As to fueling a turbo engine, common ways are remap/ chip the stock ECU, run an after market ECU, alter the input signals from sensors,
All the above are expensive & or difficult.

There is a reason those are the only ways and there is a reason they are difficult and expensive. Your approach is just too out of touch with reality to have a shot of working reliably. In modern engines you can not just throw in some additional fuel and hope things work out. Everything you do, the stock ECU is going to fight you. That is why the standard way is to change the ECU or at least blindfold it into believing an alternate reality by changing sensors.

In your scheme, if someone who has a good intuitive feel for the engine can glue their eyes to the gauges and closely control the fuel pump and the accelerator, they may be able to keep things working, but this person better not be the one driving. What is the point of such a system?

A second issue is this type of a system is going to violate all emissions regulations, even the UK ones (which are stricter than the US federal limits). You can argue that you can hide things and change the fuel type to pass the test, but the bottom line is that you are violating the standards. Go read the actual regulations on the MOT site.

Finally, when you boost the HP, you are boosting the stresses on the engine. Do you have some analysis indicating that the engine will hold up? One of the reason vehicles have become so cheap in the last 30 years is because manufacturers have figured out how to get rid of all excess stuff. This means they have a healthy design margin, but things are no longer over-designed. When you boost the HP 25% the loads may go up non-linearly and you may be very close to eating up all the design margin. When you eat up all of it, the engine will fail because it is designed to fail.

Finally, you are talking about helping others. Are you thinking of selling this? Just be warned that you could be opening yourself to some serious problems. If one of your customers gets into trouble, the authorities may come after you for selling what they will deem and unsafe and illegal product. All these problems are foreseeable. Ignorance of the law is not a defense, especially ignorance of the laws of physics!

Good luck.

---55p

 

[55pilot]

Lets assume you are using a MAP sensor that gives you a 0 to 5v signal. Do you want the pump to start running slowly at 0V and be running flat out by 5V, or start running at some other level (say 3.5V) and be running flat out by (say) 4.5V?

The desired relationship between the voltage and the motor level is highly non-linear and variable, depending upon other factors. The pump will have to start at some voltage and increase at some slope or more likely along a non-linear curve. The threshold voltage and the shape of the curve is going to change depending upon a number of factors. The most dynamic factor is going to be what the stock ECU is doing in response to what this pump is doing.

The only practical way to do it is to add a microprocessor that implements the curves and chooses which curve to apply and at what threshold based on other feedback. It is starting sound remarkable like an ECU!

---55p

 

[pxr5]

Steve, I will need the speed of the pump to run mid rpm to full rpm with a broad as possible spread.
So 3.5v to full voltage, I quess. thanks for your help!

55pilot, my A/R gauge has data plotting so no problem there.All cars etc run 'clean' at low to upper mid rpm.This is known as 'closed-loop' the ecu monitors the O2 sensor ( the bit that reads the exhaust gas ) & adjusts around the clean ratio of 14.7:1. when the throttle is floored the ecu goes 'open-loop' & ignores the O2 readings.It has to run rich.

The engine would be classed as failing any emissions test at this point, they don't do it because the engine won't operate clean whilst accelarating.
So, you are wrong, the system I will use would not fail any 'sniff' tests.
The only reason I need a methanol mix for my car is because it's a very late twin cam V6 engine that I'm runningh with carbs. ( it does run very well & quite clean, just not clean enough for the late engine emission stats.)
I will not be selling any of this stuff, it's a hobby with passion!
I final thing, I only work on well engineerd & strong vechicals, its the clutch that's seem's to be the main weakness.
My Suzuki GSXR1100 turbo, stock apart from uprated clutch makes 250bhp- will power wheelie at any speed in any gear!

 

[(*steve*)]

Steve, I will need the speed of the pump to run mid rpm to full rpm with a broad as possible spread.
So 3.5v to full voltage, I quess. thanks for your help!

Are you indicating now that you need to monitor engine revs too? What about throttle position? If the driver is at high boost and high RPMs and takes his foot off the throttle, will this device keep the pump running at high speed?

Regardless of the answers to these questions, I am beginning to agree with 55p that a uC is going to be the most viable solution. I guess that since this unit is only controlling the pump speed, it need not be especially fast since the pump cannot change speed instantaneously. Being able to set the desired speed tens of times per second should be way fast enough.

If your skill is not in writing software then you might consider a PICAXE. These are fairly simple to program (and reprogram as required) and will allow you to monitor several inputs and control several outputs. The fact that many have PWM outputs designed for controlling motor speed is a bonus for you.

The major problem with electronics of this nature is that the environment under the hood is not a good one for electronics, so you will probably want to place it inside the cabin (a la ECU).

You will need a good 5V regulator with protection from transient voltages (which are common in cars) and a power mosfet (probably just one) to drive the pump. It may be a good idea to use a proper mosfet driver to ensure the mosfet stays cool. The rest is just sensors and software (I say "just" -- they could be fairly involved too).

It would be smart to keep the programming interface connected and available through a jack to allow modification of the code without disassembly.

 

[pxr5]

No, pump RPM / speed only.
It's looking like a lot of work, may be just do it via throttle lever hook up. Or I could have say three pressure switchs connected to three trim pot's, that would give me a spread of pump RPM at the upper pump rev' range.
Not as slick as using a MAP, but if it works, so what.
I've looked at the PICAXE, very interesting, if only I hade more time!
Thanks for your help!

 

[55pilot]

pxr5: No matter how you try and justify what you want to believe, the reality is that the physics of combustion are not going to change to suit what is convenient for you: The fuel air mixture is going to run an engine only at close to isometric mixtures. You can get a little rich and keep running with excess emissions or you can get a little lean and keep running with excess oxygen, but you can not stray too far from from isometric. This means you need to tightly control how much fuel you put in based on how much air you are putting in.

The engine controller measures the weight of the air by knowing the RPM and the pressure. RPM tells you how many cylinders are filled per minute and the pressure tells you how much air is put in each of them and at what density. With that you can get the weight of air flowing into the engine. The ECU has tables that tell it how much fuel is needed based on the air flow and timing. How much the timing is retarded and advanced changes the optimal fuel.

If you decide to put additional air that the ECU does not know about (the MAP can not measure that high a pressure), you will need to make new calculations in your own computer. For that you will have to come up with fuel flow tables for various conditions and timing. Since the main controller is still running and delivering fuel, you will have to compute the additional fuel that you need to put in by subtracting how much fuel the main ECU is putting in. To achieve all of this, you will need to know the RPM, the pressure, the temperature of the air (CDT) if you do not have an intercooler and exactly what the ECU is doing including where it is running the timing. You can not do all of that with a MAP and one or 3 pots. The fact that you are even suggesting doing this based on accelerator position proves just how much you need to learn.

Regarding your misunderstanding that the engine runs "open loop" the engine NEVER runs open loop. During steady operation, the fuel flow and timing is further refined by closing a loop around the exhaust oxygen cylinder. That keeps the mixture slightly lean and ensures that the exhaust has enough oxygen for the catalytic converter to work. During rapid changes in power, the oxygen sensor is ignored and the ECU runs closed loop only on the air mass flow. Because you successfully tinker with engines does not mean you understand how they work. My knowing how to cook eggs, including a few original recipes, does not make me an expert on farming chicken.

If you do not understand anything else, understand this one thing: The guys that are selling the expensive performance mods, why are they charging so much money? If what they do can be done for a fraction of the outrageous price they are charging, a dozen others would have entered the market and undercut them on price and drove them out of business. What has it not happened? Since none of the thousands of people that are as passionate about engines as you are and know a lot more than you do have done that, that should be a huge clue that things are not as simple as you would like them to be.

This is all I am going to say. I enjoy helping people with ideas. Trying to teach someone who is reveling in his ignorance and unwilling to learn is not my idea of fun.

Bye,

---55p

 

[pxr5]

May I suggest you get hold of a copy of ' Tuning New Generation Engines' by A.Graham Bell. & read chapter, 8, Fuel Injection .
You will find your facts are incorrect.
Dont even try to suggest that the information he gives is false, including methods he suggests that I've based my ideas on are also false.
He's at the very top of this kind of work.
Imagin some knowlegeable electronics type going on to a turbo forum & then having some gear head telling him that he knows nothing about electronics?
You are a very rude & ignorant person!
In the mean time all this has been wasting my time.
Moderaters, why have you not pulled 55p up more than you have over this, I've been subjected to quite unpleasant accusations?
I've tried very hard to remain pleasant, this has not been a very welcoming experience.
Steve thanks for your input on this.

 

[(*steve*)]

No, pump RPM / speed only.
It's looking like a lot of work, may be just do it via throttle lever hook up. Or I could have say three pressure switchs connected to three trim pot's, that would give me a spread of pump RPM at the upper pump rev' range.
Not as slick as using a MAP, but if it works, so what.
I've looked at the PICAXE, very interesting, if only I hade more time!
Thanks for your help!

The picaxe is a very interesting chip, especially so in your case because once you figure out what inputs you need and have it drive a power mosfet to control the speed of your pump, the rest is in software.

If you need to tweak stuff you can do so to your hearts content (limited perhaps only by your skill as a programmer).

I must admit to still not understanding exactly what set of conditions will control the pump speed, but as long as you can get access to the output from (say) the throttle position sensor and (say) a suitable MAP sensor then I guess you could be on your way.

I've been subjected to quite unpleasant accusations

<Moderator hat:ON>

You can probably see by the L plates hanging from my hot air balloon that I'm a very recently appointed moderator and I'm not going to exercise any significant moderation until I'm happy that I understand how the other moderators deal with things. If you are unhappy with anyone's comments, you are able to flag the posts and one of the more senior people will look at it.

I think there are problems in our achieving a thorough enough understanding of how (and perhaps why) the pump speed is to be varied, as we can see a great number of variables that would impact on the design. Added to that there are issues involved in fiddling with a system that has its state maintained by a feedback mechanism. Whilst I am prepared to give you some advice without enough understanding to be able to be certain that it will be effective, p55 is being rather more conservative -- I suggest that he is loath to give advice unless he is quite sure it good and reliable. I venture that he is just as frustrated at not being able to offer you advice as you are with him pointing out what he considers to be problems with your approach.

I have been guilty of being rather less tolerant of some people than I perhaps should be, and I think we all need to carefully consider our posts when they verge toward the emotive.

<Moderator hat:OFF>

 

[Resqueline]

pxr5; I'm very sorry about the way this thread has turned out, I've found it very unpleasant too I can assure you.. You have my sympathies.
I'm also not very old in this game, still learning, and I know that 55p has been PM'ed & advised to moderate his ways but it seems we might have to do it the hard way..
I'm sorry to have to say this in public but even though he knows a lot he seems to be lacking in some departments, creating unneccessary noise around him.

I guess you've been too busy defending your position to find the time to measure the voltages I suggested earlier. If we find your present speed controller unsuitable for direct application I do have an idea about a simple and adjustable PWM circuit.