E Dogg

iTrader: (1)

Maybe ill get more opinions here than general tech....

OK first- I did search and i know the conventional wisdom....(correct me if I'm wrong)

Injector time in ms divided by [(60sec/engine speed in rpm)= ms per revolution)] =% duty cycle

for example
@ 4000 rpm/10 psi, with injector time of 6.32ms

4000/rpm = 6.32/15 gives me a duty cycle of 42%

But this takes for granted my injector is spraying throughout the entire revolution, right?

Lets say my intake duration total per revolution which would be something like degrees of intake duration multiplied by the number of times a face of the rotor moves across the intake ports per revolution of the E-shaft is 60%(just a guess for the example) so that means of our 15 ms of revolution only 9ms is available for injection, which makes the duty cycle 70%!?


is this accurate, or does the easier method mentioned first work? any ideas?

peejay

Think of it this way.

70% duty cycle means the injector is open 70% of the time. It doesn't matter if it is open for 7ms out of 10ms available before the next cycle or if it is 3.5ms out of 5ms available.

Pulsewidths are important to know for certain things (are you able to get fine enough control under certain conditions, etc) but if you just want to know if you're running out of injector, duty cycle is all you need to care about.

E Dogg

iTrader: (1)

my concern is is if the size of the injector is too small that it has to be open beyond the intake stroke(for lack of better word) to I'm shooting fuel into a closed port, even though the math will show my duty cycle is under 100.
right?

jgrewe

iTrader: (1)

Nothing to worry about, the usual problem is having injectors too big to idle well. An injector takes X number of ms to open then close and the short time to inject the extremely small amount of fuel to idle can be impossible to get with big injectors needed to feed the top end HP.

Air is compressable, for the short time the valve is closed it stacks up in the intake tract, the fuel hangs around waiting too. Beyond that there is a ton of reading you can do about Helmholz resonance, wetting of the port walls, and the speed of the air moving through the intake system. Here's a little taste; peak torque will occur at around .52-.55 Mach.

7envy

iTrader: (1)

Injector duty cycle is calculated by the injection pulsewidth / injection window.

So at higher rpms, your injector duty cycle increases as your injection window decreases. Assuming your pulsewidth stays the same.

Here's a numerical example:

Pulsewidth = 6ms. Injection window at 9000 rpm is 1/9000 (minutes per revolution) * 60 (seconds per minute) * 1000 (milliseconds per second) = 6.67 ms.

Duty cycle at 9000 rpm = 6 / 6.67 = 90%

Substituting 1000 rpm, we have injection window = 1/1000 * 60 * 1000 = 60 ms.

Duty cycle at 1000 rpm = 6 / 60 = 10%

Hope that helps.

E Dogg

iTrader: (1)

that's what i was asking about, so it appears the injection window does cover the whole revolution and not just part of it. Thanks for the help guys!

7envy

iTrader: (1)

The question is, how do you then figure out the appropriate pulsewidth based on airflow? How do you calculate the REAL injection window? The intake port is only open for a fraction of the total combustion cycle, and depending on intake porting, could be opened up a few more or less degrees.

What about the injector type? Peak and hold vs saturated? These are all factors that one must consider to properly tune a car. Going by trial and error is not an engineering approach.

peejay

You don't. The computer does that. All you do is program how much fuel you need into the map, the computer does the rest.

It's the Al Bundy algorithm. Don't know, don't care.

The fuel does not *have* to be injected while the port is open. Carburetors don't. The K-Jetronic in my VW does not. (The nozzles spray full-time, and flow is altered by changing the line pressure. Neat setup) The factory Mazda method doesn't either, at least for the earlier cars.

That's for the ECU to worry about, not you.

Yes, it is. It's scientific method. Theory is only generated by observation and empirical data, which you aquire through trial and error.

7envy

iTrader: (1)

Well... If you look at the rotary video on youtube you will see that the injectors only spray for a portion of the cycle.

How do you calculate the amount of fuel you need to input into the computer?

peejay

Because EFI injectors flow a certain amount of fuel, period, and you vary the amount of fuel by only opening them for a certain amount of time.

Thus the concern with duty cycle and such.

The computer does the calculation. If you want to know how to tune, you add fuel until it's happy, or a safer way is start with way too much fuel and take it away until it's happy.

It's really not difficult. After you have tuned a few cars you get a feel for how a map is going to want to look, and you can enter in a rough guess map (or, if you're really experienced, use one from a car you previously tuned that had similar characteristics) and then use that as a starting point.

This is why big name tuners are worth big bucks. I remember when we had an STi tuned by Tim Bailey, and on the phone I told him yeah, it's this engine with mild porting and this turbo and that intercooler and this size injectors, and he said okay, I'll e-mail you a map that worked okay on a similar car, but I'll add some fuel everywhere so you don't hurt the engine on break-in. Downloaded the map, uploaded it to the ECU, and it ran fine, if a bit rich, just like he said. (This would probably have been the eleventy thousandth GT35R-fed STi he'd tuned... ) A month or so later when he was in town, he did an actual vehicle-specific tune, which consisted of running it on the street and on a chassis dyno while he watched the gauges and made adjustments.

You know, *trial and error*...

7envy

iTrader: (1)

That's what I'm trying to get at though... Opening it for that "certain" amount of time.

You can do it trial and error, like you said, by over-enrichening and then reducing fuel until you maximize torque or power.

Or, you could calculate a theoretical number from a bunch of equations to get you very close to where you need to be, and adjust from there. I favour the second approach because it would help me understand tuning better, and it would allow anybody who has had 0 experience tuning to build a map from scratch, without even having touched any software.

peejay

The way everyone does it.

The way nobody does it. Well, the OEMs might, but they also do things like motor a mockup engine to measure airflow and cylinder pressure characteristics. And most of their equations are merely extrapolated from empirical data... that whole "trial and error" thing all over again.

"Very close" is a remarkably wide range. I tune non wideband equipped cars by making them run cleanly, then (under full load) leaning out until the O2 sensor switches lean, then adding about ten-fifteen percent fuel.

The variables involved are so numerous and difficult to quantify that by the time you find a way of measuring them, generate the equations, and measure them all, you could have tuned the car the old fashioned way about a hundred times, from scratch. And you'll be doing the final tuning *anyway*...

And a lot of the variables are difficult to quantify... Changing the rear muffler on a rotary will affect the idle mixture by quite a bit, not just how much fuel is needed but how rich the engine wants to be in order to run smoothly. This even counts if the car has a mid-muffler and a catalyst.

The *spark plug type* will affect the idle and low load drivability.

If you want to do it this way, though, be my guest. You'll be on your own as far as generating the formulas and figuring out what needs measured, though.

7envy

iTrader: (1)

I understand where you're coming from. The practical aspect of tuning is AFR based. You just target a particular AFR you are trying to reach, and key in a percentage of fuel you'd like to be added to drop/raise the actual AFR to the target AFR, where ~10% change in fuel changes 1 point of AFR.

I'm working on a couple of formulae now. I'll post my excel sheet when it's done.

7envy

iTrader: (1)

Got it. Let me know if the figures seem right.

7envy

iTrader: (1)

Added duty cycle for Primary and Secondary injs.

peejay

Let me put it this way.

My friend Alex recently put a 13B 6-port in his FB. He is using MegaSquirt. I sent him a map that I knew to work quite well on the 6-port I had in my own car. I was using a GSL-SE engine which has better exhaust port flow than the Series 4 engine he is using, but it would still be good enough to make the car start and drive.

It barely ran at all! He had to add massive amounts of fuel everywhere in order to get the car to even run.

Difference? I was using the stock GSL-SE intake manifold and a stock style exhaust, he has an IDA intake manifold with 48mm ITBs, a good header, and a 3" exhaust.

But the ports were the same size/duration...

Given how especially sensitive rotaries are to intake length and port flow, and how difficult it can be to quantify it, trying to calculate a map will get you about as close as a SWAG estimate, only for a lot more work. Take a look at the IPRA intake manifold thread on ausrotary, their unique racing class limitations have made for some great innovation in the field of intake manifold design. On the same ports that would be lucky to make 260-280hp crank with a regular intake setup, they are able to get 100hp more than that, with the main limitation being the engine's ability to stay together, not airflow.

Would that affect the tuning of the car? Oh you bet it does.

7envy

iTrader: (1)

With an IDA intake manifold, he has MUCH more flow than you. His volumetric efficiency is above 100%, while ours start off at 80%. This is why he had to add more fuel everywhere.

Thank you for that data though. It will allow me to fine tune my equation further.

7envy

iTrader: (1)

Disregard what I said, probably wrong. I'm reading up about carb tuning.

Edit: I found the answer in: Holley carburetors, manifolds & fuel injection By Mike Urich, Bill Fisher

"With the same engine speed, a larger carburetor has less velocity through its venturi (air passage). So, there is less pressure drop and a greater tendency for the fuel to to come out of the discharge nozzle in liquid blobs or large drops that are not easily vaporized. [...] The throttle-body injection system's advantage over carburetion, especially at low airflows, is that fuel is delivered at a fairly high pressure drop and there is less chance for the fuel to come out of atomization and puddle."

That said, I need to now find an equation that relates airflow to air velocity.

jgrewe

iTrader: (1)

You can melt down your computer with spreadsheets but until you can test the actual ports and intake/exhaust system before you build with a flow bench, then test the engine on a dyno after its built what you are doing is just mental ************. Don't get me wrong, I like what you've been able to come up with on the sliderule but there are eleventy million other variables that come into play.

You've got the thinking part covered, its time to get your hands dirty! Oh, and buy or build a flow bench to start.

7envy

iTrader: (1)

If I had the money (or the talent, like DelSlow - see fabrication) to actually do something other than think, I would. Unfortunately, all I can do now is think about what I would really need to achieve a certain hp goal. Saves a lot of money down the road when you realise something isn't sufficient.

jgrewe

iTrader: (1)

Been there...

I hope you can test your theories soon. There is a program called Pipemax that has been tweeked over the years with real world numbers from people after having flow bench numbers then dyno numbers it has some pretty cool features. Maybe you could start with the info from the boinger engines and perfect the numbers for the rotary.

7envy

iTrader: (1)

Thanks for the support and suggestion.

Unfortunately Pipemax costs $40... and it's no more helpful than what the user knows about his own engine. I doubt it's open source too, as in the equations used to arrive at the final HP figure are undisclosed. Anyway, my goal was to break down all the forum misinformation about "adequate" fuel setup to achieve a certain hp goal. Also to help figure out if a tuner is extracting close to the theoretical maximum hp. If anything, it's been a great learning experience for me.

I have heard flow bench numbers don't work on a rotary. Why is this?

jgrewe

iTrader: (1)

I hear that from people that have no clue what a flow bench can do. A rotary is pumped by air just like a piston engine, if you know what is able to get in a what is able to get out you can tell a lot. You can also keep yourself from going too big on ports and port volume and killing velocity and low end drivability.

The difficulty is setting up the bench with fixtures to flow parts in the way they interact on the engine.

peejay

I don't have the money or the talent, I don't let that stop me.

You don't get talent by reading books. You get talent by turning metal into chips, and later by turning engine parts (and transmission and suspension parts, in my case) into smoking wreckage.

There's a quote from "Ozy and Millie" that I use as my signature on dirtyimpreza.com.

"It is a form of wisdom to walk in the path of the wise who have come before, learning from what they got right... But to be remembered for wisdom, one must be willing to make the most spectacularly bad decisions, and learn honestly from them."

All of the greats in combustion engine history didn't learn from doing math, they learned by turning up the screw and seeing what worked, what didn't, and what broke, etc.

Walk in the path of the wise, then. It's been scienced out as far as ports/turbo sizing. Fuel system is the easiest part - big enough injectors and a big enough fuel pump or two to support the power. (Assume it will use more fuel than a piston engine of the same power, and need more CFM of turbo flow. Rotaries aren't as efficient as piston engines. Fact of life)

j9fd3s

iTrader: (3)

ive been here too. if you know bsfc, and volumetric efficiency, and the approximate power, you can come up with duty cycle numbers.

ive done it and had it work a couple of times too.

the only gotcha is that the bsfc, ve and hp numbers are educated guesses based on prior experience, and lots of reading

so its good when you are starting from scratch, and if you're realistic about it you can start tuning (by trial and error) with something that is in the ballpark.