Team Integra Forums banner

1 - 20 of 22 Posts

·
Registered
Joined
·
389 Posts
Discussion Starter #1
I was looking at some calculations involving raising fuel pressure and the related increase in fuel flow from the fuel injectors. The example I looked at used an injector duty cycle of 80% to calculate stock fuel flow rate, but when they increased the fuel pressure for dry nitrous injection they used 100% open for the injectors. Is this a valid assumption? Also, how much of a rise in fuel pressure is safe on a stock system? Does a higher flowing fuel rail (like AEM) allow you to raise your fuel pressure higher without overworking the injectors? Sorry for all the questions. once I find out how much fuel pressure we can raise, we can calculate what horsepower the stock injectors are capable of supporting.
Thanks,
Adam
 

·
Registered
Joined
·
3,953 Posts
Injector duty cycle and injector duration are two different things.

Duty cycle measures how quickly the injector can open and close. Riding the 80% redline ensures you're not over working the unit. It'sa measure of time, like RPM.

Injector duration is how open the pentil/disc is durring the duty cycle. This is similar to your throttle body and how it meters incoming air.

Changing fuel pressure doesn't affect either one of these areas, as it's computer controlled and the ECU doesn't know fuel pressure. It assumes stock fuel pressure to be the median for calculations.

When raising the fuel pressure, the extra fuel force will slow the injector's response time, but I'll bet money you would ever feel it.

Your fuel line is rated at 60 something LB MAX. Look at the fuel line and it'll have it printed on it, or at least mine does.

DO NOT piss away money on a fuel rail, unless you're making +400 whp. It's all bling-bling at our power output.

The maximum amount of power one can make isn't strickly limited to the injector, but rater the entire fuel system. It's possible to run your stock injectors at 200lb fuel pressure and have something stupid for a flow rate and support some w**ker amount of power, but you'll never find a fuel pump that will support such.Also note the injector would squeel and cry if you tried to get it to do such.
 

·
Registered
Joined
·
389 Posts
Discussion Starter #3
SouLcHiLd1719 on Sep/10/02 said:
so with a FP of 66 psi on stock injectors? how much hp are they capable of handling?
Ok, thanks for clearing that up for me Gvtec. The book I was reading really made that very unclear and actually confused me quite successfully. When I get home I'll show you the quote that threw me off.

HP = IFR * .8 / BSFC

assume BSFC = .5 and IFR is Injector Flow Rate
170 hp / 4 cyl = 42.5 HP per cyl
42.5 = IFR *.8 / .5
sloving for IFR = 26.5625 lbs/hr

F2 = (sqrt(P2/P1)*F1

Where F2 is final flow rate, F1 is initial flow rate, P2 is final fuel pressure, and P1 is initial fuel pressure. Solving for F2 = 32.5323 lbs/hr

Plug that back into the HP equation and you get 52.0517 HP per cylinder or 208.207 total flywheel HP capable with stock fuel system. This is a 22.5% increase of power from a 50% increase in fuel pressure.
 

·
Registered
Joined
·
31,217 Posts
adam5743 on Sep/11/02 said:
Plug that back into the HP equation and you get 52.0517 HP per cylinder or 208.207 total flywheel HP capable with stock fuel system. This is a 22.5% increase of power from a 50% increase in fuel pressure.
this is assuming you have optimised cylinder filling (volumetric efficiency or sufficient flow capacity, and flow quality), complete burn, and complete removal of the gases burnt from the cylinder on the exhaust stroke.

you must have first enough air to make 208 hp. The fuel controls the rate of burn to prevent instantaneous combustion or detonation.

if it only were as easy as to raise the fuel pressure to a certain point and achieve your desired hp.

the equation tells you what minimum FP you need to get that hp with the injector (and fuel pump) you currently have. You must also satisfy the other pre-requisites to make the power as well though.
 

·
Registered
Joined
·
389 Posts
Discussion Starter #5
Michael Delaney on Sep/11/02 said:
Quote: adam5743 on Sep/11/02
Plug that back into the HP equation and you get 52.0517 HP per cylinder or 208.207 total flywheel HP capable with stock fuel system. This is a 22.5% increase of power from a 50% increase in fuel pressure.


this is assuming you have optimised cylinder filling (volumetric efficiency or sufficient flow capacity, and flow quality), complete burn, and complete removal of the gases burnt from the cylinder on the exhaust stroke.

you must have first enough air to make 208 hp. The fuel controls the rate of burn to prevent instantaneous combustion or detonation.

if it only were as easy as to raise the fuel pressure to a certain point and achieve your desired hp.

the equation tells you what minimum FP you need to get that hp with the injector (and fuel pump) you currently have. You must also satisfy the other pre-requisites to make the power as well though.
I definetly understand what youre saying, my purpose here is to research every restriction on nitrous injection amounts. My means to getting the oxygen for that 208HP in the above example would be dry nitrous injection.

What would be the limit you would be willing to push the fuel pressure to for 1/4 mi passes w/ nitrous? A general rule of thumb floating out there in car land is that you should only increase fule pressure by 50% for reliability reasons, but another rule of thumb apparently based on people's experience is that you can go up to a 50% increase in Power using nitrous injection. For my LS that would be 210 flywheel hp (which I am approaching) and would require an injector flow rate of 32.8125 lbs/hr which would equate to a fuel pressure of 99psi (assuming 44 psi as P1) which is a 225% increase in fuel pressure. Something isnt adding up here Tuan, and I was hoping you could clear this up for me. Am I pushing my injectors way to hard or are these calculations not taking something important into account?
 

·
Registered
Joined
·
2,377 Posts
Gvtec on Sep/11/02Injector duty cycle and injector duration are two different things. said:
Duty cycle measures how quickly the injector can open and close. Riding the 80% redline ensures you're not over working the unit. It'sa measure of time, like RPM.

Injector duration is how open the pentil/disc is durring the duty cycle. This is similar to your throttle body and how it meters incoming air.
In general, duty cycle is the amount of time a device is active during a specified amount of time expressed as a percentage. Example : an injector is on for 1ms and off for 9ms. The ratio would be 1/1+9 = 1/10 = 10% duty cycle. on time / total time measured = duty cycle.



Specific to fuel injectors as stated by Russ Collins:

"We stated earlier that fuel flow is controlled by varying the pulse width or duty cycle of the injectors. Pulse width is the time in milliseconds that the injector is open, while duty cycle is the injector's overall percentage of open time. A 70% duty cycle means that the injector is open 70% of the injectors maximum cycling time. Great. But was does maximum cycling time mean?
You could call it the injector's redline. Maximum cycling time is how quickly the injector can open, fire a pulse of fuel, close and be ready to fire another pulse of fuel. The limiting factor here is the injector's response time. ...typical response times are in the .0015 to .002 second range. In order for an injector to fire every .010 second, i.e., every 10 ms, the duty cycle, which is the actual pulse width of the injector, can only be .008 or .0085 second, i.e., 8.0 or 8.5 milliseconds, depending on response time of the injector."

Injector duration or pulse width is the amount of time the injector is open. The injector is either open or closed (it does take it a small amount of time to go from closed to open (response time)). It is held open longer (longer duration & higher duty cycle) to inject more fuel. It is not a variable opening like a throttle body.
 

·
Registered
Joined
·
133 Posts
kelly on Sep/11/02 said:
Quote: Gvtec on Sep/11/02
Injector duty cycle and injector duration are two different things.

Duty cycle measures how quickly the injector can open and close. Riding the 80% redline ensures you're not over working the unit. It's a measure of time, like RPM.

Injector duration is how open the pentil/disc is durring the duty cycle. This is similar to your throttle body and how it meters incoming air.




In general, duty cycle is the amount of time a device is active during a specified amount of time expressed as a percentage. Example : an injector is on for 1ms and off for 9ms. The ratio would be 1/1+9 = 1/10 = 10% duty cycle. on time / total time measured = duty cycle.

Specific to fuel injectors as stated by Russ Collins:

"We stated earlier that fuel flow is controlled by varying the pulse width or duty cycle of the injectors. Pulse width is the time in milliseconds that the injector is open, while duty cycle is the injector's overall percentage of open time. A 70% duty cycle means that the injector is open 70% of the injectors maximum cycling time. Great. But was does maximum cycling time mean?

You could call it the injector's redline. Maximum cycling time is how quickly the injector can open, fire a pulse of fuel, close and be ready to fire another pulse of fuel. The limiting factor here is the injector's response time. ...typical response times are in the .0015 to .002 second range. In order for an injector to fire every .010 second, i.e., every 10 ms, the duty cycle, which is the actual pulse width of the injector, can only be .008 or .0085 second, i.e., 8.0 or 8.5 milliseconds, depending on response time of the injector."

Injector duration or pulse width is the amount of time the injector is open. The injector is either open or closed (it does take it a small amount of time to go from closed to open (response time)). It is held open longer (longer duration & higher duty cycle) to inject more fuel. It is not a variable opening like a throttle body.
thank you for writing this i was gettin geared up to explain it after i read the first few reply's but you summed it up perfectly.
another peice of valuable info is the current ramp of an inj. i've seen probs with inj. opening to slow b/c the draw from systems is so great , i've also seen drivers for the inj. gettin waisted by a faulty current limiters in the comp.
 

·
Registered
Joined
·
31,217 Posts
The safe max. limit for a fuel pressure increase is 20%. once you exceed 20%, you make the injectors work harder than they were intended for and you have thrown off the ECU fuel map calibrations. The life of the injector is drastically reduced at a 50% increase. Secondly, the fuel map is based on certain assumptions and once you push beyond these, the intended duty cycle for a given engine load input is not correct. Thirdly, you place a lot of stress on your fuel lines and fuel pump, since they now must push against this higher fuel pressure at the rail.

No-one I know would go to a 50% increase in FPR. If you datalogged your car and were able to determine the continuous changes in duty cycle as the rpms increase, you would likely find that when you need that much FP, you are exceeding 95% duty cycle and in some cases using 100% duty cycle most of the time. This is not good. You have pushed to the limit of the injector size and require a larger injector.

Remember that it is always better to get an injector that is too large and then lean it down (using an ECU reprogram or fuel controller along with lowering the FP) than to push an injector that is too small beyond 90% duty cycle to achieve your flow needs.

Not only does leaning down a larger injector put less stress on your entire fuel system, especially the fuel pump (by using a lower fuel pressure), it ensures you have finer control of your fuel delivery and you know you are delivering enough at critical moments (higher rpms or as the injector cycles up the rpm range) with plenty of reserve on tap. The chances of an injector giving up at a critical time will more likely happen when you are cranking up an injector that is too small past it's safe reserve.

So on a GSR for example, whose stock FPR range is 48-55 psi off vacuum, I would never exceed 65 psi without considering an upgrade to a larger injector first. There have been times when I have used 70-75 psi in a pinch (no pun intended) when I had no choice during a tuning session but I have Bosch Porsche 911 fuel pump that is able to deliver with plenty of reserve...most people are still working on a stock fuel pump.

I'm not the nitrous guy here but if I was using nitrous, I would not wait until I needed a 50% increase in FP before I got larger injectors that could deliver the flow for the target hp and use stock FP's. You would like to see your duty cycles drop to below that 90% threshold (preferably below 80% duty cycle) as it cycles up at the higher rpms.
 

·
Registered
Joined
·
2,377 Posts
I have learned more.

The above Russ Collins stuff is correct, but is not the generally accepted definition of duty cycle. The generally accepted definition of duty cycle is the amount of time the injector is on vs. the total amount of time available for it to be on, which varies with RPM.



An example (at 6000 rpm):


6000/2 = 3000 cycles per minute

(It takes 2 revolutions to complete the 4 stroke cycle, and fuel can be injected the whole time.)

3000/60 = 50 cycles per second

1/50 = .02 seconds (20 milliseconds) per cycle

So at 6000 rpm, the total amount of time available to inject fuel is 20 mS.

If the injector is on 10mS (pulse width) at this rpm, then 10 ms / 20 ms equals 50 percent duty cycle.
 

·
Registered
Joined
·
2,377 Posts
I haven't played with the E-manage yet. I have been helping out a friend who owns a tuning shop, and he said he had a customer scheduled for some tuning using the E-manage an would let me know when this was to happen. The customer is in the military, and because of this, sometimes his appointments get changed. I will let you know more when/if this happens.

The user manual for the unit as well as the support tool is available at their site. It pretty much sums up what the unit will do. You might want to check it out.

You can lean out by tricking the MAP (16X16), but can only add pulse width when driving the injectors directly. Since the unit only adds to the pulse width, the stock injector drivers should be doing most of the work.

If I get one, which I plan to do, I will probably only be using it to add pulse width. It does have the injector scalar function. I am not sure what tables this applies to or how it is implemented. I might well have to reduce pulse width somewhere, and I do not like the MAP trick stuff because of the timing which I could compensate for using the timing maps. Looks like I better do some more research and log some A/F ratios on my car first.

I believe you are thinking about using it to run larger injectors and would need to reduce pulse widths. Exactly why did you ask about part throttle?
 

·
Registered
Joined
·
2,721 Posts
Be interesting to create a controller board to override the injector pulses. 4 injector drivers, 4 receivers that can hear the injector pulses from the ECU. A small CPU, maybe an 8051 and some way to adjust the pulse width higher and lower than what the stock ECU had commanded.

I was under the impression this already existed but, you make it sound like this is done by tweeking the map inputs and faking the stock ECU.

If I monitor when the injectors fire, I know what the engine RPM is and have a good idea of what stroke I'm in.

Seems like it wouldn't be that hard to apply some fixed pulse width reduction based on increased injector sizes. So, when the stock ECU think's it's commanding near 100% duty cycle the actual injector duty cycle would be less by some percentage based on the injector size and fuel pressure.

I think doing it wouldn't be that hard. I might have the hardware on hand to do it. The question is, is it even worth doing?
 

·
Registered
Joined
·
3,882 Posts
i think to reduce the pulse width, there will be an inherient lag induced between the ecu's firing, and the actual firing of the injector. unless you reduce the timing of the leading edge of the pulse width, in which case you've thrown off the injector phasing, and could cause some problems.

also, by paying attention to when the injector pulses, you don't really have any idea what stroke you're in, as that's controlled by the injector phasing moreso than anything else, and noone's mapped the OEM honda ecu injector phasing. it's not something that can be adjusted, even with hondata and the like, and only a few feature-packed standalones have adjustable injector phasing.

depending on the duty cycle, i'd also say you'd be hard pressed to know for certain the engine RPM, depending on how the honda ecu works. some ecu's fire based on duty cycle alone--when you edit the pulse width in a rom editor, it's really just synthesized values that are then converted to dutycycles internally in the ecu. and again, without knowing the injector phasing, you're in the dark anyway, even if the ecu fires the injectors based on pulse width alone, and in relation to TDC--the injector phasing could throw you off by a considerable amount, especially if the engine's rpms are changing quickly.
 

·
Registered
Joined
·
2,721 Posts
You know the RPM because you know how often the injectors are firing. Knowing the injections per second and how a 4 stroke works, you get an RPM number

I'm imagining monitoring 4 pulsed injector inputs, perhaps ORing them together to give me a clock showing when the injectors are firing as well as the individual pulses from the injectors.

(ignore firing order here just injector pulses)

I then assume I'd have to run one pulse behind so, I use the timing of say, injector #1 to manage the pulse width of injector #2. I'd always be tweeking the timing one injector pulse timing behind in a continuous cycle.


1 - 2 - 3 - 4
4 - 1 - 2 - 3

So, I don't really care about phasing at this level. I leave that to the stock ECU. If #1 has a pulse width of 20ms and I want only 75% of that, I'd wait for #2 to fire, only permit the injector to remain on for 15 ms, note the real pulse width of #2 and use that to control the pulse width of #3.

I might need some kind of delay function so, instead of lopping off the end of the pulse. I delay the start of pulse. That would be harder but doable. I could use my "start of the injector on time" as a trigger to start the delay before I actually turn on the injector. Then the normal CPU off time would shut off the injector.

The question remains, is there any need for a device that would reduce all injector pulse widths by some fixed amount?
 

·
Registered
Joined
·
2,721 Posts
Just want to address your lag comment. Lag through the device of purely a function of the design. Car's are pretty damn slow when you think about it. Things happening in the ms range are very slow compared to propagation delays through gates and transister on time.
 

·
Registered
Joined
·
3,882 Posts
there's not too much slow about a car at 7500 rpm, when your windows of opportunity close to milliseconds. when spark plugs firing a millisecond or two too early will cause knocking. when milliseconds or two +/- on an injector pulse will make it too-lean.

also, the lag i was speaking of was moreso between iterations of the ecu looking up the data... say, when the motor is spinning up quickly.

if the motors spinning up quickly, and you're using the pulse from the injector before it to make judgements for the next injector, you could run into problems.

this would especially rear its head during accel pump operations. i can datalog on my aem that it sometimes takes up to 40% EXTRA fuel (in addition to the map at the higher-loaded cell) to keep a constant AFR between two differing load values. and it applies this to just a couple injector firings. if you use a firing previous to the current one for modification, problems would arise.


it would be better to lop off the end of a pulse, b/c if you mess with the timing of the rising edge, you've thrown off the injector phasing, which can cause lots of problems, especially in low-rpm operation.

and i'm still not convinced that b/c you know how often it fires, that you can certainly know the RPM. i think there's more to it than that. like when the pulse width starts to approach 90-100%. at this time, the falling edges and rising edges begin to overlap.

anyhow, to answer your question, as to if a device like this is practical and useful... in my opinion, no, not really. people pay big money for real tuning tools--if there was any use for a relatively simple device like this, i'm sure they'd already be around.

if somebody can market and sell an AIT or TPS trick (like the venom boxes and whatnot), that don't do anything... or they can develop and market VAFCs and stuff, that actually do something, and they're more complicated than the project you propose... then this would have already been tried.
 

·
Registered
Joined
·
2,721 Posts
I understand what you're saying. Still in the realm of computers, 7500 even 10,000 RPM's is damn slow. Look at the AEM EMS, it uses what I consider to be a one lung 16/32 bit hybrid CPU to do all of it's processing.

I've been inside the code for a car ECU before. It was completely sequential which meant that overlapping fire time on the injectors didn't happen. Still it would be something to watch for. The RPM thing will work up to the point the injector goes 100% duty cycle which I don't really think will happen. All I really need is one injector line to determine RPM.

10K RPM = 6ms per revolution. A 20 mhz CPU ($5-10), assuming one clock cycle per edge is running with a 50 ns instruction time. So, in 6 ms it can run 120,000 instructions. That's what I mean by slow.

It's an interesting thought experiment. I don't have time to work on it now. I think in quantity this box could be made for perhaps $60-70.
 

·
Registered
Joined
·
3,882 Posts
i'm not talking about device speed at all, when i talk about lag. i'm talking about the actual importance of each signal getting to each injector/spark plug on time, at the right time.

i'm talking about how you propose to determine the reduced injector pulse right now. you basically can't.

if you do your idea, with using the previous injector fire to determine the next, that'll certainly lead to problems. you can prove this right now by unhooking your injectors, and re-hooking them out-of-phase by one injector. it probably won't idle, and throttle response will be nill.

so that leaves you with one other option... wait for the current injector pulse to come from the ecu, somehow "know" what the injector pulse is going to be, and then delay the actual injector firing by the necessary amount of time. but, this is bad too, b/c it messes up injector phasing, which will lead to poor low-rpm performance and drivability.

also, it's known that honda ecu's run the duty cycles right up to the bleeding edge of 100% DC. they do this for better low-rpm atomization and drivability. it's the higher rpms that are going to be difficult to determine the rpm. i think if you want tach readings, you might as well tap into the distributer's TDC signal.

but by all means, if you get the time, build it. who knows, it might work by some fluke, and you could become a millionaire.
 

·
Registered
Joined
·
2,721 Posts
Lets be clear, I never said use the previous injector to determine the TIME of the next injector pulse. I said use it to guess the WIDTH (duration) of the next injector pulse.

Let me suggest this then. Say I build a box that has NO CPU in it. It has 4 inputs that turn off/on transisters to the 4 injectors. Basically it's a transister buffer. The input turns on the output. Will this cause significant delays when feeding an injector?

Now if you extend that a bit, add a CPU (though one isn't really necessary) do the very same thing but, don't pass ALL of the pulse width on to the injectors. That's basically how this box would work.


The logic flow is very easy.

1) Monitor and note pulse width of injector #1

2) When injector #2 fires actually fire injector #2 and start a timer (I think you'd agree that the time to feed an AND gate and turn on a transister is very fast so, the delay is propagation delay through the AND and turn oN time in the transister).

3) When the timer expires at some percentage of #1's pulse width, turn off injector #2. (drop one leg of the AND gate which disables the output transister)

4) Note Injector #2's actual pulse width. (When injector #2's finished, you have to add the 1 to the AND gate for the next injector pulse).

5) When injector #3 fires, repeat the process using #2's actual pulse width as the reference.


You've said it won't work but, you have no numbers or really any good reasons why it won't work. You're the EE here. What propagation delays are there through a logic gate and a transister? You can see the stock ECU STILL controls the timing of the beginning of the injector on time so, there is NO affect on phasing.

So explain to me again why this can't work? I'd like to see some numbers if possible.

Almost 100% isn't 100%. All I need is some edges to work with. I don't really need RPM for this to work anyway.
 

·
Registered
Joined
·
3,882 Posts
the delay and lag i'm talking about comes from your logic flow. i'm not talking about lag from propogation delays in the nano-second regions.

basically, i don't think those 5 steps will work very well. i don't think it will work because you're using the ecu's output to injector #1 to determine the signal to injector #2.

now, this is how i'm reading what you're saying. if i'm wrong, then perhaps some more explanation is in order.

this means, that, say, RIGHT NOW, the ecu wants to inject 5ms of fuel... RIGHT NOW, based on the MAP, rpm, accel pump, etc etc.

but, you propose to intercept this 5ms message, reduce it by some predetermined percentage, and then pass it along to injector #2 when the time comes. meanwhile, the message getting sent to injector #1 is a reduced value calculated from the previous #4's pulse width.

this means that the signals reaching the motor are out of phase, by one firing, with what the ecu is sending out.


combine that with the fact that the ecu is consantly readjusting the pulsewidths based on the o2 feed back, the accel pump data, the decel pump, etc etc, and i think you're going to run into problems.

my biggest concern that i can back with numbers is the accel pump functions. if you're crusing around at, say, 20" of vacuum, and you smack the throttle open to 5" of vacuum, the ecu immediately dumps a lot of extra fuel... but it only does this to the next few injector pulses.

in my car, smacking the pedal from ~18" of vacuum to 6psi of boost, my aem ems sends out a pulse width 40% greater than the actual 6psi value. but, it only does this for around 3 injector pulses, at crusing rpm. i have my accel pump set to bleed off the extra fuel entirely within 200 revs--but it bleeds off exponentially.


so, now, with your modifier box (i'll use 1-2-3-4 firing order to illustrate)... say you were crusing along (20" vac, 10ms inj pulses), smacked the throttle (to 2" vac, 20ms inj pulse), yadda yadda. so, the ecu sees that its first opportunity to do its accel pump function is injector #2, whose firing is coming up. so, it applies the accel value (say 30%) to that 20ms pulse, and sends it along to injector #2.

but, with this box, that extended pulse doesn't get to #2, it goes to #3. what gets to #2 is still something from the 20" vac crusing part of the map. that's where you're going to see some leaning-spikes.

so, that goes on for a little while, 20 revs or so (i think, in the stock ecu), until the motor wants to start pulling the accel pump, and throttle the injector pulses back to the 20ms. it tries to send this first reduced pulse to, say, #3, but with this box, it doesn't make it to #3, it goes to #4. what gets to #3 is still an accel-pump modified value... this is going to cause a rich-spike. (probably not that big of a deal)

it only takes one especially lean cycle to cause a good hit of detonation, and cause trouble. and this would be especially lean, b/c that poor cylinder #2 is only getting the equivalent of a 10ms pulse, when it really needs the equivalent of a 26ms pulse. that's 1/3 the amount of fuel it really needs.


i'd been wondering why you needed an rpm signal anyway, but i figured it was b/c you'd probably want to be able to reduce the pulse width by different amounts at different rpm ranges. like, knocking out 10% at idle will cut out more fuel than knocking out 10% at 7000 rpm, because of how slow injectors open.
 
1 - 20 of 22 Posts
Top