CIS injector flow limitations?
 

Some of us have been debating if the injector is a limitation to how much fuel we can deliver.

I tend to believe they are not.

With EFI injectors, gross flow is determined at the injector based on injector size (20, 30, 55lbs...), fuel pressure, and how long the injector is open (duty cycle).

With CIS flow is determined by the CIS fuel distribution head. The upper lower pressure differential acting on the diaphragm and springs, the system-pressure level, and the internal orifice size -- mostly being what determines our gross fuel delivery capacity.

So, as long as the injector orifice is not smaller than the heads internal orifice, I would not expect a larger injector to be a source of increasing gross fuel delivery.

However, my "theory" is not enough to convince even me. What we need to do is some type of test.

What would this look like.

Maybe test a fuel head on an injector bench with the metering pin at the max delivery position without an injector.

Then to test a CIS injector at the same pressure and make a comparison.

I might suggest a pressure just under the system pressure regulator level to take that out of the equation and ensure the head orfice and injector see the same pressure.

What do we think?

I contacted an Injector Test and cleaning company with some questions to see if he could help.

This is his reply:

If we can get together the different 930 injectors (US, Euro, and 3.6) I would be willing to pay to test them against each other. This is not the best test but it might give us some clues.

If Pelican could source a new one of each for the cause, it might result in a bunch of sales should say the 3.6 injector proves to have more gross flow potential. If so, it might be a better fit with a HF fuel distribution head.

I can maybe contribute some here...
Recently when flowmatching injectors on a customers car,i found several failed spray patterns...Swapped his failed standard 1988 930 euro injectors with a few spare euro standard 3.3 964t ones i had and i am 99% sure if memory dont fail me with the 964t ones in place they flowed quite a bit less fuel!

Different orifice size and part number on the injectors also.

Spray pattern and cracking pressure seem to be the recommended tests for CIS injectors.

I guess if they do not all "crack" at the same point, the week one is going to drain down with shut down.

Is there a way to just measure the orifice size on different CIS injectors?

I have not had one in my hand for a long time.

Keith,

I buried with work right now and can't contribute much. You might consider calling John at specialitauto.com. He has more Porsche CIS parts and info than anyone in the US.

Cole

Keith

That's a good idea to create this thread ! :)

Ok, on my side I have 6 pces of 3.6T injectors that I bought with the a complete 3.6T kjet body & head.

As you know, my clutch is failing and I need to replace it. This will be done in few weeks.

As the car will be at the workshop and as I scheduled to change my injectors ( using the new 964T 3.3 injectors) in the mean time , I will take some time to make comparison.

It will be the same head , same fuel pressure, same everything .... I will push the metering plate at full excursion in both case. So I will just switch from 3.3 to 3.6 injectors.... Then we will see if the flowed fuel volume is the same or not.

In the mean time, as my kjet will be fully accessible, I will measure exactly the cone OD /shape and compare it with the 3.6 cone.


How do you think ?

Can't ask for more than that ;)

Thierry,

You are the man!!! :D

Might test with half the injectors 3.3 and half 3.6 at the same time then poor the gross from each side into its own measuring beaker.

Or what ever you think is best.

Just use cheap set of digital scales (£10/$16) and there is no need to measure anything in a beaker..Far quicker and more accurate!
1 gram = 1 ml

I the next few weeks i could find out the difference between standard injectors on a euro 1978 3.3 turbo i have coming in and then test with 1992 3.3 964 turbo injectors in place.

Dude, your post on measuring CIS injector flow is really good so I'm pasting the link to it.
http://forums.pelicanparts.com/911-930-turbo-super-charging-forum/461843-meter-head-flow-adjustment-question.html

Using those common 12 oz. ribbed drinking water bottles from a 6 pack is good for quick comparing the fuel delvery/level form each injector.

J,

Agreed. Everyone should consider doing this. Maybe even with there valve adjustments.

I love your suggestion for testing the cam spray bar for blockage. (Lost a couple lobes my self once.)


T,

I do not know if it makes any sense to measure fuel delivery without any injector at all against the 3.3 & 3.6 units on the end of the injector line as a reference?

Exciting stuff!

Thanks Cole and JF for the refural to John at Special T. They specialize in CIS FD heads and WURs for Porsche and other cars.

I asked his opinion of the injectors being a limiting factor.

Here is his reply:

The specific gravity of water is 1.000 g/ml. The specific gravity of unleaded gasoline is 0.739 g/ml. Picky point I know but noteworthy.

Cracking pressure and spray pattern are the parameters I use to reject a bad injector. Flow rate is used to adjust the fuel head.

Just a thought, addational pressure is probably the best way to get more fuel.

There are a lot of potental restrictions along the fuel delivery path that could become a bottle neck.

Enlarge the metering slits and we may just get fuel earlyer but not any more. Ream the orfice's and the injector might not let it pass.

Bumping system pressure and playing with the upper and lower head pressure differential keeps these issues from arising.

MFI runs at around 230psi so we should have room if we can get fuel pumps that can accommodate us. Seems two 044's linked can get us to maybe 160psi. Also seems the stock pumps start to top out around 100psi or so.

Just a thought.

Not picky at all, Brian. ;)

I remembered water's density to be 1 metric tonne per metre cubed, and that oil floats on water (i.e. is less dense). But didn't have the exact figures, so let is pass.

However, relative masses were all that was necessary for Stu's fine work on Cylinder Fuelling ratios.
:cool:

Not sure what the S/G of water has to do with "CIS injector flow limitation"?

Here is the guy that builds the HF heads replyed about injector flow:

The injectors don't seem to be a restriction from my experience.

I've mentioned this experience before:
When I first put one of Larry's M007 flowtech modified USA fuel heads on my car after installing an HFS turbo the car would not run above 4000rpm at all because it was getting so much fuel.

The Brian Leask adjustable control pressure regulator allowed me to raise boost control pressure alot higher, and steady cruise control pressure a little higher than what Brian set it up for before sending it back to me.
After doing that the modified fuel head works quite nicely on my car which has 964 cams, B&B headers, Garretson longneck intercooler, and 1 bar wastegate spring.

The M007 flowtech modified fuel head did not make the idle mixture any richer than the stock fuel head. The idle CO did not change much at all after I installed it. This is why I think he tapers the metering slits in the control plunger cylinder to a larger diameter towards the top when he CNC enlarges them. I havn't removed it and measured them so I don't know though.

I forgot to mention in an earlier post on these Flowtech M007 fuel heads that he also adds a thin shim to the shim stack under the spring in the fuel heads system pressure regulator to raise system pressure to 97psi on the fuel head I have.

JFairman,

Good heads up warning of the need to rebalanced the WUR / control pressures with a HF FD.

I knew we would have issue with the AFR curves with the first head I talked Larry into doing for Brent930. It flowed 17% more fuel at all points along the metering pin travel. No issue getting idle right, cruse was a little issue, however first on boost was the biggie. We tried getting control of it with the WUR but could not get the first boost enrichment where we needed it. He set his WUR to delay enrichment till a higher boost level but it blew through and it only delayed enrichment a small amount. He adjusted the on boost enrichment CP higher and could make improvement but lost the high rpm enrichment we needed. We ran it w/o any enrichment and got a near perfect AFR curve except it went lean at the top but what a nice on boost AFR through the mid range even without lower CP on boost. That was when I suggested using an rpm switch to clamp the boost signal until much later in the RPM range.

Imagine my surprise to come back to the boards 5 years later to see such a large adoption of the HF fuel distributor and the MSD switch/ solenoid enrichment clamp.

We got some 17% increase if fuel with I think just readjustment of the spring tension and the height at the internal orifices along with higher system pressure (which was not a big secret at the time) without any machining or hard mods.

I have no idea what he and Stephen came up after the first head to get more fuel except for modifying the Lambda head to work like a non Lambda head. That was brilliant!

My dream at the time was to try building a HF FD that kept the Lambda function and dial it in with a 25% duty cycle at the frequency valve. Then bump it up to 75% to get even more fuel than the non Lambda head. Unfortunately, my turbo head was changed to a euro head and I could not find a reasonable C2 Turbo core at the time on the cheap.

Have you confirmed any of the other hard mods like machining to the slits or any other? Stephen or Larry did say they had not done anything to the pin but not sure about the other.

If the slits are larger in gross than the internal orifice that could just accelerate fuel delivery without adding to the total. If the orfice can flow more, such a mod might help w a restriction.

A perfect taper on the slits it could tune and fatten up the later part of the pin travel so that we might not have to play w the WUR.

Seems system pressure and reducing restriction from the fuel regulation function around the orifices is what has gotten most of our increases. Getting more might require a larger opening in the orifices if they are the bottle neck. Maybe then the injectors would become the bottle neck if they are not already.

However, I could be wrong on this but that is how I learn.

----

Many experts have told us fuel is not the ultimate restriction, air is.

It was not until I started thinking about your comment about the turbo snorkel restriction that I stared thinking about this and start to understand that we are approaching the air flow limits of the CIS. Appreciated!

Do or anyone know how large the exit on the metering assembly?

It seems that will set the limit of how large of Turbo intake we can go up to and only if we do not have restriction in the path way to the turbo.

Unless come up with some type of air bypass around the metering system that should set the ultimate limit as Turbo efficiency drops off quickly when it tries to suck air through a to small straw increasing the pressure-differential about the compressor wheel.

Anyway, back to the topic at hand. It seems the injector has up until now, not kept increases in fuel we have found at the head from making it through the injectors.

I am worried that at the higher flows, if the injectors are near becoming a restriction, they may contribute to unbalanced fuel delivery. If the 3.6 injectors are bigger they might be a good insurance. Your encouragement of doing the injector balance test is a very, very good idea.

K

Dear Jim

When I look the very small injector outpout hole and compare it to the size of the metering slits, I really suspect the injectors to be much more restrictive. it doesn't means that the metering slit are not restrictive at all. There is certainly a little margin which can be used in order to increase fuel flow under middle and full metering deviation. I speak about middle deviation, because in your in valuable expérience it is possible that the tuner have increased the slits on the bottom direction which could cause too much fuel at 4000 RPM but which doesn't means you could get tons more fuel at full deviation of the metering arm.

When I see the tons of fuel that the head is able to supply when the injectors are removed ( as Stu did in his matching thread) it makes no doubt to me the injectors are huge part of the equation.

The 3.6T and 3.3T are using exactly the same head distributor. The arm is the same, the metering slits and pins are the same, the valves are the same, the input and output line are the same, same fuel pumps too, but the control pressure is about 0.4 bar higher on 3.6T while the metering cone is a bit larger........ where do you think they get more fuel ? ;)

Maybe the K-jetronic ECU is programmed differently in the 964 turbo to make the lambda valve bleed off more lower chamber pressure at large throttle switch positions. That would send more fuel to the injectors.

I don't know as I've never even seen a 964 turbo up close and don't know what all the subtle differences are in state of tune.

I remember someone posting pics of the 964 turbo and 930 fuel system specs from a book and I remember they were all the same accept for boost control pressure and I don't know where that post was and I think I remember reading it the other way around - I thought I read that the 964 turbo WUR was set to lower boost CP than the 930 one.
If it was thats probably where the extra fuel is coming from under boost.

I have no idea what thread that was in and don't feel like trying to find it right now.

I think area of the slits are not directly what determis fuel flow.

That is, how much that flows through the slits is not equal to the flows to the injectors.

The slits effect the "differential pressure " between the upper and lower chambers. Increasing the differential pressure is like loosening the spring tension in a fuel pressure reg and effects how much the regulator membrane can distort and thus how much fuel passes.

I t makes sense the injector would act to restrict to some level.

Thus, making it larger could increase flow.

The main function of the injector in a CIS is probably to maximize atomization.

With that comes pressure build up, any increase in pressure indicates restriction.

We can overcome that restriction by increasing pressure or by reducing the restriction.


I truly do to know if the injector is "the" restriction. I very much look forward to seeing results of testing.



The rest of this is just educated guessing among friends.

qoute-

"but the control pressure is about 0.4 bar higher on 3.6T while the metering cone is a bit larger......."

Thierry,,,,my data qoutes the control pressure on the 964t 3.3 and the 964t 3.6
to be the exact same on both models! 4.5bar +- 0.2

Keith, if this were true, all of our turbo engines would go rich at max rpm, not lean.
Also, the compressor is sucking double the air across the metering plate than a n/a engine does.

That's true that if you are not familiar to 964T, it a bit difficult for you to follow me. ( As it is sometime difficult for me to follow you on 930 territory ;) ).

To be honnest with you I have the full original Porsche workshop files on my hand. ( both 3.3T and 3.6T)

The 3.3 T kjet ECU supply the same stable 50% duty cycle at WOT to the freq valve ( which is the same as 3.3T). There is absolutely no lambda regulation at WOT.

Fuell pressure system is the same ( 6.1 to 6.8) The warm control pressure is the same 4.5 b . On boost, the control pressure is 1.1 bar shifted on the 3.6T while it is 1.3 shifted on 3.3 T ( Sorry I made mistake it is not 0.4bar less but 0.2 bar only )

By the way, while I tuned my KJET-BOOST controller system on my car . With the stock pumps (before to switch to 044 front fuel pump) , I made the test to trigger the lambda freq valves at stable 75% duty cycle on all RPM. As you can see on the following curve, it brings a lot of fuel ( the car almost stalled) in the mid rpm)....but it doesn't bring same fuel increase at high rpm

http://img697.imageshack.us/img697/6...mbdatrig75.jpg

Sorry, I made mistake a said before. :(

The warm control pressure is 4.5 bar on both 3.3 T and 3.6 T

But the control pressure on boost (300 mb) is lowered by 1.1 bar on 3.6T and 1.3 bar on 3.3T. Thus the boost control pressure is 0.2 bar higher on 3.6T ...

Paul,

Sorry, the key word is 'ultimate'. Yes we are more fuel limited at this time.

I think we can get a lot more fuel out of your CIS systems as they are just mechanical devices. Heck, we could have a HF head built on a V8 core that supplies 33% more than the biggest HF FD we could conceive.

However, I think the tube out of the metering assembly is only about 3" or so. (Dose anyone know the actual size?) Most the turbo's that make over 500hp have 4" inlets.

One way around this would be to convert to a blow through system and having the inter cooler exit hook to the top of the metering assembly and the metering exit to the throttle body. This would remove all restriction from the turbo inducer side (reducing lag) and at 1 bar double air flow we can get through the metering system. (Think 3.2 Carrera blow through turbo conversion making 475whp through a small little AFM opening). The plumbing would be easy with a custom IC. However, it would probably take a lot of work to then get the fueling right.

Crazy, I know.

This is what I thought. CP on boost is lower for the 3.6 than the 3.3. This would give the 3.6 more fuel on boost.

Thierry,

Did you ever test the factory FV at 75% with the increased pump capacity?

Increasing the Lambda FV capacity I think is like making the slits bigger. There may be a limit to how much differential pressure we can generate or how much we can get past the internal orifices. That is find a way to create a higher differential than the Lambda heads can alow. Maybe a more complaint membrane, a larger "unsupported circle on the bottom head section", and or more spring tension around the internal orifice.

The US Lambda head was designed to allow a higher pressure differential than the non which might be its key to supplying more fuel as a HF head than the others.

If that is the case, a larger orifice and or higher system pressure might be our path to more fuel.

---

JF,

I think the main difference between the 930 and 964 Lambda system is the 964 has a provision for acceleration fuel for a short time after start up when at low temp. There is also a relife function built into the 964 head to keep vaccum from building up behind the control pin after shut down and cool off that I do not know if the 930 received. Could be the 930 got both these but I am not sure.

The real difference was on the ignition side.

qoute- However, I think the tube out of the metering assembly is only about 3" or so. (Dose anyone know the actual size?) Most the turbo's that make over 500hp have 4" inlets.

I have spare 964t one on my bench and i am pretty sure by memory it appears more than 3" inches.......in fact even though its about -10 below freeeezing just now here in scotland its bugging me and i will check!!

Gee!
That is restrictive and unfortunately has quite a thin wall,,so not much chance of opening up in size(inside dimensions)

2.85" approx (72.5mm approx)

Your sentence allow me to explain my idea about the full behaviour.;)

Actually, I think it is wrong to think the intake air flow increase considerably at high RPM.
Actually past the max TQ band , we can see the torque is dropping dramatically. This drop tell us directly that barrel filling becomes very poor. ( due to the cams, turbo, intake and exhaust limit...). Somewhere step by step, air can not go in and exhaust gas can not go out.

So past the max TQ band, the flow is mathematically increased by RPM but is decrease by filling efficiency. So I think the air flow doesn't increase so much .

Then be back to the so called Kjet behaviour (stock), I think the max fuel capacity of the Kjet is already achieved at little above max TQ band. Past this point the air flow increase slowly but as the flowed fuel is already at it max, the AFR is increasing step by step either.

Then when we switch to more agressive cams, we never get any problem with mid RPM ( since the kjet is self tuning ) but past it fuel capacity we hit the limit in sharper way. No matter higher or a bit lower RPM is ....

Dropping the pressure control doesn't change anything... the metering plate is fully moved ( or very nearly).

Acting the lambda freq valve doesn't change anything either...For max capacity it is even much better to turn it off and to increase the valves adjs. screws. ( Remember this FV returns an important amount of fuel to the tank)

Increase the fuel system pressure ( by shimming the head) bring almost nothing it self.


....

Sorry keith but no !

3.6T on boost CP = 3.4 bars ( 4.5- 1.1)

3.3T on boost CP= 3.2 bars (4.5 -1.3)


So CP on boost is higher on 3.6T !

I have both Porsche doc in my hands right now...

Thank you much!

Our cars seem to respond well to about a 60mm inducer. That is about 2.4". Most turbo of that size have and a minimum a 2.75" tube connection on them.

I suspect at some point we are asking the turbo to not only compress air but to suck it through the upstream restriction. Thus, boost on on side of the compressor wheel and vaccum on the other.

If someone could figure a way to put a vac gauge on that section and check it at peak HP it would probably yeald some very interesting results.

With this happening it should reduce the efficiency of the turbo significantly at some point.

Thus, the compressor maps for most turbos are not going to be very accurate in our situation.

Thanks for the correction on the CP for the 3.3 & 3.6 heads.


I like your thoughts on where gross air flow is.

Peak TQ is where the motor is running at its most efficient level.

Peak HP is where it is probably using its max air and fuel.

The metering plate moves quickly at first with changes in air flow. Partially to the ratio of the arm to the metering pin but mostly due to the shape of the metering cone.

As we get to higher air flows, the metering plate dose start to slow as air starts finding a more direct path off the end of the metering plate and the plate extends past the metering assembly exit. I say it stalls but to say it slows is probably more accurate.

There is about one full point of added fuel available by reducing CP below stock. This has been the corner stock of the Andial WUR sense the beginning. I believe after about 1.9 bar there is nothing to be gained.


So, after we hit peak HP and air flows fall off, do we want the fuel level to fall off to or do we want to maintain our goal AFR.

If we can get the metering function to compensate for changes in air flow accurately at our expanded air flows it will be self compensating.

Maybe the key is lower restriction injectors and a larger metering cone to a degree.

At some point increasing the flow of the FD is needed but we need to do so in relation to changes in air flow.

Here is a lead on the Gold CIS injector: http://forums.pelicanparts.com/911-930-turbo-super-charging-forum/453683-cis-monster-16.html#post5144489

My guess is they are nothing more than a Porsche 930 injector in a Volvo body.

It looks like a 930 injector might be flowing about 500cc/min at about 6.4 bar.

However, the Gold Volvo Race injector seems to flow 600+ at 8 bar!

It looks like it takes about 600cc under boost to support 100hp.

If we make about 3000cc/ min we might have enough fuel in a stock head for up to 500fwhp if we can keep the air sensor plate from stalling. (another belief of mine).


I am believing the injector is a way to increase fuel.

However, so is a much higher system pressures than we are running as long as we have the fuel pumps to support it.

Thus, the 1994 3.6 Turbo CIS injector might be the king of injectors.

It would be interesting to test a 3.3T against the 3.6T injector at stock System Pressure and see how much of an advantage it might yield.

Fun stuff.

OEM CIS injector capacities aren't an issue. They allow MUCH more fuel than required - with the IA fuel head & the Leask setup I can adjust the control pressure to allow tons more fuel than required....much more than required, top to bottom, for my 514HP 930.

Craig, part of the reason the injector question keeps coming up is because in some cases there may not be enough flow, and switching injectors would be easier than modifying the fuel head. For example, I have an SC CIS which is not flowing enough for my supercharge addition. If other injectors did the trick, then I'd be set. Otherwise, I have to modify my fuel head or get another one from a 930. Looks like you've got a great fuel setup and the HP to boot.

Given the dearth of CIS injector flow volumes, it's worthwhile to add data points on this whenever possible.

Very doubtful, actually -- for a 930. I'd just like to help others not chase down different injectors for a 930 when not needed.
I ought to dig up my flow video and my numbers - they allow HUGE amounts of fuel.

A modified fuel head provides a shedload of fuel,; I am at 1 bar and I use the Leask device to significantly dial back the fuel.
I'd bank on being able to run 1.2 bar and have plenty; at one time I adjusted things to 1.1 bar and still a TON of fuel available.

+1 Craig, it's in the fuel head. The injectors have no metering function, They operate at a constant pressure. The needle valve in the injector vibrates at different rates as the flow changes thus maintaining the constant system pressure. In other words if you can't get the flow through the fuel head don't expect more through different injectors.