[Sazerac]

Hi All

So I was getting my list of activities for the Summer ready and one of them was solving my slightly lumpy idle issue on my 91 GT. (I know that GTs idle a little lumpier than the S4s... But the 91 GT is perceptibly lumpier than my 90 GT. Considering all aspects, I have concluded it is likely an injector problem.)

So, I thought to get the LLI on order only to find out that the original fuel injectors appear to be basically NLA. Fortunately, there is an update to a Design 2 which even has the 4 holes on the way. But, I saw a note on Roger's website about them being 14 Ohm. I was actually under the impression the originals were 15 and change.

That observation was EXACTLY the key concern that I had about the use of non-OEM injectors. Any changes in the electrical properties like resistance or inductance could really have an effect on the performance of the electromechanical system as a whole. (!) (Sure it could be corrected with shark tuning but I am preferring not to go that route right now.). Additionally, I suppose these will have to be exchanged in complete sets now...

Has anybody else had to deal with this problem yet here? Or am I the lucky one to test and report back?

[hans14914]

The driver FET in there can probably handle that just fine, its on a mini heatsink to help dissipate any transient peak temperatures.

The part, even with SharkTuning that will get you, is the dead cycle time, small pulse adder, and flow characterization. That stuff is baked into the hard code. You can apply some correction factors, but it wont allow you to go in there and build a table.

The old Ford truck injectors worked for me back in the day.

I intend to send a pile of S4 and S3 injector cores our to be cleaned, refurbished, and bench tested, then take the "mean" injector samples and have them characterized for the sake of getting that data out there.

Its not a high priority at the moment, but will happen in the next couple of months. You may be able to find something that closely matches, but the new EV6/EV14 injectors have a lot less internal mass, so its going to be difficult to find something.

[worf]

... the dead cycle time,

aka Latency, Open Latency? Or something else?

small pulse adder, and flow characterization.

Please explain the first one - I'm not familiar with the term - and the context of the second one.

[hans14914]

Injectors are not exactly linear devices. There are many factors that ultimately have an offset from calculated flow. Here is a short article on the topic:

https://injectordynamics.com/articles/g ... erization/
https://support.haltech.com/portal/en/k ... idth-adder

Essentially, to get a good metering model of your injector, you need a table that plots flow against duty cycle, manifold pressure, and fuel pressure.

Forces, such as vacuum or boost in the intake also effect the internal valve operation, as does pressure in the fuel rail. (These values are not as important on a 928, as there is no mechanism to feed that back into the fuel calculation).

What is baked in the hard code is the assumption of the injectors behavior with short pulses, as this is where the injector is incredibly non-linear. The compensation for this is the short-pulse adder.

If you can match an existing production injector to the old Bosch, as measured on modern equipment, you should have a decent running car. There are a couple of factors you can change in ST2, so finding an injector with similar attributes that cannot be changed in ST2, where overall flow can be compensated as a scaling mechanism could be better. IE a 30lb injector that has similar latency and effective PW run at 63.3% would be a better running option than a modern 19lb injector with "better" linearity...

[Sazerac]

...the new EV6/EV14 injectors have a lot less internal mass...

Many thanks for the insightful comments, Hans. You caught me in an "incomplete reasoning loop" of sorts. I was thinking that if the electrical properties are the same, the behavior will be the same. But, of course, I even said they are "electromechanical" systems, so the mechanics would have to be at least representatively the same... It sounds like the internal mass is different though. So, my observations are completely meaningless...

Ok, I will also check out the links this weekend. Very intersting. Thanks for posting!
PS: send me an email/messaging telling how much I owe you for the amp and hardware for mounting and interconnects.

[Sazerac]

Hans, again, many thanks for the last post.

What I gathered from the articles is that the main problem is non-linearity of fuel delivery in the initial phase of the pulse width. I assume this comes from the electro-mechanical properties of the injector. For instance, a heavier needle in the injector might open more slowly. The static rate they talk about seems to be the steady-state flow during the pulse. So, once the needle is open, the flow is known. I assume there is also a non-linearity when closing like bounce or lag or something.

The only way the non-linearity which needs compensating for could be the same for both the old injectors and the new injectors would seem to be by keeping the exact same design of the mechanical as well as the electrical parts (coils). For instance a lighter needle would need a smaller coil, but the inductance would be different. So, it becomes a bit of an approximation as soon as you change anything, which require tons of iterative testing, because I don't think these non-linearities are really easily calculatable. You can estimate by calulation but then you have to test, modify and test.

Knowing how the car manufactures really don't care about spare parts and only want to make lots of money with the sales of them, I hardly believe the new injectors have been tuned by an iterative test cycle to fit as well as the old ones in the existing system which was tuned with the old injectors. We all see how the quality of spare parts is decreasing from Porsche. Also, it is much easier to design a system and code it with a set of identified compensating factors rather than trying to match a new design to old compensating factors. The latter would be an order of magnitude more expensive for Porsche than the original design actually.

So, yeah, what to do with injectors now? I will think about it.

[worf]

So, yeah, what to do with injectors now? I will think about it.

Shark Tuner and PEMs.

[maddog2020]

an interesting fact about fuel injectors is that if you watch the voltage with an oscilloscope, you can see a hump when the pintle actually moves so you can determine if an injector is stuck without even removing it by looking at the wave form. for more info search youtube for "pintle hump"

[Mrmerlin]

I dont have an O scope BUT I do have a long screwdriver,
this works perfectly to listen to each injector as its clicking away, OR not.
Just place the tip onto the injector then the handle to your ear.

NOTE when fitting new injectors.
I suggest you test them with 12 v to see if they open,
after the test spray some PB blaster into them ,
then open a few times to cycle the PB.
this should prevent a sticky injector from being stuck after you have fitted it.

NOTE with a stuck injector spray with PB into the inlet,
then tap the inlet squarely onto a block of wood then retest.

[maddog2020]

I dont have an O scope BUT I do have a long screwdriver,
this works perfectly to listen to each injector as its clicking away, OR not.
Just place the tip onto the injector then the handle to your ear.

the injector can still click and have a stuck pintle at the same time. the only way to tell is with a scope or have the injector out and see it spray in an injector tester.

[Mrmerlin]

I thought we were talking about injectors that were already cleaned and could be stuck. But I concur the best way is to flow test them,
The PB blaster will keep them from getting stuck after they have been cleaned and sitting on a shelf .

[worf]

You guys should re-read the OP. He's focused on different electrical properties of the newer non-OE injectors (e.g. resistance and latency.)

The original injectors will eventually, regardless of cleaning, be sufficiently worn to change flow and pattern properties.

[Sazerac]

You guys should re-read the OP. He's focused on different electrical properties of the newer non-OE injectors (e.g. resistance and latency.)

I am always interested to learn things, of course, but yes the subject of this thread is the different electro-mechanical properties of the new redesigned version vs. the original version injectors.

Hans provided a great theoretical basis with his links about the potential differences and problems. The next thing would be people with real-world experience in installing the new re-designed injectors. If anybody has any experience, I would love to hear about it.