Index
Home
About
Blog
From: John De Armond
X-Source: The Hotrod Mailing list
Date: Mar 1992
Subject: Fuel Injection poop
Note: I've been consulting to the Rutgers Formula SAE team regarding
engine management for some time. I thought bits of our conversations might
be of interest to the list. I'll post some info on Formula SAE later.
Enjoy
---------------------------------------------------------------------------
I've done some injector evaluation today and the results may be of
interest to the team. It's really nifty to reread an old document
(in this case J1832) and re-evaluate it in terms of new resources (in
this case, the fluke scope :-) I decided to set up some of the tests
specified in the standard and look at the results.
The test setup is as follows:
The injector is from a Datsun Z car. It is a low impedance (2.3 ohm,
6.5 millihenry)) injector and is driven by a saturated driver through a ballast
resistor in its native environment. In the test setup, A 13.5 volt power
supply supplies power to one terminal of the injector. The other terminal is
sunk to ground with a darlington transistor pair (2n3055 pair connected
as a darlington.) The base is driven from a function generator that is
driven from another function generator (so that width and frequency can
be varied completely independently.) A 0.1 ohm resistor is in the
emitter lead of darlington pair to measure current.
An Endevco piezo accelerometer is attached to the injector with a rubber
band and serves to detect the mechanical opening and closing of the
injector. (A cheap substitute for the Endevco is a phono cartridge
with a #8 or #10 lead shot glued to the stylus.) The endevco is attached
to one channel of the Fluke scope and the other channel is connected either
to the resistor for current or to the collector of the darlington for
voltage.
The first test involved driving the injector saturated at full current,
about 4 amps. The pulse width was increased from zero until the
injector opens. The injector started lifting at 1 ms and exhibited full
opening at 1.25 ms. Pretty much what we expected. Here's the interesting
part. The closing time is over 1 ms! This is with no flyback diode -
the kick goes to over 40 volts.
I next simulated a peak and hold driver by current limiting the power
supply and adding an output capacitor which supplies the peak. With
the hold current set to 0.5 amps (low as I can easily go), the closing
time is only 0.25 ms. I suspect that 250 ma would suffice as a hold
value and would cut the closing time even more.
Since the Electromotive is a saturated driver, allbeit to high impedance
injectors, this closing delay may have a significant effect on the
ability of the engine to idle. I've gained almost 10% in dymanic range
by going to peak and hold. It would be fairly trivial to add peak/hold
to the Electromotive system via an external driver stage with a
one shot.
I'd suggest you measure the closing time on the car (use a phonograph
needle if you can't find an accelerometer) and see if those injectors
behave similarly. If you see an extended closing delay, you might
want to build an external peak/hold driver. I can suggest a design if
you like.
From: John De Armond
X-Source: The Hotrod Mailing list
Date: Mar 1992
Subject: Re: Fuel Injection poop
>>about 4 amps. The pulse width was increased from zero until the
>>injector opens. The injector started lifting at 1 ms and exhibited full
>>opening at 1.25 ms. Pretty much what we expected. Here's the interesting
>>part. The closing time is over 1 ms! This is with no flyback diode -
>>the kick goes to over 40 volts.
>
>Two questions come to mind just off the top of my head John, what part
>does fuel pressure play on opening and closing (any?)
Da books say none. I can sorta confirm that and will be able to
fully confirm it as soon as I get my flow stand finished (day or two.)
We looked at opening and closing times of the injectors on the race
car but did not have an accelerometer. When you look at current draw,
you see a little jig in the rise time when the armature strikes home.
You see a similar jig in the flyback on release. I could not see any
difference using this measure. With the accelerometer, I'll be able
to tell for sure.
>and without
>digging out an ol' transistor book, what is the collector breakdown
>voltage on the that transistor??
Oh, 60 volts or so. The flyback pulse is real sharp so I don't think
the tranny is breaking down. Even if it did, I'd expect only some
zener conduction considering how low the coil inductance is. While
I had my bridge out, I measured a convenient relay just for grins.
1.5 henry. The few millihenrys of the injector don't look so bad.
The three other systems that I've put my scope on (BMW, Toyota and the
Electromotive system) all switch in the same manner, ie, they allow
full flyback voltage to develop.
Amongst the other experiments I'm going to do is one where I drive the
injector from a bridge and apply 500 us or so of reverse drive at
release time. That should kill the electrical delay. The only remaining
lag would be the mechanical inertia of the pintle valve. The accelerometer
output looks like that might be as low as 100 us. With the gain turned up
on the accelerometer charge amp, one can see the noise of the pintel
moving before the actual touchdown is seen.
I've ordered the optical printer interface for the Fluke Scope so I'll
be able to make screen dumps of this stuff available to anyone who
wants it. Maybe even as a GIFF.
John
From: John De Armond
Date: Tue Mar 31 22:38:32 1992
Newsgroups: wiz.hotrod
Subject: Fuel injection
>John,
>
>Just got off the phone with my brother-in-law who owns Pacific Auto
>Wrecking here in Seattle. He says a fairly low mileage <30K, late
>model ~88^ Camaro or Firebird 350 TPI with sensors, wiring and the
>ECU and related items such as the fuel pump, should run around $1400.
>That may vary according to the area you're in, but provides you with
>a baseline.
Vic,
Thanks for the poop. That and the other one with all the model numbers
was most helpful. I'm now armed and ready to go shopping.
>FYI, my nephew and I dropped a 350 short block into his 305 TPI with no
>other modifications. Turns out the injector size is critical as well as
>a few other things. Aftermarket chips, repeated trips to local experts
>and so on did nothing to improve the situation. Finally rebuilt the 305
>and swapped back in for the 350 and now all is okay.
Did you try buying an ECU for the 350? I'd imagine that what bit you
was simply the difference between the engines in the quantity of fuel
needed for a given MAP. A $chip$ would not help the situation.
>I know you're down on carbs, but aftermarket injection is still way
>off from being cost effective and easily integrated for the average
>enthusiast.
yeah, I know. I'm trying to remedy that problem.
>I'm glad to see someone with your smarts is working on it. It gives me
>hope that an effective solution may be forthcoming.
Thanks a bunch. I'm gonna try to demystify fuel injection for people.
Following is the first installment, an exercise on calculating
fuel injector sizing.
John
---------------------------------------------------------------------
This calculation was done for an engine that had been originally equipped
with carburators. Since this engine was fuel injected from scratch,
it is a prototypical example.
The target engine is a 600 CC 4 cylinder 4 stroke motorcycle engine being
used in a small race car. My job was to fuel inject it. Peak power
is designed to come at 12,000 rpm.
First thing to do is to calculate the amount of air required at full
power:
600 cc * 12,000 rpm * 0.5 (every other stroke) = 3,600,000 cc per minute.
each cylinder uses 900,000 cc per minute.
Assuming a 0.8 volumetric efficiency, that works out to 720,000
Note: 0.8 is a good VE for stock engines. For better manifolding and
a good exhaust, use 0.9. For turbocharged/supercharged engines, use
a VE equivalent to the absolute pressure ratio minus 0.1. For example,
14 psi gauge manifold pressure is a ratio of 2:1 absolute. A good
VE would be 1.9.
Air density at STP is 0.00129317 gm/cc so the engine consumes
931.2 grams per minute.
For an air/fuel mass ratio of 11:1 (good full power mix), we need
931.2/11 or 84.7 grams/minute of gasoline. The racing fuel we use
has a density of 0.775 (from the data sheet) so that means we need
109.23 cc/minute of fuel. The nearest off-the-shelf fuel injector
flowed 120 cc/minute so that's the one we chose. [1]
An engine revolution at 12000 rpm takes 5 ms so a whole cycle takes 10 ms.
That entire 10 ms is available for injection. According to the above
calculation, the injector should be open 10 ms * (109.23/120) = 9.1 ms.
On the dyno, the engine made peak power at 11,500 rpm with an
injector open time of 8.776 ms. Not bad calculating, huh? :-)
BTW, this engine went from 55 to 70.5 hp simply from adding the FI
and a tuned intake. No other changes.
I usually like to keep the open time below 80% at these engine speeds
because a typical electronic injector takes 1 - 1.25 ms to open
and some non-zero (around 0.5 ms) time to close. Also, I need a bit
of time over 100% so I can do acceleration enrichment.
Observe that the injector is flowing practically all the time. This is
OK because the closing of the intake valve delimits one cycle from the
other. This is necessary in order to give enough dynamic range to
allow the engine to idle. For example, this engine at 2000 rpm needed
an open time of under a millisecond. Since the injector starts flowing
before it is completely open, it is impossible to calculate the timing.
it must be done experimentally.
If a standard injector size is just a bit low in flow capacity or you
determine that you under-allowed for acceleration enrichment, you
can simply increase the fuel pressure a bit. Over the narrow pressure
range of interest, flow and pressure can be treated as linear.
There is a tradeoff in that with increased pressure it becomes more
difficult to regulate idle mix. You get to the point where the
injector is not completely opening before being closed. This is
not the way to get consistent idle.
------------------------------------------
[1] Getting data sheets on fuel injectors is an exercise in futility for
the average hotrodder. I'll address how to actually measure the
delivery in another article.
Date: Thu Apr 30 18:53:01 1992
X-Source: The Z-Car Mailing list
Subject: Fuel Injectors
From: John De Armond
carpenterj@asuvax.eas.asu.edu (Jim Carpenter) sez:
>*Question*: I need some new fuel injectors. Does anyone
>know of a source of good inexpensive (less than $60 each)
>injectors. Even JC Whitney seems to be expensive!
>I am having a problem with engine miss at hot start that seems to
>be injectors sticking until they cool down. I need to
>investigate this further, but I may need to replace some of them.
I've been looking too and $60 seems to be about typical. You might want
to check with a couple of places that rebuild injectors. Typically
this will be a diesel injection service that has branched out.
I have one near my house that I plan on visiting in the next couple of
days. The guy at the car parts place said they sell 'em pretty cheap.
You might want to consider cleaning the ones you have and even testing
them. I've designed and built a computerized testing and cleaning
system. I've yet to get an injector that did not have gross mechanical
or electrical problems that did not respond to cleaning.
Cleaning is very easy. The Z injector uses a removable spray shaping
cup that fits over the end of the injector. Remove this cup, clean
the inside very good with carb cleaner. Then clean the pintle assembly
and nozzle tip.
Internal cleaning involves energizing the injector and pumping cleaning
solution through it. My assembly consists of a small gas can, a
spare fuel punp, a spare pressure regulator, a standard Z fuel filter
and a can with a drain tube soldered in for the injector to squirt into.
This is hooked up just like in the car. Energize the injector with
12 volts through a 10 ohm power resistor. You may have to momentarily
jumper the resistor to get a sticking injector open. Then pump fluid
through the injector until the spray pattern is uniform.
I use Chevron Techron undiluted as my cleaning solvent. Works very
rapidly.
It helps to operate the injector while cleaning. I've written some
code that runs on a PC and drives a power transistor from a bit on
the parallel port. A common darlington transistor connected
through a 1kohm resistor in the base to a bit on the parallel port
is all that is required. Ground the emitter, hook the collector
to the injector and connect 12 volts to the other injector terminal.
Do not use a snubbing diode. What I do is tell the program to
pulse the injector 5 ms on, 10 ms off. This pulsing helps clean internal
wiping surfaces. You can safely drive two injectors using the
TIP120 NPN Darlington transistor, p/n 276-2068 from Rat Shack.
To clean all 6 at once, simply grab three bits off the parallel port
and make three drivers.
If you don't want to fool with it and can stand the shipping delay,
get with me in email and we can work something out where I could
test the injectors.
John
Date: Tue May 19 15:36:57 1992
Subject: Re: fuel injectors
>You could try to listen to the injectors through a screw driver or a
>socket extension to see if they are working while the engine is
>running rough. I wouldn't expect that such an electro-magnetic device
>would be so sensitive to such a low temperature threshold.
I've been doing a bit of research in this area recently. I could
readily believe excessive temperature could cause the pintle-style
injector used on Z engines to bind or take longer to open.
A bit more elegant way than a screw driver to monitor the injectors
is to attach an accelerometer to the body of the injector with
a bit of superglue. I use an Endevco accelerometer but I've also
made almost equally good accelerometers by glueing a #6 lead
shot to the needle of a cheap phonograph cartridge. The shot is the
reference mass. Simply connect the output to either an audio amp
(Radio shack, ~~$10) or preferably a scope.
I'll be looking into the high temperature issue a bit more in
coming days.
John
Date: Thu May 21 16:00:34 1992
Subject: Re: fuel injectors
>This is an interesting approach. Is it possible to evaluate the
>opening of the injector with the impact of closing as the primary
>forcing function? It might also be interesting to look at the
>frequency spectrum of the accelerometer signal.
>
>Please post any conclusions you can make.
Absolutely. In fact this is the method specified in SAE J1832, the
specification for testing electronic fuel injectors. With the Endevco
accelerometer, a good charge amp and a digital storage scope, I can
see the slight motion of the injector as the pintle starts moving in
the other direction and when the pintle hits home, a large damped
resonance is seen. Same thing upon closing. I have correlated this
initial motion signal with a precision laser position gauge looking at the
physical motion of the pintle. I have not tested it yet but I believe
the phonograph cartridge method will be as sensitive.
I have designed and constructed a fuel injector tester that implements
most of SAE J1832. It cost less than $200 and is completely software
driven through the parallel port of a PC. Construction of this
device and its use will be the first feature series of my new magazine
"Performance Engineering". In the meantime, anyone interested in
playing with this stuff can get a copy of the source code (requires
turbo/borland C) via email to me (jgd@dixie.com). Source is necessary
because a machine specific compensation factor is necessary to
compensate for the execution speed. Actual timing is done from the
PC real time clock but manipulating the port bits takes a few microseconds.
Drawings and a brief description of the unit can be had by sending a
large SASE with sufficient postage for about 20 pages and $5 to cover
my reproduction costs to the address at the end of this article.
You should also purchase a copy of SAE J1832. Join SAE while you're
at it.
My address is
John De Armond
PO Box 670386
Marietta, GA 30066
Anyone interested in a sample issue of "Performance Engineering" can
email me their postal address.
John
From: John De Armond
X-Source: The Hotrod Mailing list
Subject: Re: injectors
Date: Monday, Jul 27 1992 11:46:52
X-Sequence: 1735
> I was looking through a '90 motorcycle magazine. There was a short
>article describing a Haltech injection kit for a BMW motorcycle. In the
>sidebar, they mentioned the kit used Lucas rotary disc valve injectors
>instead of ordinary pintle type Bendix injectors.
>
> I did some flipping through my (quite elderly) references, but
>couldn't find anything about a Lucas rotary disc valve injector. How
>exactly does it work, and what would make it superior to a pintle type?
Go look at a late model GM injector. Lucas and GM build to the same
specs. After they figured out that a wide spray pattern was not
necessary in most cases, they figured out they could also get rid of
the pintle valve and all its monkey motion. I've not cross-sectioned
a GM injector yet but from the pictures, it looks like a disk
that seats over an orfice. The disk is actuated by the coil.
I know from testing GM-style injectors that they open faster, make less
noise (actually a design goal from what I've heard) and appear to stay
clean longer.
john
From: Vince Reed <emory!jsrvx1.ucsf.EDU!vince>
X-Source: The Hotrod Mailing list
Date: Feb 1993
Subject: No Subject Line
I think it was on CNN last night--they ran a story about an
inventory in Nevada who had mixed 1 part water with 1 part
gas, added some detergent, and ended up with a white liquid.
He said that after an engine controller chip change and mods
to the fuel system, this fuel was a drop-in replacement for
gas in his Porsche 944.
The engine did sound smooth running on this (assuming there
wasn't a 20 gallon hidden gas tank in the trunk with the real
stuff :) ).
He said emissions fell 90%, and oil companies are approaching
him in large numbers. He also said he has no idea how it
works :)
Mike
--
mike%jim.uucp@wupost.wustl.edu
[This thread could easily degenerate into a magic elixir discussion which
is against the charter of this list so let's keep this focused on
hotrodding.
I have problems with the concept right off the bat because I've seen
bosch injectors rust literally before my eyes after getting a bit of
water in them. I have one here in which, being lazy, I dipped the tip in an
aqueous solution in the ultrasonic cleaner. The pintle rusted and stuck
literally before I could dry it. JGD]
Date: Wed Mar 24 18:13:59 1993
From: "Jon \"yes that's right\" Gross" <jong@halcyon.halcyon.com>
Subject: Help on Cleaning Fuel Injectors
To: Z-Car <z-car@dixie.com>
I recently have had three injectors go bad. Is there any way to
clean them by hand (not the fuel additive stuff) that gets them really
clean?
HISTORY --
I checked the fuel pressure at the rails, and it was fine
checked spark at the plugs it was fine. So I turned it over a few times,
and then pulled all the plugs. Three of them (3,5,6) were dry and smelled
like hot metal. The others smelled like gasoline and were faintly wet.
This leads me to beleive that it's the injectors. I welcome other diagnoses!
Thanks for your help.
Jon
jong@halcyon.com
[I'd check the valves and compression in addition to the injectors.
Cleaning injectors is easy out of the car. The best way is to pull the
tips off and ultrasonic clean 'em in a solution of injector cleaner.
While they're in the cleaner, stand each on its tip and GENTLY push down.
This depresses the pintle, opening the valve and allowing cleaning
solution to go inside.
Flip the injectors over and let the solution clean in the inlet screens.
After crud quits coming out, apply about 6 volts to the injector to open
it and blow clean, dry air backward from the pintle end through the injector.
Do NOT allow water or moisture to enter the injector. The pintle will rust
and freeze in milliseconds. Reinstall the plastic tips and you're ready
to go.
An alternative if you've built the fuel injector flow bench described
in my magazine is to mount the injectors and run 'em at about 50% duty
cycle with half and half gasoline and injector cleaner for 10 minutes or
so. A few minutes of ultrasonic pre cleaning is advantageous to keep
the solvent from getting so dirty. The flow bench will also let you
test your work, which is the proof of the puddin' :-) JGD]
Date: Fri Apr 9 19:28:41 1993
Subject: Re: Ford Fuel Injection Specs
X-Source: The Hotrod Mailing list
>> A friend of mine said that he saw a listing of Ford Fuel Injector
>> Specifications posted here. I would really appreciate it if somebody who
>> has this information could post it again. Thank you very much.
I happened to save these when they were posted, hope it's what you're looking
for... (Both were originally posted by Dan Malek )
=====================================
Since John asked, here is my current list of stock fuel injectors from Ford.
I only work with Ford stuff because I used to work there, and I can't keep up
with these details for every manufacturer (I wish I could). Explanation and
further information follows the table.
Flow PSI Resist Mfg. Part Color Application
1.76 (14) ?? 2.25 Bosch E3EE-BA Blue 1983 1.6L
1.76 (14) ?? 2.35 ND E4EE-AA Blue 1984 1.6L
1.76 (14) 30 16.20 ND E59E-AB Gray 1985-6 2.3L Truck, 5.0L SEFI
1.76 (14) 33 14.50 Bosch E67C-AB Gray 1986 2.9L, 3.0L
1.76 (14) 33 14.50 Bosch E67E-BB Gray 1986 5.0L SEFI
2.45 (19) 40 2.25 Bosch E6EE-AB White 1985-6 1.9L
2.45 (19) 32 16.20 ND E6TE-AB Gold 1986 5.0L HO SEFI
2.45 (19) 33 16.20 DKK E5TE-AB Gold 1985-6 5.0L Truck
2.45 (19) 33 14.50 Bosch E5TE-BB Gold 1986 5.0L Truck
3.00 (23) ?? 2.40 Bosch E4EX-AA Black 1984-5 1.6L Turbo
3.86 (30) 35-45 2.40 Bosch E3ZE-BA Green 1983 2.3L Turbo
3.86 (30) 35-45 2.35 DKK E4ZE-AA Green 1984 2.3L Turbo
4.41 (35) 35-45 2.35 DKK E5ZE-AB Gold 1985-6 2.3L Turbo
4.54 (37) 40 2.25 Bosch E3VE-A1A Green 1984-6 3.8L CFI
4.54 (37) 40 2.0 ND E3VE-A2A Green 1984-6 3.8L CFI
5.81 (46) 33 2.40 Bosch EOSE-A1A Blue 1980-3 5.0L CFI
5.81 (46) 33 2.00 ND EOSE-A2A Blue 1983-4 5.0L CFI
5.81 (46) 33 2.25 Bosch EOSE-A1A Blue 1984 5.0L CFI
6.60 (52) 32 2.25 Bosch E4ZE-CA Gray 1984-5 5.0L HO CFI
7.00 (56) 16 1.40 Bosch E43E-AC Blue 1985-6 2.3L HSC
8.00 (64) 16 1.40 Bosch E53E-AB Green 1985-6 2.3L HO HSC, 2.5L HSC
The flow rate is gm/sec and lb/hr in parenthesis. The resistance is obvious.
The manufacturer is next, I think ND is Nippondeso (sp?), I can't place DKK.
The part number is constructed as follows. All injectors carry a basic part
number 9F593, which is prefixed with a production release, and followed by
a revision/application code. For example E3EE-BA from the table will be
E3EE-9F593-BA to your local dealer. The tops of the injectors are color
coded, as a quick check to the technician to make sure an engine contains the
"same" injectors. I included some sample applications which may help the
dealer if you are buying new, or yourself if scrounging in a junk yard.
I don't know at what pressure the flow rate was measured. The value in the
table is listed as the operating pressure for the engine. The turbo engines
use a regulator that increases pressure as the boost increases, hence the
multiple values in the table (I assmume the flow is at the lower pressure).
I don't know what was used for the density of the fuel. I typically use
0.79 (rounded up to 0.8 for "quickie" calculations), and I have seen John
use 0.775 for "racing" fuel. From this you could probably come close to
a ml/min value.
The last two big injectors on the list are classified as "low pressure"
in the manual. Since these are on engines I have never seen, nor have a
manual for, I don't know what they look like. These engines are used in
the Tempo, and I am guessing that they may use a single central injector.
I am expecting a 1993 electronics manual for my new Explorer I recently
purchased (I really need more garage now :-), which will contain a new
table. When it arrives (week or so), I will update and re-post this table.
=====================================
also:
=====================================
I promised an injector update when I got my new shop manuals. Well, I got
them, and here is the info:
Flow Part Color Application
14 F1ZE-B4C Gray 1.9 - 3.8L
19 FOTE-D1A Yellow/Orange 4.6L SOHC
24 F2LE-B2A Lt. Blue 4.6L DOHC
25.3 F3DE-A2B Blue/Green 3.0L Flex Fuel
30 F1SE-E1A Red 3.8L SC
Flow is lb./hr. for gasoline (I don't know how the flex fuel was measured).
All have impedance of 11-18 Ohms. The basic part number is 9F593, so a
complete part number would be F1ZE-9F593-B4C.
These are the newest injectors. The 5.0L/5.8L/7.0L/7.5L V-8 use "older"
part number 19 and 24 lb./hr. injectors.
Now here is the interesting part. These injectors are described as
"Deposit Resistant, Top Feed Stage III". The further description and
cut away diagrams show an injector with multiple plungers and the following
words (from memory): Fuel metering controlled by adjustable top plunger,
lower pintle and orifice designed to control injection timing, spray pattern,
and minimze deposits.
I just love words like "adjustable". Wouldn't it be cool if these were
all the same injector, adjusted differently. I think I need to purchase
a few and take them apart, just to see how adjustable they really are. Or
is this the way injectors are designed and I just was not aware of this
information?
From: emory!gnu.ai.mit.edu!lusky
X-Source: The Hotrod Mailing list
Date: May 1993
Subject: O2 sensors, M85, stoichiometry
X-Sequence: 5251
Anyone here ever used a zirconia O2 sensor with M85?
Should stoich still be approx .5v (how bad is the hydrogen
shift with methanol)? Right now my engine is running like
crap, and won't run at all unless O2 sensor voltage is > .8v.
Could I have somehow damaged the sensor in such a way that it would
be reading high? I know my voltmeter isn't the problem, since
I get the same numbers from an oscilliscope. Anyone know the
stoiciometric A/F ratio for M85? Hydrogen/carbon ratio &
oxygen/carbo ratio?
Also, anyone have a suggestion as to how low I can go with the pulse
width @ 40 psi to Rochester 31lb/hr injectors before I start having
problems?
[I have tested a Rochester ball-and-seat type injector on the flowbench.
The dynamic flow starts deviating pretty badly at around 2 ms. The
minimum usable (defined as the narrowest pulse that will cause a
recognizable dollip of fuel to emit while viewed with a strobe light)
is about 1.2 ms. If you'd like to send me those injectors, I'd be
happy to run 'em through the bench and see what they'll do.
24 hour turnaround is possible. JGD]
--
--=< Jonathan Lusky ----- lusky@ccwf.cc.utexas.edu >=--
\ "Turbos are nice, but I'd rather be blown!" /
\ 89 Jeep Wrangler - 258/for sale! /
\ 79 Rx-7 - 12A/Holley 4bbl /
\________67 Camaro RS - 350/4spd________/
From: emory!gnu.ai.mit.edu!lusky
X-Source: The Hotrod Mailing list
Date: May 1993
Subject: Re: O2 sensors, M85, stoichiometry
X-Sequence: 5255
> [I have tested a Rochester ball-and-seat type injector on the flowbench.
> The dynamic flow starts deviating pretty badly at around 2 ms. The
> minimum usable (defined as the narrowest pulse that will cause a
> recognizable dollip of fuel to emit while viewed with a strobe light)
> is about 1.2 ms. If you'd like to send me those injectors, I'd be
> happy to run 'em through the bench and see what they'll do.
> 24 hour turnaround is possible. JGD]
Hmmm... that might explain why we haven't been able to get this thing
to run under 4,000 rpm, although our seem to spray a nice little shot
at 1.6ms, sometimes(had cranking injection time set at 1.6, and spun the
engine with the starter).
[What is the duty cycle at WOT? If it is not the lesser of >90% or 1 ms
closed time, the injectors are too big. It will therefore be impossible
to get the motor to idle down because the idle mixture would require such
a small open time.
It IS possible to control the injector in the sub-2 ms range but it is
tedious. Your ECU must have the capability of non-linear mapping so that
you can move more map points down low and it must have supply voltage
compensation. The dynamic flow deviates negatively, that is, the delivered
flow is less than indicated, at short times and is repeatable so it works in
your favor if your ECU can take advantage of it. What you have to do is
map every .25 ms or so between the minimum interval and about 2 ms.
The minimum interval (the lowest time where any repeatable discharge is
observed) must be determined experimentally. It is real easy to do
with a synchronized strobe light. The FITFOTO software available from
my server is designed to do just that. It is designed to fire an injector
and then after a keyboard-settable delay, fire a strobe light. It is
specifically designed to fire a General Radio Strobotach.
BTW, if your ECU gives you the option, peak-and-hold drive is a big, big
win in situations like this. Not only does it knock the injector open
faster and close it faster, it is also much less sensitive to supply voltage.
JGD]
--
--=< Jonathan Lusky ----- lusky@ccwf.cc.utexas.edu >=--
\ "Blowers are great but it'd be damn nice if I could get one
to last more than a day. (grenaded the Camden vane blower
for the Nth time)
From: emory!gnu.ai.mit.edu!lusky
X-Source: The Hotrod Mailing list
Date: May 1993
Subject: Re: O2 sensors, M85, stoichiometry
X-Sequence: 5269
> [What is the duty cycle at WOT? If it is not the lesser of >90% or 1 ms
> closed time, the injectors are too big. It will therefore be impossible
> to get the motor to idle down because the idle mixture would require such
> a small open time.
Hmmm, I'm not sure... I'll have to talk to Brian about it in the
morning, and look at some of the maps.
> It IS possible to control the injector in the sub-2 ms range but it is
> tedious. Your ECU must have the capability of non-linear mapping so that
> you can move more map points down low and it must have supply voltage
> compensation. The dynamic flow deviates negatively, that is, the delivered
> flow is less than indicated, at short times and is repeatable so it works in
> your favor if your ECU can take advantage of it. What you have to do is
> map every .25 ms or so between the minimum interval and about 2 ms.
I think the breakpoints in my map have to be evenly spaced, but the ECU
does do interpolation between map points. Also have supply voltage
compensation, but I dunno if the curve is correct.
> BTW, if your ECU gives you the option, peak-and-hold drive is a big, big
> win in situations like this. Not only does it knock the injector open
> faster and close it faster, it is also much less sensitive to supply voltage.
Yeah, weve got peak and hold drivers, and the injectors are 1.8ohms.
SOmething strange happened when we tried to hook up the scope, though...
When I hook up either lead from the scope to the ground wire of the injector,
I get a small arc from the clip to the wire and the injector hangs open.
This is with and without the other lead from the scope connected, and w or w/o
the car not connected to the lap, battery charger, or anything like that.
Strange, eh?
[The way almost all these things work is there is battery voltage on one
side of the injector and the other side is switched to ground through
a transistor to fire the injector. If your scope does not have isolated
inputs (most don't) and you try to connect the scope directly across
the injector, the ground side of the scope will either short the battery or
fire the injector depending on which side you're on, assuming there is
a ground path back, which there usually is. Easily solved. Simply
connect the ground lead to ground and probe each injector pin with the probe
until you find the pin that is switched. 50% probability :-). You'll
see that lead go to ground when the injector is fired. The waveform will
look quite similar to the scope shots in my FI flowbench article.
This is an area where portable scopes like my Scopemeter or the Tek
portable really shine. Being battery powered and having isolated inputs,
I don't have to think about ground loops. JGD]
The ECU is the "low-end" NOS/EFI Technologies speed density unit.
--
--=< Jonathan Lusky ----- lusky@ccwf.cc.utexas.edu >=--
\ "Turbos are nice, but I'd rather be blown!" /
\ 89 Jeep Wrangler - 258/for sale! /
\ 79 Rx-7 - 12A/Holley 4bbl /
\________67 Camaro RS - 350/4spd________/
From: emory!gnu.ai.mit.edu!lusky
X-Source: The Hotrod Mailing list
Date: May 1993
Subject: Re: Higher RPM unstability
X-Sequence: 5283
>
> Here is a quick question refering to my Mustang's 302 (Now 306).
> At about 4800-5000 RPM, I suddenly experience a rapid loss of power, surging,
> and a strange noise (kinda airy). The equipment on this engine consists of
...
> occurs (but made getting there much less responsive). I have tried a
> different carburator too (Holley) with no change. Can anybody help me
> become less perplexed? It has always done this.
Off the top of my head, it sounds like you are sucking the float bowls dry
due to inadequate fuel flow. Tee in a fuel pressure gauge at the carb
to verify you are maintaning fuel pressure up top. I have a little 1" Holley
fuel prssure gauge with a couple of feet of hose that I just clip under
a windsheild wiper for this type of testing.
<cut here>
While I'm on the subject of fuel pressure, is referencing the fuel pressure
regulator to MAP necessary for custom/racing EFI applications? It seems
to me that it would be very important, especially on a blown engine,
in order to keep fuel flow a function of pulse width instead of pulse
width and MAP. Our F-SAE guys had been running with the regulator referrenced
to atmosphere, and I think that would at least partially explain the trouble
we've been having getting the engine to run at low RPM and/or high MAP.
Will hopefully have the engine put back together tomorrow to test this
theory. Am I correct in my reasoning, or does it really not make that big
of a difference (ie.. something that could easily be corrected for in the fuel
map)?
[Referencing atmospheric pressure is hideously wrong. If your fuel pressure
is referenced to atmospheric while the injector discharge is seeing
manifold pressure, the differential pressure across the injector and thus
the flow is REDUCED with increasing boost - just the opposite of what you
want. If they've sized the injectors large enough to run properly at
WOT under boost, they're terribly, grossly oversized for idling.
This is worse than you might at first think because flow across an orfice
varies as the square of the differential pressure. If you know the
flow F1 at a pressure P1, you can compute the flow F2 at another pressure P2
with the following formula:
F2 == F1 * sqrt(P2/P1)
This holds over a reasonable range of pressures. A large change in pressure
might edge the velocity in the orfice close to the critical flow, at which
point all computational bets are off. But you get the idea.
A not-so-obvious related problem is the fact that most injectors' opening
and closing delays change with fuel pressure. Sometimes significantly.
This stands to reason since the fuel pressure is almost always working
with the spring to hold the valve shut.
Reference your fuel pressure to whatever the injector discharge port
sees so the pressure across the metering orfice is constant. JGD]
--
--=< Jonathan Lusky ----- lusky@ccwf.cc.utexas.edu >=--
\ "Turbos are nice, but I'd rather be blown!" /
\ 89 Jeep Wrangler - 258/for sale! /
\ 79 Rx-7 - 12A/Holley 4bbl /
\________67 Camaro RS - 350/4spd________/
Date: Thu Jun 17 14:36:27 1993
From: jsnyder@td2cad.intel.com (Jan Snyder)
Subject: FI replacement question
Help. Ouch. Money...
I'm caught at work w/out my Z manual and need a second opinion...
Is it true that one leaky fuel injector requires ALL 6 FIs to
be replaced? They're a unit, not separate?
[Absolutely false. While *I* would check the replacement injector on my
flowbench before installing it, you risk little by just plugging one in.
Injectors are NOT matched from the factory so anyone who tells you
otherwise is blowing smoke.
Tell you what. I have some extra stock early model Z injectors.
I've been wanting to test an actual leaker (they're REAL hard to find
in the real world). You send me your injector along with return postage
and I'll send you one of mine, complete with test data from my flow bench.
What a deal, eh?
JGD ]
I'm told it's $580 for new injectors and seals (including labor,
I think he said...).
This on top of the $1,400 worth of P&L already determined, so I
may have to delay a few lesser items.
We're deciding on this rather soon, so email replies would be
much appreciated!
-Sallijan
Date: Thu Jun 17 14:59:14 1993
From: jsnyder@td2cad.intel.com (Jan Snyder)
Subject: Re: FI replacement question
I have more information now, having talked to my mechanic again.
It's not whether they CAN be replaced separately, but whether
they should be. At 135K miles, with one spectacular failure and
the possibility that another one could go in the next few months,
it's reasonable to do the whole set now. Saves labor cost in
the long run. Greg Scott (of Scott Performance in Santa Clara)
has had too many instances of serial FI failures so he makes a
policy of doing all or none. I could wait, but the problem was
costing fuel and causing the car to act like it's flooded when
started warm (in fact, #5 cylinder WAS flooding).
Well I just learned more than I knew about EFI on a 78 280Z...
plus a lecture on the different FI system the Germans use.
Love this guy! He does good work, too.
[But he's STILL blowing smoke regarding injector failure. In the last
6 months, through my free injector cleaning/characterization offer, I've
tested literally hundreds of injectors. I've yet to find a single worn
out injector! I'd be VERY interested in testing your injectors.
I'd bet a hamburger that there is nothing wrong with 'em. JGD]
Any discussion here will still be welcome information.
-Sallijan
and ailing 'Baby'
(If anyone is interested in the laundry list of repairs and
costs, I'll be happy to share it.)
Date: Fri Jun 18 12:45:18 1993
From: jay%vitec.UUCP@mathcs.emory.edu (Jay Thompson)
Subject: Re: FI replacement question
----- Begin Included Message -----
Tell you what. I have some extra stock early model Z injectors.
I've been wanting to test an actual leaker (they're REAL hard to find
in the real world). You send me your injector along with return postage
and I'll send you one of mine, complete with test data from my flow bench.
What a deal, eh?
JGD ]
----- End Included Message -----
I may have missed a previous post about your flowbench; is this
a homebrew item? Care to elaborate? I've dinked around with fuel
injection on an Audi, and built a homebrew testing system with glass
medicine bottles and electrical tape--Enought to see the injectors
working and get a basic average on the flow rate and cone spread.
Nothing fancy.....
--JT
[This fully computerized flowbench is the feature construction article
in the free sample edition of my magazine, Performance Engineering Magazine.
How to build it is in the sample. How to use it is in the first full
sized issue being printed right now. The software, which runs on a PC, is
available from my file server, listserv@dixie.com. To get it send mail
to this address and include the line "send fit.exe" in the body.
This flow bench will measure static flow - what you measured with your
lashup. It also measures dynamic flow over the whole injector envelope.
It measures opening and closing delay. And if you have some sort of
strobe light (a timing light will do), you can freeze the discharge pattern
at any time, settable to 0.1 ms, after opening.
Anyone can obtain the free sample edition by sending your postal address
to perform@dixie.com. Do NOT mail to me or to this list.
JGD]
Date: Mon Jun 21 13:40:52 1993
From: mikef@rosevax.rosemount.com (Michael Foerster)
Subject: Injector Question
To: z-car@uunet.UU.NET@dixie.com
John, can you tell me what the pattern of flow that should be comming
out of the fuel injectors for the (78) zcar?
I picked up a fuel pump from a junk yard last fall, and jury riged a
tester (I didn't go as far as building the test jig that you designed)
and when thru cleaning and checking flow test on the injectors, when
I had my engine pulled this winter.
I found that the flow on all (but one bad one) was a straight stream,
not a cone pattern (as one other guy mentioned).
I suspect that I was not getting 30 lbs of pressure (the pump is quite
old and may be bad).
Is a straight stream the right pattern for these injectors?
Mikef
After having tested several sets of Datsun injectors and many more non-Datsun
pintle-type injectors, I'm not convinced there is a correct pattern. At
low pressure the discharge is generally a stream. At operational pressure
(my test pressure is 38 psi), the discharge gets ragged and may break up
into several smaller streams. I've decided that as long as none of
the stringers diverge too widely from the main discharge, the injector is OK.
BTW, a typical Z injector discharge pattern is featured in the next issue
of my mag currently being printed. The pattern is photographed at 1 ms
intervals after opening by using a high speed strobe light. JGD]
Date: Sun Jul 11 22:55:33 1993
From: John De Armond
Subject: Fuel injector followup
To: z-car@emory.mathcs.emory.edu (z car list)
A couple of weeks ago Jan Snyder told us about his 78 Z that had running
problems. His mechanic told him it had a leaking injector and that it
needed a whole new set of injectors. I suggested it probably was NOT
related to all the injectors. Jan had to get the car fixed quickly so
he let the guy replace all the injectors. He then sent the old ones to
me to evaluate. Following are the results of my evaluation. This data
was collected from my fuel injector flowbench.
The visual inspection indicates the injectors have been very hot. Jan,
you might want to check to see if your cooling fan is working. (if you
have one.) The rubber bushings were welded to the body. The spray forming
tips were all cracked and brown, though the discharge pattern is unaffected.
All the inlet screens are clean.
The first thing that surprised me is that one injector actually leaked.
This is the first injector I've ever gotten my hands on that actually leaked.
It drips about 2 drops per second off the end of the injector. Ultrasonic
cleaning and backflushing did not remedy the leaking. The inlet screen
is clean. I plan to dissect this injector to determine why it is leaking.
The rest of the injectors look pretty normal. There is a deviation in
mass flow below 4 ms but that is typical of these old-design pintle
injectors. Production injectors are typcially matched within 5%.
These certainly are.
The one thing not tested in this evaluation is performance at operating
temperature. I plan to construct a heated injector mount for the
flowbench in order to look at hot performance. I really don't expect
to find much but I've been surprised before.
The test data below are generated from the program fit.exe, the software
that drives the flowbench from a PC. The software executes a simple
script language. The test script tests the injector for static flow and
for dynamic flow at 2 ms intervals from 2 to 18 ms. This simulates
typical operating conditions. The script was configured to drive the
injectors with a peak/hold scheme. The injectors on the car are driven
by a ballasted saturated current driver but these injectors perform
similarly with either type.
My conclusion is 5 of the injectors are in good shape, consistent with
what I see from other good running cars with 100k miles. The leaker
has sufficient leakage (as noted by the high flow at 2 ms for #1) to cause
very poor idle and cruise.
John
-------------
78 Datsun 280Z 2+2 #1 - This is the leaker
----
peak On Cycle Total Tot.On Total Density Volume Mass
Time Time Time Cycles Time Flow Rate Rate
ms ms ms sec. cc gr/cc cc/sec lbs/hr
---------------------------------------------------------------------
Static Flow 146.754 400 0.770 2.73 28.09
2 2 20 8146 16.633 51 0.770 3.07 31.60
2 4 20 4084 16.507 44 0.770 2.67 27.47
2 6 20 4017 24.270 64 0.770 2.64 27.18
2 8 20 2007 16.140 44 0.770 2.73 28.10
2 10 20 2032 20.405 56 0.770 2.74 28.29
2 12 20 2028 24.421 67 0.770 2.74 28.28
2 14 20 2001 28.098 78 0.770 2.78 28.61
2 16 20 2003 32.068 89 0.770 2.78 28.61
2 18 20 2017 36.066 100 0.770 2.77 28.58
78 Datsun 280Z 2+2 #2
----
peak On Cycle Total Tot.On Total Density Volume Mass
Time Time Time Cycles Time Flow Rate Rate
ms ms ms sec. cc gr/cc cc/sec lbs/hr
---------------------------------------------------------------------
Static Flow 140.626 400 0.770 2.84 29.32
2 2 20 8155 16.652 45 0.770 2.70 27.86
2 4 20 4032 16.297 41 0.770 2.52 25.93
2 6 20 4005 24.198 63 0.770 2.60 26.84
2 8 20 2038 16.389 44 0.770 2.68 27.67
2 10 20 2008 20.164 55 0.770 2.73 28.11
2 12 20 2028 24.421 67 0.770 2.74 28.28
2 14 20 2023 28.407 79 0.770 2.78 28.67
2 16 20 2013 32.276 90 0.770 2.79 28.74
2 18 20 2014 35.831 101 0.770 2.82 29.05
78 Datsun 280Z 2+2 #3
----
peak On Cycle Total Tot.On Total Density Volume Mass
Time Time Time Cycles Time Flow Rate Rate
ms ms ms sec. cc gr/cc cc/sec lbs/hr
---------------------------------------------------------------------
Static Flow 137.493 400 0.770 2.91 29.99
2 2 20 8010 16.356 44 0.770 2.69 27.73
2 4 20 4090 16.531 42 0.770 2.54 26.19
2 6 20 4004 24.192 64 0.770 2.65 27.27
2 8 20 2042 16.422 45 0.770 2.74 28.25
2 10 20 2036 20.445 57 0.770 2.79 28.74
2 12 20 2009 24.192 67 0.770 2.77 28.55
2 14 20 2022 28.393 81 0.770 2.85 29.41
2 16 20 2021 32.324 93 0.770 2.88 29.66
2 18 20 2005 35.885 103 0.770 2.87 29.58
78 Datsun 280Z 2+2 $4
----
peak On Cycle Total Tot.On Total Density Volume Mass
Time Time Time Cycles Time Flow Rate Rate
ms ms ms sec. cc gr/cc cc/sec lbs/hr
---------------------------------------------------------------------
Static Flow 140.279 400 0.770 2.85 29.39
2 2 20 8026 16.388 45 0.770 2.75 28.30
2 4 20 4092 16.540 41 0.770 2.48 25.55
2 6 20 4035 24.379 64 0.770 2.63 27.06
2 8 20 2019 16.237 43 0.770 2.65 27.30
2 10 20 2027 20.355 55 0.770 2.70 27.85
2 12 20 2013 24.240 67 0.770 2.76 28.49
2 14 20 2005 28.154 79 0.770 2.81 28.92
2 16 20 2020 32.373 91 0.770 2.81 28.97
2 18 20 2018 35.903 102 0.770 2.84 29.28
78 Datsun 280Z 2+2 #5
----
peak On Cycle Total Tot.On Total Density Volume Mass
Time Time Time Cycles Time Flow Rate Rate
ms ms ms sec. cc gr/cc cc/sec lbs/hr
---------------------------------------------------------------------
Static Flow 139.312 400 0.770 2.87 29.59
2 2 20 8174 16.691 45 0.770 2.70 28.53
2 4 20 4100 16.572 42 0.770 2.53 26.12
2 6 20 4038 24.397 64 0.770 2.62 27.04
2 8 20 2016 16.212 44 0.770 2.71 27.97
2 10 20 2019 20.275 56 0.770 2.76 28.47
2 12 20 2031 24.457 68 0.770 2.78 28.66
2 14 20 2020 28.351 80 0.770 2.82 29.09
2 16 20 2004 32.132 91 0.770 2.83 29.19
2 18 20 2007 36.030 101 0.770 2.80 28.89
78 Datsun 280Z 2+2 #6
----
peak On Cycle Total Tot.On Total Density Volume Mass
Time Time Time Cycles Time Flow Rate Rate
ms ms ms sec. cc gr/cc cc/sec lbs/hr
---------------------------------------------------------------------
Static Flow 140.219 400 0.770 2.85 29.40
2 2 20 8001 16.337 42 0.770 2.57 26.50
2 4 20 4080 16.491 41 0.770 2.49 25.63
2 6 20 4025 24.319 63 0.770 2.59 26.70
2 8 20 2039 16.397 44 0.770 2.68 27.66
2 10 20 2002 20.104 55 0.770 2.74 28.20
2 12 20 2023 24.361 67 0.770 2.75 28.35
2 14 20 2014 28.280 79 0.770 2.79 28.79
2 16 20 2003 32.036 90 0.770 2.81 28.96
2 18 20 2018 35.633 102 0.770 2.86 29.51
Date: Mon Jul 12 17:39:09 1993
From: jsnyder@td2cad.intel.com (Jan Snyder)
Subject: Re: Fuel injector followup
John,
Just got this and will share with Greg Scott at Scott Performance.
Only quibble I have with your report is that you used the wrong gender
pronoun. I'm a "her" not a "his." :-)
Sorry 'bout that :-)
>A couple of weeks ago Jan Snyder told us about [his] 78 Z that had running
>problems. [His] mechanic told [him] it had a leaking injector and that it
>needed a whole new set of injectors. I suggested it probably was NOT
>related to all the injectors. Jan had to get the car fixed quickly so
>[he] let the guy replace all the injectors. [He] then sent the old ones to
>me to evaluate. Following are the results of my evaluation. This data
>was collected from my fuel injector flowbench.
The main cooling fan (radiator fan) seems fine. Car doesn't run hot
normally, though it heats up a bit on hills and in slow traffic on hot days.
Possibly the oil loss problem I was also having contributed to overheated
injectors? Is there a fan other than the main radiator fan? Never noticed
another fan under the hood.
[Some Zs have a little electric fan on the right side with a duct that goes
over the top of the engine with outlets pointing at the injectors. This
is designed to cool the injectors. Not sure exactly when they started
putting this fan on Zs. I know my brother's 82 had it.
It is a jerry-rig at best. JGD]
Gotta run to a meeting...Thanks for the data! BTW, the bad injector came
off the #5 cylinder.
-Sallijan Marie Snyder
a pretty decent woman driver but not a mechanic. :-)
-
Sallijan Snyder | "My employer shares my opinions. NOT."
Santa Clara, California | (408)765-9068 work (408)629-5909 home
Internet: jsnyder@td2cad.intel.com | =>Internet Z-Car Club Member #42<=
From: Dan Malek <emory!gatech!westford.ccur.com!dan>
X-Source: The Hotrod Mailing list
Date: Jul 1993
Subject: Fuel injectors and MSD
X-Sequence: 5761
One complaint and some questions.
Complaint: The price of fuel injectors has gone out of sight over the
past six months! I ask why and the answer is always "New pintle
design that reduces deposits." Damn, so much for dragging my feet
when building projects.
So, where are people buying injectors these days? I mean big ones, like
50 lb./hr. or so. There are lots of used ones in the under 30 lb./hr.
category, maybe I should set my sights lower (no way), or use two per
cylinder (I would rather not, although my pulse times are real short
at idle and I may have to).
[At the local junkyards. My program of offering free injector
testing/cleaning is paying off. I've tested OEM injectors all the way up to
200 lbs/hr (!) As soon as I get the magazine in the mail (I pick it up
tomorrow I think. Yea!) I'll tabulate some part numbers, car make and models
and flow capacities and post them.
The injector situation is getting out of control. The industry seems to
be in a feeding frenzy. Talking about killing the golden goose. BTW,
you have to really watch these guys. They're quoting mass flows
at widely varying pressures. Jacking the pressure up can magically
convert a 40 lb/hr unit to a 50 lb/hr one. Problem is, on all injectors
I've ever tested the opening and closing delay lengthens with increased
fuel pressure. Jacking fuel pressure is counter-productive too because
the flow is proportional to the square root of the pressure. The flow
rapidly flattens. Example: For a 40 lb/hr injector at 40 psi, the
following flows apply as computed on a little spreadsheet I whipped
together:
10 psi 20.02 lb/hr
20 28.32
30 34.68
40 40.00
50 44.77
60 49.05
70 52.98
80 56.63
90 60.07
100 63.32
JGD]
Another totally unrelated question about MSD ignition. I want to use the
stock tach with the MSD on a late model Mustang. The tach uses the
typical battery voltage sensed at the coil to determine RPM. How to
I build something (rather than pay MSD $$$ for their module), to drop the
MSD voltage output, and manage the multiple sparks, so I can attach the
tach again? Yeah, I know, buy a real tach....but I want to keep the stock
look (and its inaccuracy :-).
[Can't you just run the "tach" output from the MSD box to this terminal?
JGD]
Thanks.
-- Dan
Date: Mon Oct 4 16:42:59 1993
From: bobw@PROCASE.COM (Bob Weissman)
Subject: Fuel injector cleaning
What do you folks think about having fuel injectors professionally cleaned?
This is where they disconnect your fuel system and pump some solvent or
other through the injectors.
My first impression was that it's just another way for the mechanics to
extract dollars from my wallet, but a local fellow Z owner swears by doing it
every 30k miles.
Thanks,
--
-- Bob Weissman, Procase Corp.
-- Internet: bobw@procase.com
-- UUCP: ...!{uunet, voder, veritas}!procase!bobw
--
[I've offered a program for quite some time whereby I'll clean and calibrate
injectors at no cost except for postage to and from my shop. The purpose
of my program is to gather technical specs on various OEM injectors AND
to get a feel for the condition of injectors out there. I've tested
well over 100 sets of injectors. So far I've found ONE set of injectors
that needed cleaning and those looked like they'd been dug out of a
garbage dump. I define "needs cleaning" as the flowrate varying among
the set of injectors by more than 5% and/or any injector improving its
flow by 5% or more after cleaning.
I use a couple of methods of cleaning injectors including using a 50-50 mix
of paint thinner and Techron in my flow bench and ultrasonic cleaning with
trichloroethane. Invariably firing the injector through 10,000 cycles
with the paint thinner/Techron mix cleans the injector such that the
subsequent ultrasonic cleaning removes only cosmetic dirt and has
no effect on flow. Note that these were injectors that showed <5%
improvement and thus didn't meet my criteria for cleaning but did respond
to cleaning. I should also note that a typical OEM spec for injector
flow is +-5%; thus two injectors could vary by as much as 10% and
still be in spec.
In case you perceive that I consider injector cleaning services to be a
fraud, you would be correct. The MOST you should ever need to do is dump
in an occasional can of Techron on a quarter tank of gas and then run it
almost empty. JGD]
Date: Mon Oct 18 15:35:36 1993
From: rsiatl!emory!gain.com!Carlos.Chang (Carlos Chang)
Subject: Re: Fuel injector cleaning
> Techron sounds like a brand name, what actually is it,or what else is
> it called?
>
> [It is a fuel injector cleaner made by Chevron. I recommend this one
> because it works, at least on my flowbench. JGD]
>
Doesn't Chevron gas come preloaded w/ Techron? What kind of gains are you
seeing from the flowbench? At about $5 a pop, how often should it be
used?
[yes Chevron gas has Techron. But because any of these cleaners has
an adverse effect on oil (detergents tend to do that), the quantity is
limited. Techron merely allows you to use more when you (think you) need
it.
Since I've ever only had one set of injectors that needed cleaning from
a performance perspective, my experience is limited. Techron, one can
to one gallon of mineral spirits, very rapidly cleaned each injector
to the point the flow was back to OEM specs within a couple of minutes.
With all the other injectors I've tested, I can only say that the Techron
does a real good job of cosmetically cleaning.
I'd only pop a can in when there is some indication of need. Since
dirt in or on an injector invariably results in either leaking or
reduced flow, the indications are either flooding after sitting or
leanness. JGD]
Date: Tue Oct 19 08:40:45 1993
From: glenna%zx.UUCP@mathcs.emory.edu (Glenn Arbitaylo)
To: z-car@dixie.com, z-car@uunet.UU.NET@dixie.com
Subject: Re: Fuel injector cleaning
> Since I've ever only had one set of injectors that needed cleaning from
> a performance perspective, my experience is limited. Techron, one can
> to one gallon of mineral spirits, very rapidly cleaned each injector
> to the point the flow was back to OEM specs within a couple of minutes.
> With all the other injectors I've tested, I can only say that the Techron
> does a real good job of cosmetically cleaning.
>
When you say "Techron, one can to one gallon of mineral spirits," do you
mean that you take the injectors OUT of the engine and you clean them in the
"Techron, one can to one gallon of mineral spirits" solution?
Also can you tell me how you clean them (ie. do you just let the injectors
"take a bath" in that solution OR do you use a brush to scrub the injectors).
-Glenn Arbitaylo
[In the first couple of issues of my magazine, I describe how to build a
PC-driven fuel injector flowbench that can measure all important
fuel injector parameters. I normally use mineral spirits in the bench
because it is much less flammable than gasoline while being close enough
in viscosity that an equivalence factor can be developed. Techron is
mostly petroleum distillate of the same general character so it blends
right in. To clean an injector, I simply mount it on the flow bench
and run it at 50% duty cycle for 10,000 cycles or so.
If the injector is gunked up - not uncommon but apparently not of
consequence to the performance - I clean it using an ultrasonic cleaner
charged with 1,1,1 trichloroethane. This is the SAE recommended method.
JGD]
Date: Thu Oct 28 18:58:44 1993
From: Jim Nagle <jwnagle@loglady.ninds.nih.gov>
Subject: Re: Fuel injector cleaning
What's DG...oops, I think I meant BG....anyway it's a rather expensive
($16/can) gas treatment which is for cleaning injectors. Our local
car repair shop swears by the stuff and I was just curious if it was
the same or similar to the Chevron product you mentioned. It _seems_
to work pretty well but I might be imagining a positive result after
spending > cost of fill up to treat a tank of gas. Any one else out
there use this stuff I'd be interested in your comments.
-jim
[I'll bet they love it cuz they're probably paying $1.98 a can for it.
Injector cleaner isn't black magic. It's just detergent. Based on
what I've seen here in my injector testing program, I seriously doubt
anyone who uses brand name gasoline EVER needs injector cleaner. There ARE
dirt-related problems such as fuel filter punch-through and injector
inlet screen plugging but no amount of magic-in-a-can will fix these
problems. JGD]
Date: Mon May 2 12:41:55 1994
Subject: Re: Z-car Mailing List Digest...
[[article deleted]]
[I must correct a mis-conception. The fuel pressure is referenced to
the intake manifold so that the injectors will always be subjected
to the same differential pressure(DP). Given a constant DP, the
delivered fuel is directly proportional to the injector open time.
This makes the ECU's job very easy. It is programmed with the
injector's delivery rate so all it has to do to deliver the proper
fuel for a given condition is to compute (or look up in a table)
the proper pulse width to the injectors. So in your case, even
though your manifold vacuum is less, the pressure across the injector
is still exactly the same because it is referenced to the manifold, and
the delivered flow is the same.
The reason your engine feels "fat" at low and part throttle is the same
reason you have low vacuum - the wider lobes and more overlap allow
more reverse flow back into the intake at low speed where no ram or
resonant effects are at work. This backflow contains some fuel
already injected and some residual exhaust. THAT makes it rich.
The same effect is at work with a carburator except that if the reversion
is severe, the air passes over the metering jet 3 times (in, out, in again)
and thus gets even richer. With EFI, one can compensate for reversion
in the calibration map but obviously this is something that must
be done in the EFI programming. JGD]
Index
Home
About
Blog