[Diy_efi] Measuring Pulse Width/Duty Cycle of an Injector?

Tue May 24 05:51:22 UTC 2005

At 01:24 PM 24/05/05, you wrote:
>Anyway.  Getting 2 significant digits from a pulse width meter would be
>tough enough, 3 or 4 as you've indicated below is right out.  The
>injector probably spends half a millisecond somewhere between open and
>closed every time it's fired and it's flow rate is undefined during this
>period of time.  With that kind of uncertainty in fueling getting a
>bunch of decimal places on the pulse width measurement probably isn't
>that useful.

I would have thought for a median level of reliable injector operation
there would be a probabilistically normal distribution correlated with
drive. - Your classic bell shaped curve - well, more like a narrowish peak.

>It should work fine on P/H circuits, as long as the voltage to the
>injector doesn't go below the forward voltage drop of the diode, which
>is probably something like 1.2 volts?

The LEDs in a opto are similar to the normal visible leds of around 2v
forward voltage at around 20mA. Though you might want to desensitise
so stray ignition or ground currents dont fire the opto transistor as
even optoisolators are subject to capacitance and ground current
influence on the other (amplified) side etc.

>Remember that your scanner is telling you what the computer wants the
>injector to do, which isn't necessarily what the injector is doing, if
>there's a defect somewhere in the circuit or one of the injectors.

Hence the normal distribution issue and the  processor could also look at
variation in std dev - if there were a reference for overall fuel flow, that
would then be able to point to an injector unreliability. Somewhat like
the cycle by cycle injector/O2 sensor systems which can indicate
which particular cylinder is running rich or lean - ie. Dither PWM on
an injector to calibrate the flow on an injector by injector basis - probably
bit too sophisticated for mosy DIY EFI but doable nonetheless...

Mike

>--steve
>
>
>> -----Original Message-----
>> From: diy_efi-bounces at diy-efi.org 
>> [mailto:diy_efi-bounces at diy-efi.org] On Behalf Of Marcello A. Belloli
>> Sent: Monday, May 23, 2005 10:56 PM
>> To: diy_efi at diy-efi.org
>> Subject: RE: [Diy_efi] Measuring Pulse Width/Duty Cycle of an 
>> Injector?
>> 
>> Steve,
>>      With your setup how accurate can you get?  All the 
>> scanners I've used
>> in the past to look at cars gave readings in the x.xx ms range.  At
>> idle I've seen cars that started running rich, or lean based on the
>> o2 reading and seen no change in injector pulse width on the scanner.
>>  I see the computers short term / long term trim values change, and
>> yet the IPW reading never changed?  I am guessing I need another
>> order of accuracy like x.xxx ms to see these changes at idle.  Do you
>> think I could get that out of your circuit.  And how does it handle a
>> Peak Hold type injector?  This is the most important piece of data
>> I'm grabbing to try and do what I want to do in the end.  I'm going
>> to go grab the details to your project right now, and check it out.
>> 
>> Thanks,
>> 
>> Marcello
>> >
>> >> -----Original Message-----
>> >> From: diy_efi-bounces at diy-efi.org
>> >> [mailto:diy_efi-bounces at diy-efi.org] On Behalf Of Marcello 
>> A. Belloli
>> >> Sent: Saturday, May 21, 2005 10:31 PM
>> >> To: diy_efi at diy-efi.org
>> >> Subject: [Diy_efi] Measuring Pulse Width/Duty Cycle of an Injector?
>> >>
>> >> Hello Everyone,
>> >>      I'm still working on a datalogger project.  Been doing a
>> >> lot of the
>> >> software while I've been on vacation in Thailand. I'm still on
>> >> vacation, but can't keep my mind off of this project.  I'm using a
>> >> PIC18F452 as the Micro behind my datalogger.  I've got just about
>> >> everything working.  The one place I'm having problem is the
>> >> understanding of how to read pulse width of a fuel injector.
>> >>      Is there a way of looking at all types of injectors,
>> >> that will allow
>> >> for an accurate account of timing?  I've been looking at 
>> two type of
>> >
>> > For the pulse width meter I built (see the diy-efi.org 
>> projects page) I
>> > used an opto-isolator to both insulate the measuring 
>> circuit from the
>> > injector circuit, and make it able to hook up across the injector no
>> > matter which side of the injector is driven, or what the voltage
>> > waveform looks like.  This particular opto-isolator has 2 diodes in
>> > parallel, so the polarity of how it's connected doesn't 
>> matter -- if the
>> > injector fires, then one of the diodes will light up and 
>> complete the
>> > circuit.
>> >
>> >> injector firing.  Low resistance injectors, which use a driver that
>> >> duty cycles the injectors after the main firing to keep the current
>> >> within limits.  And Higher resistance ones where the 
>> resistance alone
>> >> keeps the current down.  Now how do you look at firing time of an
>> >> injector when the driver starts to duty cycle it after 1.2ms?
>> >
>> > I haven't see all driver circuits out there but I think most of them
>> > don't duty cycle the injector, they just reduce the voltage 
>> applied to
>> > the injector.
>> >
>> > --steve
>> >
>> >
>> > _______________________________________________
>> > Diy_efi mailing list
>> > Diy_efi at diy-efi.org
>> > http://lists.diy-efi.org/mailman/listinfo/diy_efi
>> >
>> 
>> _______________________________________________
>> Diy_efi mailing list
>> Diy_efi at diy-efi.org
>> http://lists.diy-efi.org/mailman/listinfo/diy_efi
>> 
>> 
>_______________________________________________
>Diy_efi mailing list
>Diy_efi at diy-efi.org
>http://lists.diy-efi.org/mailman/listinfo/diy_efi

Regards from

Mike Massen
Perth, Western Australia
VL Commodore Fuse Rail that wont warp or melt !
http://niche.iinet.net.au