PWM to analogue.
 

I've fitted a new fuel injection system to the old car - the original was
analogue and not too brilliant.

It's working well but has an interesting side effect. There is an OBC
which calculates fuel consumption etc which is now wildly out. It worked
by counting the time one injector was open. The new system uses PWM with
feedback current limiting to drive the injectors which obviously confuses
it.

Any ideas?

--
*Cover me. I'm changing lanes.

Dave Plowman London SW
To e-mail, change noise into sound.

PWM to analogue.
 

Dave Plowman (News) wrote:
I've fitted a new fuel injection system to the old car - the original was
analogue and not too brilliant.

It's working well but has an interesting side effect. There is an OBC
which calculates fuel consumption etc which is now wildly out. It worked
by counting the time one injector was open. The new system uses PWM with
feedback current limiting to drive the injectors which obviously confuses
it.

Is it reading well under on the fuel consumption? Response time of the
OBC to detect the injector openness could be slow, and hence a bit too
lazy if the the thing is PWM'ing...

Just a guess?

--
Adrian C

PWM to analogue.
 

In article ,
Adrian C wrote:
Dave Plowman (News) wrote:
I've fitted a new fuel injection system to the old car - the original
was analogue and not too brilliant.

It's working well but has an interesting side effect. There is an OBC
which calculates fuel consumption etc which is now wildly out. It
worked by counting the time one injector was open. The new system uses
PWM with feedback current limiting to drive the injectors which
obviously confuses it.

Is it reading well under on the fuel consumption? Response time of the
OBC to detect the injector openness could be slow, and hence a bit too
lazy if the the thing is PWM'ing...

Yes. With the ECU set to two squirts per cylinder cycle (same as the
original) it reads something like 1/2 the true consumption. Set to 4
squirts per cycle approx 2 times. So what I'm after is some way of
converting the pulses into a signal it can recognise. After all the amount
of petrol being injected will be the same.

Just a guess?

--
*Arkansas State Motto: Don't Ask, Don't Tell, Don't Laugh.

Dave Plowman London SW
To e-mail, change noise into sound.

PWM to analogue.
 

On Thu, 22 Oct 2009 10:17:58 +0100, Dave Plowman (News) wrote:
It's working well but has an interesting side effect. There is an OBC
which calculates fuel consumption etc which is now wildly out. It worked
by counting the time one injector was open. The new system uses PWM with
feedback current limiting to drive the injectors which obviously confuses
it.

Hmm, interesting. From your description, it sounds like you need something
to convert PWM to a proportional voltage, then a second stage to
essentially carry out voltage to frequency conversion (giving pulses of
fixed width but varying frequency according to variation in the injector
pulse width - a VCO)? Possibly with some sort of adjustment between so you
can fine-tune things...

In the early 90s I'm sure there were circuits in the Maplin catalogue to
do both the PWM-analogue and the VCO part, but of course the catalogues
got less comprehensive over the years (and the one I'm remembering is
unfortunately stuck in storage).

PWM to analogue.
 

On Thu, 22 Oct 2009 11:39:48 -0500, Jules wrote:

On Thu, 22 Oct 2009 10:17:58 +0100, Dave Plowman (News) wrote:
It's working well but has an interesting side effect. There is an OBC
which calculates fuel consumption etc which is now wildly out. It worked
by counting the time one injector was open. The new system uses PWM with
feedback current limiting to drive the injectors which obviously confuses
it.

Hmm, interesting. From your description, it sounds like you need something
to convert PWM to a proportional voltage, then a second stage to
essentially carry out voltage to frequency conversion (giving pulses of
fixed width but varying frequency according to variation in the injector
pulse width - a VCO)? Possibly with some sort of adjustment between so you
can fine-tune things...

In the early 90s I'm sure there were circuits in the Maplin catalogue to
do both the PWM-analogue and the VCO part, but of course the catalogues
got less comprehensive over the years (and the one I'm remembering is
unfortunately stuck in storage).

A simple RC filter would probably be good enough, just Google PWM to
analogue - it'll only cost you pennies to try and it'll probably be good
enough.

SteveW

PWM to analogue.
 

On Thu, 22 Oct 2009 17:16:09 +0100, Dave Plowman (News) wrote:

It worked by counting the time one injector was open.

Is it reading well under on the fuel consumption?

Yes. With the ECU set to two squirts per cylinder cycle (same as the
original) it reads something like 1/2 the true consumption. Set to 4
squirts per cycle approx 2 times. So what I'm after is some way of
converting the pulses into a signal it can recognise.

The latter description of just counting pulses doesn't match the
orginal description of measuring the duration of a pulse...

If it's just a pulse counter and assumes a fixed duration you just
need to stretch the first pulse to the correct duration (and possibly
ignore following ones for a period afterwards). Try a google on
monostable. The good old 555 chip can do the first part but will
retrigger straight away.

--
Cheers
Dave.

PWM to analogue.
 

In article o.uk,
Dave Liquorice wrote:
Yes. With the ECU set to two squirts per cylinder cycle (same as the
original) it reads something like 1/2 the true consumption. Set to 4
squirts per cycle approx 2 times. So what I'm after is some way of
converting the pulses into a signal it can recognise.

The latter description of just counting pulses doesn't match the
orginal description of measuring the duration of a pulse...

If it's just a pulse counter and assumes a fixed duration you just
need to stretch the first pulse to the correct duration (and possibly
ignore following ones for a period afterwards). Try a google on
monostable. The good old 555 chip can do the first part but will
retrigger straight away.

I *think* the original analogue system just measured the time the injector
was open for. So just a variable width DC pulse. The fuel rail pressure
is varied by the manifold depression so the fuel flow per unit of time
open would be relatively constant.

--
*When everything's coming your way, you're in the wrong lane *

Dave Plowman London SW
To e-mail, change noise into sound.