From: "Christopher R. Carlen"
Subject: How to drive a solenoid fuel injector?
Date: Fri, 20 Dec 2002 11:48:40 -0800
Organization: Sandia National Laboratories, Albuquerque, NM USA
NNTP-Posting-Date: Fri, 20 Dec 2002 18:47:34 +0000 (UTC)
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.1) Gecko/20020826
X-Accept-Language: en-us, en
Looks like I may have to make a driver for a fuel injector, the solenoid
driven needle type.
The existing system comes from Fiat, and it can only pulse the injector
once or twice per cycle. We want to be able to pulse any number of times.
I have not yet been informed of the numerical specs. needed. At this
point, I am interested in some conceptual descriptions of how to drive
solenoid actuated injectors.
My first idea is to use a current source that is able to sink and source
current, and swing the output voltage positive as well as negative.
The current source can be commanded with a logic level pulse input to
produce the desired needle lift and hold current. Since it is a current
source, it will swing immediately (limited of course by its risetime) to
it's maximum compliance voltage, and then the current will ramp up in
the solenoid coil until it hits the desired level, at which the current
source will level off it's voltage output so as to satisfy V=IR for the
coil DC resistance.
When the voltage input current control command switches to zero, the
current source will swing to it's maximum negative voltage, until the
coil current ramps down to zero.
There doesn't seem to be anything one can do other than increase the
coil voltage to speed up the current risetime, so I will have the
compliance voltage of the source as the main parameter controlling
overall lift time performance. The injector drivers that we have which
are actual "under the hood" injector electronics packages, seem to
charge a capacitor to a high voltage (about 50V), and connect it to the
injector. But they can't give multiple pulses because they need too
much time to recharge.
Our research injector driver need not be small to fit under the hood.
So if it has a bit of oversized hardware in order to implement a "brute
force" current sourse, that would be Ok.
The main desires here are relatively simple circuitry that can be tested
and put in service quickly, as opposed to requiring lengthy theoretical
investigations in order to understand how to even get started.
What other approaches might I consider other than what I'm picturing? I
am even considering making this thing linear, instead of PWM. I have
had squirly experiences with PWM drive techniques in my limited
experience, and so I'd like to do more labs on that for developing my
abilities before putting PWM into something that has to be used to do
real work. I am reasonably comfortable with making high power linear
The other thing I don't know at this point besides numbers, is if it is
necessary to have the pull-in current be one level, and then if the
injection pulse is much longer than the pull in time, should there be a
lower holding current value, and how I might go about implementing this
all, while presenting the user with the simplest interface possible,
just a TTL level inject pulse. I suppose I can figure this out myself.
I'm mainly interested in if there is a better way than a current source.
Christopher R. Carlen
Principal Laser/Optical Technologist
Sandia National Laboratories CA USA