AMAZING ARTICLES #25:
"reality is never what it appears to be."
Our previous article, A24, deals with a true, practical,
engineering problem, only it is still . . . too general. We need to present you
something specific, in order to prove that we are not just utopian idealists living in the realm of
Of course, we do not have to prove anything to anyone, because the
readers who understand our
theories do not need these minor and insignificant examples. However, if you look beyond
the technical aspects
presented here, you could see a few,
major, social-psychology implications!
Particular to our days is the great desire for "dream jobs": the employees want new, better-paid jobs, and
the employers want to hire the best trained employees.
Now, when an employer buys the latest machinery, say
BH345T--just a fake name--he also wants to hire someone having at least 10 years
of working experience on that particular machine. Please note this: the
mentioned machine is only a few months old! Unfortunately, such
incidents happen quite often, lately.
Anyway, while working with hardware and firmware we noticed an incredible fact:
electronic engineers "know" and
use electricity differently than electrical engineers do! We hope our words sound sufficiently absurd to stir
up your curiosity. Things are this way.
A few years ago we had to design an automotive
Injectors Controller, and we decided to do it
by-the-book; of course, that was, by the electronic hardware books. So, an automotive injector may draw
between 1A to 4A--the ones we worked with required 4A. Particular to those injectors is, they need to open very
fast, and they draw a lot of extra current while opening.
For example, a 4A injector takes about 16A to open. After 1ms (generally) injector's current falls back to 4A. Now, a good driver must supply all
the current the injector
needs in order to open it as fast as possible. The total "open injector time" is a variable ranging between 2.5 ms
to 30 ms.
That particular current variation, when the injector is opened, is handled by professional hardware designers in
1. with "PWM"
2. with the "Peak-and-Hold" circuitry
We decided on the Peak-and-Hold hardware implementation, because PWM was a dangerous source of EMI for the automotive environment.
Consequently, we searched for a factory-built IC to do the job we wanted. There are many options available because
driving automotive injectors is nothing new. The most tempting one was LM1949 IC built by National Semiconductor
were told the LM1949 IC had been used in similar applications by Ford]. The NS recommendation of circuit implementation was
something similar to the following:
driver circuit - approximation of manufacturer's recommendations
was perfectly fine, and the only problem we had was R1 (0.05 ohms, 5 W). It was a "sense resistor", both expensive
and very difficult to find. In fact, R1 was so difficult to procure that we gave up using it, and we modified the
above schematic to:
injector driver circuit
problem with R1, the sense resistor, is mentioned in
LEARN HARDWARE FIRMWARE AND SOFTWARE DESIGN. What we did was,
we used a lot higher value for R1 (1 Ω, 5 W), only a lot easier to procure and way cheaper. The new problem
was, LM1949 IC required lower voltages, therefore we had to divide the voltage developed on R1 with the help of a
programmable potentiometer MCP41010. As it is mentioned in
LHFSD, this circuit worked incredibly well.
Even more, by using a programmable potentiometer, MCP41010, the above schematic is able to handle a wide range of
primary currents. Implicitly, it also handles various injector types without modifications on the PCB.
We built the first version of the controller, and then we tested it: it worked perfectly well. However, the
interesting aspect was, we understood that electronic engineers handle electricity differently than electrical
engineers do. Here is why.
Both methods used to control Ip, the primary current, [this is, PWM and Peak-and-Hold] are able to only REDUCE the primary
current. In the injectors driver circuit case, any reduction of the primary current is not wanted/needed! In fact,
the electronic designers of the LM1949 IC try to follow injector's natural current curve with their IC. As for
control, the only control they could implement with LM1949 is to reduce the primary current, and that is, again,
To any electrical engineer things are very clear: the injector will draw its peak value
current, then the current drops down to
normal values BY ITSELF, and no electronic circuitry is needed! Please understand this:
we need no PWM and no
Peak-and-Hold circuits to drive the injector. To help you understand this, think of the electrical circuit used to
wire the bulbs in your house. On one distribution line are connected a certain number of bulbs. Now,
electrical wires are designed to handle, say, 20A of current. When we switch the bulb to ON, it will draw 10-16
times more current than its nominal current for a short period of time, and that is named by electricians the
"inrush current". Next, the inrush current drops by itself to the nominal value.
All it takes to control that bulb (or a coil/injector, or a simple motor)
are power lines to carry sufficient current, and a
reliable switch. We do not need any PWM or Peak-and-Hold circuitry. Exactly the same thing happens with the
automotive injector, and the entire process is just a basic electrical application. The primary circuit in all
pictures above is an injector in series with a switch.
Incredibly, the hardware designers manage to implement the most complex circuits possible. We have seen an
injector "driver" built with "hardware logic": it had about 300 (three hundred) electronic
components on a 4"x4" PCB area. It was so dense with surface mount components, on both sides, that it
took us a long time to realize what were we looking at. That is just beyond any reasonable logic!
maximum circuitry needed to drive automotive injectors
This note was added later, because we received a few objections from one reader. He
said that by using PWM or Peak-and-Hold circuitry, the electronic designers try to reduce the temperature developed
on the switching transistor. That is not true. The transistor generates the minimum amount of temperature when it
works in the saturation mode. On the other hand, by using PWM or Peak-and-Hold we change in fact the switch (the
transistor) with a resistor, as an equivalent circuit. That resistor will lower the primary current Ip, and that
will result in a longer time for the Peak period.
Implicitly, any reduction in the primary current will result in
MORE temperature generated by the transistor-switch. In addition, even the sense resistor is another source of
heat; by eliminating it, we also get less dissipated heat on the PCB.
The "injector closing time" is handled orderly by the the following:
1. by a sudden drop in current, naturally;
2. by a very strong mechanical spring, carefully and specifically built
by the manufacturer inside the injector to perform this job.
The flyback diode employed (which has nothing to do with
either the PWM or the Peak-and-Hold circuits) is needed only to
reduce/annihilate the induced EMF parasitic currents.
By the way, the injectors do not burnout due to high currents or due to
high temperatures; they are specially built to work at very high
temperatures (around 250 C) generated by the running engine block. We hope this helps.
The PWM method has a certain specific advantage, considering the
"electrical primary circuit", which is of the "inductive" type due to
injector's coil. It happens that an inductive circuit forces the current
to "lag" the voltage. This means, there is a specific time delay until
the current reaches its maximum. As a result, electrical engineers use
to "chop" an AC sine wave, using PWM, in order to bring the current a
bit "closer" to the voltage (in phase).
Naturally, for each PWM pulse, the current still lags the voltage in
inductive circuits; however, the resultant current-vector (as the sum of
very many small PWM pulses) is almost identical to the voltage-vector.
This is the reason PWM is so efficient: it "pulls" the current-vector
closer to the voltage one. However, in automotive environments that PWM
frequency is a real danger, therefore it requires expensive shielding.
[By the way, in case there is some confusion, an inductive DC PWM
circuit behaves very much like an AC one; this means, there is some circuit
inductance/reactance to consider.]
schematic works as follows. The ON/OFF command signal comes from
the microcontroller on the "CONTROL" line as
a +5V/0V digital signal, and the Darlington pair, TIP121, closes or opens the primary circuit accordingly.
That is all. As you can see, in order to drive the injector, we need only one good transistor (of 0.25 USD), and one
ordinary current limiting resistor Rc (of 0.01 USD)!
The diodes D1 and D2 replace the "flyback" Zener diode; this is another "issue" mentioned in
LEARN HARDWARE FIRMWARE AND SOFTWARE DESIGN, and you can also see the simulation models for each case in our
"Diode Page"[of Corollary Theorems]. Diode DZ in the previous schematic is an expensive component--about
2 USD--and its main "qualities" are:
1. it is incredibly inefficient and unreliable;
2. it heats up a lot!
By using two (0.1 USD) ordinary diodes, the protection function is greatly improved and heat dissipation is
at a minimum.
Well, this is the entire "mystery" about driving automotive injectors. Please experiment for yourself, because we
suspect everybody could afford: 1 transistor, 1 resistor, + 2 ordinary diodes. Use the best oscilloscopes you
can find, and change the injector to any type you can get. The resulting conclusion is going to be that
you do not
need anything more than the above circuit to drive any automotive injector.
Please be careful with TIP121 because it is not the best Darlington pair for all
current ranges. TIP121 is manufacturer's recommendation for injectors working at 1A-2A. Other transistors are
way better suited for that, but you will have to discover them
yourself--we worked with 2N6045 at 4A only for
testing purposes. Again, other transistors are a lot better. Try to discover an 8A or 10A continuous DC
transistor in a TO220x package.
In addition, the sense resistor in the above schematics needs to be carefully calculated. We used the value of
0.05 ohms for exemplification, only. The calculated value at 4A is 0.07 ohms. Please be very careful when
dimensioning your electronic components, and never consider the schematic circuits
have correct values for all
possible situations. Always recheck and recalculate those values yourself.
Please be aware the transistor will heat up a lot: it could easily reach 250 Celsius degrees or even more. It
could become so hot that it may unsolder itself from the PCB in an instant. It did happen to us, because we had a
fault in the control circuit and one transistor remained ON for more than 100ms without any heat sink.
not hold your transistor ON for more than 10ms until you have proper heat
dissipation means in place. Once your transistor is well protected
against heat, experiment gradually with increased ON times greater than 10ms.
You could do yourself a lot of good if you take a look at a few "professional" automotive injector
drivers available on the marked. You are going to be stunned by
how inventive people could be
when there is absolutely no need or reason for that.
This example, and many others, come to motivate a few topics in our
Amazing Articles. Incredibly, many still refer to them as being
First published on August 04, 2005
© SC COMPLEMENT CONTROL SRL. All rights reserved.