Over the last four weeks we’ve looked at systems that switch on and off
(Performance Electronics, Part 1), systems that use pulsed signals for control
(Performance Electronics, Part 2)
and
(Performance Electronics, Part 3), and the modification of analog signals like those from most airflow
meters
(Performance Electronics, Part 4).
This week, in the last of the series, we’re going to look
at one of the electronic problems that can occur when intercepting pulsed
signals.
Frequency Signals
In Parts 2 and 3 of this series we discussed variable duty cycle signals.
These are used to control solenoid valves in systems as diverse as boost control
and the fuel injection. In addition to variable duty cycle signals, there’s
another form of pulsed signal. It often keeps the same duty cycle (eg 50 per
cent) but varies in frequency.
One of the easiest of these signals to understand is a speed sensor signal.
There are different ways in which speed signals are generated but let’s look
first at the simplest. Some older speed sensors are built into the speedo of the
car and comprise a reed switch that is closed whenever a magnet passes. If the
magnet spins with the speedo drive, the switch will open and close rapidly as
the car is driven along. The faster the car is travelling, the faster the switch
is turned on and off.
This diagram shows the system. Each time the magnet gets near the reed
switch, it closes, sending a 5V to the ECU. Then, when the magnet moves away
from the reed switch, the switch opens, turning off the 5V. The results is an
on/off 5V signal – a square wave.
ECU Input Pull Up and Pull
Down
Now, let’s take a closer look at what happens inside the ECU. The arrangement
could be like this – the signal goes straight into a microcontroller that looks
at the signal’s frequency and then works out how fast the car is travelling.
When the reed switch is closed, there’s 5 volts being fed into the micro input.
But what happens when the reed switch is open? Then there’s nothing – the input
is just floating! Any electrical noise on the input could be seen as a signal –
not good.
So to avoid that problem, here’s what is done. A resistor is wired between
the input and ground. When the reed switch is closed, 5 volts is available on
the micro’s input – the resistor is too high in value to prevent much current
passing through it to ground. But when the reed switch opens, the resistor can
pull the micro input to ground. Now the input is no longer floating because it’s
tied to ground.
So that’s a pull-down resistor – but what about a pull-up? It’s very much the
same idea but this time the other side of the reed switch is connected to ground
and the pull-up resistor is tied to 5V. When the reed switch is open, the input
to the micro is pulled-up to 5V. When the reed switch is closed, the input to
the micro is pulled-down to ground. (Note the voltage of the pull-up doesn’t
have to 5V – it could be 12V.)
Sensor Output Pull-Up and
Pull-Down
Some sensors need pull-up or pull-down resistors before they’ll produce an
output. For example, a Hall Effect sensor usually requires a pull-up resistor
before it will have an output. Let’s take a closer look at this.
The action of the Hall Effect device can be thought of as being like a switch
to ground. Here the switch is open and when the output is off, the meter shows
zero voltage output.
But when the switch is closed, there’s still zero voltage to ground! Hmm, no
signal output, whether the Hall Effect sensor is ‘open’ or ‘closed’.
But if we add a pull-up resistor, the situation changes. Now when the sensor
is ‘open’, the output is 5V.
When the sensor is ‘closed’, the output drops to 0 volts. (There’s not a
short circuit between the 5V supply and ground because of the resistor in
between.)
In some cases sensors need pull-down resistors to operate – but it’s
just a reverse of what was shown for a pull-up sensor.
The Implications
As we said, a few parts ago in this series, all this used to fall into the
“yeah-it’s technically-trick-but-so-what?” area of knowledge. Pull-ups,
pull-downs – who cares? Well, if you intercept a frequency signal and want to
modify it, it all becomes very important! For example, if you put a speedo
interceptor into the circuit and the correct pull-ups and pull-downs are not put
into place, the system will stop working.
Intercepting a frequency signal is much trickier than intercepting an
analog voltage of the sort coming out of an airflow meter!
The main trouble with intercepting frequency circuits is that until you make some measurements, you don’t know
what you’ve got. And furthermore, after you’ve
made those measurements, the module that you’re doing the intercepting with
needs to be able to take all configurations into account. Before we get into the
possible problems, let’s take a look at the Silicon Chip Speedo Corrector
interceptor kit. (This can alter the speedo reading in 1 per cent increments,
either up or down. It can also be used for tachos.)
This diagram shows how you would never have the system set up! In fact
it shows pretty all the options in place simultaneously. But let’s take it step
by step. On the input to the speedo corrector (so what the sensor sees as its
output) we have the options of a pull-up resistor or a pull-down resistor or no
resistor. The pull-up resistor has the further options of being either pulled up
to 5V or pulled up to 12V. As you would expect, the pull-down resistor goes to
ground.
On the output (so what the ECU sees on its input) we have the same pull-up
options of 5 or 12V (if it’s selected for the input it’s also selected for the
output), and also a pull-down is available.
If the pull-up or pull-down is wrong on the input to the Speedo Corrector,
it’s likely the sensor won’t work – it will have zero output. However, what if
the pull-up or pull-down is wrong on the Speedo Corrector output? Hmm, now it
gets really tricky.
If the pull-up or pull-down resistor is wrong on the output of the Speedo
Corrector (ie it’s the opposite of the input to the ECU) then the resulting
signal will depend on the battle of the values between the two resistors. In
other words, as shown here, with a pull-down at the output of the Speedo
Corrector and a (hidden!) pull-up in the ECU, the result will be that the two
resistors act as a voltage divider and the signal voltage ends up somewhere
between 0V and 5V. So it’s important the output pull-up/pull-down resistor
matches what’s inside the ECU.
OK, it’s not a very easy subject, is it? So let’s take a look at how you can
work out what’s needed to successfully intercept a frequency signal, just by
making some measurements with your mutlimeter.
Measuring a Working System
The first step is to measure what’s going on when the standard system is
working correctly – that is, before an interceptor is put into the circuit.
Since the system is working we know
that there will be a signal travelling between the sensor and the ECU. But what
sort of signal is it? By far the best way of finding this out is to use an
oscilloscope
(see Using Oscilloscopes on Cars, Part 1)
but even if you have just a multimeter, it’s possible to work out what’s
happening.
So far we’ve described systems where the signal is a square wave – an on/off
signal. As can be seen on this graph, as time passes, the voltage goes
high-low-high-low. This is a 5V square wave.
But another possibility also exists – the signal might be a sine wave like
this. For example, the ABS sensors in most cars output a sine wave signal. A
sine wave signal is measurable by a multimeter set to AC volts, but when set to
DC volts, the meter will show little or no voltage (because the signal spends
the some time above 0V as it does below it).
So the first question to answer is: is the signal a square wave or sine
wave? If measurement of the working signal shows an AC voltage, it is a sine
wave. If measurement of the working signal shows no (or very little) DC voltage,
it is a square wave.
A sine wave output sensor doesn’t need a pull-up or pull-down resistor on the
input to the interceptor – like an AC bicycle generator, this type of sensor
will work without being pulled up or down.
But what about on the output of the interceptor? How is that configured? In the case of
the Speedo Corrector and most other interceptors, sine waves get turned into
square waves by the action of the interceptor. This is the case simply because
it’s easier to electronically engineer an interceptor that does this, and since
the ECU isn’t looking at the shape of the wave but only how many up/downs it has
per second, the system still works fine. We know the ECU wouldn’t have had a
pull-up or pull-down resistor inside it (because there was no need to have one)
but for the Speedo Corrector output to work, it needs a pull-down or pull-up
resistor on its output. So with a sine wave signal, an output pull-down resistor
is fitted, pulling down from 5V.
But what if the signal is a square wave? A 50 per cent duty cycle square wave
working on a 0-5-0-5V waveform will have an average DC voltage of 2.5V. So if we
can measure a 2.5V DC voltage on the working signal, we can be fairly sure we’re
dealing with a 5V square wave. If that’s the case, we need to pull the signal to 5V, or pull the signal from 5V. Either way, we know we’re
dealing with a 5V square wave.
If we measure a DC voltage higher than 2.5V (eg 6V) we know we’re dealing
with a square wave that works on a higher voltage – like 12V for example. So
we’re going to need to pull the voltage from (or to) a higher voltage, like
12V.
So at this stage we don’t know whether we need pull-up or pull-down
resistors, but at least we now know we’re dealing with a square wave that works
0-5-0-5 or 0-12-0-12 volts. To find out whether pull-up or pull-down resistors
are needed requires that we disconnect the sensor and directly measure its
output.
Measuring the Sensor
The next step is to measure the sensor output signal without the sensor
connected to the ECU. If we can measure a voltage that’s relatively high (eg
above 2V) we know that the ECU must use a pull-down resistor, in order that the
voltage falls to zero in the ‘low’ parts of the waveform. Therefore, a high measured voltage will
require the installation of a pull-down resistor on the interceptor’s input.
If we can measure a sensor voltage that is relatively low (eg below 2V) we
know that the ECU must use a pull-up resistor, so the voltage is pulled to a
much higher value than this on the ‘high’ parts of the waveform. Therefore, a
low measured voltage will require the installation of a pull-up resistor on the
interceptor’s input.
(Important: when doing these measurements, just the signal output of the
sensor should be disconnected from the ECU. Make sure power and earth are still
available to the sensor!)
The output of the interceptor is then configured in the same way as the input
– ie an input pull-down uses an output pull-down, and an input pull-up uses an
output pull-up.
Steps Summarised
So by making measurements with a multimeter, first of the working system and
then of the sensor output, we can find out the following:
- Is
the signal a sine wave or square wave? A sine wave does not require a pull-up or
pull-down resistor on the input to the interceptor, but the interceptor output
signal will need to be pulled down from 5V.
- Is
the square wave working with 5V or 12V on its high points? The pull-up resistor
that we might need to fit will need to pull-up to this voltage.
- Is
the signal output of the disconnected sensor high? If so, a pull-down resistor
will be needed.
- Is
the signal output of the disconnected sensor low? If so, a pull-up resistor (to
whatever pull-up voltage we measured above in Step 2) will be needed.
|
Series Conclusion
That’s it for this series. In it we’ve covered a lot of ground – from
hysteresis to pull-up resistors. We hope you’ve enjoyed the ride, brow-wrinkling
though it has been. One thing’s for certain, understanding these concepts will
make electronically modifying your car far more likely to be successful.
Did you enjoy this article?
Please consider supporting AutoSpeed with a small contribution. More Info...
|
More of our most popular articles.
|
|
|
Email a friend
Print article
|