Need help designing an automotive engine run detector.

[[email protected]]

Hello all,

I am trying to develop a new type of aftermarket product that will
connect to a vehicles battery. Ground return will be through the
chassis. I desire no other electrical connections. This will help
simplify installation for the typical consumer. Do you know many
people who can't find their vehicles battery? Well, I'm sure you
could name a few. But, the majority of us can. What I want to avoid
is telling someone to splice into some signal within their vehicle.
This can get difficult. The type and location of such signals will
vary from one manufacture to another. Who knows, it may not exist at
all in some vehicles...

Oh yeah, one more thing. This product will be mounted in the engine
compartment so any electronics specified will need to be available in
extended temperature ranges.

When activated this product will draw about 5 amp. I'm trying to
devise a way to allow activation of this product only while the engine
is running (for obvious reasons). Therefore the product will need to
sense engine running. If this criteria is met, it will be allowed to
switch in its load. It will also need to detect when the engine has
stopped. At this point it will disconnect its load and go into a low
power mode. It will sit powerd up indefinitely in low power mode.
Therefore, whatever sensing scheme we come up with will need to draw
minimal current. Preferably in the neighborhood of 500uA or below. I
don't want to drain someones battery!

I have already tried electrical system voltage sensing with no
success. One would think that when a vehicle is at rest the battery
would sit at approximately 12.6VDC. When the engine is running,
system voltage climbs up to charging levels (typically 13.0VDC or
greater). However, what I find is that system charging voltages range
all over the map from one vehicle manufacture to another. For
instance, one vehicle manufacture in the USA typically charges at
14.00VDC. However, they also have a mode that will drop charging
voltage down to approximately 10.8VDC in cold weather. Another US
manufacture charges at 13.2VDC and also has its own version of low
charging voltage mode. Some European vehicles charge at 12.8VDC.
Then there is the inconsistancy in rest voltage of a vehicles
battery. A fellow co-worker has a brand new vehicle who's battery
rests above 13VDC! As you can see, this is all over the map. Where
would I set a comparator to allow activation of my circuit?

This got me thinking about alternate options. One thing I do know is
that the raw power feed on a vehicle is REAL noisy when the engine is
running. Is there a way I can use this to my advantage? Essentailly
what i would need is a noise detector. At rest the VBAT power buss is
quite. While the engine runs there is all kinds of crud superimposed
on it. One interesting nibblet is the pulse generated on the power
buss whenever a cylinder fires. Could I detect and latch on in the
presence of ignition pulses? A potential problem... The shape and
amplitude of these pulses varies from vehicle to vehicle. In one
vehicle we tested, pulses were about 600mV high. The highest peak was
about 2uS wide. The period between pulses was 50mS with engine at
idle. In another vehicle, the pulse was approximatley 1V in amplitude
but only a few nanoseconds wide! The period between pulses was about
25mS with engine at idle.

Can you think of a highly reliable way to detect engine noise, or
better yet engine running, through sensing of the vehicles battery
feed? Something that perhaps would throw a logic signal high in the
presence of engine noise and then bring it low when the engine is shut
off? This circuit would need to reject random EMI coming from other
sources than the vehicle. I don't want someones ipod setting off my
product while the car is parked in an airport long term parking lot!

Any help you can offer would be greatly appreciated.

Ge0

 

[Martin Griffith]

On Mon, 08 Oct 2007 13:03:43 -0700, in sci.electronics.design

Hello all,

I am trying to develop a new type of aftermarket product that will
connect to a vehicles battery. Ground return will be through the
chassis. I desire no other electrical connections. This will help
simplify installation for the typical consumer. Do you know many
people who can't find their vehicles battery? Well, I'm sure you
could name a few. But, the majority of us can. What I want to avoid
is telling someone to splice into some signal within their vehicle.
This can get difficult. The type and location of such signals will
vary from one manufacture to another. Who knows, it may not exist at
all in some vehicles...

Oh yeah, one more thing. This product will be mounted in the engine
compartment so any electronics specified will need to be available in
extended temperature ranges.

When activated this product will draw about 5 amp. I'm trying to
devise a way to allow activation of this product only while the engine
is running (for obvious reasons). Therefore the product will need to
sense engine running. If this criteria is met, it will be allowed to
switch in its load. It will also need to detect when the engine has
stopped. At this point it will disconnect its load and go into a low
power mode. It will sit powerd up indefinitely in low power mode.
Therefore, whatever sensing scheme we come up with will need to draw
minimal current. Preferably in the neighborhood of 500uA or below. I
don't want to drain someones battery!

I have already tried electrical system voltage sensing with no
success. One would think that when a vehicle is at rest the battery
would sit at approximately 12.6VDC. When the engine is running,
system voltage climbs up to charging levels (typically 13.0VDC or
greater). However, what I find is that system charging voltages range
all over the map from one vehicle manufacture to another. For
instance, one vehicle manufacture in the USA typically charges at
14.00VDC. However, they also have a mode that will drop charging
voltage down to approximately 10.8VDC in cold weather. Another US
manufacture charges at 13.2VDC and also has its own version of low
charging voltage mode. Some European vehicles charge at 12.8VDC.
Then there is the inconsistancy in rest voltage of a vehicles
battery. A fellow co-worker has a brand new vehicle who's battery
rests above 13VDC! As you can see, this is all over the map. Where
would I set a comparator to allow activation of my circuit?

This got me thinking about alternate options. One thing I do know is
that the raw power feed on a vehicle is REAL noisy when the engine is
running. Is there a way I can use this to my advantage? Essentailly
what i would need is a noise detector. At rest the VBAT power buss is
quite. While the engine runs there is all kinds of crud superimposed
on it. One interesting nibblet is the pulse generated on the power
buss whenever a cylinder fires. Could I detect and latch on in the
presence of ignition pulses? A potential problem... The shape and
amplitude of these pulses varies from vehicle to vehicle. In one
vehicle we tested, pulses were about 600mV high. The highest peak was
about 2uS wide. The period between pulses was 50mS with engine at
idle. In another vehicle, the pulse was approximatley 1V in amplitude
but only a few nanoseconds wide! The period between pulses was about
25mS with engine at idle.

Can you think of a highly reliable way to detect engine noise, or
better yet engine running, through sensing of the vehicles battery
feed? Something that perhaps would throw a logic signal high in the
presence of engine noise and then bring it low when the engine is shut
off? This circuit would need to reject random EMI coming from other
sources than the vehicle. I don't want someones ipod setting off my
product while the car is parked in an airport long term parking lot!

Any help you can offer would be greatly appreciated.

Ge0

Somewhat confusing on the first reading, but if it's a petrol engine,
you could make a capactive sensor around the spark plug lead, that
could possibly trigger a monostable/ microprocessor to come out of
sleep mode, for diesel, a guess, a current transformer across the fuel
injector cable


Martin

 

[[email protected]]

Somewhat confusing on the first reading, but if it's a petrol engine,
you could make a capactive sensor around the spark plug lead, that
could possibly trigger a monostable/ microprocessor to come out of
sleep mode, for diesel, a guess, a current transformer across the fuel
injector cable

Martin- Hide quoted text -

- Show quoted text -

What, do you think I'm a little wordy ?

I've learned from experience that I need to lay most of my groundwork
up front. That way I'm not spending days answering questions.
However, there is a fine balance between not enough content, and too
much content. I haven't mastered that one yet .

If I were to run an external sensor, or pickup wire as you suggested,
my options would open up considerably. However, in this instance, I'm
trying to work out something using only the power and ground feed
coming into my product. My intent was to constantly monitor for noise
vs. a clean DC signal. This is under the assumption that the DC
voltage buss in a vehicle gets real noisy once the engine has started,
and is nice and clean when the engine is not running.

Ge0

 

[Martin Griffith]

On Mon, 08 Oct 2007 13:52:16 -0700, in sci.electronics.design

What, do you think I'm a little wordy ?

Yes, a small table, may have been simpler

I've learned from experience that I need to lay most of my groundwork
up front. That way I'm not spending days answering questions. agreed
However, there is a fine balance between not enough content, and too
much content. I haven't mastered that one yet . yep

If I were to run an external sensor, or pickup wire as you suggested,
my options would open up considerably. However, in this instance, I'm
trying to work out something using only the power and ground feed
coming into my product. My intent was to constantly monitor for noise
vs. a clean DC signal. This is under the assumption that the DC
voltage buss in a vehicle gets real noisy once the engine has started,
and is nice and clean when the engine is not running.

Ge0

What you seem to have proved is that the variables mentioned are
extremely variable, I would not trust them. You could however use a
motion sensor, to see if the vehicle is moving, say a magnet on a
string and a reed relay.


Martin

 

[PeterD]

Hello all,

I am trying to develop a new type of aftermarket product that will
connect to a vehicles battery. Ground return will be through the

Your observations on charging are off base... but...


Use a microphone and *listen* for a running engine.


I'd monitor battery voltage, that will always tell you when the
battery is charging, and it is safe to draw a load from it.

Another alternative is an indcutive pickup fastened to the alternator.

 

[Jim Stewart]

Your observations on charging are off base... but...


Use a microphone and *listen* for a running engine.


I'd monitor battery voltage, that will always tell you when the
battery is charging, and it is safe to draw a load from it.

Another alternative is an indcutive pickup fastened to the alternator.

Or a thermocouple on the exhaust downpipe.

 

[me]

Or a thermocouple on the exhaust downpipe.

That will tell you its running for an hour after it is off.

Just pick the signal off the distributor (i.e. tachometer signal).

 

[Glen Walpert]

Can you AC couple the alternator ripple through a capacitor, rectify,
filter, then compare to a reference with a low power comparator?

 

[Jim Thompson]

Can you AC couple the alternator ripple through a capacitor, rectify,
filter, then compare to a reference with a low power comparator?

If you happen to have a GM alternator where they bring out the
"neutral" pin, there's a nice BIG ripple there.

This output is interesting, because its duty cycle and DC offset can
pin-point faults in the alternator.

...Jim Thompson

 

[Boris Mohar]

What, do you think I'm a little wordy ?

I've learned from experience that I need to lay most of my groundwork
up front. That way I'm not spending days answering questions.
However, there is a fine balance between not enough content, and too
much content. I haven't mastered that one yet .

If I were to run an external sensor, or pickup wire as you suggested,
my options would open up considerably. However, in this instance, I'm
trying to work out something using only the power and ground feed
coming into my product. My intent was to constantly monitor for noise
vs. a clean DC signal. This is under the assumption that the DC
voltage buss in a vehicle gets real noisy once the engine has started,
and is nice and clean when the engine is not running.

Ge0

Batteries have a temperature coefficient. Account for it. Also Google for
automotive load dump and be prepared to take on some 80V spikes.



Regards,

Boris Mohar

Got Knock? - see:
Viatrack Printed Circuit Designs (among other things) http://www.viatrack.ca

void _-void-_ in the obvious place

 

[Jeff Liebermann]

[email protected] hath wroth:

I am trying to develop a new type of aftermarket product that will
connect to a vehicles battery. Ground return will be through the
chassis.

Bad idea. You're bypassing the entire fuse and breaker protection
mechanisms in the vehicle. You would need to attach a fuse directly
to the positive terminal of the battery, a non-trivial exercise.
You're worried about the GUM (great unwashed masses) doing a proper
install, yet you don't seem to be worried about the same person is
going to get a wrench between the positive terminal and ground, with
the resultant smoke, fire, burns, and litigation.

Ground return through the chassis is possible, but be advised that
many vehicle bodies are now mostly plastic and fiberglass. Finding a
suitable place to handle your 5A of continuous load may be a
challenge.

This product will be mounted in the engine
compartment so any electronics specified will need to be available in
extended temperature ranges.

Yep. That's the worst case scenario for electronics. Roughly -40C to
+105C. Don't forget about water proofing and high voltage spikes on
the power lines.

When activated this product will draw about 5 amp.

75 watts? Where are you going to empty out the heat?

I'm trying to
devise a way to allow activation of this product only while the engine
is running (for obvious reasons).

Ignition switch. If that's offensive, try the cancer stick igniter,
which only has power when the engine is running or the ACC switch is
on.

 

[D from BC]

Hello all,

I am trying to develop a new type of aftermarket product that will
connect to a vehicles battery. Ground return will be through the
chassis. I desire no other electrical connections. This will help
simplify installation for the typical consumer. Do you know many
people who can't find their vehicles battery? Well, I'm sure you
could name a few. But, the majority of us can. What I want to avoid
is telling someone to splice into some signal within their vehicle.
This can get difficult. The type and location of such signals will
vary from one manufacture to another. Who knows, it may not exist at
all in some vehicles...

Oh yeah, one more thing. This product will be mounted in the engine
compartment so any electronics specified will need to be available in
extended temperature ranges.

When activated this product will draw about 5 amp. I'm trying to
devise a way to allow activation of this product only while the engine
is running (for obvious reasons). Therefore the product will need to
sense engine running. If this criteria is met, it will be allowed to
switch in its load. It will also need to detect when the engine has
stopped. At this point it will disconnect its load and go into a low
power mode. It will sit powerd up indefinitely in low power mode.
Therefore, whatever sensing scheme we come up with will need to draw
minimal current. Preferably in the neighborhood of 500uA or below. I
don't want to drain someones battery!

I have already tried electrical system voltage sensing with no
success. One would think that when a vehicle is at rest the battery
would sit at approximately 12.6VDC. When the engine is running,
system voltage climbs up to charging levels (typically 13.0VDC or
greater). However, what I find is that system charging voltages range
all over the map from one vehicle manufacture to another. For
instance, one vehicle manufacture in the USA typically charges at
14.00VDC. However, they also have a mode that will drop charging
voltage down to approximately 10.8VDC in cold weather. Another US
manufacture charges at 13.2VDC and also has its own version of low
charging voltage mode. Some European vehicles charge at 12.8VDC.
Then there is the inconsistancy in rest voltage of a vehicles
battery. A fellow co-worker has a brand new vehicle who's battery
rests above 13VDC! As you can see, this is all over the map. Where
would I set a comparator to allow activation of my circuit?

This got me thinking about alternate options. One thing I do know is
that the raw power feed on a vehicle is REAL noisy when the engine is
running. Is there a way I can use this to my advantage? Essentailly
what i would need is a noise detector. At rest the VBAT power buss is
quite. While the engine runs there is all kinds of crud superimposed
on it. One interesting nibblet is the pulse generated on the power
buss whenever a cylinder fires. Could I detect and latch on in the
presence of ignition pulses? A potential problem... The shape and
amplitude of these pulses varies from vehicle to vehicle. In one
vehicle we tested, pulses were about 600mV high. The highest peak was
about 2uS wide. The period between pulses was 50mS with engine at
idle. In another vehicle, the pulse was approximatley 1V in amplitude
but only a few nanoseconds wide! The period between pulses was about
25mS with engine at idle.

Can you think of a highly reliable way to detect engine noise, or
better yet engine running, through sensing of the vehicles battery
feed? Something that perhaps would throw a logic signal high in the
presence of engine noise and then bring it low when the engine is shut
off? This circuit would need to reject random EMI coming from other
sources than the vehicle. I don't want someones ipod setting off my
product while the car is parked in an airport long term parking lot!

Any help you can offer would be greatly appreciated.

Ge0

The High Tech Way...
Ever use voice recognition software?.. Last time I used that, I had to
talk for 30 minutes to train it to my voice.
Similarly, a computer (uC, DSP) may need to be trained to identify the
noise off the battery when the engine is running.
Your gizmo may need a training mode.

The Low Tech Way
Hall effect sensor (current clamp) on ignition coil primary wire..


D from BC

 

[Tom2000]

What, do you think I'm a little wordy ?

Nope. In fact, it's refreshing to see such a comprehensive problem
statement.

The typical statement reads something like "I'm trying to fix a
system that runs all the automation in a factory, but I only read 2.3
volts. Should I change R17?"

Tom

 

[Tom2000]

Nope. In fact, it's refreshing to see such a comprehensive problem
statement.

The typical statement reads something like "I'm trying to fix a
system that runs all the automation in a factory, but I only read 2.3
volts. Should I change R17?"

Further on the above... you also presented some good real-world
information on charging systems that was new to me, and I appreciate
it.

Now, on to your problem. I'm thinking that anyone who has the
skills and confidence to find some clear sheet metal in the engine
compartment for a mounting location, drill the mounting holes, and
install the battery cable taps could find the LV ignition wire that
would be used as a goof-proof signal to turn your unit on and off.

However, I can appreciate your desire for the simplest installation
possible, so I'll add one more idea to the collection. How about
little a piezo vibration sensor that the installer would attach to a
convenient mounting point on the engine assembly? Valve cover?
Wiring loom mounting bracket? Something like that.

Good luck with your endeavour.

Tom

 

[Jan Panteltje]

Hello all,

No electical connections? Forget it.
Engine sensing I have seen done by taking ignition coil primary,
and using that impulse to discharge a capacitor with a transistor.
Something like this was in my car that ran on LPG (liquid propane),
it closed the valves if engine stoppped:


+12
|
R1
from ign. |----------------R --- 0V if engine runs, else 12V
coil prim. c |
- R -- C ---b NPN |
| e C1
R | |
| | |
/// /// ///

I know this, because I had to fix it....
Solder joint (can blow you up too).
Ooops, homeland security, ban all LPG cars....
LOL
;-)

 

[Barry Lennox]

Hello all,

I am trying to develop a new type of aftermarket product that will
connect to a vehicles battery. Ground return will be through the
chassis. I desire no other electrical connections. This will help

snip

A microphone and detect acoustic noise. I'm aware of one "engine
detector" in a product that uses this.


Barry

 

[PeterD]

If you happen to have a GM alternator where they bring out the
"neutral" pin, there's a nice BIG ripple there.

This output is interesting, because its duty cycle and DC offset can
pin-point faults in the alternator.

Many GM alternators also bring out one phase as well. Frequently used
to drive tachometers for example, and to provide a low power AC
voltage.

 

[Ken S. Tucker]

Hello all,

I am trying to develop a new type of aftermarket product that will
connect to a vehicles battery. Ground return will be through the
chassis. I desire no other electrical connections. This will help
simplify installation for the typical consumer. Do you know many
people who can't find their vehicles battery? Well, I'm sure you
could name a few. But, the majority of us can. What I want to avoid
is telling someone to splice into some signal within their vehicle.
This can get difficult. The type and location of such signals will
vary from one manufacture to another. Who knows, it may not exist at
all in some vehicles...

Oh yeah, one more thing. This product will be mounted in the engine
compartment so any electronics specified will need to be available in
extended temperature ranges.

When activated this product will draw about 5 amp. I'm trying to
devise a way to allow activation of this product only while the engine
is running (for obvious reasons). Therefore the product will need to
sense engine running. If this criteria is met, it will be allowed to
switch in its load. It will also need to detect when the engine has
stopped. At this point it will disconnect its load and go into a low
power mode. It will sit powerd up indefinitely in low power mode.
Therefore, whatever sensing scheme we come up with will need to draw
minimal current. Preferably in the neighborhood of 500uA or below. I
don't want to drain someones battery!

I have already tried electrical system voltage sensing with no
success. One would think that when a vehicle is at rest the battery
would sit at approximately 12.6VDC. When the engine is running,
system voltage climbs up to charging levels (typically 13.0VDC or
greater). However, what I find is that system charging voltages range
all over the map from one vehicle manufacture to another. For
instance, one vehicle manufacture in the USA typically charges at
14.00VDC. However, they also have a mode that will drop charging
voltage down to approximately 10.8VDC in cold weather. Another US
manufacture charges at 13.2VDC and also has its own version of low
charging voltage mode. Some European vehicles charge at 12.8VDC.
Then there is the inconsistancy in rest voltage of a vehicles
battery. A fellow co-worker has a brand new vehicle who's battery
rests above 13VDC! As you can see, this is all over the map. Where
would I set a comparator to allow activation of my circuit?

This got me thinking about alternate options. One thing I do know is
that the raw power feed on a vehicle is REAL noisy when the engine is
running. Is there a way I can use this to my advantage? Essentailly
what i would need is a noise detector. At rest the VBAT power buss is
quite. While the engine runs there is all kinds of crud superimposed
on it. One interesting nibblet is the pulse generated on the power
buss whenever a cylinder fires. Could I detect and latch on in the
presence of ignition pulses? A potential problem... The shape and
amplitude of these pulses varies from vehicle to vehicle. In one
vehicle we tested, pulses were about 600mV high. The highest peak was
about 2uS wide. The period between pulses was 50mS with engine at
idle. In another vehicle, the pulse was approximatley 1V in amplitude
but only a few nanoseconds wide! The period between pulses was about
25mS with engine at idle.

Can you think of a highly reliable way to detect engine noise, or
better yet engine running, through sensing of the vehicles battery
feed? Something that perhaps would throw a logic signal high in the
presence of engine noise and then bring it low when the engine is shut
off? This circuit would need to reject random EMI coming from other
sources than the vehicle. I don't want someones ipod setting off my
product while the car is parked in an airport long term parking lot!

Any help you can offer would be greatly appreciated.

Ge0

At first I thought your question was silly and easy but
in 2nd thought I see it's more subtle than that.
We'll need to 1st define what "engine run" is defined
by, off hand I'd say a Power Stroke Combustion, PSC.
If so, what's a cheap easy rugged way of confirming
PSC? And then do we use one or all cylinders?

I know how to do it, but it ain't cheap.
Ken

 

[René]

I did once play with the thought of making an RPM indicator / shift
light based on the alternator whine always present on the battery DC.
I never built it, but should not be too difficult with some adaptive
filtering and a simple cpu. (and some calibrating obviously)

No matter how low the Ri of the battery, the rectified pulses from the
alternator are measurable. Car radio manufacturers spend a lot of
engineering to get this unwanted sound out.

If only the *presence* of the characteristic whine needs to be
detected...., for added securety corellate frequency with DC -
delta...hmm.

 

[Rich Grise]

When activated this product will draw about 5 amp. I'm trying to devise
a way to allow activation of this product only while the engine is
running (for obvious reasons). Therefore the product will need to
sense engine running.

How do they do it in standby gensets? They auto-start; there must
be some way to detect "engine running" so it knows when to release
the "start" button, right?

Good Luck!
Rich