Maker Pro
Maker Pro

General engine tach. electrical characteristics?

Hi there,
I'm considering designing a data aquisition/storage box for some
personal use on my car. The car in question is a 1992 Nissan 240sx,
pre-ODBII. Because I don't have a working scope at my disposal(long
story), I'm looking for general information on the tachometer signal.
The car is a distributor-based 4cylinder, with an external coil and
power transistor. There is a crank position sensor(I assume it to be a
hall-effect type), integrated into the distributor housing. The tach
signal to the gauge cluster is supplied by the ECM/ECU.
I'm basically looking for general information about this signal:
what I can expect. I'm assuming it will be fixed-pulse width, 5V(TTL)
or 12V peak-to-peak signal, with varrying frequency. What is the upper
range for the frequency? Is the frequency a *direct* indication of the
crank/engine RPM, or is it divided up/down?

Thanks for any insite: I'll need to get my scope working eventually,
but I just need to know about the tach signal to select an apropriate
microprocessor(capable of timing the signal, and if need be, reduce the
pk-pk voltage).

Regards,
-Jamie
 
M

Mike Deskevich

Jan 1, 1970
0
if you're going to be making a daq for your car, your best bet is to
get a factory service manual. they're usually pretty cheap on ebay.
i'm working on the same project as you are for my 1989 toyota (also
pre-OBDII) and the FSM has really helped. it has pinouts for the ECU
and all the info on voltages and everything. one place you might want
to check out is www.autoshop101.com. they have tech manuals on all
kinds of neat stuff. their manuals are primarily toyota, but i
wouldn't be surprised if the ECUs were similar (toyota and nissan use
aisin as a parts supplier).

but to answer your question about the CPS, here's what i do know. it's
a +5 (or maybe -5) signal that happens once per revolution (no
multipling or dividing of the signal). it's primary use is to tell the
ECU when to fire the injectors. my toyota (4cyl) has multiport
injection, however all 4 injectors are fired at the same time, so the
CPS only needs to give the signal once. i remember that my '95 nissan
had TBI injection, so i wouldn't be surprised if your '92 also used
TBI. so it's the same principle, the CPS just tells the ECU when to
fire the injector.

from your question, it seems like you want to read the tach signal
coming from the ECU to the gauge. i can't help you there, but my guess
is that it's an analog signal and the tachometer is just a voltmeter
(just like the speedometer). in my project, i'm not trying to read
anything being sent to the gauges, i'm just trying to read the output
of sensors that are going to the ECU and the output of the ECU going to
the injectors. that seemed to be the easiest way to do it.
 
B

BobG

Jan 1, 1970
0
The alternator has 3 humps per alternator rev (its 3 phase), and if you
measure the diam of the alternator pulleys, you can get the crank rps
from the pd of the alternator pulses (cap couple and use a schmidt
trigger)
 
Bob, that's quite a clever way of determining the RPM! That would be
great on an older vehicle with no dedicated CPS output. In my case
however, I plan to use the sensor already in place.

Mike, I do indeed have a FSM, and it has certainly been an invaluable
purchasse. The engine in my Nissan is also multiport fuel injected. In
my Nissan the engine will fire sequentially normally, but will fall
back to simultaneous when in 'safe' mode or starting.

You brought up a good point that I had not considered: the tachometer
gauge signal is probably analog. I would rather read the signal going
to the ECU, as you have planned. Good call there.

What frequency range have you observed in your Toyota?

Thanks to both of you for your responses.
-Jamie
 
M

Mike Deskevich

Jan 1, 1970
0
You brought up a good point that I had not considered: the tachometer
gauge signal is probably analog. I would rather read the signal going
to the ECU, as you have planned. Good call there.

What frequency range have you observed in your Toyota?

there are three signals i can choose from to read the rpm on the
toyota. there's the ignition signal that comes from the distributor
and that fires for each spark (2 times per revolution - 4 cylinders on
a 720 degree cycle). i have that hooked up to a cheap aftermarket tach
(my gauge cluster doesn't have a tach). then there are two AC signals
that go to the ecu. one of them has a period of 180 degrees on the
crank and the other has a period of 30 degrees. i have looked at both
of these, but haven't used them yet, i'll probably just put a low pass
filter one one of them and read the voltage with an A/D.

i'm doing my project in two steps. the first step is just to make some
cool gauges that i can use to monitor as i trim fuel delivery, the
second stage will be to hook up a PIC or something similar and do
actual data collection. since engines are relatively slow compared to
electronics (rpm signal is only 10-100Hz), i can get away with cheap
electronics.

my first project is to read the signal from the ecu to the injectors.
i put a low pass filter which gives me an analog signal that is
proportional to how long the injectors are open. hooking that up to a
voltmeter gives me a gauge of how much fuel i'm burning per unit time.
if i had a speed sensor in my car, i could divide the signal by speed
and get an analog instantaneous gas milage gauge. i also need to hook
up a gauge to my o2 sensor (1-2Hz frequency) so that i can watch how
rich/lean i'm running. after i do that, i'm going to add a trim pot
between the air flow meter and the ecu so that i can tell the engine
that it's getting more air than it really is. this way the ecu will
lean out the mixture a little more. i can then use the o2 sensor gauge
to make sure i don't lean it out too much and melt my pistons. the
reason i'm doing this, is that in the primitive efi in the toyota, any
time i run more than 1/3 throttle, the ecu goes into open loop mode and
ignores the o2 sensor. the default programming is to run rich at WOT.
i'd rather save a little fuel and go a little lean at WOT. especially
since with my little 100HP motor i'm running WOT anytime i'm on the
highway.

after i get the fun gauge project done, it'll be easy to hook up a
couple of A/Ds and send data to my laptop. haven't decided exactly
what i'm going to do with the data yet. but it sure would be fun to
have.

mike
 
D

Dwayne

Jan 1, 1970
0
BobG said:
The alternator has 3 humps per alternator rev (its 3 phase), and if you
measure the diam of the alternator pulleys, you can get the crank rps
from the pd of the alternator pulses (cap couple and use a schmidt
trigger)

In theory yes, but if the belt slips then it would mess up the system.

Dwayne
 
D

Dwayne

Jan 1, 1970
0
Dwayne said:
In theory yes, but if the belt slips then it would mess up the system.

Dwayne

And you have no way of telling when any of the pistons are at TDC (or any
other reference point).

Dwayne
 
Mike,
I follow what you are doing with the fuel trim. I'm suprised to
hear that your O2 sensor has a frequency output. Is it a 4-wire
type(with heater)? Mine is the single-wire output, so it must varry
resistance. Typically the higher end wideband O2 sensors are a direct
0-5V output.
I'm not sure I understand how a 10-300Hz frequency correlates to
the RPM. If 100Hz has a period of 10mS, then that would be
0.000166666minutes, correct?
 
M

Mike Deskevich

Jan 1, 1970
0
I follow what you are doing with the fuel trim. I'm suprised to
hear that your O2 sensor has a frequency output. Is it a 4-wire
type(with heater)? Mine is the single-wire output, so it must varry
resistance. Typically the higher end wideband O2 sensors are a direct
0-5V output.

my o2 sensor is a 2 wire (your's probably is too, the other "wire" is
probably the ground from the exhaust pipe). afaik all o2 sensors
output a voltage that's a function of o2 content in the exhaust. .45v
means that it's running at 14.7:1 less than .45v means it's running
lean and mre than .45v means it's running rich. the goal is to have
14.7:1 mixture averaged over time, however, for the catalyitic
converter to work right it has to have a slight excess of CO (rich) for
a short time and then a slight excess of NOx (lean) for a short time.
so what the ecu does is run the mixture slightly lean until the o2
sensor reads less than .45v then it enrichens the mixture until the o2
sensor reads more than .45v and then it leans out the mixture and so
on. the voltage coming from the o2 sensor oscillates in the 1-2Hz
range. the rate is really a function of how fast the o2 sensor can
react to change in exhaust o2 concentration (pretty fast), and how fast
the ecu can change the fuel delivery rate (pretty slow).

the difference between 2 and 4 wire o2 sensors are just that the 4 wire
ones use the extra two wires to run a heater to get the o2 sensor up to
operating temperature (700F) more quickly.
I'm not sure I understand how a 10-300Hz frequency correlates to
the RPM. If 100Hz has a period of 10mS, then that would be
0.000166666minutes, correct?

100Hz = 6000/min

6000/min with 2 pulses per revolution = 3000 rpm

mike
 
Ok I get the first part... 1Hz=1Pulse/sec, so 100Hz=100Pulses/sec..
Therefore, (60secs/min) * 100Pulses =6000Pulses/min

Why is it 2 pulses per revolution though? One turn of the crank means
the 1st cylinder is at TDC again, for the second time, correct?

Also FWIW, my sensor is 'technically' not a crank position sensor is
it? (gear driven off timing chain components).
 
M

Mike Deskevich

Jan 1, 1970
0
Ok I get the first part... 1Hz=1Pulse/sec, so 100Hz=100Pulses/sec..
Therefore, (60secs/min) * 100Pulses =6000Pulses/min

Why is it 2 pulses per revolution though? One turn of the crank means
the 1st cylinder is at TDC again, for the second time, correct?

it just happens that my cps is driven off the timing chain (probably
the same gear that runs the camshaft), so it spins at 1/2 the rate of
the crankshaft. it has 4 places during it's revolution that sends a
pulse to the ecu, so it's 4 pulses per 720 degrees of the crankshaft,
or 2 pulses per revolution.

Also FWIW, my sensor is 'technically' not a crank position sensor is
it? (gear driven off timing chain components).

it probably is. as long as it's in phase with the crankshaft, you're
good. if you care about exact timing, then you'll just need to know
the phase between the pulse and TDC of 1 piston. for example, in mine
i get the 4 pulses per 720 degrees. i haven't checked, but i bet that
each pulse comes when one of the pistons is at or near TDC. my guess
is that toyota was planning on using each pulse to tell the ECU to fire
an injector, but during production decided to just fire all 4 injectors
when piston #1 is at TDC. so rather than change the engine design
(expensive modification), they just leave the 4 pulse CPS in there so
that in the future when they upgraded to a better ECU (cheap
modification) they could fire the injectors sequentially. the
particular engine that was in my truck was available from 1984 to 1995
(back to 1960's if you just count minor changes to the block), so i
wouldn't be surprised if the later models had a faster ECU that could
run the injectors sequentially. that would be a cheap way for them to
increase performance for the marketing people.
 
M

Mike Deskevich

Jan 1, 1970
0
Why is it 2 pulses per revolution though? One turn of the crank means
the 1st cylinder is at TDC again, for the second time, correct?

oh, one more thing, i missed in my previous reply. one turn of the
crank is piston from TDC back to TDC once. each cycle of the 4 cycles
in 180 degress of crank angle.

maybe this little ascii art of a 2 cylinder crankshaft will help.
piston #1 is as BDC and piston #2 is at TDC. rotation of 360 degrees
of the crankshaft will return them to the same location in space, but
only 1/2 through the 4 cycles.

X
|
|
|
X |
| -----
| | |
--- | ---- ------->to transmission
| | |
 
Top