GM3.1L no spark

In article >,
Lawrence Glickman > wrote:


> If he can buy a new one at a dealer and install it himself...that is
> the magic question. He needs that 5 volt output, or darn close to it.
>
> I would call anything less than 3.8 or 4 volts not good enough.
>
> Lg

On the other hand, the OP hasn't yet mentioned what model year
this 3.1 is, IIRC there are a couple of versions of crank sensor
depending on when it was built...

On Sat, 01 Jan 2005 17:50:32 GMT, Neil Nelson >
wrote:

>In article >,
> Lawrence Glickman > wrote:
>
>
>> If he can buy a new one at a dealer and install it himself...that is
>> the magic question. He needs that 5 volt output, or darn close to it.
>>
>> I would call anything less than 3.8 or 4 volts not good enough.
>>
>> Lg
>
>On the other hand, the OP hasn't yet mentioned what model year
>this 3.1 is, IIRC there are a couple of versions of crank sensor
>depending on when it was built...

This is True.
There are many *generations* of logic, too many to mention, that have
differing requirements regarding VCC and input signals.

That is -a- problem trying to diagnose things via long distance, where
you can't be there to see things for yourself. It is a universal
problem. I mean, the OP would best be served by going to a dealership
or somebody and asking a tech "What is the output voltage for this
sensor _supposed_ to be."

That is maybe how I would approach it. But if it is an older vehicle,
or the dealership isn't going to make any money off you, they might
not bother to go to the trouble to look up the specs.

OTOH, if you tell them you need a new part at the parts counter,
-then- you can ask the question and expect a proper answer. At least
this is my experience. When you are flashing a *wad* of money in the
air at the parts counter, looking eager to hand it over, THEN they go
ask the tech for you

Lg

This is the 1990 version of the 3.1. I read the output pulses of the
sensor on the DC scale of my Fluke 79 to see if they were consistant
(the 79 has the bar graph on the display and reacts very quickly to the
input).

Now that I think of it....when troubleshooting another problem a few
months back on this vehicle I read the output of the crank sensor and
it was the same as it is now. All was well until about a week ago when
it just refused to even fire one day after running great the night
before.

As far as the sensor needing to put out a logic level sine wave, I
doubt this was the plan. They are probably amplifying the wave at the
module then diode clipping it to get a useable digital pulse.

Travis






Neil Nelson wrote:
> On the other hand, the OP hasn't yet mentioned what model year
> this 3.1 is, IIRC there are a couple of versions of crank sensor
> depending on when it was built...

On 1 Jan 2005 11:50:36 -0800, "boardjunkie" >
wrote:

>This is the 1990 version of the 3.1. I read the output pulses of the
>sensor on the DC scale of my Fluke 79 to see if they were consistant
>(the 79 has the bar graph on the display and reacts very quickly to the
>input).
>
>Now that I think of it....when troubleshooting another problem a few
>months back on this vehicle I read the output of the crank sensor and
>it was the same as it is now. All was well until about a week ago when
>it just refused to even fire one day after running great the night
>before.
>
>As far as the sensor needing to put out a logic level sine wave, I
>doubt this was the plan. They are probably amplifying the wave at the
>module then diode clipping it to get a useable digital pulse.
>
>Travis

That all makes *sense,* but we can come up with 1,000 scenarios that
make sense, that have nothing to do with the solution to the problem
at hand.

This is why I like to build a *library* of engine parameters on my
known good engine, so I can see if something is *drifting out of
tolerance.*

I have the Motorcraft.com data sheets on my engine parameters, but I
much rather have the real, than theoretical values they give. IOW, I
measure what things are when all is running correctly, and use that as
a baseline to which to compare future readings.

I am able to do that with my Carchip e/x, which reads engine
parameters under -dynamic- conditions, rather than at idle in a shop
bay. I build the library, and save it to a database, and in the
future, I have -real- data to refer to, rather that the hypotheical
*stuff* motorcraft offers up at www.motorcraft.com, although I
wouldn't call their publications useless.

My car has it's own parameters/tolerances, those are what I want to
know about. At 5 second monitoring intervals, I can get 25 hours of
driving time on to a chip before I have to download it.

But some parameters, like you crankshaft sensor voltage, that would
have to be measured with my Fluke 97 50 MHz Scopemeter. I don't envy
the day I may have to do it, then, otoh, I may never have to do it, as
my powertrain is warranted for the life of the vehicle as long as I am
the original owner. At least, that's whay they -say-.

I even have it in writing, but their condition is to have ALL routine
service done at their shop, at 10x what I can do it for myself, so,
there is always the *catch 22.* They'll screw you one way or the
other, so it is best to learn how to do some of these things for
yourself.

Lg

Here's some answers given you still haven't told the group what model
you're working on -- info will be general to the 90 3.1 V-6.

Can a crank sensor damage an ICM?
- If you throw it hard enough. Not a problem because yours is
functioning. 0.7 volts as measured with a RMS DVOM is good at cranking
speed. You've done the correct step to verify good CKP input to the
ICM.

Do I have a decent input from the ICM to the ECM?
-Check the Dref high wire (Purple / White) with positive lead and Dref
low (Black / Red) with negative lead AT THE ECM. Your RMS Fluke should
show maybe about 3 volts (it's a 5 volt square wave). Check signal
frequency in Hz if you can for cranking speed reading. For your engine
---> (Hz * 20) = RPM. So perhaps about 15 to 25 Hz would be right.

Do I have a spark command from the ECM to the ICM?
-Check the EST wire (White) at the ICM for voltage. This is trickier.
Under aprox. 400 RPMs it should be a 0.5 volt square wave. Above 400
RPMs it should be a 5 volt square wave. You can check Hz on this wire
to verify a pulsing signal. (Hz * 20) = RPM.

Whats the deal with 400 RPMs?
-Thats the point that the ECM tells the ICM that its taking over spark
timing control. It does this with the Bypass Wire (Tan / Black). This
wire should read aprox. 0 volts under 400 RPMs and aprox. 5 volts above
400 RPMs. If it stays at 0 volts at high cranking speeds (with battery
charger assist) suspect a grounded wire or bad ECM. Of course this is
an OBD I GM, so suspect a bad ECM. But test the wire anyway for
educational purposes.

Essentially, a good EST signal at the ICM would suggest a bad ICM.
Check for B+ to your coils (Pink / Black) and ground to both modules
(Black / White) if all else fails. Do this while cranking to load the
circuits.

Toyota MDT in MO

Excellent info Comboverfish! That was exactly the response I was
looking for. Now I know exactly what to look for and can isolate what
is not operating as intended.

Sure would be nice to have a scopemeter to check out how all the
control pulses look. I have several bench scopes but don't really want
to haul one outside to putz with this thing. The meter should let me
know what's going on good enough.

Thanks again. I'll post my findings when I get it all checked out.

Travis





Comboverfish wrote:
> Here's some answers given you still haven't told the group what model
> you're working on -- info will be general to the 90 3.1 V-6.
>
> Can a crank sensor damage an ICM?
> - If you throw it hard enough. Not a problem because yours is
> functioning. 0.7 volts as measured with a RMS DVOM is good at
cranking
> speed. You've done the correct step to verify good CKP input to the
> ICM.
>
> Do I have a decent input from the ICM to the ECM?
> -Check the Dref high wire (Purple / White) with positive lead and
Dref
> low (Black / Red) with negative lead AT THE ECM. Your RMS Fluke
should
> show maybe about 3 volts (it's a 5 volt square wave). Check signal
> frequency in Hz if you can for cranking speed reading. For your
engine
> ---> (Hz * 20) = RPM. So perhaps about 15 to 25 Hz would be right.
>
> Do I have a spark command from the ECM to the ICM?
> -Check the EST wire (White) at the ICM for voltage. This is
trickier.
> Under aprox. 400 RPMs it should be a 0.5 volt square wave. Above 400
> RPMs it should be a 5 volt square wave. You can check Hz on this
wire
> to verify a pulsing signal. (Hz * 20) = RPM.
>
> Whats the deal with 400 RPMs?
> -Thats the point that the ECM tells the ICM that its taking over
spark
> timing control. It does this with the Bypass Wire (Tan / Black).
This
> wire should read aprox. 0 volts under 400 RPMs and aprox. 5 volts
above
> 400 RPMs. If it stays at 0 volts at high cranking speeds (with
battery
> charger assist) suspect a grounded wire or bad ECM. Of course this
is
> an OBD I GM, so suspect a bad ECM. But test the wire anyway for
> educational purposes.
>
> Essentially, a good EST signal at the ICM would suggest a bad ICM.
> Check for B+ to your coils (Pink / Black) and ground to both modules
> (Black / White) if all else fails. Do this while cranking to load
the
> circuits.
>
> Toyota MDT in MO

Cool! I'm layed up here at home as sick as a dog so all I can do is
type right now. Wish I could be doin' the testin' instead! <g>

Its good that you have the scope as a backup -- but if you can get RMS
voltage and a reliable Hz reading out of your Fluke then thats 50% of
the info you would glean from a waveform anyway.
Have fun,

Toyota MDT in MO

On 1 Jan 2005 15:06:01 -0800, "Comboverfish" >
wrote:

>Cool! I'm layed up here at home as sick as a dog

I've had bronchitis going on for about 3 weeks now.
Nice to have your company. Maybe it will turn into pneumonia?

> so all I can do is
>type right now. Wish I could be doin' the testin' instead! <g>
>
>Its good that you have the scope as a backup -- but if you can get RMS
>voltage and a reliable Hz reading out of your Fluke then thats 50% of
>the info you would glean from a waveform anyway.
>Have fun,
>
>Toyota MDT in MO

Ok, just got done with testing....here's what I found.

Reference pulse was not there, same for the EST control signal.
However, when I looked for the EST bybass, it was stuck at battery
voltage. This leads me to believe that the logic IC that is responsible
for putting out that signal has shorted to the supply.

Now I ask this....is that output that should be at logic level likely
to have damaged the ignition module's input? Not knowing if the input
is TTL or CMOS makes me wonder. Perhaps if it is cmos powered from 12v
it may have survived. If it is TTL it is surely damaged.

So I'm off to find a replacement ECM tomorrow and I just may get a
replacement module just for good measure.

Travis






Comboverfish wrote:
> Cool! I'm layed up here at home as sick as a dog so all I can do is
> type right now. Wish I could be doin' the testin' instead! <g>
>
> Its good that you have the scope as a backup -- but if you can get
RMS
> voltage and a reliable Hz reading out of your Fluke then thats 50% of
> the info you would glean from a waveform anyway.
> Have fun,
>
> Toyota MDT in MO

boardjunkie wrote:
>
> Ok, just got done with testing....here's what I found.
>
> Reference pulse was not there, same for the EST control signal.
> However, when I looked for the EST bybass, it was stuck at battery
> voltage. This leads me to believe that the logic IC that is responsible
> for putting out that signal has shorted to the supply.
>
> Now I ask this....is that output that should be at logic level likely
> to have damaged the ignition module's input? Not knowing if the input
> is TTL or CMOS makes me wonder. Perhaps if it is cmos powered from 12v
> it may have survived. If it is TTL it is surely damaged.
>
> So I'm off to find a replacement ECM tomorrow and I just may get a
> replacement module just for good measure.
>
> Travis

I would suspect that the input is TTL feeding a CMOS output - but I
don't know for sure. The last one that burned out on my own car I
opened up and looked at the parts. The only things that appeared to
be CMOS were the output transistors/ic's.
Also, I would think that it's possible that a constant high burned
out it's brain.