ECT Sensor replacement

[jzbeaver]

After removing the air filter housing, I finally located the ECT sensor, behind the lower intake manifold.

The Chilton manual drawing shows it with the manifold removed- is that necessary,
or can it be replaced just with removal of the cowling?

1998 Windstar 3.8 engine
211,000 miles

[wiswind]

There are 2 engine coolant temperature sensors

1 sensor is for the temperature guage ONLY (green in my picture)
The other sensor is for the PCM (computer). (grey in my picture)

I have them labeled as they are marked in a diagram in the factory manual.
It seems to line up with the colors (green and grey) for the Airtex/wells parts in Rockauto.




I took this picture with the flex hose and air filter / MAF box removed from the vehicle.
You may be able to access them this way without removing the upper intake manifold.

[jzbeaver]

Update: Problem not solved. I had a local mechanic replace the sensor- as he was doing some other things for me. It's the grey one, for PCM, and I showed him Wiswind's photo, and I took the MAF box and filter off, so it could be seen. He had no trouble doing it, but did not partially drain the radiator first, and the Haynes manual indicates that's not necessary.

Before the replacement, I had been getting a PO117 code, "engine coolant temperature (ECT) sensor circuit low input", and I had assumed that the code was caused by a bad sensor. But after driving off with the new sensor, it came back within less than a mile, when before it might take 30 miles or more before appearing.

My first thought was that the new sensor might also be faulty, esp. since I bought it
at Auto Zone. But then I tested the removed one for changes in resistance. I tested it in water at 3 temperatures: room, hot water tap, and boiling water, and it tests OK, on all 3, based on the figures in the Chilton manual.

So now I'm thining that the problem may not be with the sensor, but somewhere else. I will test the voltage to the disconnected cable in the next day or so, but am also wondering if I should get a Ford sensor from Rock Auto, just to rule out that
factor. I don't think it would be the PCM itself-- how would only that small section of it be bad?

I thought that low input referred to the temperature, not to the supply voltage, but
maybe that's not the case. If the normal 5v power to the sensor- is low, what could be causing that? Might there be anything where the sensor gets screwed in that could prevent actual temperatures from registering accurately?

Any ideas/suggestions anyone has about this will be much appreciated.

1998 Windstar
3.8 Engine
211,000 miles.

[wiswind]

I think that the low input code is indicating that the input to the PCM FROM the sensor is too low.
The PCM is expecting a input between a minimum voltage and a maximum voltage.
If the input is lower than the minimum, it knows that there is a problem.
The input to the PCM will be controlled by the resistance of the temperature sensor, if the resistance is greater than the maximum value, then the voltage will be too low.

As you have a new sensor, I would suspect a wiring problem.
PCM is possible, but much less likely than a wiring issue.

I would do as you mention....verify that the terminals in the connector are good.
One wire is a Light Green with Red Stripe wire that comes from Pin 36 on the PCM......inside the PCM it there is a resistor that connects it to a reference voltage.
The Reference voltage is from INSIDE the PCM.

The other wire, Grey with Red Stripe took me a while to track down....goes to splice #123 (connects a number of wires together) which goes to splice #124 (which connects some more wires together).....from splice #124, there is a wire that goes to Pin 91 on the PCM....which goes through a resistor (Inside the PCM) to ground (inside the PCM).
These splices (123 & 124) seem to connect the "circuit common" for all, or nearly all, the sensors together to a ground....but isolated by the resistor from chassis ground.
This would be to reduce electrical noise in the sensor circuits.

I just looked at the electrical diagram, and the ECT is a 2 terminal device....the body of the sensor to the intake manifold physical connection is not a part of the electrical circuit.

Note that I am looking at the electrical diagram for my '96, your '98 will be slightly different, but VERY close.
A '99 and newer will have more differences as that was a major change year.

[olopezm]

This might be a long shot but how's your thermostat working? Have you noticed any overheating or the engine taking longer than usual to reach normal temp?

A bad t-stat could trip some other dtc's from time to time...

Oscar.

[jzbeaver]

Thanks again, Wiswind and thanks, olopezm for your info.

Re the suggestion that the thermostat may not be working properly, there hasn't been any overheating at all, or taking too long to warm up. If anything, its the reverse- - the guage indicates a very quick warming to normal operating levels.

But I'm speculating that this follows the logic of this sensor: If "low input" is a lower than "expected" minimum voltage signal to the PCM, that would result from a higher resistance. Since resistances are higher when the coolant is cold, I assume that the PCM would try to compensate-- I am guessing here- with faster idle speed and/or different fuel/air mixture or other adjustments to accelerate the warming. This would seem to be consistent with the trouble code.

I haven't had a chance to check the voltage yet, and I don't have much experience with wiring problems- haven't had any, other than finding bad grounds, so any suggestions in that area would be appreciated. Do you have to take the splices apart
to isolate the sensor wires?

Thanks

1998 Windstar 3.8
211,000 miles.

[wiswind]

You should be able to verify the voltages.....and continuity.......I would only open up a splice point if you have verified that there is a problem with it.

[jzbeaver]

I checked the voltage at the ECT connector, and it reads 4.7. Haynes says
it should be approximately 5v, so I think this qualifies. At this point I'm not sure
how to proceed to find the cause of the code, which appears regularly soon after I
erase it. There seems t be continuity, correct voltage, and a new sensor- so what else could be wrong, or am I missing something?

Some new information: it seems I was mistaken about the relationship between
resistances created by the sensor and the voltages produced. Instead of producing
lower voltages, as I had thought, , the higher resistances that take place at colder coolant temperatures produce higher voltages- inconsistent with the "low input" code.

I found this out from the the Ford Service Manual, P. 3-03-11 . This is a test for
the operation of the thermostat, and Ford makes a gizmo that you install between the connector and the ECT sensor, which you can then attach a VOM meter to, and
read the voltages as the coolant warms up. Examples: At 71F you are supposed to get 3v; .75v at 180F (the correct thermostat opening temp) and .4v at 221F.

The PCM also seems to be getting inaccurate voltages from the get-go, at low temperatures, so low that the radiator fans, which are controlled by the PCM, are on as soon as the engine starts, when the coolant is cold, and the reisistances and voltages are supposed to be high. So, if an initial voltage reading is below .75, corresponding to 180F, the PCM may think that the temperature is above 180- and trigger the fans.

When I get the code, I also get a freeze-frame reading on my scanner which shows it came on when the temperature reached 340F- which would be in response to a really tiny voltage, and consistent with "low input." But I don't believe that it ever gets that hot, or anything near it. The gauge doesn't climb above the midpoint, and if it was that hot, I would expect to find a noticeable increase in the level in the overflow reservoir, and that's not happening. So I think the thermostat is operating properly after all, it's not controlled by the PCM, only by the actual temperature.
Last week I was driving in stop and go traffic for 2 hours when the outside temperature was over 100F, and no problems (except for the CEL)

All a bit puzzling. Any ideas?

Thanks

1998 Windstar 3.8l
211,000 miles

[wiswind]

I don't know what to offer.
However, 195 degrees F is the OE temperature for the 1998 Windstar with the 3.8L motor.
180 degree is a "alternate temperature" thermostat.
I would prefer to use the temperature that was installed at the factory.

[jzbeaver]

Wiswind- This came from the Ford Service Manual for the 1998 Windstar; not clear if it was for a 3.0 or 3.8 engine, but the text reads: "if the water thermostate opening voltage is LESS than .75 volts (greater than 82C 180F),water thermostat is good.." This would also cover a 195F thermostat as well. I didn't mean to focus on whether the thermostat is the correct one or not- just wanted readers to see why the .75v figure for the ECT was used. I don't know which is installed, and without the Ford tool, I can't tell, but it has never been changed, and it seems to be operating OK. But this doesn't help solve the ECT problem.

[olopezm]

You don't need the ford tool, you can back probe the connector with a pair of paper clips and take the readings that way with your DMM.

I'll check the diagrams to see if there's anything that might be giving you problems and will post back ASAP.

Oscar.

[jzbeaver]

olopezm--
Thanks for this. Not clear how you could get paper clips into the female connector end, and then
push it onto the ECT terminals, without possibly damaging the connector (#144) Even if I could do so, it's besides the point- whether it opens at 180 or 195 wouldn't cause the ECT code. If it's not opening at all, the engine would be overheating (and that might produce the low input code) but the gauge would indicate that, and it's not overheating. By getting a good voltage reading at the female, I have already established that there is continuity on both circuits from the PCM to that point.

[wiswind]

Correct, the thermostat, 180 or 195, is not the ECT issue.
However, the ECT will read a voltage based upon the temperature that it is at.....regardless of the thermostat.
There is a scale......resistance to temperature for the resistance of the ECT.
A reading of insufficient sensor input means that the input to the PCM is below any value that would make any sense for any temperature that makes sense.
Even if you had no thermostat, it would not cause this code.

Yes.....340 degrees F would be very bad......and you did not have that.....but it would go along with "low input from ECT"......as the voltage was lower than what makes sense.

It is possible that you have a defective PCM....but I would like for you to diagnose or have diagnosis done before springing for a new PCM.
The PCM does not fail very often....(sensors or connections are most common)....but they can fail.

[jzbeaver]

wiswind- I'm also reluctant to mess with the PCM; I've never touched it, and never had any problems with it. Perhaps one of the internal resistors in the ECT circuit is bad.
Thought of another, perhaps remote, cause: What if, when the temperature- water, and air and metal under the hood-- gets really hot continuity is interrupted in one of the wires or the other connector (not found on your '96) somewhere between the PCM and the sensor- resulting, instead of low voltage, no voltage at all-- which might get translated by the PCM as low input, taking place at 340F? I was not getting this code in the winter! It may not be possible to troubleshoot this when everything is so hot. I'm also tempted to spring for 80 bucks for the PCM Manual from Helm, and see what it says about this problem. Thanks again for continuing to
follow up on this thread with me.

[jzbeaver]

Clues: Still haven't solved the problem, but have seen a pattern with the trouble code:
The pending code appaers (on my code reader) after driving for a while, but the CEL
comes on-- almost always-- after stopping and parking, for a while. The freeze
memory function shows it comes on at the exact same temperature: 338F, and the same rpm or close to it-- idle at 668 rpm. Any ideas, anyone?

1998 Windstar 3.8 engine
211,000 miles