[bmorgoch]

Last August, my wife used the rear end of a SUV to stop the Astro when the ABS unit decided to permanently open the dump valve. The insurance company of course wanted to write it off, but I love this beast so I decided to fix it myself. I acquired an 89 for $150 that was headed to the scrap yard and got all the parts I needed. MY rad was of course toast, but even though the condenser was bent 45 degrees across the middle, I salvage all my R12 refrigerant! I cut off the bent nose and took the beast to the body shop to have the replacement installed.

When it returned, I decided now would be a good time to replace the water pump as it was still the original (365,000 kms). I installed a premium brand new (not rebuilt) unit (lifetime warranty). As it was early December now, I decided to do the AC in the spring, "just in case" I ran into some problems. At least that was a good decision.

When I fired the beast up and took it for a spin, the temperature gauge (digital dash) went all the way to the top (260 F or 125 C). It would then drop to 230 then back to 245 and toggle between these 2 numbers. Output from my heater was normal, about 115 F measured with my AC temp probe. I have done the following without changing the symptom:

Verified that the water pump is a reverse rotation design (removal of the rear plate to look at the impeller) and that the serpentine belt is routed correctly (diagram on the power steering reservoir) The pump is turning counterclockwise as viewed from the front. The guys at Parts Source then convinced me to replace it as a warranty issue just in case it wasn't circulating the antifreeze properly. Same result.

The thermostat has been replaced 3 times, 2 with a 195 (a Stant and a clone) and the last was a 180 (Stant). All of them were tested prior to installation and after removal in a pot of antifreeze on my stove while I stirred the liquid with my wife's candy thermometer. They all work as advertised. Same result.

The fan clutch works, but I replaced it with the unit from my donor vehicle. Same results.

The rad shop verified good flow through the donor vehicle rad and could see no issues. We did a rad flush anyways. No change in flow characteristics.

I removed the t stat and went for a test run. The temp gauge got to 1 notch below normal (middle of the gauge 95 C or 200 F). Air from the heater was cold.

All these test were conducted when the outside temp was 0 C or 32 F and colder (its winter here!)

I currently have the 180 t stat installed (don't want to tell you how many gaskets I've used)

Using a digital temp probe attached to a multimeter (calibration verified with snow and boiling water), I read about 150 F on the aluminum intake housing next to the t stat. Its difficult to get good thermal contact but I scraped the area and used Wakefield Thermal Joint Compound for this test (white stuff used to mount transistors to heat sinks)

The hose from the t stat to the rad feels hot to the touch, but does not burn you. The rad feels warm on the drivers side but stone cold on the passenger side.

The past couple of days have been pretty mild, mid 40s and the temp gauge still reads the same. I am interpreting this as the t stat working properly.

The Stealer tells me that there are 2 temp sensors, 1 for the ECM and 1 for the gauge. The ECM temp sensor has 2 leads attached to it . It is located by the t stat on the intake manifold. As the engine seems to be running fine and has not "boiled over". I suspect it is fine. The sensor for the temp gauge has a single lead. He could not tell me where it was located. It is a non stocked, special order item.

I am now 99.9% sure I've been chasing a faulty temp sensor (what can I say, I've earned every gray hair I have)

Does anyone know where this temp sensor lives. The Stealer believes it is in one of the heads someplace. If so, any ideas?

This is a 4.3 TBI engine, nothing fancy. I am the second owner of this vehicle (original, a retired Porsche mechanic, was the father of one of my best friends). Other than the ABS failure, its been a dependable work horse and I've done my best to look after it (usual stuff brakes, alternator, etc).

I would appreciate the following:

- location of the temp gauge sensor
- expected voltage reading at normal operating temp
- any info from anyone else having a similar experience

Thanks for reading my rant, but I want to resolve this so that I can drive the beast again with my eyes on the road instead of the temp gauge!

Bob :canada:

 

[gtkane]

Bob, let me tell you that you have done an excellent job of tracing this problem. Most people would not have gone to the lengths that you have to sort this problem out. Major props to you!

The sensor for the guage is a single wire sensor, located on the cylinder head, just above the exhaust port line. It is a brass sensor usually mounted on the drivers side, but may also be mounted on the passenger side head.
This sensor is what is called a thermistor, in that when the temperature goes up, the resistance goes down. I don't have the numbers right now, but think in the lines of 100k ohms for ice cold, and 5 ohms for meltdown.(I hope I got that one right!)
Disconnecting the wire should move the needle to cold, and shorting to ground will show hot.(unless I am "bass akwards")
If you do have to replace it, do not put thread sealer or teflon tape on the threads. It is made of brass, and will conform to the threads to make a leak-proof seal.
The only other thing I can think of is to check the wire going to it carefully-they have been known to corrode/burn, thus giving faulty readings.

I hope this helps.

 

[deeno]

your'e looking for the coolant temperature sender. It's located on the drivers side of engine, in line with the spark plugs and exhaust manifold, in between the middle and rear cylinder.

 

[bmorgoch]

Thanks guys for your replies. Sorry for the inactivity, but I've been pretty much unconscious due to a severe head cold for the past while. Back on my feet now.

Deeno, you were exactly correct. The sensor was on the driver's side, between the 2nd and third cylinder between the spark plugs, attached to a dark green wire (as best I could see amongst the engine grunge). Kind of dumb to run a low voltage, high impedance digital input circuit directly against the spark plug wires, but hey thats just one person's opinion.

Gtkane, thanks for your info as well. Here is what I have been able to determine:

- connection at the sensor was good, no corrosion, etc.
- the sensor is indeed a thermistor (resistor that's value changes dependent on the temperature)
- it has a negative temperature co-efficient, i.e. the resistance goes down as the temperature rises.
- at about 32 F ( 0 C ) the resistance is about 3,750 ohms with the wire disconnected and measured with a digital volt ohm meter
- with the wire disconnected, ignition on, but engine not running, I read 12.14 volts on the wire and the temp gauge reads nothing (i.e. coldest value)
- with the wire grounded, the gauge reads its hottest value. NOTE: the gauge took a full minute to climb to the highest value, indicating that it has a built in delay circuit so that it reads the LONG TERM AVERAGE TEMPERATURE, not the instantaneous temperature. I found this interesting/surprising

With the wire reconnected to the sensor via alligator clips, I attached my volt meter and made the following measurements:

- initial voltage reading of 5.75 (engine at ambient outside temp of about 32 F (0 C) I then started the engine
- there are a total of 12 "bars" on the digital gauge.
- first bar illuminated at 4.6 volts
- both second and third came on at 3.4 volts
- both fourth and fifth came on at 2.35 volts
- both sixth and seventh came on at 1.73 volts

- eight came on at 1.26 volts THIS IS THE MIDSCALE POINT AND WHAT SHOULD BE THE NORMAL READING

- ninth came on at .98 volts
- tenth came on at .88 volts
- eleventh came on at .79 volts THE OVERTEMP INDICATOR STARTS BLINKING

The engine did not get hot enough to cause the 12th segment to illuminate.

From this point on, I saw the voltage vary from as low as .87 to as high as 1.30. My guess is I was watching the thermostat introduce cool antifreeze into the system as required. The temperature gauge did not actively follow the voltage swings in real time, but seemed to average things out over the long term.

Is there any chance I could get someone out there to make 2 measurements for me:

1) disconnect the wire from the sensor and measure the voltage to ground with the ignition on but the engine not running

2) with the wire connected to the sensor,the engine running and up to normal operating temperature, the voltage at the sensor

Does the sensor actually go into the water jacket, i.e. do I have to drain the coolant before I remove it (I suspect so, but you never know)?

I am still trying to determine what has changed since the accident. As the sensor seems to be functioning, I am somewhat hesitant to replace it "just for the hell of it" (Stealer wants $65.00 plus $9.10 taxes). I have worked with several different thermistors in the past and have never seen one change its characteristics. When they fail, they go open circuit.

The only other thought I have is something I do not want to face. As they were welding on the new rad cradle onto the vehicle, I'm wondering if this caused some damage to the digital gauges, specifically, is there a voltage regulator in there that has been compromised? I am hoping the first reading will answer that question. I measure 12.14 volts on my vehicle which is essentially the battery voltage, not a regulated value.

My oil pressure and volt meter seem to be working normally, but my gas gauge seems to be a little "off". This is only a subjective feeling at this point. During the winter, I do not like to let the tank get too empty so as to avoid any condensation issues.

Thanks for all your efforts so far!

I will continue to update as things develop.

Bob :canada:

 

[deeno]

If it's just your coolant temp gauge you suspect, you can eliminate that gauge and install an aftermarket temp gauge [iequus,auto meter,etc.]. I recommend iequus gauges. they're pretty cheap and very accurate. You can drain some coolant or just remove the sender- some coolant will pour out, but you wont empty the entire system. place a catch pan underneath. The gauge kit comes with threaded adapters. Don't wrap the sender with teflon tape, silicone, antiseize compound, etc. Just install it bare, metal to metal. You'll need to mount the gauge backing plate somewhere. Under the edge of the plastic lower dash panel near the engine cover is a good spot. It's close to the sender [less wire routing] and it doesn't disturb the look of the stock dash. Next, drill a small hole in the floor below the engine cover flange and pop in the rubber grommet supplied in the kit. Connect the signal wire from gauge through the grommet to the sender. Connect the power wire from the gauge to the fuse block [ignition] slot, crimp on blade connector works well. Splice all ground wires from gauge and ground the wire to a bolt connected to the chassis. Connect the gauge lighting curcuit to the parking light switch in dash. Use a multimeter or test light to backprobe the wires behind the switch. Test each wire by switching on/off the parking light. When you find power in the on position, tap into that wire. Now your gauge will light up with your instrument cluster. For less than the price of stock sender, you'll have way more accuracy and it looks trick too!

 

[LSkrabut]

Ever find out what your issue was?

I had a very simular issue with my digital dash/1992 GT Safari. After about 8 month of every type of troubleshooting I could figure out. Going to 4 shops, nothing would seem to fix my issue, they are were either saying, Live with it or get a new dash. Well, I stumble on the fix for my issue, loose battery ground cable, located under the battery. Clean it up, tighten it down, and I found out it fixed my water temp display issue. Not saying your battery ground is loose or bad, but maybe only ground is?

So good luck figuring out your incorrect displaying of water temp.