temperature gauge escalates to max reading

[Jerry Whitt]

I have a 65 Corsa convertible with 4 carburetors and it runs very good. New problem showed up. While driving, the temperature gauge begin to rise in the normal way. After it hit about 200 degrees, the needle then shot over to the 600 degree range.

I checked the voltage from the alternator, thinking maybe the voltage regulator stuck. Showed about 14.2, 14.3. Seems ok.

Lifted the rear of the car and looked at the wire going to the thermister, thinking perhaps it grounded out. No. The wire is fine.

Disconnected the wire from the gauge in the engine compartment, found the voltage a little lower than the 14.2 but close. Checked the wiring diagram to see if there was supposed to be a resistor in the system. Nothing noted.

Pulled the thermister, held it over a flame with a special pliers, and measured resistance. When cool, open circuit. After warming, the resistance then varied. Checked another thermister here at my shop, similar results.

Any ideas?

Any help would really be appreciated. Thanks in advance

[bbodie52]

It would appear that your analysis and troubleshooting process seems to be logical. The only resistor (current limiting device between the gauge and ground) in the circuit IS the thermistor. When you disconnect the wire from the thermistor at the engine, does the gauge return to the left side of the display? Possibly a faulty gauge?

[66vairguy]

I've worked on the thermister/gauge system in the Corvair.

While I can't say what your issue is, I can say the gauges are very well built and typically are not an issue, but anything is possible.

The most likely culprit is the thermistor. A common mistake is to heat the thermistor and measure the resistance with a modern DVM that conducts very little current. When hooked to the gauge the thermistor conducts more current than a DVM provides and this changes how the thermistor works, it's called the "Self Heating Effect" and the engineers designed the system to deal with it. As thermistors age they can reach a point were they cascade above a certain current/temperature (current increases as the engine and thermistor heat up). This runaway cascade effect is what you MAY be observing and will not show up testing with a DVM.

Bottom line - Try another thermistor in the engine and see what happens. If a different thermistor works then let us know

[Jerry Whitt]

When the wire was disconnected from the thermister, the gauge swung back toward cool, but not all the way. Left the car overnight, and then with a cool motor, attached the wire to the thermister. The needle then immediately moved to the full cool end of the gauge.

Tried warming up the engine again, once again normal movement up to around 200 degrees, then pinned the 600 plus end of the scale.

Seems like this is an example of the "runaway cascade effect" mentioned by 66vair guy.

Will try to locate a new thermister. Thanks again for your help

[bbodie52]

I have a theory... I believe that the cylinder head temperature gauge in the Corvair functions very much like the fuel level gauge. In the fuel gauge, there are two electromagnetic coils. One coil is a constant pull, provided with power whenever the ignition key is ON. The other coil has more windings, so it is capable of a stronger magnetic pull when fully energized. This coil is wired to the fuel tank sending unit (variable resistor). With the cylinder head temperature gauge, the variable resistor is the thermistor. When the engine is cold, the thermistor provides maximum resistance, and as the engine heats the thermistor resistance decreases, causing current in the variable gauge coil to increase, which strengthens the magnetic field it generates. The more the engine heats, the more the thermistor resistance decreases — pulling the gauge needle to the right as the variable electromagnetic coil generates a stronger pull as the electric current through the thermistor increases.

The left-side electromagnetic coil in the cylinder head temperature gauge is supposed to be fully energized whenever the ignition key is turned on. 12 VDC is applied to the gauge via the 20 B/P (20 Gauge - Black with Pink Stripe) wire. This voltage is applied internally within the gauge to BOTH electromagnetic coils. The coil that pulls the gauge to the left is like a constant spring — current passes through the coil from the voltage source and then directly to GROUND, which places its magnetic field at maximum. Voltage is also applied to the other coil, but it does not have a direct connection to GROUND. Instead, the coil on the right is grounded via the 20 B/LG (20 Gauge - Black with Light Green Stripe) wire that is connected to the thermistor at the engine cylinder head. With a cold engine, the thermistor resistance is at maximum, which limits the current flow through the gauge coil on the right so that it has a weak pull. As the engine heats, the thermistor resistance is reduced, which gradually increases the current flow through the gauge coil on the right — moving the needle from the low end of the scale to toward the high end.

Now what if your gauge housing is poorly grounded? The fixed-pull coil on the left would have little or no current flow, so the "spring" pull to the left would also be weak or non-existent. You stated: "When the wire was disconnected from the thermistor, the gauge swung back toward cool, but not all the way." Without a magnetic coil pulling on the left, or with a weak pull (poor ground) the needle might "float" somewhere between the left and right side of the gauge. You also said: "...attached the wire to the thermistor. The needle then immediately moved to the full cool end of the gauge." It is possible that grounding the gauge via the thermistor provided a weak ground path for the coil on the left, so that it was energized and was able to pull the needle to the cool side of the gauge.

I am suggesting that your thermistor may be OK, but you may have a poor instrument panel ground. The second diagram below shows that the Corsa instrument panel (which is plastic) is grounded via a single metal strap attachment that is held by a screw to the metal dashboard frame (below the speedometer). If this attachment is loose or corroded your instrument panel may be poorly grounded. Even if the primary ground connection is OK, that same metal strap criss-crosses along the instrument panel, where it is attached with four screws to the tachometer, upper pod (temp gauge and clock), lower pod (fuel gauge), and speedometer mount.You should check the ground connection to the upper pod to ensure that the connection is not loose or corroded, which would make a poor ground to the temp gauge.

Left-click each image with your mouse to enlarge for better viewing.

You should test the gauge and sensor! To test the gauge i used a potentiometer from Radio Shack (0-1000 Ohm, 0.75 watt, 15-turn) and tested, and validated the temp gauge!



What follows is the Resistance to Temp curve. Set the pot at 450 ohms and the gauge should be at 300 Deg F, 250 ohms and the gauge should read 400 Deg F... Put the thermistor in the oven at 300 for 20 minutes - it should read 450 ohms, etc.. If all that is correct - what is left is wiring, connections and ground plain... Dash must have good ground to body, body to engine, engine and body to battery negative (and of course sensor to head!)... Put all of this together and it will work!



Good luck with your project!

[66vairguy]

Yes the gauge is easy to check, but note you need a good sized potentiometer (at least 2 Watt).

Remember the well circulated chart is NOT accurate except for DVM use, but close enough to test a gauge. The gauges are most accurate from 350F to 550F based on my experience. At least you can tell if the gauge is responding in a proper curve.

[66vairguy]

Bodie - yes the temperature gauge is magnetic and used the balance bridge concept of a fixed coil to compensate for voltage variations in a cars electrical system. But - it's not like the fuel gas that is linear, the temperature gauge responds on a resistance curve, so the coil setup is different.

You make a good point - always make sure connections are clean and there is a good ground with any gauge. A bench test will tell if the gauge is O.K. or not.

[Jerry Whitt]

Wow! Great information from MR. B.

Will be visiting local Radio Shack for a potentiometer. Will run the tests suggested, and verify good ground of gauges.

Thank you!!!!!!

[66vairguy]

While testing the gauge keep in mind the "chart" that has been widely published is inaccurate as it a static test of the thermistor NOT the gauge values. Not a big deal as you just want to test the gauge to see if it responds properly as the resistance drops. Also - calculate the wattage - you'll need a 2 Watt potentiometer, anything less will overheat and fail.

[BobWitt]

Wow - great post, Brad you amaze me - and it very cool to be quoted!

I did the above set-up to test the gauge on my Corsa - it was very correct.. The potentiometer that I used was from Radio Shack, before the local store went out of business - never got hot and I have used a number of times after.

I went searching through the boxes of test equipment - and being anal pays off - the RS part number follows...

I think Brad's ground theory makes a lot of sense - not sure why the ground would drop would be repeatable at 200, but... The Clock also shares this ground - if your clock works it would tend to stop shortly after the loss of ground. Would not help me - my clock does not work! (afterthought - would also affect the backlighting! do your lights go out?)

See the picture below for the part I used...

Good luck Mr. Whitt...

[BobWitt]

And yet another thought - if your backlighting and clock continue to work with the Gauge at 600 - it could be the gauge to housing loss of ground. The three studs are identified in the following photo.

With age and corrosion, if the ground stud from the gauge is NOT getting a good connection to the housing - this could be the problem. CAUTION! hold the stud as you loosen and tighten (maybe add a 'star washer') to assure the connection between the stud and housing. If the stud rotates it can break the winding wire that is soldered to the stud within the Gauge - then you have real problems. (don't ask me how I know!)...

Continued luck...

[Jerry Whitt]

Went to the local RadioShack today. Clerk never heard of a "potentiometer" and when trying to locate it in the computer really had a hard time trying to spell the word so the computer could locate it. With luck, the picture and part number supplied by Bob Witt, (distant relative???) helped him locate the part, although at a different store. Got it!

Will be playing with the new tool tomorrow and report results. Thanks for the help, everyone!!!!

[bbodie52]

A pathetic statement on the nature and current condition of Radio Shack!

Quite a few of these "rare" items listed on Amazon.com in a wide variety...

https://www.amazon.com/s?ie=UTF8&page=1 ... entiometer

[Jerry Whitt]

Struggling with this process. Some of the helpers suggest that a "DVM" wont work. I hope the translation is digital volt meter, multimeter

If this is correct, then what tool would I use?

Another note suggests setting the "pot" to a specific reading. Hope this is reference to the potentiometer.

When I used a digital multimeter and attached to the potentiometer, an ohm reading would pop up on the screen, and only last one or two seconds.
Tried turning the small screw, same result.

May be an operator problem. Called my son, an electronic engineer, he has worked on black boxes in airliners for Honeywell. He chuckled a little
and said he would give a call a little later.

As you can tell, frustration is creeping in. May take a break for a day.

Thanks for your patience and help

[BobWitt]

Jerry - you CAN use a DVOM on the Potentiometer (aka pot) - select the center leg and either outside leg. Turn the screw to dial into ~450 ohm.

Then connect these two legs to the gauge ground and sensor leads. Connect battery + to the pink (power) lead and then connect the ground. Gauge should move to 300 - disconnect power. Repeat at 250 ohm and 125 ohm respectively for 400 and 500 degrees.. To avoid overheating the pot, pull the power as soon as the gauge needle stops moving...

Good luck and report results!

[Jerry Whitt]

Had a much better day after taking a break. Yesterday, friend Charley Lee came over. Went through the various methods previously discussed.
We placed my thermister in the oven at 300 degrees for twenty minutes.
The result from heating showed the resistance at 282 ohms.
Heated the thermister to 400 degrees and measured 180 ohms.
Heated the thermister to 500 degrees and measured 112 ohms.
These tests did not match the printed curve, but did indicate the thermister is operational.

Removed the dash from the car. Visual inspections did not spot any problems. I did take off the ground strap to the gauges, cleaned all of the connections with a wire brush, then reassembled. Made sure all the screws were tight., Charley then soldered some wires on to the potentiometer. We attached the pot to the temperature gauge with alligator clips. We added a ground strap and a hot lead from the battery.

We then moved the screw in the pot and watched the gauge move. The gauge moved through the full range.
This test indicates the gauge is operational. Hmmmm!

Placed the gauge panel near the dash, hooked the big wire harness plug into the gauge panel. Added a ground strap to the cars metal dash
frame. Started the car. (I had previously installed the thermister.)

Watched the gauge. After a few minutes the gauge slowly began to move. Eventually moved to 300 degrees. This was at idle on a nice day,
outside temp about 80 degrees. The gauge system is now working. The only actual repair was that of cleaning the various connections and ground
straps and verifying clean connections. (also learned the DVM works better if I turn the switch to the right position)

Now have to reassemble the dash panel, get ready for a show this Saturday. Thank you all for the help!!!!!!!