Coolant temperature gauge mystery

[gw33gp]

Last weekend I did some work on my Ranger that I had been wanting to do. One thing was installing an aluminum thermostat housing. I had a couple typical issues with this install. One was the holes for the sensor and sending units were larger than stock and I had to find larger O-rings. I could not find any that looked suitable so I took ones that were too thick and sanded them down in a lathe to a size that worked.

The biggest problem was electrical. Initially the temperature gauge would not work, it just stayed on C when running the engine. With some investigation, I found the connector was bad. Apparently, it got damaged during the install. I ordered a new one and installed it. I tried to connect the wires the same way they were, but I was not that careful since I read somewhere the sender was just a variable resistor and it would make no difference which way it was wired.

Once installed, I started the engine and waited for the gauge to come up. It did and I took it for a drive. The needle got up to the normal position at just below half way. After about 10 minutes, the gauge suddenly went to H. I began pulling over and it went back down to normal. It continued this up and down maneuver and I eventually got off the freeway. When I checked the engine, I could tell it was not over heating and thought it was just a bubble in the cooling system from refilling. I continued on and eventually added more coolant, but by then the gauge was way past H at full lock and would not come back down.

When I disconnected the connector from the sending unit, the gauge would go back to below C with the ignition on. Connecting would send the gauge back to above H. I was baffled, because if there was a short in the wiring it should have stayed above H when disconnected. I got out my multimeter and started investigating. I could tell the sensor was still good but I had no values to know what is good on all the other readings I got. I began to suspect the new connector and cut the wires to check. It was good too. I tried changing polarity on reading the resistance on the wires that go into the loom. I found slightly different readings and decided to try connecting the wires to the sensor in reverse of what I did originally.

Reversing the wires resulted in the gauge going to C with the ignition on and the gauge continued to operate normally on a long drive. So, apparently it does make a difference as to how they are connected.

The mystery is, why did the gauge work correctly for a short time with the wires backward and then suddenly go bad?

 

[RonD]

Its actually a 12v Sender for the dash gauge
In any case no it wouldn't matter which wire was which with the PLASTIC housing
But the metal housing changed things, lol

Originally these 12v senders were grounded by their threads when screwed into a metal engine part, so just had 1 wire connected that ran to the gauge
So the sender is a variable resistor that measures resistance to engine/vehicle ground

As more and more plastic parts were being used the senders got 2 wires, a ground wire and a Gauge wire
Red/white gauge wire
Black/white ground wire usually

I would expect that one terminal of the two on the sender would read 0 OHMS to the metal part of sender, that would be the ground terminal, black/white wire

As to why gauge acted that way, well the path to ground may not have been direct with o-ring and aluminum
I would expect H almost instantly when wires were connected, if the ground was good, but it may not have been right off the bat, but a little warm up could have made it a better ground
Just guessing

I have never tested if one terminal is connected internally to the metal part of the sender

 

[gw33gp]

Well it looks like the sender does not have a great ground to metal on one terminal, but close. One terminal measured 2 - 100 ohms and the other terminal 2.67 K ohms when measured to ground on the engine with the sensor mounted in the installed thermostat housing. The sensor measured 2.63 K ohms across the two terminals.

Maybe the coolant is the grounding path. Do you think coolant could have a resistance of 40 ohms? Maybe I should measure it. I am sure the distance apart on the measuring electrodes would make it vary considerably.

 

[RonD]

So yes, I would say it could be hooked up backwards

 

[Eddo Rogue]

I had a similar problem. Changed the sending unit and sensor, twice (the second time with motorcraft units). Also burped the coolant running it parked uphill for a good long time. Still had some air in it, so I installed a bleeder valve at the heater hose line. problem solved.

 

[gw33gp]

I got rid of the air in the cooling system, but it took reversing the wires to the connector to get the gauge to work properly.

I think I have figured out the mystery of why it worked OK for a while after installing the new connector. The O-ring keeps the sensor in the center and there is no metal to metal contact in the hole. However, there is a clip that keeps the sensor from popping out when the coolant is under pressure. I may have gotten lucky (or unlucky) and got the position of the clip in such a way that it was in contact with the sensor but not in contact with the thermostat housing. When I was driving, the pressure buildup in the cooling system probably moved the sensor up slowly until the clip hit the top of the slot and made an electrical ground path to the block. With the wire connected wrong, it sent low ohm resistance to the gauge which the gauge shows as overheating. It eventually made enough contact that it got permanent ground contact and the gauge was stuck on hot as long as the connector was connected and the ignition was on..

I think the aluminum thermostat housing is a good upgrade but care must be taken to make sure the wiring polarity is correct. Since the plastic thermostat housing can't create a ground path, the factory wiring may be not right for the aluminum housing. I know some people have no problems with the change and others like me do.

 

[Uncle Gump]

Your assessment of what happened certainly sounds plausible.

Glad you got it figured out.