Coolant Temp Gauge [More than you ever really want to know]
abnormal
New Wagonist
For a couple years now, my temp gauge has never really worked properly. I had problems where the needle would quickly jump to hot. This symptom sounds like a short to ground on the temp sender, but I checked the wiring from the gauge to the sender and it all checked out. Even if this wire did short to ground the gauge would smoothly go to hot, but instead it bounced to hot and back down. This would happen randomly while driving the car.
Watching the gauge spazz out all the time quickly got old, so I got a new temperature and fuel gauge cluster and this seemed to have fixed the problem. All seemed well for a while, but then I see the needle jump to hot. It only did this once in about a year, so I just ignored the one incidence.
About a month ago, I noticed after driving for a long time the needle would move up to a little over the half way mark, and then slowly it goes back down to normal. I started with the simplest solution I could think of: getting a new radiator cap. This seemed to work for a week, but then the same symptom reappeared. The car didn't act like it was overheating (fans are on and no coolant boiling into overflow).
I decide to flush the cooling system and check the thermostat. This proved to be a waste of time. All the cooling passages were clean and the thermostat was operating properly(tested in pot of water). However, upon refilling and bleeding the system, I noticed the fans would not turn off after they turned on. They would only turn off after shutting the engine off for a while.
To fix the fan problem, I picked up a new fan switch (the one that screws into the back of the engine block). When I bled the cooling system, the fans didn't come on and the temp gauge read about 3/4. I had also purchased an infrared temp gun. The temp gun reading at the t-stat housing contradicted what the gauge in the car read. It was only about 87 deg C. I decided to run the car longer and finally the fans turned on and shut off like they are supposed to. With the cooling system working properly I turned my attention to the pesky gauge.
First, I tested the sender unit. I connected a DMM to the sender in order to read the resistance. 142Ω is cold, 49-32Ω is normal(1/3), and anything lower is overheating. After warming the car up, the lowest reading I get is about 47. If anything the gauge should be reading on the low side of normal; not overheating.
I still had the old gauge cluster sitting on the workbench, so I took it apart and draw a wiring diagram:
I figured out the reason it would spazz out. It is the result of a failing zener diode. I went up to the electronics store and got a new 3.0V 1W zener diode. I tested the gauge and it worked flawlessly on the workbench using the 12 V line out of an old computer power supply.
At this point, I figure the new gauge maybe blew the same part, so I swap them out. After testing the repaired gauge in the car, the gauge still doesn't read correctly (3/4 up when fans come on).
Frustration is setting in at this point. The only thing that is different on the car to the workbench test is the sender. For the test, I used some resistors to simulate the sender unit. I get out my DMM and set it to measure current. Maybe the sender fails under load. I measure the current on the test setup ad it is about 137 mA at normal operating temperature. Then I go out to the car, warm it up, and measure the current. Seemed normal about 145 mA (larger due to 13 V source compared to 12 V test source). Then when the needle started heading north, the current jumped up pretty high about 180 mA. This suggested to me that the sender starts to fail after a certain threshold current. The next day I install a new sender and my problems are over. It finally works perfectly and hopefully it stays this way.
I'm not sure if anyone has had this many problems with their temperature gauge circuit, but hopefully, this will help someone.
Watching the gauge spazz out all the time quickly got old, so I got a new temperature and fuel gauge cluster and this seemed to have fixed the problem. All seemed well for a while, but then I see the needle jump to hot. It only did this once in about a year, so I just ignored the one incidence.
About a month ago, I noticed after driving for a long time the needle would move up to a little over the half way mark, and then slowly it goes back down to normal. I started with the simplest solution I could think of: getting a new radiator cap. This seemed to work for a week, but then the same symptom reappeared. The car didn't act like it was overheating (fans are on and no coolant boiling into overflow).
I decide to flush the cooling system and check the thermostat. This proved to be a waste of time. All the cooling passages were clean and the thermostat was operating properly(tested in pot of water). However, upon refilling and bleeding the system, I noticed the fans would not turn off after they turned on. They would only turn off after shutting the engine off for a while.
To fix the fan problem, I picked up a new fan switch (the one that screws into the back of the engine block). When I bled the cooling system, the fans didn't come on and the temp gauge read about 3/4. I had also purchased an infrared temp gun. The temp gun reading at the t-stat housing contradicted what the gauge in the car read. It was only about 87 deg C. I decided to run the car longer and finally the fans turned on and shut off like they are supposed to. With the cooling system working properly I turned my attention to the pesky gauge.
First, I tested the sender unit. I connected a DMM to the sender in order to read the resistance. 142Ω is cold, 49-32Ω is normal(1/3), and anything lower is overheating. After warming the car up, the lowest reading I get is about 47. If anything the gauge should be reading on the low side of normal; not overheating.
I still had the old gauge cluster sitting on the workbench, so I took it apart and draw a wiring diagram:
I figured out the reason it would spazz out. It is the result of a failing zener diode. I went up to the electronics store and got a new 3.0V 1W zener diode. I tested the gauge and it worked flawlessly on the workbench using the 12 V line out of an old computer power supply.
At this point, I figure the new gauge maybe blew the same part, so I swap them out. After testing the repaired gauge in the car, the gauge still doesn't read correctly (3/4 up when fans come on).
Frustration is setting in at this point. The only thing that is different on the car to the workbench test is the sender. For the test, I used some resistors to simulate the sender unit. I get out my DMM and set it to measure current. Maybe the sender fails under load. I measure the current on the test setup ad it is about 137 mA at normal operating temperature. Then I go out to the car, warm it up, and measure the current. Seemed normal about 145 mA (larger due to 13 V source compared to 12 V test source). Then when the needle started heading north, the current jumped up pretty high about 180 mA. This suggested to me that the sender starts to fail after a certain threshold current. The next day I install a new sender and my problems are over. It finally works perfectly and hopefully it stays this way.
I'm not sure if anyone has had this many problems with their temperature gauge circuit, but hopefully, this will help someone.
Comments
Thanks !!
Aftermarket thermostat vs. Honda thermostat (brownish). The Honda one is beefier, but the other difference is the air bleed is reversed in the aftermarket one.
Picture of the diode I replaced.
This is something I did while troubleshooting the temp gauge problem. I noticed there was a slight resistance caused by the dirty copper traces. I went ahead and cleaned all the contacts including the tach and speed pulser not shown in picture. Cleaning the contacts did improve the gas gauge reading. Before, when I was near E the tank really was almost empty (there's supposed to be 2 or 3 gallons left).
Here is a picture of the temperature sender wire. This is where i did most of my troubleshooting. If this wire is de-pinned, the resistance to ground and current can be easily measured. Measure the resistance with the engine and ignition off. Measure the current by connecting the ammeter to the lower screw in the temp gauge and the other lead to the depinned sender wire.
I still believe the sending unit is to blame, so I got a couple at the junkyard. I still need to try replacing it. In the mean time, I did some circuit analysis on the circuit diagram. I found out the way it is connected will not work. In other words, I messed up and drew it wrong.
I post up my results.