My first post. I know it is long… don’t shoot me. Hope someone finds this interesting. It is about the fuel gauge on my nearly 11 year old RTV 900. I will describe my RTV 900 failure, do some failure analysis, describe my fix and describe how to make your fuel gauge custom calibrated and more reliable.
My fuel gauge died a week ago, so I did some trouble shooting and believed the problem to be the ammeter/gauge in the dash. I put in an order for a new gauge through Missick’s and decided to dig into the broken gauge to understand what failed. The gauge can be taken apart by “expanding” the bezel crimp around the plastic housing. What I found was that a resistor inside had been very hot for some time and had eventually caused the solder to migrate and oxidize to the point that there was NO solder in the hole in the printed circuit (PC) board where the resistor lead goes. The resistor lead was not touching anything in the hole. There was even a little splash of solder on the back side of the display.
After measuring the values of the resistors, and meter I calculated the power in the meter and in that particular resistor (56 ohms) that was so hot. With a full tank of fuel and battery voltage at 13 volts when running, the current in the meter (plus the two resistor) is about 139 milliamps and the power in the 56 ohm resistor is about 1.09 watts. It is getting VERY hot and even though the resistor is tough and did not fail, it caused the PC board to be discolored from the high temperature and eventually the solder connection failed. I suspect some people have also experienced this failure mode and that tapping the meter may make the meter work for a while.
I fixed mine by using several (7) resistors that added up to 56 ohms. They are spread out and have more surface area so that they can transfer the heat to the air and there is a much longer thermal path to the PC board traces. It works fine now.
Another solution would be to move the resistance outside the meter. This would be very easy since the 56 ohm resistor is in series with the Ignition switch battery wire. The meter would need to have a piece of wire in place of the resistor internal to the meter and an external resistor stays in cooler air. Just cut the wire and put the resistor in series once the internals of the meter were modified.
I got the new meter from Missick"s and did some measurements on it since it was physically different. It did have higher total resistance. The new meter had about 130 ohms total resistance and the original had 93 ohms. Looks like Kubota has been trying to solve the problem and did so by increasing the resistance. I did not dig into a brand-new meter so I don’t know if it is just resistor changes or the meter movement/coil.
After more head scratching, I concluded that with a few modifications outside the meter, I can fix the calibration problem and I could have a gauge that really was accurate when full and empty and it would operate at lower current. Since the meter is reading higher than it should and is more sensitive than desired, it takes just two resistors to fix the calibration problem. One resistor sets the full reading and another sets the empty reading.
You just need to select resistor values when the tank is what you consider full and empty. When full, set the resistor in series with the ignition switch wire so that the needle just reaches the F mark and is not “pegged” very much. I just did that and the value was 27 ohms The smaller the resistor value, the needle will indicate more full. Now when the tank is empty, based on your definition of empty, set the other resistor that attaches to the ignition wire and the sensor wire to a value indicating empty. This resistor value I came up with is 470 ohms. A smaller resistor will make the needle indicate more empty. I set mine up so with a half inch or so fuel in the bottom of the tank, it points to the edge of the white just below the Red mark. I had to siphon it… Yuck!
This modification can be done in wires going to the plug that attaches to the meter. Heat shrink tubing and good soldering will make this a tidy and reliable modification. Besides better calibration of the meter, the current when the tank is full has dropped due to the added 27 ohm resistor. This lowers the power internal to the meter. So, the gauge is now calibrated AND should be more reliable. I should mention that the meter indication will change from ignition switch on and motor running. The alternator pumps the battery voltage up so the meter will indicate higher. Just something to be aware of.
Fun project that turned out well…. So far. If anyone wants more details, let me know.
Mike Wilson
My fuel gauge died a week ago, so I did some trouble shooting and believed the problem to be the ammeter/gauge in the dash. I put in an order for a new gauge through Missick’s and decided to dig into the broken gauge to understand what failed. The gauge can be taken apart by “expanding” the bezel crimp around the plastic housing. What I found was that a resistor inside had been very hot for some time and had eventually caused the solder to migrate and oxidize to the point that there was NO solder in the hole in the printed circuit (PC) board where the resistor lead goes. The resistor lead was not touching anything in the hole. There was even a little splash of solder on the back side of the display.
After measuring the values of the resistors, and meter I calculated the power in the meter and in that particular resistor (56 ohms) that was so hot. With a full tank of fuel and battery voltage at 13 volts when running, the current in the meter (plus the two resistor) is about 139 milliamps and the power in the 56 ohm resistor is about 1.09 watts. It is getting VERY hot and even though the resistor is tough and did not fail, it caused the PC board to be discolored from the high temperature and eventually the solder connection failed. I suspect some people have also experienced this failure mode and that tapping the meter may make the meter work for a while.
I fixed mine by using several (7) resistors that added up to 56 ohms. They are spread out and have more surface area so that they can transfer the heat to the air and there is a much longer thermal path to the PC board traces. It works fine now.
Another solution would be to move the resistance outside the meter. This would be very easy since the 56 ohm resistor is in series with the Ignition switch battery wire. The meter would need to have a piece of wire in place of the resistor internal to the meter and an external resistor stays in cooler air. Just cut the wire and put the resistor in series once the internals of the meter were modified.
I got the new meter from Missick"s and did some measurements on it since it was physically different. It did have higher total resistance. The new meter had about 130 ohms total resistance and the original had 93 ohms. Looks like Kubota has been trying to solve the problem and did so by increasing the resistance. I did not dig into a brand-new meter so I don’t know if it is just resistor changes or the meter movement/coil.
After more head scratching, I concluded that with a few modifications outside the meter, I can fix the calibration problem and I could have a gauge that really was accurate when full and empty and it would operate at lower current. Since the meter is reading higher than it should and is more sensitive than desired, it takes just two resistors to fix the calibration problem. One resistor sets the full reading and another sets the empty reading.
You just need to select resistor values when the tank is what you consider full and empty. When full, set the resistor in series with the ignition switch wire so that the needle just reaches the F mark and is not “pegged” very much. I just did that and the value was 27 ohms The smaller the resistor value, the needle will indicate more full. Now when the tank is empty, based on your definition of empty, set the other resistor that attaches to the ignition wire and the sensor wire to a value indicating empty. This resistor value I came up with is 470 ohms. A smaller resistor will make the needle indicate more empty. I set mine up so with a half inch or so fuel in the bottom of the tank, it points to the edge of the white just below the Red mark. I had to siphon it… Yuck!
This modification can be done in wires going to the plug that attaches to the meter. Heat shrink tubing and good soldering will make this a tidy and reliable modification. Besides better calibration of the meter, the current when the tank is full has dropped due to the added 27 ohm resistor. This lowers the power internal to the meter. So, the gauge is now calibrated AND should be more reliable. I should mention that the meter indication will change from ignition switch on and motor running. The alternator pumps the battery voltage up so the meter will indicate higher. Just something to be aware of.
Fun project that turned out well…. So far. If anyone wants more details, let me know.
Mike Wilson
