Fuel Level Switch

[GuzziMoto]

I would think your issue is the LED light takes much less electricity to light then the bulb does, and even when cool the thermistor can flow enough electricity through it to allow the LED bulb to light. Adding a resistor in parallel to the LED bulb probably won't change that, but maybe it will. I am not an electrical guy. Perhaps it would give the small amount of electricity that is lighting the LED bulb an easier path through the resistor so it doesn't go through the LED. Electricity does like to take the easiest path. That is the only way I can see a resistor in parallel help. But I am not sure it will. I know if you used a relay that would switch when 1.2 watts of 12 volt power was applied to it, and put the LED on the output of that relay, that could work. But that is starting to get too complicated and I am not sure about the availability of a relay that would switch reliably at 1.2  watts / 12 volts. What is that, a tenth of an amp?

I know this isn't the answer anyone wants, but I would keep the stock bulb. It is tried and true in this application. Anything else is just going to introduce reliability issues, additional points of failure.

[Tinus89]

2 hours ago, Weegie said:

The TWA 800 remark was made with my tongue firmly in my cheek and I agree.

The idea of mixing components with current passing through them and a potentially explosive atmosphere might sound utterly mad, but could still be safe, it's dependent on a few different variables.  I've never heard of any scenario of a tank exploding due to either a fuel pump or low fuel warning sensor failure either

@Tinus89 am I missing something here?

AFAIK its the properties of the thermistor itself that alters resistance, resulting in a higher current flow for the fixed Voltage across the circuit it has nothing to do with the bulb resistance. The bulb rating is chosen to allow it to be bright enough to be observed, but also to balance/limit the total circuit resistance. If the bulb resistance is too high (i.e. lower wattage) it won't allow enough current to flow and would never light. Too low, would result in too much current and burn out the thermistor.

The Thermistor's resistance lowers as it gets hotter, when it's dunked in fuel it's cold and the resistance is high. When it gets uncovered, fuel below the sensor, the thermistor starts to get hot. When it gets hot the resistance lowers, that will also result in self heating (by how much I don't know).

I don't understand the difference between using a bulb and an LED plus equivalent resistance, in a purely resistive circuit

If I'm talking complete bollocks or missed something feel free to educate me (because I've never been accused of being "the brightest bulb in the building")

 

 

 

I am no electrics guy, so I may not understand, but what I meant is essentially what GuzziMoto wrote:

 

49 minutes ago, GuzziMoto said:

I would think your issue is the LED light takes much less electricity to light then the bulb does, and even when cool the thermistor can flow enough electricity through it to allow the LED bulb to light. Adding a resistor in parallel to the LED bulb probably won't change that, but maybe it will. I am not an electrical guy. Perhaps it would give the small amount of electricity that is lighting the LED bulb an easier path through the resistor so it doesn't go through the LED. Electricity does like to take the easiest path. That is the only way I can see a resistor in parallel help. But I am not sure it will. I know if you used a relay that would switch when 1.2 watts of 12 volt power was applied to it, and put the LED on the output of that relay, that could work. But that is starting to get too complicated and I am not sure about the availability of a relay that would switch reliably at 1.2  watts / 12 volts. What is that, a tenth of an amp?

I know this isn't the answer anyone wants, but I would keep the stock bulb. It is tried and true in this application. Anything else is just going to introduce reliability issues, additional points of failure.

I think that an added resistor (instead of the bulb) will protect the thermistor and therefore safeguard the intrinsically safe system in the tank, but I think even with the high resistance of thermistor & resistor, the current flowing is still enough to light the LED. As the LED lights with the thermistor cold (and ~ 1800ohm - I have not measured yet), it will most likely also light with the 1800+150ohm resistance...

However, would switching the LED through a 100mA relay not work? Something like this (https://www.diverseelectronics.com/upload/documents/SARA.pdf), which can only close at ~100mA (which almost lights the incandescent bulb). I would suppose the coil in the relay would still allow some current to flow for the thermistor to heat up & keep it's function.

 

Also, are we confident this fuel gauge works with a thermistor as well? Because it has a small "float" as can be seen on the picture:

 

Keeping the stock bulb is not an option. I have a set of LEDs nicely built into the handlebar clamp, the incandescent bulb simply does not fit there. Making an electronic solution is less work than going back to the original bulb...

[Weegie]

Ok now I understand, the reduced current which flows through the circuit when the tank is full and wouldn't light a bulb is still sufficient to illuminate the LED.

A resistor in parallel wouldn't work, the easiest solution I can think of then is to use Roy's solution of a resistor in series to limit the current to the point that the LED won't iilluminate. Resistors are cheap, it doesn't cost much to try it and Roy even suggested a value which gives you a starting point. If it kept the LED lit on a full tank, step up the value.

A relay in theory would also work, but matching the voltage to get the coil to pickup in the relay sounds like a bit of a nightmare, you'd also have to ensure it dropped out at an appropiate voltage too.

I haven't read through the whole thread, but when I looked at that petcock you referenced on the SD page, my first thought was it looks awfully like a float switch, but I don't know

Apologies for the misleading information, I'd not understood the problem fully.

[GuzziMoto]

I could see a resistor in series with the LED light being able to reduce current flow enough that the LED doesn't light until the thermistor has heated up enough. But you may need to try various values for the resistor to find one that is suitably high enough to stop the LED from lighting when the thermistor is cool while allowing enough current through to allow the LED to light up when the thermistor heats up.

Although I really don't know. I would rather go with a relay or a fuel level sensor that is straight up on or off. The thermistor does not seem well suited to an LED light, as the properties of the incandescent bulb seem to be an integral part of the circuit.

[Lucky Phil]

The advantage to the current system thats also been used in the automotive world for many many years in millions of cars is that it has it's own in built hysteresis. A magnetic float switch is subject to the vagaries of fuel sloshing and false indications or flashing warning light unless I guess you include a latching relay of some sort. The thermistor system is also ultra reliable and very compact and simple. The bulb lasts the lifetime of the vehicle essentially but the thermistor itself tends to drift or age to the point of failure after 25 years or so under the influence of fuel. The V11 issue is the instillation style of the thermistor due to the inaccessible fuel tank on the early bikes. It's tricky to design something other than whats original and the thermistor itself is pretty much impossible to replace on it's own. The later in tank bikes are easy and the thermistors are a dime a dozen, easily available and easy to replace.

 

Phil