# Unstable input to trigger LED after 10-15 seconds (12 V)

I have a small excavator that, when fuel levels get low, the warning light flashes irregularly (due to movement of the machine) until it stays on as fuel level gets really low. It's very annoying and loses the effect of a goofy light as it loses importance in my concentration.

I have looked at the sensor and it is a simple On-Off float switch, not a more complex resistive wiper float switch.

What I'd like to do is to have a circuit that only triggers when the input trigger is on for ~10-15 seconds (ideally adjustable to 30-60 seconds or so - I'd have to work this out in the field) and the warning output stays on until input is reset by turning machine off, as one does when refueling (usually!!)

The output signal ideally would be a bright LED (even flashing ~ 1 second interval is OK) with possibly an audible alarm as well.

I have occasionally mucked around with, and have some, 555's since they first came out ;( and can read circuits OK. I am assuming that these can be used satisfactorily, but happy for any ideas.

I'd appreciate some help here. I searched for similar, but the answer was not quite suitable. Thanks

simulate this circuit – Schematic created using CircuitLab

• Do you want it to latch-on after detecting 10-15 seconds of solid on-time? – Daniel Dec 6 '15 at 22:29
• Yes. I would need to be able to vary the time to 'latch-on' until I found the right level though. The sensor is a float switch that simply drives a goofy-light. Depending on the work I'm doing eg on inclined ground and being a small machine it is like riding a buck-jumper so this light starts flashing when tank is still half full. Damn annoying. As the fuel gets lower the light stays on longer. – Clive Dec 8 '15 at 4:27

This circuit should do what you want:

It uses a 556 IC, which is a dual 555. The first timer is configured as a missing pulse detector (monostable). With the pot set at 15% of its value (45K of the 300K), the time output period is 15 seconds. With the pot set at 90% of its value With the post set at 60% of its value (180K of the 300K, the timeout is 60 seconds. Note: these values are approximate, mostly due to the tolerance of the capacitor.

The voltage divider made up of R5 and R6 cut the 12V input down to 5V. When the input goes high, it starts the timer (monostable). As additional pulses come through, they retrigger the timer. As long as the input is 0, or the first timer is running, the $\overline{\small{RESET}}$ lead of the second timer is held at 0 keeping the LED from flashing.

When the first timer expires and the input is high, the second timer, configured an an astable, is pulled out of reset. It flashes the LED at a 1 Hz rate (500, ms on, 500 ms off). (If you want to change this timing, you can use this 555 astable calculator.)

Because the output of the 556 typically limited to 10 mA or less, and you will probably want more than that for your LED, I added an inverter with an open drain to the output which can sink up to 40 mA. To save parts, I used another gate in the same package for the other inverter, even though it doesn't need the extra output current. Because it's open-drain though, I added the pull-up resistor to +5V.

If you don't want the LED to flash, you can discount the input of the inverter IC2A and connect it to the input of the NOR gate IC3A, and then replace the 556 with just a 555 (the pin numbers will of course change, but the names will be the same).

If this circuit were to be placed in a vehicle, like a car, it would be a very bad idea to connect it up to the 12V battery directly. First of all, the 12V battery is nominally about 14V. In addition transients can be nasty on a vehicle's 12V system, with voltages rising as high as 125v for 10 ms during a load dump.

This power supply circuit provides protection against negative voltages in addition to the positive spikes due to load dumps, noise, and jump starting, and regulates the 12V down to a filtered 5V:

No matter what circuit you end up choosing, if it's electrical system is such that transients like this could occur, then you should include a similar conditioning circuit.

• Op has confirmed it is not a resistive wiper float switch, it is a simple on-off momentary one. – Passerby Dec 9 '15 at 2:37
• Thanks @tcrosley. That would be stretching my talents somewhat to make up that!! Good points though on the voltage spikes. The machine is very small (it will fit easily through a standard doorway) so electrically doesn't need a high performance system, as such. Probably similar to a larger ride-on mower. – Clive Dec 9 '15 at 2:48
• @Clive thanks, I didn't know what size your excavator was. Thought it might be a big CAT. I would still keep C1, L3, C2, D3, C3, IC1 and C4 as my power supply since there still is probably some nasty noise on the 12V line. As a minimum you must at least have C2, C3, IC1 and C4. – tcrosley Dec 9 '15 at 3:20

The simplest solution to this I can imagine is a 555 in monostable mode, using the "output" signal as a reset signal to the timer. That way it will only activate after some amount of time, but will reset and start over if the "output" signal blips out for a moment.

This tutorial shows how to make a monostable 555 circuit.

The signal from the excavator is likely a common automotive fuel pump setup. A float sensor varies the resistance to a regular incandescent bulb. Tank gets low, sensor resistance lowers, bulb gets more relative voltage, bulb turns on. It's most likely not a digital signal. I know this happens in my 99 Camry when I placed an led bulb instead of an incandescent bulb. The much lower current needs of the led bulb left them lit even when the tank is full. The resistance is calculated for an incandescent over the full range of the tank. Which is why they slowly fade on, or turn on and off when moving.

Of course this may change for newer setups but 10 bucks says this is how it's on your excavator. A Haynes or factory manual could confirm.

Simplest solution in this case is to change the bulb for one that turns on with a lower voltage or current.

Otherwise a common microcontroller with an ADC can be used. Measure the input, trigger at the voltage you want.

• The sensor is a float switch that simply drives a goofy-light. Depending on the work I'm doing eg on inclined ground and being a small machine it is like riding a buck-jumper so this light starts flashing when tank is still half full. Damn annoying. As the fuel gets lower the light stays on longer. – Clive Dec 8 '15 at 4:32
• @clive the same type of setup is used for the light most of the time. It's rarely an on/off type. – Passerby Dec 8 '15 at 4:38
• In this case it is simply an off/on float switch. I took the sender/sensor unit out of the machine to check it. I am not sure how to add a photo here. – Clive Dec 9 '15 at 2:32
• No need, we will take your word for it. – Passerby Dec 9 '15 at 2:35
• Thanks @Passerby. It's a simple yet effective system. It literally hangs on a small rod from the top of the tank and the switch closes at a certain level. No springs, resistors, etc. Thanks for your help – Clive Dec 9 '15 at 2:42

A low pass filter followed by a schmitt trigger is not exactly what you asked, but it would fit the application nicely. The indicator only turns on after the input has been high for a while, and vice versa.

simulate this circuit – Schematic created using CircuitLab

Note that if your lamp draws more than 100 mA, it should not be driven directly by the opamp. If the float in the fuel tank actually drives a potentiometer, the above circuit should still work with the slight modification of removing R7, diregarding the switch and connecting the potentiometer wiper to R6.

• Is it easy enough to make the float switch time variable so that I can set it for optimum trigger time? Say 15 seconds to 45 seconds. It's not easy to even estimate the time interval I'll need and will need to be on a bit of a 'suck it & see' principle !! I can handle making up this circuit :) – Clive Dec 8 '15 at 4:37
• Use a led and 1k resistor for your lamp or use a transistor buffer .If you dont do something your lamp wont light properly or the TL081 will die young. – Autistic Dec 8 '15 at 11:30
• @Autistic The TL081 can source 100 mA, thus the "150 ohm" lamp. I however have no idea about the current draw of the average automotive indicator light, so you might be right. – jms Dec 8 '15 at 22:36
• @Clive You can replace R6 with a 1MΩ potentiometer or trimmer resistor. Increasing the capacitance of C1 or increasing the resistance of R6 will increase the delay. Keep in mind that at some point the input impedance of the opamp or the leakage current of the capacitor will prevent the light from ever turning on if you just keep increasing the delay. – jms Dec 9 '15 at 0:31
• Thanks again. I plan to use a higher intensity LED as the main warning light. The original goofy light I'll disconnect because it is down around my ankles and rather useless. Poor design – Clive Dec 9 '15 at 2:58