Fuel Gauge Signal Latching

[xrotor]

Hello
I have built a circuit described in the forum using the LM 3914 display driver and a resistor chain for a fuel gauge I am building. It uses a float switch to activate 10 reed switches to switch in resistors attached to the input as it moves down (10 in all).
I need some help in working out a way of latching the output signal till another switch is activated, it would remove the need for very careful reed positioning
Can I feed the output of the display driver to a latch ic or ic`s and then to a uln2003s to drive the LED output. As the new signal from the reed is received it would need to the reset all the other outputs and wait for the next signal.
I am a little unclear of logic latches and if they can be used, output from the lm3914 would have to have a pull resistor take it to high and stop it to floating .
Any help would be appreciated its a long time since I did any electronics do they still do OC71 transistors!!

 

[dougy83]

Post/link your schematic. Depending on your cct, you may get away with using a capacitor to hold the current level (i.e. sample-and-hold); no digital latch required for that. But can't tell until I find out what your circuit is.

 

[xrotor]

Latching Required

Hi
I think latching will be required, I have tried using a cap to hold the signal but the delay from one switch being activated to another could be a long time, depending on fuel use. I have attached a standard lm3914 circuit, The design is close to the standard lm3914 drawing but the input is supplied by a resistor being switched into the input . 10 different to light up each individual led . The problem with which I am pulling my non to abundant hair out is, how to instead of lighting a led, latch that signal then pass that latched signal onto say a unl2003 to drive the led because the signal to the lm3914 would disapear till the next resisitor is switched in . The latch circuit would have to cancel the first latched led and light the next led.
Hope this is not as clear as mud.

 

[dougy83]

'A picture is worth a thousand words'... or so they say. Unfortunately none of those 1000words help me understand what the input setup is...

Please explain how the input is driven. A relevant picture would be useful. You say you 'switch a resistor to the input', is this resistor part of the resistor divider cct? Also, is each of the reed switches wired up to your LED board? Or are they all on the same 2 wires?

While you can latch the output, you can also use a buffered capacitor (which will take hours or days to discharge). To do either of these, the input configuration is required to be explained (you don't need to be lengthy).

 

[3v0]

. It uses a float switch to activate 10 reed switches to switch in resistors attached to the input as it moves down (10 in all).

If I understand it correctly he has 10 reed switches. Each is connected to a resistor that will supply a unique voltage to the input of the lm3914 chip.

If you were going to use a latch it would have to be one that would latch a analog voltage level, not logic.
------------------------------------------------------------------

A better way would be to use a strip magnet such that all switches that have been turned on stay on.

Note that you can position the magnetic strip to turn more or less switches on as the float moves. It is just a matter of positioning.

Choose resistor values that work in parallel to provide the desired voltage.

(Why not use a normal variable resistor style sender unit ?)

 

[xrotor]

Latching

Hi,
I have attached some pics of circuit and resistor input, also the float switch.
I cannot fit a normal resistive fuel sender due to lack of space, I am building for my very very light aircraft with a 9ltr tank.
The sender unit only requires two wires to it as the float magnet switches in each resistor in turn.After the float activates each switch there is a gap before the next is activated.
I think I need something on the output of the lm3914 to latch the last led signal output till the next reed is activated, it then needs to latch that led on and cancel the one preceeding it.
Thanks for help

 

[xrotor]

Circuit Correction

sorry the circuit is wired in dot mode not bar. This might make my problem clearer.

Thanks Neil

 

[dougy83]

OK, that info's useful. Tell us what you think of the attached cct.

Analogue latch: The top left opamp detects when a reed switch is closed and turns on the transistor. When the transistor is on, the lower left opamp charges the 10uF cap to the input voltage. The other opamp just buffers the signal to reduce the discharge rate of the cap. Cap should be low leakage and possibly a fair bit bigger than 10uF...

Digital Latch: The other cct (which is the digital latch cct you were after), with the transparent latch (74HC373 or 74HC573) has the same switch closed detector to enable the latch. When the switch opens, the latch is closed. A ULN2007 is not required, as the latch is capable of directly driving up to 35mA. You may need pull-ups on the LM3914 outputs, and be sure to use a current limiting resistor from VCC to the LED anodes (1 resistor shared between all leds is OK, as only 1 led is driven at a time).

 

[xrotor]

Latching Signals

Hi,

Many thanks for the help. I can see why you were stearing me towards an analog solution, it seems a little more direct and simpler solution. I will breadboard both of these and let you know how I go on.

Many Thanks Again
Neil

 

[xrotor]

latching Output

Hi,

Both circuits were breadboarded, the analogue worked but with two bigger lag on the output display which was made worse if a meaningfull hold on the signal was to be achieved.
The one using the transparent latch worked great, I am now buiding this up to a working unit to fit the tank and I will post results finished circuit and photos.

Many Thanks
Neil