LED stays on when it shouldn't. pulling my hair out.

[Martin Graham]

Hi all,

I hope you can help. I have put together a small pumping unit with a pair of LEDs that can charge and also work from the mains. The problem I am having with it is that when I disconnect the mains the "Charging" LED stays on for some time and never actually goes out, it just dims really low. please help as it is causing me a right ball ache.

 

[chopnhack]

I know this is obvious, but I will state it anyway. If the LED stays lit and dims over time, yet stays lit, it is receiving current and is the low point to drain it away.

With mains disconnected, can you easily remove the charging LED from the circuit?
This way you can test what current runs through that connection, perhaps with a diode, you can learn from which direction the current originates with the circuit off. I suspect that there is leakage from the UPS unit.

 

[Martin Graham]

Thanks for the reply. I am a bit of a noob.

If there is a leak from the UPS how should a address it?

 

[chopnhack]

Thanks for the reply. I am a bit of a noob.

I am too, I suspect a diode in series would solve the problem. I would wait for further advice from some of our masters around here.

While you are waiting, I would use your multimeter and start taking some readings like I mentioned before.

 

[dorke]

Welcome to EP,
Please measure the voltage between pin-2 and pin-6
of the PSC, with the led fully on and dim

 

[Gryd3]

I found the datasheet here : http://www.mouser.com/ds/2/260/PSC-60-SPEC-806374.pdf
Pins 1 and 2 of the output side are Active Low pins for 'AC OK' and 'Battery Low'.
They recommend a separate power source for these 'Alarm' outputs in the manual.
Pins 1 and 2 appear to be merely a NPN type transistor between the respective Alarm Pin and Pin 6 which has a max current of 30mA.
If the LED stays on and simply slowly fades, I would assume leakage as mentioned above. It could be a design flaw or something could have broke.
When you poke with a multi-meter, please keep the black probe on pin 6 and provide voltage measurements for the output pins before you disconnect AC.
Once you disconnect power, please re-measure as the LED is fading.
Let it rest, and once the LED is out, please provide measurements once more.

 

[Martin Graham]

thanks for your help. I will got on to this today.

 

[Martin Graham]

hi,
When all is plugged in and charging the measurement across pins 2 & 6 is 35.9 mv.

when i unplug the mains and measure again as it is fading the meter shows open circuit 0L.

 

[dorke]

hi,
When all is plugged in and charging the measurement across pins 2 & 6 is 35.9 mv.

when i unplug the mains and measure again as it is fading the meter shows open circuit 0L.

Please measure the voltage in the fading state ,
like you did in the charging state.

 

[Martin Graham]

I have. The meter shows open circuit between pins 2 & 6

 

[dorke]

What kind of meter are you using?
As far as I know meters show open circuit (or O.L) only in the resistance range.
So looks like you are not measuring voltage.

Anyways do this:
connect the resistor as shown and repeat the measurement(pins 2 to 6 voltage).

 

[Martin Graham]

thanks, I'll try again.

 

[Martin Graham]

Dorke, you're a legend. thanks. As soon as AC is disconnected the led fades away within about half a second. I can now sleep tonight.

Next I'll be working on how to integrate into the circuit a battery full charge indicator.

 

[chopnhack]

Dorke, you're a legend. thanks. As soon as AC is disconnected the led fades away within about half a second. I can now sleep tonight.

Next I'll be working on how to integrate into the circuit a battery full charge indicator.

Good job Dorke! Looks like the resistor is enough to decrease the current to below the threshold required for the LED to light.
Please note Martin that this does not mean that there is zero voltage present, just that it's below the LED's threshold to light. That may be all you need though.

 

[Gryd3]

Considering how those 'alarm' pins are wired, it would be wise to put a resistor in series with the LED's as well wouldn't it?
The AC_OK and LOW_BATT outputs are active low with a 30mA capability and will sink whatever current you provide which may cause damage to the transistor shown in the schematic for that part #.

 

[dorke]

Considering how those 'alarm' pins are wired, it would be wise to put a resistor in series with the LED's as well wouldn't it?
The AC_OK and LOW_BATT outputs are active low with a 30mA capability and will sink whatever current you provide which may cause damage to the transistor shown in the schematic for that part #.

Yes,
but that LED already has an internal resistor it is a "12V LED"