Alec_t, I have just two questions. If we will change the way the acknowledge and the lamp test get the supply voltage to the positive side, what would be the circuit. Another thing is the relay which I connected as shown, can it function the same way like your circuit? Can you please simulate and modify the circuit to work as designed? It is very important to change the polarity of the Test Lamp and the Acknowledge because it is connected to a BUS or a network of cards.

 

Your circuit won't work, because
1) You now have a P-FET instead of an N-FET,
2) the relay coil current will drop excessive voltage across R1,
3) C4 should be 10μF,
4) Q1 and Q2 don't provide a latching function.

'Lamp test' will also actuate the relay. is that what you want?

I'll take a look to see what modifications would make it work as you want, but it may not be today.

 

Is the 48v supply regulated?
What is the max supply current?

 

Ok, here's a modified circuit. I have assumed (a) that the Alarm signal can't supply enough current to drive the relay directly, so the relay is switched by a MOSFET M1, (b) that the Alarm signal can supply the ~11mA required by the rest of the circuit apart from the LEDs, (c) that the relay is energised only when the Alarm signal is present, (d) that the relay has a 48V coil.

 

Thank you Alec_t, I'll implement the circuit and see the actual performance. It may take some time depend on the availability of the components.

 

You haven't said if the Alarm signal can provide enough current to drive the relay directly. If it can, then M1 and its associated D1, R1, R2 can obviously be omitted.

 

Yes Alec_t, the alarm signal can provide enough current to supply the relay. I am facing difficulties on the selection of the LEDs. I could not find the specs in DIGIKEY. Also can we connect these LEDs in parallel? If one LED fails at least there will still be one lit. Regarding the Mosfet, 100 ma would be enough to drive the LEDs. I found only SMD devices in Digikey # FK3506010LCT-ND.

 

Led current depends on color. Most of the time 10mA is good provided the resistor is rated properly. You can parallel leds but each led should have it's own resistor, Not a common resistor to all the leds

 

I am facing difficulties on the selection of the LEDs. I could not find the specs in DIGIKEY.

The sim used randomly selected white LEDs. I assume for an alarm indicator that red or amber might be preferred? There are heaps to choose from, for example these. Just apply the filters (specs) you want when selecting from the Digikey (or other supplier) product catalogue.
LED failure is most unlikely, but parallel some if you wish (with appropriate dropper resistors). The dropper resistors will, of course, dissipate more power if the LEDs are in parallel, so should be rated accordingly. However, what if something else in the circuit failed? It would be difficult to allow for every eventuality.

I found only SMD devices in Digikey # FK3506010LCT-ND.

If you want through-hole components (FET etc) I'm sure Digikey has plenty, including the venerable 2N7000. What about these? Perhaps a bit OTT for driving LEDs, but useful if for some reason you needed to drive incandescent lamps.

 

Alec_t, I have tested the circuit in a board and the only thing that is not functioning is the acknowledge latching. When the acknowledge is press, the lamps goes steady on, when you released the button, it should remain on until you reset then it goes back to flashing mode until the alarm is gone.

 

Which transistor types are you using for Q1 and Q2?
What is the voltage at Q1 collector when the Acknowledge button is kept pressesd?

 

I have used BD680 for Q1 and TIP31C for Q2. The voltage at Q1 collector is 11.5V when pressed.

 

I suspect the problem is due to the BD680, which has a collector/emitter cut-off current greater than the current through R7/D5/R11. Power transistors are not ideal for Q1 or Q2. Try low power signal transistors instead, such as a 2N3906 for Q1 and a 2N3904 for Q2.

 

Alec_t, I have used 2N3906 as Q1 and 2N3904 as Q2 as you suggested and I got 11.5 Volts on the collector of Q1 if the Acknowledge is pressed. But when I released the Acknowledge button it did not latch. Why in your simulation it is latching, in the actual circuit it is not?

 

Hmmm. Why indeed? Try reducing R9 (the Q2 collector resistor) to 4k7.

 

I have reduced the R9 to 4K7 and still the same.

 

Then you can return R9 to its 10k value.
Make sure that R7 is the correct value and making good electrical contact with Q1 collector and Q2 base. If R7 were open circuit or very high value it would explain the non-latching. If your circuit is on a breadboard then connections can be unreliable.

 

Alec_t, I actually removed the R7 because if it is connected, it will latch automatically after one flash. That is why R7 should not be there. If I removed the R7, the circuit flashes and then when I pressed the Acknowledge, the LEDs remain lit. If I releases the Acknowledge it goes back to flashing mode. I think Alec_t, we'll go back to my previous drawing that uses an SCR to latch the LEDs.

 

Hmmm. Why indeed? Try reducing R9 (the Q2 collector resistor) to 4k7.

Hi Alex_t

I think the latch circuit in post #44 is missing a (1uf) cap from the collector of Q2 to +V.

 

I agree a cap is missing. I must have posted an early version of the circuit. Apologies. I see that my current version has a cap from Q2 base to ground, so it's worth trying the 1u from Q2 collector (or Q1 base) to +V or from Q2 base to ground.

I think Alec_t, we'll go back to my previous drawing that uses an SCR to latch the LEDs.

Ok. Your call.