Oh! In that case, adding one shouldn't be very difficult. Let me quickly add it.
Yeah, I got it.

microswitches -> photodiode -> Relay ->555 timer right?

I think I have to redraw from first. The circuit is a bit complex than it was before.

Edit: Are there any free or low cost soft for circuit simulations? Is Proteus Free tier good?

take care when incorporating a circuit like this. This makes it subject to safety worthiness and can become a liability.

Can the circuit fail to an unsafe state...

 

Oh! In that case, adding one shouldn't be very difficult. Let me quickly add it.
Yeah, I got it.

microswitches -> photodiode -> Relay ->555 timer right?

I think I have to redraw from first. The circuit is a bit complex than it was before.

The photodiode element is necessarily separate from the watchdog...

The watchdog is necessarily separate from your interlock.

However... you can use the interlock switch to control the 'safety' relay. This means that switch is no longer carrying mains AC, so can be smaller, with a lower current requirement.

 

A very reliable system to indicate that the LEDs are on would be a small window with a phosphor that will glow a visible color when they are lit, and be clearly visible from the outside. Not as effective as the interlock, but close to failure proof.

 

Here's one way to do it, there are many others...

 

A simple indicator light on the outside panel near the switches or door latch can serve as an indicator that power is applied to the inside lights. And if it is a Neon or LED light it should outlast the rest of the system. The goal is to have the system be "error resistant", not idiot-proof.

 

A very reliable system to indicate that the LEDs are on would be a small window with a phosphor that will glow a visible color when they are lit, and be clearly visible from the outside. Not as effective as the interlock, but close to failure proof.

Yep, i think tinted glass front panel might also be a good choice. Also makes the box look good.

Here's one way to do it, there are many others...

View attachment 212164

Thank a lot. Since I already have a ULN2003A can I connect the output from 555 to the one of the free inputs? I can cut down on the 2N7000 if that's possible.

A simple indicator light on the outside panel near the switches or door latch can serve as an indicator that power is applied to the inside lights. And if it is a Neon or LED light it should outlast the rest of the system. The goal is to have the system be "error resistant", not idiot-proof.

Can't agree more on the last point.

 

Thank a lot. Since I already have a ULN2003A can I connect the output from 555 to the one of the free inputs? I can cut down on the 2N7000 if that's possible.

Yes, that would work.

Although not strictly necessary, you might consider a diode across the relay winding...

 

Yes, that would work.

Although not strictly necessary, you might consider a diode across the relay winding...

I will add it. I already got one on the other relay.

 

take care when incorporating a circuit like this. This makes it subject to safety worthiness and can become a liability.

Can the circuit fail to an unsafe state...

I don't think so. The probability of everything failing at once is minimal and even it does, the relay is in NO position so the supply will be cut-off as soon as the microswitch is opened.

 

Yes, that would work.

Although not strictly necessary, you might consider a diode across the relay winding...

Can I use a single terminal relay to drive the AC_Live and wire the AC_Neutral directly?

 

Can I use a single terminal relay to drive the AC_Live and wire the AC_Neutral directly?

Its generally considered bad practice to do so, the fail-safe is intended to be isolating - if L and N were reversed, which does happen, it would be unsafe. Anyway DPST NO relay contacts are pretty standard, there's little value in skimping.

 

Its generally considered bad practice to do so, the fail-safe is intended to be isolating - if L and N were reversed, which does happen, it would be unsafe. Anyway DPST NO relay contacts are pretty standard, there's little value in skimping.

Thanks, I just decided on the 7805 IC. The datasheet recommends 0.33uF on the input, so I guess I have to replace the 0.1uF I have. Do I need to check anything while selecting the transformers? I expect my system to consume at 500mA, does it make sense to go with a 1A transformer or can I stick with 500mA?

 

Same reason why main switch is DPST on both L and N. Incidentally, unless you have a specific reason not to, I'd recommend using an all-in-one IEC socket/fuse/switch/output ( eg RS part 311-8075) for your mains connection, its much neater and easier to install & wire.

 

Thanks, I just decided on the 7805 IC. The datasheet recommends 0.33uF on the input, so I guess I have to replace the 0.1uF I have. Do I need to check anything while selecting the transformers? I expect my system to consume at 500mA, does it make sense to go with a 1A transformer or can I stick with 500mA?

The input capacitor isn't that critical, but never hurts to follow data sheet.

How have you arrived at 500mA?

 

Same reason why main switch is DPST on both L and N. Incidentally, unless you have a specific reason not to, I'd recommend using an all-in-one IEC socket/fuse/switch/output ( eg RS part 311-8075) for your mains connection, its much neater and easier to install & wire.

Thanks. I will check it out.

The input capacitor isn't that critical, but never hurts to follow data sheet.

How have you arrived at 500mA?

Alright.
The microcontroller even if all pins are powered consume 250mA, LCD display -- I assumed 100mA and two relays each consuming 25mA. So total ~400mA. But I am not really sure about the calculations.

 

Its generally considered bad practice to do so, the fail-safe is intended to be isolating - if L and N were reversed, which does happen, it would be unsafe. Anyway DPST NO relay contacts are pretty standard, there's little value in skimping.

The intention of the interlock is to switch off the lights, it IS NOT an interlock to protect against somebody sticking their finger in an empty light socket and getting a shock.
Once again, the intention of the safety system is to make the device error-resistant, not idiot proof.

 

The intention of the interlock is to switch off the lights, it IS NOT an interlock to protect against somebody sticking their finger in an empty light socket and getting a shock.
Once again, the intention of the safety system is to make the device error-resistant, not idiot proof.

Agreed.

But making it idiot proof doesn't cost any more - there's no big saving on single pole v double pole.

 

Thanks. I will check it out.


Alright.
The microcontroller even if all pins are powered consume 250mA, LCD display -- I assumed 100mA and two relays each consuming 25mA. So total ~400mA. But I am not really sure about the calculations.

Where did the LCD come from? Don't recall that on the schematic. Atmel CPU is <50mA + output currents of which you're currently using 2 @ a few uA - depends on clock speed, do you need to run at 16MHz for this? LCD backlight is either 25mA or 100mA+ depending on version, always best to check datasheets. Relays are 47ohm coils so 100mA each but you could choose lower (80mA) current versions, again check specs.

So could be ~350mA or as low as ~200mA

But sticking with 400mA DC, a rule of thumb is AC current is DC/.7 = 0.57A so a VA rating of 9 x .7v= 6.3VA, a 10VA one should be sufficient.

 

Where did the LCD come from? Don't recall that on the schematic. Atmel CPU is <50mA + output currents of which you're currently using 2 @ a few uA - depends on clock speed, do you need to run at 16MHz for this? LCD backlight is either 25mA or 100mA+ depending on version, always best to check datasheets. Relays are 47ohm coils so 100mA each but you could choose lower (80mA) current versions, again check specs.

So could be ~350mA or as low as ~200mA

But sticking with 400mA DC, a rule of thumb is AC current is DC/.7 = 0.57A so a VA rating of 9 x .7v= 6.3VA, a 10VA one should be sufficient.

I haven't started interfacing the LCD yet, I would do it after I choose the controller. Right now I'm confused between ATmeaga808 and 1609. Since it is powered directly from main, I don't think if the current consumption reduced by underclocking the CPU will be making any difference.

Do I need to worry about the temperature rating of transformers? They are like 40 degC, isn't that too low?

 

I haven't started interfacing the LCD yet, I would do it after I choose the controller. Right now I'm confused between ATmeaga808 and 1609. Since it is powered directly from main, I don't think if the current consumption reduced by underclocking the CPU will be making any difference.

True, still it's something to consider in a production environment where the difference between a 10VA and a 7.5VA transformer over a 100,000 units could be significant. Just passing on useful thought processes based on experience.

Do I need to worry about the temperature rating of transformers? They are like 40 degC, isn't that too low?

Well, it depends on whether you think your box will hit 40degC internally. I doubt it but again this is a decision to be made from experience. Or consider the mitigation if that proves to be wrong.. It'll mean ventilating the box, a few strategic holes, or, worst case, adding a small fan. This is why you build engineering prototypes for production.

On this, I'd take a wait and see approach. Remember it's not going to suddenly fail at 40degC ambient, all that means is that it's MTBF (mean time between failures) will be reduced. There's a lot of fancy math you can do, if you're NASA say, to assess the impact, but here a suck it and see approach is justified.