Hey y'all, here's my take on an instant-off relay timer, used to mute line level inputs and outputs, either by shorting or switching the signal. (In most cases, shorting the signal is preferred.)
It will run off a wide range of supply voltage, from 12-24VAC, from which it derives a regulated 12V for the timer and relay.
Loss-of-AC detection circuit pulses Q1 into RC network R3/C3, to bring RESET voltage above 1.8V in normal operation. If more than one cycle is lost, the timer resets and cuts power to the relay in less than 10ms.
The CMOS versions of the 555 timer, the ICM7555 or TLC555, have an actually useful RESET line that sources in the pico-amp range, so allows for high-impedance control. The bipolar NE555 has to be pulled down much harder, so most designers chose another method to reset it. More on this topic here.
I'm specifying very high-impedance control circuitry to avoid electrolytic timing capacitors, or large polys.
Here's the SPICE.
Instant-Off Relay Timer
And you'll need this model and assembly file for LTSPICE. Verify the spice install file path in the ASY file.
ICM7555 Model and ASY
It's not been tested physically yet, just in LTSpice. I will be bread-boarding it soon.
See any issues?
If it's good, I already have a board design done in Eagle, and I'd be happy to share that as "Don't care-ware".
It will run off a wide range of supply voltage, from 12-24VAC, from which it derives a regulated 12V for the timer and relay.
Loss-of-AC detection circuit pulses Q1 into RC network R3/C3, to bring RESET voltage above 1.8V in normal operation. If more than one cycle is lost, the timer resets and cuts power to the relay in less than 10ms.
The CMOS versions of the 555 timer, the ICM7555 or TLC555, have an actually useful RESET line that sources in the pico-amp range, so allows for high-impedance control. The bipolar NE555 has to be pulled down much harder, so most designers chose another method to reset it. More on this topic here.
I'm specifying very high-impedance control circuitry to avoid electrolytic timing capacitors, or large polys.
Here's the SPICE.
Instant-Off Relay Timer
And you'll need this model and assembly file for LTSPICE. Verify the spice install file path in the ASY file.
ICM7555 Model and ASY
It's not been tested physically yet, just in LTSpice. I will be bread-boarding it soon.
See any issues?
If it's good, I already have a board design done in Eagle, and I'd be happy to share that as "Don't care-ware".
Made some minor tweaks to R1 and R2, driving the transistor a little harder to bring the reset voltage up to a minimum 2.6V or so, just to allow for a 20% timing capacitor variance.
Just a couple notes:
The AC detect transistor exists to keep any voltage off of the RESET pin, before the power supply comes up, or the timer might latch up. This is warned of in the datasheet.
The open-collector switching transistor driving the relay could probably be deleted in favor of sinking up to the timer's rated 100ma. Just flip the time-constant R and C to flip the logic.
But I like the switching transistor because I feel it's robust and proven, and sinks 200mA. I've used this arrangement for relays for years with zero issues.
Just a couple notes:
The AC detect transistor exists to keep any voltage off of the RESET pin, before the power supply comes up, or the timer might latch up. This is warned of in the datasheet.
The open-collector switching transistor driving the relay could probably be deleted in favor of sinking up to the timer's rated 100ma. Just flip the time-constant R and C to flip the logic.
But I like the switching transistor because I feel it's robust and proven, and sinks 200mA. I've used this arrangement for relays for years with zero issues.
Last edited:
That looks really good 
The 555 in all it varieties is a really useful chip... it makes a great 'power buffer' for driving relays and motors as well 😉
Thoughts... it's good... the only quick thing I can think of is that the 'one cycle dropout' is great in theory and certainly proves the design, but it may not be practical in many parts of the world where supply brownouts are common.
Perhaps make an option to give much wider tolerance on that. Most audio equipment would be OK with many cycles going missing before the rails dropped out.
The 555 in all it varieties is a really useful chip... it makes a great 'power buffer' for driving relays and motors as well 😉
Thoughts... it's good... the only quick thing I can think of is that the 'one cycle dropout' is great in theory and certainly proves the design, but it may not be practical in many parts of the world where supply brownouts are common.
Perhaps make an option to give much wider tolerance on that. Most audio equipment would be OK with many cycles going missing before the rails dropped out.
Thanks Ketje! That is much simpler! And lower parts count.
I had to increase C3 to 200nF to get it to work as I believe you intended—Q1's base was ticking up to 0.6V and needed more capacitance to keep it down, pulsing to maximum 200mV. To keep parts count down, I went with a 1uF/470K arrangement, which gives about 40ms delay after loss-of-AC.
Mooly, to your point about the single-cycle dropout. I'm a little torn here. If I let the delay go long enough, the relay just cuts out once the power supply droops anyways. I want it to click off sharply. Would a brown-out actually result in missed cycles or just low voltage? The circuit seems to work down to 8VAC, giving 9V to the relay.
Even though this will work with the regular NE555, I think I'm still going to go for the CMOS version, for it's low quiescent current, and no need for a bypass capacitor hanging off the unused control pin.
Ketje did you delete the reverse protection diode around the relay because it's not needed? It was my thought that the flyback voltage of the relay could exceed reverse Vbe?
Here's the new Spice, using the normal 555 model so no need to create a model for the ICM7555.
Instant-Off Relay Timer
I had to increase C3 to 200nF to get it to work as I believe you intended—Q1's base was ticking up to 0.6V and needed more capacitance to keep it down, pulsing to maximum 200mV. To keep parts count down, I went with a 1uF/470K arrangement, which gives about 40ms delay after loss-of-AC.
Mooly, to your point about the single-cycle dropout. I'm a little torn here. If I let the delay go long enough, the relay just cuts out once the power supply droops anyways. I want it to click off sharply. Would a brown-out actually result in missed cycles or just low voltage? The circuit seems to work down to 8VAC, giving 9V to the relay.
Even though this will work with the regular NE555, I think I'm still going to go for the CMOS version, for it's low quiescent current, and no need for a bypass capacitor hanging off the unused control pin.
Ketje did you delete the reverse protection diode around the relay because it's not needed? It was my thought that the flyback voltage of the relay could exceed reverse Vbe?
Here's the new Spice, using the normal 555 model so no need to create a model for the ICM7555.
Instant-Off Relay Timer
I'm not really sure on the brown out issue as fortunately it is not something I encounter here, but it is something I have seen mentioned in the past.
Design is always a compromise and the time constant of the rail feeding the circuit is something you have to decide on. I would have thought missing up to (say) 10 cycles or so should not cause any issues.
Design is always a compromise and the time constant of the rail feeding the circuit is something you have to decide on. I would have thought missing up to (say) 10 cycles or so should not cause any issues.
I can probably get rid of the reverse protection diode around the regulator though. No big capacitance on the regulated side, and no way for a relay field collapse pulse to get through.
The way I usually do it is with an LM311 and look at the rail voltage. For my last project running on +/- 18 volts, I switched the relay when the rail voltage dropped to 16 volts. This got rid of the turn-off thump. The reason I often use this method is that it's super convenient and easy. I'm going to have +/- rail voltages on the board anyway, but I probably won't have the AC voltage there as well.
By the way Chris, your Adcom retrofit boards are very cool and I have to applaud you for your work- it's good to see people chipping in to keep older equipment serviceable.
By the way Chris, your Adcom retrofit boards are very cool and I have to applaud you for your work- it's good to see people chipping in to keep older equipment serviceable.
Hmm, yeah that's a good way to do it! I suppose you could also use a comparator to detect when your regulator falls out of regulation and click the relay out then. That could work for turn-on and off with no need for a timer.
Glad you like the Adcom stuff. There's a lot of these amps out there and I'm trying to keep them out the landfill! They're pretty amazing when refurbished properly.
Glad you like the Adcom stuff. There's a lot of these amps out there and I'm trying to keep them out the landfill! They're pretty amazing when refurbished properly.
The main reason I don't use the AC detection method (which is how a lot of equipment does it) is because it's fewer leads to cable manage if I go off the power rails.
I actually used a 555 based circuit many years ago and that one did not use AC detection but did aim for quick drop out and fool proof operation if the supply was quickly cycled.
A universal 'cookbook' relay delay using an LM555 timer.
A universal 'cookbook' relay delay using an LM555 timer.
You can also make R3 2M2 for a longer delay and perhaps take a schottky diode for D2.
The reverse current pull the emitter negative, turning on Q2.So no turn-off spikes.
Mona
Cool! Even lower parts count.
Basically, I'm trying to make this a handy board that can be added to pretty much any project, as long as 12-24VAC is available. Having its own power supply means very little interaction with the audio power supply.
40x32mm!
Basically, I'm trying to make this a handy board that can be added to pretty much any project, as long as 12-24VAC is available. Having its own power supply means very little interaction with the audio power supply.
40x32mm!
Ack! I bread-boarded the circuit. It works, except the timer doesn't actually reset once the AC power is removed, so if power is lost momentarily and comes back, the relay clicks in immediately. It would be better if the timer would fully reset whenever the AC detect is lost.
Any suggestions on how to reset the timer completely on loss of AC?
Any suggestions on how to reset the timer completely on loss of AC?
I realize one could just use a small input filter capacitor so that the power supply collapses quickly, but that puts more ripple on that capacitor and I'm going for long-term reliability.
Use a transistor to short the timer capacitor for reset instead of the reset pin. This will restart the timer. That's how I use the 555 as a reset/delay control circuit.
Thanks Fast Eddie, that's the trick!
Do I need D1? Without it, I'm getting -12V from B-E and I presume that would blow the transistor up.
Here's the SPICE.
Do I need D1? Without it, I'm getting -12V from B-E and I presume that would blow the transistor up.
Here's the SPICE.
I'm not sure about D1 but here's a couple tips.
1) Put a 100 ohm resistor in series with the 1 uF capacitor. This limits peak discharge current (for what it's worth) and may lead to longer transistor and capacitor life. The 100 ohm resistor is effectively in series with R-TC so it does not effect the timer operation.
2) You might have a problem cutting off Q1 in the real world circuit. With the values you have chosen for R2 and R-TC I don't think you will, but the 555 doesn't swing all the way to ground; It has a Darlington output so it only goes down to around 1 volt above ground, or more. It might be a good idea to put a bias resistor between the base of Q1 and ground to pull it down. Try different values between maybe 220 K and 1 meg and choose the largest value that works. You might have to adjust c3 upwards to make it work like you want.
Rod Elliott has a DC detect circuit that works very well. It has a resistor in series with the detect capacitor; the cap has to charge up to bias the transistor on for DC detect.
1) Put a 100 ohm resistor in series with the 1 uF capacitor. This limits peak discharge current (for what it's worth) and may lead to longer transistor and capacitor life. The 100 ohm resistor is effectively in series with R-TC so it does not effect the timer operation.
2) You might have a problem cutting off Q1 in the real world circuit. With the values you have chosen for R2 and R-TC I don't think you will, but the 555 doesn't swing all the way to ground; It has a Darlington output so it only goes down to around 1 volt above ground, or more. It might be a good idea to put a bias resistor between the base of Q1 and ground to pull it down. Try different values between maybe 220 K and 1 meg and choose the largest value that works. You might have to adjust c3 upwards to make it work like you want.
Rod Elliott has a DC detect circuit that works very well. It has a resistor in series with the detect capacitor; the cap has to charge up to bias the transistor on for DC detect.