As I said, it needs tuning.

Diodes - The intent of the diodes is to assure that the opto is off when the input is at 1.8 V, but one single Vf drop might not be enough. D1 - D8 should be replaced with 5.1 V zener diodes. This raises the start-of-conduction up to the 6-7 V range. I didn't put them in last night because I don't have a horizontal decal in my library - yet.

Current limiting resistors - I picked 4 mA as an operating current design point as something large enough to guarantee the opto is fully ON, but not drawing unnecessary current from the alarm circuits. In a previous post the TS suggested 15 V as the off state threshold voltage. With 5.1 V zener diodes in series, this makes the current limiting resistors 2.0 K. The opto will start to conduct at a much lower current, and the output will cross the CMOS transition level at a secondary current of approx. 0.6 mA. Depending on the actual opto device used, the input current at 15 V probably can be reduced even more.

ak

 

OK, this is about as simple as it gets. This is the circuit referred to in post #33. U1 powers up in a random state. If Q14 is high, the oscillator stops immediately. If Q14 is low, the circuit runs until it goes high, a max. of 5 minutes, then stops with Q14 high.

U2 and U3 form an 8-input wire-NAND gate. If any input voltage goes low,its output trannsistor opens and the combined output goes high and turns on Q1. If all inputs are ok (above approx. 7 V), all U2 and U3 output transistors are saturated, pulling the Q1 gate down to approx. 0.8 V (0.1 V x 8).

Q14 provides the gate drive voltage for the output transistor Q1. When SW1 is pressed and released, resetting Q14 pulls the Q1 gate low through R11, removing gate drive and turning off the buzzer no matter what the input condition is. In this way, Q1 is part of an open-drain wire-NAND circuit - if either Q14 or the input transistor stack is low, the Q1 drain is not pulling low and the beeper is off.

Note - I used the optos shown because they are convenient on the schematic. Almost any standard optocoupler will work, although it should have a minimum CTR (current transfer ratio) of 50%

ak

 

Use the optos listed in the schematic. (MOSFET optos were to insure the following transistor was turned off in my design)

If current draw from the detect lines is the real concern, and not the supply, then I would suggest dropping an 8 channel Darlington array in front of the optos to drive them.

With Darlingtons driving a small load such as the optos you could drop an additional 10k or so resistor on the inputs, and greatly reduce the load on the detect lines...shifting it to the supply proper.

The ULN2804 already has a 10500 ohm resistor installed in the base input line, an additional 10k would reduce the current needed at each line to about 1mA.

With some trial and error you may even get it even lower.

 

Actually…I was being VERY conservative, you can probably get the current requirements on the 8 lines to well below 1mA each.

At 4 mA collector current the Darlington gain is likely to be 1000 or more.

But keep in mind this idea only has merit if the supply can handle some extra load, and the draw from the 8 lines is critical.

 

Just an abstract thought

reed relays,

old, reliable and very robust

 

Checking Digi-Key for 24 V reed relays -

The lowest operating current is 9.8 mA, considerably higher than the other options being discussed. Also, that part is $22.

The lowest cost is $1.27, with a coil current of 11.1 mA. Minimum operate voltage is 16 V, which is fine. But the minimum release voltage is 2.0, which is too high for an input that drops to "around" 1.8 V. I think there is a tradeoff there - Lower operating energy means a lower current needed to activate, but also a lower current needed to maintain. Also, many reed relay datasheets do not list a minimum holding voltage.

ak

 

Drilling down into the circuits I have posted, there are potential issues with each of them. No big problems, just control quirks.

For example, once the 5 minute silence timer starts, there is no way to reset it and re-enable the beeper. So if at minute 1 the original alarm signal clears, then at minute 2 a different one (or the same one, or a group of three) goes low for one minute and then clears, you miss all of that activity because the beeper was disabled. The thread title says "simple", and that is what we have so far. But in this case simple means it is hiding things, maybe important things. Only the TS knows for sure.

The circuit in #42 covers the basic requirements. And if the connecting-grounds question goes the other way, #42 can be modified to use the diode tree in #22 (without the CD4001 flipflop) for the same basic functionality at lower cost and complexity. From there, we can add more sophisticated input comparators, more logic to capture hidden alarm trips, etc.

ak

 

I like AK's designs.

There are timer modules available. An example is this one. https://www.macromatic.com/uploads/files/pdf/instruction_sheets/901-0000-281.pdf
Previously, I mentioned SSAC.

So, for < $45.00 USD (Allied Electronics) you save the labor of putting together the delay module. 12 V is generally harder to come by.

You did mention space, so that's another consideration.

The reason that multi-function modules are used is if you did a lot of in-house engineering, then the multi-function module means less parts to stock and a shorter time to repair because it's likely stocked in-house.

Most of these timers require you to remove power (this one requires 150 mS) to reset. Some, like this one offers retriggerability and reset capability.

OPTOMOS relays can get the current down to about 1 mA per channel. They are essentially optoisolators and MOSFETS and are small.

The LT6700 series parts would probably draw the lowest power from the system.

If you could replace the incandescent 24 V lamps with LED versions of the standard lamps, you'd end up with plenty of current to spare.

 

As mentioned see here the pictures of the prototype ready for testing.

 

The original lamps are these T5.5 24V telephone lamps: http://www.lampbank.net/telephon.htm
I have found LED versions of them but they are very expensive, will see if I can order directly from China where they are much cheaper and replace the current lamps with the LED version so as stated having more power to spare for the alarm circuit.

 

What is the buzzer / beeper / alarm device? Piezo? Electro-mechanical buzzer? It makes a difference, and you might have to add a diode to protect the output transistor.

ak

 

What is the buzzer / beeper / alarm device? Piezo? Electro-mechanical buzzer? It makes a difference, and you might have to add a diode to protect the output transistor.

ak

12V Piezo buzzer and I used 3.9k resistors as input resistor and 1N4148 diodes no zener. I will test this in next 2 days. I can build the simpler circuit with zener and 2k resistors later as well and test. This now becoming really fun. Just a question about the zeners. if I use 5V Zeners will this not affect the 21V to the lamps and draw this voltage down to 5V? Sorry for my limited knowledge on this.

 

Just a question about the zeners. if I use 5V Zeners will this not affect the 21V to the lamps and draw this voltage down to 5V?

No. In round numbers, the voltage drop across the zener is 5 V, and the drop across the input diode of the optocoupler is around 1.5 V, for a total of 6.5 V that varies very little with changes in current. Everything else is dropped across the resistor, which is how the resistor sets the current for the string. Using Ohm's Law:

At 24 V, the peak current through the string is (24 V - 5 V -1.5 V) / R.

At 21 V with a 1N4148 instead of a zener, the current is (21 V - 0.6 V - 1.5 V) / 3.9 K = 4.85 mA

You can try different diodes and resistors and see what the current is; or rearrange the equation, put in a fixed current value, and calculate the required resistor.

ak

 

Is the CD4060 at the bottom inserted incorrectly a problem?

 

Is the CD4060 at the bottom inserted incorrectly a problem?

yes, I was under a little time pressure to make this prototype and I made a counting mistake in no of pins for the CD4060 and I noticed only after I had soldered the IC holder and was placing the IC, so I had to shift the IC by 1 pin in the holder to make it work. I will of course fix this in the next version.

 

Final Update.
Sorry for delay but I finally managed to setup the unit at the client.
The 1st design with the flipflop somehow I couldn't get it to work. The buzzer kept beeping.
And the connection did have an impact on the input voltage. The voltage dropped from 21V to about 19.5V and the low voltage raised from 1.8 to 2.4V. This did not trigger the circuits.

So I changed the circuit to version 2 with the 5.1V Zeners and the 2K resistors in the input circuit and removed the flip flop circuit. I found that the 5 min delay for the buzzer with the set components was more like 8.5 min but manageable. Added a LED to the power circuit so at least to see that the power is on, and added a switch to the input lines (they all group together to ground). The switch was more to allow me to connect the 8 input connections and ground safely, and switch on the circuit. If anything would go wrong I could quickly switch it off.

During testing found that the cheap china powersupply was having too much of noise and so added a 1500uf 25V capacitor to the power circuit just to remove the noise. This worked. Tested the circuit with batteries (2x9V in series). At home all was working.

Installed the system last week and tested it with the customer. And it all worked... pffff that was a big relieve, although tested at home i was still concerned about the impact of the circuits on the input signal. and if the customer would accept the 8.5 min delay at startup and when silencing the buzzer. Luckily he was ok with it all.

 

I hereby want to thank all those All About Circuit Members that have contributed to the design, discussions and the success of this project. Without you all I would have not been able to do it. It is so great to know that people from all over the world can contribute to a small problem I posted her from Singapore.

A million thanks to you all and may you all be blessed with more knowledge and the ability to help others in their quest to solve electronic challenges.

 

Glad it all worked for you. Which schematic (post #) did you end up using?

ak

 

Glad it all worked for you. Which schematic (post #) did you end up using?

ak

Yours dated 13th May, with some minor changes.
thanks again.

 

As wonderful as the 4060 is, it does have a relatively low-precision oscillator section. Now that you have a reference point defined by R9, C1, and the 8.5 minute delay period, you can decrease either the R or the C to speed up the oscillator and shorten the delay. Changes should be directly proportional, so if you change R9 to 62K, the delay should shorten to 5.27 minutes (-ish). 59.0K 1% tolerance should get you very close to 5.0 minutes.

ak