Delay on a mains-powered relay?

I never cease to be amazed at the questions asked in this NG from people who do not have a clue from grade school experiments about electronics but want to design or modify industrial situations where some manager should be putting up the money to employ a competent engineer skilled in the particular area.

Relying on poorly informed advice is asking for trouble and a sure course to the courts for a damages claim.

--
John G

Wot's Your Real Problem?

Look for a spring powered timer then. They use an AC clock motor to hold them at, say, 5 minutes. When the power goes off the spring unwinds until the clock goes to zero and the contacts close/open. When the power is restored the motor rewinds the timer to 5 minutes. Looks a bit like a cheap cooking timer on a frame.

N

| It's an AC relay. Just how do you intend to attach that "big capacitor?"

Obviously use a light (0.5W) DC relay!

N

Refer to the post by Tony Williams. There might be a cheaper way. As he mentioned, the capacitor is expensive - you'll pay over $15.00 dollars for one. The alternative described below would cost a lot less. The super caps would cost you $1.70 for 2. The opto costs

50 cents. The voltage regulator costs another 50 cents. You can get 2 10 K pots for a dollar. Add another dollar for the 1000uf cap and resistor. I assume same price for the transformer as in the relay design. You'll save at least $10.00 per sensor, if the idea works. Since you have a number of sensors, it seems worth at least trying one as proposed below. I do not know if it is practical for you, or if it will work, since I don't have specs on the existing setup.

You *may* be able to use an optoisolator, a super cap, and a potientiometer instead of the relay. The potientiometer would allow you to adjust the delay time. You would need to use a lower voltage transformer and a regulator IC. The cost would be a lot lower. The key is the current sense loop - without specs, I don't know if an optoisolator will work. Here's the details if you decide to try it.

Here's Tony's circuit, modified for an optoisolator:

View in fixed width font. ____ +----+ +---|AC +DC|--+--7809---+---+-------+ | | | | | | +|C2 | __|__ )||( | | +|C1 | === R1 | + |--+Sense 120 )||(12 |Rect. | === | | | |Opto | Vac )||(Vac| | | | +|C3 P

Look for a TDDO (Time Delay Drop Out) relay.

--
Paul Hovnanian     mailto:Paul@Hovnanian.com
------------------------------------------------------------------
Reject nihilism!

You're right. They appear to do it all for you for $85.00. Notice the total power consumption is 180mW, so the 'relay' must be quite low power to be included in the budget.

RL

Here's an even better mod to Tony's circuit. You need

4 parts - all available from Allelectronics: 5 volt regulated wall wart p/n DCTX-512 $3.50 Minature 4.5 volt relay p/n RLY-639 $0.75 1N4001 diode (1 needed) p/n 1N4001 15 for $1.00 1 farad 5.5 volt super cap p/n CBC-12 $3.50

D Wall + ------|>---+-----+ Wart +|C | === Relay | | - -----------+-----+ I suspect the relay will stay energized longer than 5 minutes. You can add a second diode to reduce the time. You might need to add resistance in parallel with the super cap to reduce it further. This depends on the characteristics of the relay - I don't have one to test.

Ed

It is likely that your wireless device has an input circuit with a dc voltage across the inputs when the input is open (that is how the open contacts are sensed). One side is likely common with one terminal of the battery.

You should try putting a power FET device across the input (source and drain) and using a capacitor from source to gate to hold the FET on for a period after your relay opens. A resistor across the capacitor sets the discharge period. I would start with 470 MF and

1meg resistor for 4-5 minutes delay (depends on gate voltage on turn-off, device specific).

If the battery negative is common to one of the input terminals, use a N-channel FET (I did something similar with an IRF520, way over kill on current, but you are not likely to damage it). If the wireless device voltage is only 3 volts, you will need a FET designed for logic level gate voltages.

Connect as follows: Source of FET to - input terminal Drain of FET to the other input terminal Capacitor negative (use 16V cap) to Source Capacitor positive to Gate 1 meg resistor across capacitor 100K resistor to battery +, other side of 100K to one of your relay contacts Other relay contact to Gate.

Use a grounded soldering iron as the FET gate can be damaged by static electricity.

If the battery positive is common to one of the input terminals, use a P-channel FET and reverse all polarities listed above.

When the relay contacts are closed, the 100K resistor uses the battery to charge the capacitor. Voltage on the capacitor keeps the FET ON.

When the relay contacts open (on power fail) the capacitor holds a charge while the 1 meg resistor discharges the capacitor. At some low voltage on the Gate, the wireless sensor will detect an open circuit and send the alarm. Experiment with resistance and capacitance to get an acceptable time delay.

If you shop the parts, likely cost per circuit is USD $2.00-4 .00. Bill Kaszeta Photovoltaic Resources Int'l Tempe Arizona USA snipped-for-privacy@pvri-removethis.biz

Here's what you do. Put a cheap answering machine on the same power circuit that you're trying to monitor, and connect it to the phone line that the call gets made on.

Tell the guys that when they get a call (they'll still be awakened, so perhaps this is not a perfect solution) to call back to that number. If the answering machine answers, it has power, and hence so does the equipment. If the answering machine fails to answer, get in the truck.

;-)

--
The e-mail address in our reply-to line is reversed in an attempt to
minimize spam.  Our true address is of the form che...@prodigy.net.

The 85 cent ones (CBC-125) are 1 F, 2.5 V. You'd need four in series to handle 9 V. The manufacturer may recommend a resistor in parallel with each cap, which I think makes a voltage divider to keep the voltage about the same on each cap. Digi-Key has Cooper PowerStor capacitors that come already packaged as two 2.5 V caps in series to get a 5 V rating for $4.20 to $10 quantity 1.

Your second idea, using the CBC-11/CBC-12 1 F 5.5 V memory back-up capacitors, may or may not work. Some of these capacitors are made for very low current (microamps, CMOS memory) and not the tens to hundreds of milliamps you'd need for a relay or the 10-20 milliamps you'd need for an optoisolator or solid state relay.

Matt Roberds

power

cap?

1000

Just put a delay circuit in between the relay contacts and whatever (assuming low voltage) signal the control to the alarm panel. A 555 IC driving a small low voltage relay should do the trick.

capacitor?"

But it sounds like the 120v coil is used to detect mains failure, power goes out, relay loses power, contacts open/close, triggering the alarm. The relay can be left alone, you only have to mess with the low voltage signal it switches.

| But it sounds like the 120v coil is used to detect mains failure, power goes | out, relay loses power, contacts open/close, triggering the alarm. The relay | can be left alone, you only have to mess with the low voltage signal it | switches.

If he wants to use a cap to delay it the lower power the better. Even a reed relay will work OK and keep the cap size down. That's the key - run from a rectifier circuit.

N

It seems to me that the OP would prefer to use what hardware exists now in order to minimise expense and the complexity involved in a complete re-design of what he now has. The relay he is using now is a

3 pole changeover unit so depending upon what is currently connected to the relay contacts will have some bearing on what he can do in the easiest and most cost effective manner.

However, using a reed relay to detect the mains failure would involve using an additional step-down transformer in order to supply the ac sense input to the reed relay/rectifier/capacitor arrangement. While a reasonable delay might be achieved it certainly would not be in the order of more than a minute at most even if the reed relay coil was

10,000 ohms and a 10,000uF capacitor was used.

Assuming the OP did not want to alter the existing KUP relay or wiring, the reed contact would be connected in series with the existing 120Vac coil of the KUP relay so that the latter acted as a slave off the reed relay. This would also mean that a separate no-break 120Vac supply would have to be present to maintain the KUP relay operated until after the reed relay contact had opened at the end of the delay. Clearly, the need for a separate no-break 120Vac supply was not on the OP's agenda, although a UPS could be used for such a function.

Reed relay contacts are not intended to switch high voltages such as

120Vac and they don't like to switch ac current in inductive circuits. The contacts quickly weld together or become pitted very quickly so a reed relay contact would not be suitable to open the existing KUP coil circuit. Even if the OP altered his existing KUP relay coil for a low voltage dc unit capable of being switched by the reed contact, the dc supply to hold the KUP relay operated AFTER the 120Vac mains had failed would have to come from somewhere. Again, this would require a separate no-break dc voltage source so this is even more complex than using a UPS with the existing KUP relay. In this case a rechargeble battery could be used but this would require that a battery charging circuit would be required to maintain the battery in a fully charged condition. All in all, it would be more expensive and there would be no real advantage over sticking with the existing KUP relay and providing a UPS.

If the OP really needs a delay exceeding 1 minute then the only way is to incorporate a true delay-off relay as the ac sensing device and slave the existing relay/s from this. He will still have to provide a separate no-break 120Vac source (eg, UPS) to keep his existing KUP relay operated though.

I would suggest the simplest method to achieve what the OP requires is to first get a UPS of say 150 - 300VA rating and use this to supply the existing KUP coils via the contact of a true delay-off relay such as

A 4390A-E would give him an adjustable 30 - 300 seconds delay and the allow its input to sense the 120Vac mains supply. Several KUP relays (assuming more than one) could be handled by the delay-off contact since it has an 8A rating.

| It seems to me that the OP would prefer to use what hardware exists | now in order to minimise expense and the complexity involved in a | complete re-design of what he now has. The relay he is using now is a | 3 pole changeover unit so depending upon what is currently connected | to the relay contacts will have some bearing on what he can do in the | easiest and most cost effective manner.

....

I remember small, inexpensive delays that used an AC clock motor to wind them up. When the power went off they ran down and switched contacts. I suggested that previously.

N

No, the CBC-125 is not the one I had in mind. I should have made it clearer. It is CBC-131, which is 85 cents in qty < 10 and is .1 (point one) uf at

5 volts, not 1 uf at 2.5 volts. The CBC-125 is $1.00 each in qty < 10

Some optos will work as low as 50 ua, from what I have read. The question that is not answered is what his sense loop needs (how much current at what sense voltage) and of course whether the opto will permit that current over the full delay time.

That's good to know. I gave a quick try searching for super caps in Mouser and came up empty. (Doesn't mean they don't have them - I didn't give it much of an effort.)

I agree. I don't know that it will work for sure, but I strongly suspect it will. It is guaranteed to give *some* delay - the question is, how much. At least it avoids the question of the sense loop voltage and current requirements. (We know it has to be small from his brief description, but not whether an opto's vce would prevent operation.) What I didn't know is implied in your next sentence:

What happens if you place a 10 to 30 ma load on one of those caps? The relay I specified will draw about 30 mA at ~ 4.4 volts and should continue to stay energized down to 10 mA and possibly below that as the voltage decays. The relay I *meant* to specify is RLY-635 (not 639). I had them both written down on my scratch sheet when I looked up the parts. I think the 639 will work - but the

635 would give more delay. The 635 relay has a 500 ohm coil and would draw 8.8 mA to start, and would probably still be energized when the voltage dropped to 1.67, where it would draw about 3.3 mA.

Time will tell. As is typical with OP's we may never hear how he makes out. But I already ordered all the parts I specified for the second idea, and will test to see what kind of delay I can get. We just have to wait a bit for the shipment to arrive.

See

No manual "wind up" required.

Granted, such a device will work but it is hardly an elegant technical solution. The whole idea of technology in industry is to make the operation and maintenance of equipment reliable as possible without having to worry about remembering to "wind up that timing relay".

PS. Not having seen one of these wind up relays I suppose it would be possible for it to be automatically wound up again by a motor when power was restored.