relay switching logic

[Adrian Brentnall]

HI all It's that glass kiln build again...

Looking for suggestions on how to cause a relay to be energised when two separate inputs are 'low', but to drop out if either (or both) of them go high.

It's part of a safety circuit - the two inputs indicate 'thermocouple fail' and 'over-temperature' - and the relay 'dropping' will un-latch a further relay circuit, which kills power to the main contactors.

The two inputs are low-power themselves, so a driver stage is required to switch sufficient current for the relay.

One plan would be to have a pair of relays with n/c contacts in series, one switched by each of the inputs. When either one is energised it cuts power to the latching circuit...

..but there may be a less 'clockwork' solution - anybody ? This circuit does the relay driving

- but how to get the logic for the two inputs ? Diode & 2k2 resistor for each ??

I learnt all of this too many years ago... but have since forgotten it!

Thanks in advance Adrian

[Martin Brown]

If it is DC then diode logic so that any positive high voltage can activate the relay coil but the inputs cannot see each other.

A ----|>|-----o------|

[Dave Plowman (News)]

First, what is used to drive the relay coils? Eg, 12v DC?

[The Natural Philosopher]

what you want is a NAND gate

In the circuit you have shown us, connect two transistors whose emmitters are both earthed, and whose COLLECTORS both go to the BASE of the existing transistor.

Put a couple of 10k resistors in series with the bases and use those as your NAND inputs.

a high on either will short the input to the driver transistor and cause the relay to drop out.

[Adrian Brentnall]

OK - thanks.... No need for duplication of the input resistor, then > Yes - it's a DC voltage round-about 11v - with a 'hi' on either i/p to drop the relay

Thanks Adrian

[Adrian Brentnall]

HI Dave

Yes - all of the logic & control switching is @ 12v - a final 12v 'slave' relay switches 240v to the contactors.

[Adrian Brentnall]

Thanks! (also to Martin)

I've put both circuits up on the web - Martin's with the diodes & TNP's with the two transistors..

The 'latching' relay has a manual "start" pushbutton - so that if either of the alarms go off then a manual action is necessary to re-start things ...

Any pro's or con's to either circuit (assuming that both of them will function as drawn..?

Thanks

[Martin Brown]

The base resistor is only there to limit the maximum current into the transistor you should ensure enough to saturate the transistor when it is supposed to be "on".

[The Natural Philosopher]

No you havent. You have reversed out of a NPN transistor to a PNP and the one I suggested is implemented wrong.

well mine wont as drawn.

sigh.

The other one wont either.

Its an OR as drawn, not an AND.

THIS is what I meant

[Adrian Brentnall]

The circuit using the PNP (according to wherever I lifted it from on the web) would achieve the 'input goes high, relay picks' function...

..though I might have got that wrong ?

I'm glad it's not just me Sometimes it's the simplest things, isn't it ?

OK - I see what you're doing there. (I think!)

Push the switch and the relay picks, because the third transistor is normally biased 'on'

If either of the inputs go 'high' then the 10k b-e resistor is shorted out, and the relay drops out...

Aha - light-bulb moment....

Even if the inputs subsequently go low, _because_ the push-button hasn't been pressed, the relay stays unpicked...

I think you've nailed it, Sir ! Many thanks Adrian

[SteveW]

For a safety circuit, you should really have a relay that de-energises on thermocouple failure and an overtemperature relay that de-energises on overtemperature; both of these in series with one contact on a third relay should feed the third relay's coil, with a pushbutton in parallel with the third relay's contact; another contact on the third relay should feed the main contactors. When power is first applied, you would need to press the button to energise the third relay, which would then hold itself in until either power is removed or either of the first two relays drops out. This ensures that everything is fail safe - ie: any failure in the system (other than a welded contact) causes the kiln to be powered off.

Depending what the risks are and how serious the consequences of failure could be, you could use proper safety relays, that have multiple contacts, forced contact separation and feedback to confirm healthy operation, but you'd only be looking at that if the consequences of failure were severe.

You can buy 12V and 24V dc control relays with built in transistor inputs to simplify things.

SteveW

[Adrian Brentnall]

That was my original thought (I used to work on process control / micro circuitry - and the rule was that things 'failed safe')

My latest attempt is here

but I'm not convinced that the relay will drop in response to either one of the inputs going high....?

I might wire an external alarm to the relay contacts - at least that way I'd be aware that 'something's failed'.

Didn't know that - but I'm fine for relays at the moment - just need to drive them with a transistor.

Thanks Adrian

[meow2222]

no. If you want that, use an npn tran followed by an invertor tranny driving the relay. Both npn.

NT

[Adrian Brentnall]

Thanks I'm leaning towards cmos gates at the moment, as a fairly elegant way of isolating the inputs from the AD594, being un-critical of the absolute input voltages, and allowing easy inversion of signals that are the wrong way up!

The safety engineer in me wants to use as few 'components' as possible, though I do realise that a single 4011 isn't a single component internally...

Thanks Adrian

[tony sayer]

Yes it's a delicate balance;!..

Whatever you use I'd bypass the emitter base junctions with a small ceramic cap something like .1 nf just to keep any nasty HF switching spikes out of its delicate RF area. And put a few across the supply lines and across the power supply pins any CMOS devices you might use...

And a reverse EMF diode across any DC supplied relay for any switching spike suppression..

[Adrian Brentnall]

True...

Yup - the 12v coils on my din-rail mounted relays have their own diode connected across them - but a bit of decoupling local to the cmos and the AD594 wouldn't go amiss. I've also got some 350v MOV's on order to sit across the coils on the

240v contactors, just to snub any nasties. Was planning on having two 'enclosures' - one with the low-level stuff in and 'tother with the power switching - to keep it all separate. Long-term, I'd hope to substitute the proprietary controller with a home-grown pic controller - but that's _long_term

Indeed Thanks Adrian

[tony sayer]

Thats a very good idea,. I'd have an opto-coupler in there somewhere too;-)....

[Adrian Brentnall]

...and explosive bolts so the control box can jettison the power box in case of malfunctions.. ?

Hopefully, the relays will provide sufficient isolation between the 'clean' and 'dirty' boxes, and the main load is only a long bit of resistance wire, rather than anything hugely inductive...

[meow2222]

4011s are highly static sensitive, which can translate to a safety issue. The circuit in question offers no static protection. Also they're not un-critical of V_in, trannies are less fussy. But either can do the job.

NT

[meow2222]

erm... :)

slightly inductive though, might be worth an RC snubber to maximise relay life, since its a safety issue area.

NT