relay switching logic

[Adrian Brentnall]

I'd always worked on the principle of 'static (relatively) safe once in-circuit" ... - are you suggesting some hi-value resistors from input to gnd on each gate ?

Thanks Adrian

[Adrian Brentnall]

I was only joking!

Actually across the load ? I've not worked out how inductive - it'll be in the region of 2 parallel coils, each of of about 400 turns @ 22mm i/d over 16 ft (nice mix of metric & imperial there!)

I could spare a couple of MOV's - if that would help.

The commercial kiln it's replacing has little in the way of these refinements - doesn't even have 'thermocouple fail' or 'I've turned on the heat but the temperature isn't increasing.....' intelligence.

Thanks Adrian

[mick]

Shouldn't you be looking at this the other way up? Your input signals should be "Temperature controller OK" and "Not over temperature". Both contacts should close when the detectors are powered up and drop open under power fail or fault conditions. It's easy then. A N.O. contact on the contacts in series - preferably controlling the contactor coil directly.

[Adrian Brentnall]

- quite possibly - we seem to have looked at it from every other angle!

What I've actually got in terms of signals, from the IC that's looking at the thermocouple, is

Thermocouple-OK - open-collector which the application notes show connected via a led / 270r to V+ .... darn - just realised that the 'alarm' condition pulls the pin low so 'good' = logic low

(This isn't actually 'Temperature Controller OK' - there's a separate uP controlling temperature, with it's own thermocouple) - this signal indicates that the independent 'safety' thermocouple is connected.

Temp-OK, (which is actually 'low' while the kiln temperature is below a 'maximum' setpoint - about 875c), swings to a volt away from VCC when in the 'alarm' state..

SO - the logic required of the latching relay is 'de-energise if TC-OK is low or Temp-OK is high'

- which isn't quite what I originally said....

These two inputs are intended to drop out a 'latched' relay in the case of error conditions. The relay can only be latched by pressing a manual pushbutton, so, once an alarm condition has been detected then, regardless of what the TCOK and TempOK do, the power remains disconnected from the main contactor.

In actual fact, there's an isolating contactor (that's controlled by the safety circuit mentioned above) - this has the ability to kill power to the actual Control relay, which is, in turn, switched by the uP controller...

Hope this makes sense ? Thanks Adrian

[Adrian Brentnall]

See the (latest) final circuits here

A

[meow2222]

yup. They might or might not improve relay reliability a bit

small fire & failure risks

NT

[meow2222]

not really that simple

better to have a series input r, then a diode to each power rail, + & 0, for each input

NT

[mick]

A possible variation? Based on your latest circuit:

tc OK signal goes via 10k resistor to base of NPN transistor. This has control relay, latching contact & reset button circuit in its collector and grounded emitter. If this input is high, transistor is on and control relay can be reset. (Don't forget to put a suppression diode across the relay coil!)

temp OK signal goes via 10k resistor to base of second NPN transistor. Grounded emitter. Collector goes directly to base of the above transistor so that if this input goes high the transistor turns on it will turn off the other.

Voila! 4011 not needed. This saves soldering 11 extra pins. :)

[Adrian Brentnall]

HI Mick!

The relay's a din-mounted thing, with it's own indicator led and suppressor diode..

My fault - but I should have said that the names of these signals are misleading... see below..

The two outputs from the AD594/595 are TC-OK - open-collector which the application notes show connected via a led / 270r to V+ .... and the 'alarm' condition pulls the pin low - not high.....

Temp-OK, on the other hand, swings to a volt away from VCC when in the 'alarm' state

Not sure that works! The 4011 seems like an easy way of getting the input signals the right way up - and also gets around the fact that the temp-OK signal doesn't go all the way to VCC...

The "separate transistors" approach was explored further up this thread... and we moved towards the current circuit after some discussion..

I guess if you're counting 'soldered joints' then the 4011 solution might just come out on top - but that's getting pedantic!

FWIW _ just discovered a forgotten stash of components, all in their original CPC packaging, that happens to include quite a few CD4011bf's - as well as loads of other cmos ics and quantities of 7-segment LED displays - all dating from some ten years ago. I can't remember what I was going to build with them - but it was going to be AMAZING! Now we'll never know...

Adrian

[mick]

It works. :)

Logic is: Relay energised if TC-OK high and Temp-OK low.

Temp-OK: With temp-OK high the transistor is going to turn on with 0.7V on it's base, the base current drawn will be Ic/hFE. That will be pretty small with most reasonable transistors. Just a guess: The transistor driving the relay from the TC-OK signal will have a base current of about 1mA. So, base current of Temp-OK transistor will be 1mA/hFE. That will be somewhere less than 10uA if hFE is 100. That will give a volt drop of

0.1V across the 10k base resistor. The transistor needs 0.7V to turn on, so the signal needs to source 0.8V at 10uA. That's nowhere near Vcc and you can expect hFE to be higher with a subsequent lower base current.

TC-OK should have a pull-up resistor (say 470R-1k) to Vcc if you aren't using the LED and series resistor. That will ensure that you get about Vcc out for "OK" and about 0V for thermocouple fault alarm.

Personally I'm not a fan of using an IC where 2 (or 3) conventional transistors will do the job. I'm not saying that the ICs are unreliable, but they are definitely not as rugged. Also, from a servicing point of view, I would much rather change a transistor than an IC. :) If you were to use a medium-high power darlington and your contactor coil is suitable then you could drive it directly and omit the control relay altogether. Don't do it unless the coil is specified for driving from a solid state output though.

I like safety cutouts to be as simple as possible, using the minimum number of components.

Strangely enough, I have various little component collections like that... lol

[meow2222]

no, its whatever you let flow into it. The b-e junction is a diode.

NT

[mick]

Correct, but my figures are correct for a base resistor of 10k and an hFE of 100. This defines the point where the relay will have enough current to operate. Perhaps I should have explained it better and said that the base current required to close the relay will be Ic/hFE. The base current would reach about 1.1mA (using a 10k resistor) if the source can manage it. The resistor could be increased providing that Ic/hFE is satisfied, allowing the source voltage to go higher if required.

[meow2222]

I don't know how much you know... tranny gain often falls way below 100, it varies according to the operating conditions of the moment

NT

[mick]

I was basing it on something like the BC337-16 which has a stated minimum hFE of 100 at 100mA Ic with a Vce of 1V. I've used the BC337-25 for a long time now, but that may not be so typical with a min hFE of 160 under the same conditions. They were cheap at the time though. :)

I realise that the gain varies (quite a bit!) as conditions change, but in this case the Onsemi data sheet should be ok as it shows hFE as being over 100 from 5mA to 500mA Ic.