Hi all,

Ive been asked to make a circuit that operates a relay for a selectable amount of time.

Using dials like that from a decade box i hope to be able to set the time for how long the relay is 'on' before returning to its 'off' state, it should be operated using a push switch regardless if the switch is still pressed down.

The trouble is I cant find any type of circuit where a timer is controlled by resistance, or multiple resistance dials, hopefully i can make it in the range of microseconds to 10 seconds.

any help please?

 

Old faithfull 555, & Bill's blog.

 

+1

Bill Marsden's 555 monostable circuit article

There is a complete circuit and really good description of how it works. It can be easily modified to have an adjustable time delay and operate a relay instead of LED's as in his schematic. One problem that I initially noted with your requirement is the "microseconds". Just the act of pressing the start button will take a few milliseconds then there is the switch bounce which will take 10 - 20 milliseconds to stabilize. The latter is dependent on the quality of the switch. I think that a time delay of a couple of seconds to 10 seconds is a little more reasonable.

 

Ive been asked to make a circuit that operates a relay for a selectable amount of time.

I posted a reply yesterday but it seems to have gone lost.

First, you are aware that off-the-shelf lab timers already do this, and they are not expensive?

Do you hope to have accurate control? I mean, do you want to be able to dial in "96µs" and have it pulse the relay for that amount of time, ± some acceptable error? That degree of precision will require an accurate clock, for instance using a crystal oscillator, divider and counter, and a digital approach throughout.

The 555 (analog) solution is great if you just need adjustment over the wide range you mentioned. But the timing is not precise - it will drift - and there will be no way to get accuracy without difficult calibration.

Working with relays in the microsecond range will be tough because they are mechanical devices and have finite switching times.

 

I " feel " that you want discrete step control which can be achieved with logic switching different R values with a 4066, analog-digital switch & a 555. Might need to switch a few different C values. More information is always helpfull.

 

Hey all, thank you very much for your replys!

Ive now been told not to use a 555 because theyre hard to calibrate and i should try for a harder project to make as 555 is deemed too basic, PICs and TTL counters are the route i should take!

It has to be pretty accurate, maybe to 10's of microseconds so the switching of the relay will be a factor so maybe i could start counting after the relay makes contact!

Im now thinking of other ways to control the time other than resistance wheels and may likely use an oscillator

Im not very good at programming PICS either so the simpler the better really

@bernard ill look into it thanks

any thoughts?

 

Some stray thoughts: Start counting with relay contact closure presents bounce problems which lasts 10-20ms, but there may be work-arounds.What would be the repitation rate of relay activations ?
A 4060 IC with crystal can give an accurate time base.
Consider Hg wetted reed relay for no bounce.

 

Again thank you for your replys.

ive talked it over with a few people but the majority seem to think that going down the PIC route may be easiest. i will also use solid state relays to take away the switch time although i see they are still not perfect but i suppose ill get that sorted at a later date

An alternative someone else suggested was an oscillator, TTL counters/dividers, latches, BCD switches and signal gates but i dont quite understand the whole picture, i gathered they were counting each pulse and dividing it until the required amount of pulse occurs calculated from 1/t etc...

But as stated i will try for the PIC route so if you could suggest;

1)what PICs look at first and why

2) how to set the time/counter from an operators point of view i.e. from switches on a box that operates the required count/time. im just unsure on whether i need BCD counters or decimal counters.

i.e. http://uk.rs-online.com/web/p/thumbwheel-switches/4250136/ or http://uk.rs-online.com/web/p/thumbwheel-switches/4250120/

any thoughts

 

2) how to set the time/counter from an operators point of view i.e. from switches on a box that operates the required count/time.

A thumbwheel switch is very intuitive. You literally dial in the time you want. You'll have to decide whether to have many dials to cover your whole range or maybe some sort of range switch, so that the dial is displaying units of, say, 10µs per digit at one switch setting and ms or 0.1s at other settings.

 

Thumbwheel switches take me back to the 1970's, when customer wanted me to convert from thumbwheel sw's to key pad. How many discrete steps are desired?
As for a hard wired counter, BCD, Binary or decade, might pick decade such as a 4017 IC, as all outputs are already decoded; with 7 , any count from 10 μs to 10s with decade sw's & AND gate can be realised. Say set a FF at start & reset at selected count.

 

Thumbwheel switches take me back to the 1970's, when customer wanted me to convert from thumbwheel sw's to key pad. How many discrete steps are desired?
As for a hard wired counter, BCD, Binary or decade, might pick decade such as a 4017 IC, as all outputs are already decoded; with 7 , any count from 10 μs to 10s with decade sw's & AND gate can be realised. Say set a FF at start & reset at selected count.

i was thinking of having 3 thumbweel switches and a seperate switch that will either move the decimal place to the left or right i.e start at 0.001mS next is 0.0001mS etc... or it will either make the 3 switches operate in x10-3, x10-6 and x10-9 ranges.

i suppose it doesnt matter how precise the timing is i.e. 1.2525mS is not really needed, straight forward times like 1.2Seconds or 0.36mS should be enough.

im finding it hard to find a decent tutorial on a 4017, youtube sometimes can be a gem for this but im not seeing alot of things out there that make a lot of sense. you say the outputs are already decoded, what does that mean?

thx for your reply!

 

Try putting 4017 in " find box " this page, upper right, or try Bill's Index, same place.
4017 is a Johnson ring counter with 10 outputs. At reset, 0 is hi, at each count all other stages go high in sucession. Cascading 4017's gives a divide by 10 for each IC, so with a 100 kHz clock & 7 IC's , every possible time from 10 μs to 9.999,99 s is available. 7- 10 position Dip Sw's & a 7 in And gate " should " give any time desired. Lower order times can be left at 0. It would not be amis for all of this to be reviewed .