[Apheared]

Hey,

This is part of a thread that existed at my other home, but now... anyway...

Imagine a 555 timer that controlled an electric chair that you personally sat in. No, that's not the application but I need it to be on your mind.

This 555 timer sometimes triggers on power up, sometimes on power off. It's running off a power supply that's well regulated (<1mV ripple) and locally bypassed; it behaves like it should, except for this annoying habit of triggering when power is applied/removed, and only sometimes.

For some applications this is just annoying, but what do do if it's something that can kill someone or destroy multi-thousand dollar equipment?

Where do you learn the skills and techniques to build devices that will never trigger when they aren't supposed to? I'm talking like medical equipment certification techniques... imagine if a implanted defibrillator just started shocking the crap outta you constantly because of a failure? That kind of assuredness.

Help?

 

[aeberbach]

Do you have all the pins tied to something? Could a pull-down resistor on the trigger pin help?

Can you post a picture of the schematic maybe?

 

[rickcr42]

I can throw you a couple of links but the rickster is fried from reading data sheets all day,I can't even focus on the question

(lesson kiddies,information overload leads to zero learning curve,better to step away and reload

)

http://www.westminster.org.uk/intran...ronics/555.htm all about the 555 timer including software

http://www.epanorama.net/lights.html some control circuits,maybe some useful information

If I can focus I will see what I can find , but all those false triggers sounds like a noisy layout,simple ferrite beads may cure it.Put them on all the V+ leads of each active device.A possible cheap solution

 

[thomas]

Does your circuit have its own on/off switch, or do you simply control it via the power supply?

I sometimes get wiered effects with logic circuits if it is switched on/off at the power supply. As the PS caps charge up/get discharged, the voltages are really crazy, and you often get wierd outputs...So make sure you power up your supply for a few seconds first to allow voltages to stabilize, then switch on your circut.

Or how about adding an AND gate to the output, that will only enable the output when you tell it to? why not leave the timer running all the time, and just enable the ouput at the AND gate when you need it?


Finally, check out the Ti NE555's datasheet, they have a few suggest decoupling pin5 to ground as well...

And although this doesn't sound like your problem, you really should bypass all logic chips, as they can add large spikes to the power rails whenever they switch...

 

[Apheared]

Yea get this; I wanted to control an AC motor with one in monostable pushbutton... this worked like gangbusters once I found the info I needed; 555 to optoisolator to snubbed triac, badda-bing badda-boom. Hasn't false triggered once. (that I've noticed)

So now I'm trying to control an AC reciprocating pump and an AC motor separately but in stages; this was my O/T flip-flop question on HeadWize... right now I've built the same ps as the first, it's rock stable... I've got the output of one timer cascaded to the input of another trying to stagger them on HI/LO pairs to rig up a timed on/off/on/off config... right now it's just a visible LED for load, not tryin to mess up my apartment with a pump that powers itself up...

edit

Quote:

I sometimes get wiered effects with logic circuits if it is switched on/off at the power supply. As the PS caps charge up/get discharged, the voltages are really crazy, and you often get wierd outputs...So make sure you power up your supply for a few seconds first to allow voltages to stabilize, then switch on your circut.

Ah. Well, I'm not gonna do that. What is this, a tube amp? Do I need to power up my main supply first and let it stabilize? fsck that.

Hmm... maybe a way to do it so that the caps slowly charge up, like the "slow turn-on" schematics in LM317 datasheets?

 

[rickcr42]

Ah Ha ! Switching reactive loads with the triac dude ! Follow the triac with a mechanical relay with suppression caps of course,should do ya

 

[thomas]

hmmm, how about using a few j-k (master-slave) flip flops? not sure exactly what you mean, but you can use it to divide your clock frequency in half. And you'll always have an inverted and a non-inveted output, so you can get whole number of clock configurations... leave the 555 running all the time, and use a pushbutton that toggles another j/k, connected to the preset/clear pins of the first j/k's to enable/disable the output...

 

[rickcr42]

I was thinking along that track earlier thomas but my thought is,with the false triggers it would still activate the flip-flops unless they were mode locked in some fashion,but then the auto sequence would not work unless manually triggered.Or maybe I'm just too tired and I am WAY WRONG

 

[thomas]

oops, i'm the one that's way wrong

spend all day checking out universities, too tired to think

yeah, the auto sequence would be messed up by the method i stated.

but, if you locked the clock input of the j/k with the AND gate i mentioned originally, there shouldn't be any problems with the sequence...

And its much harder to get a logic gate to false trigger than a flip flop!

 

[rickcr42]

Kind of off on a different tangient-maybe ****can the 555 timers and go with shift registers and cmos clock,just an option

 

[Apheared]

guys, the first one works fine; and the 2nd one doesn't have a real load yet - just LEDs. I unwired the optos and triacs after noticing this crap, working backwards to figure out where it's goin nuts...

I'm using Fairchild MOC3032s optos, Fairchild 555s with bypasses, and ST's BT "snubberless" 16A triacs for inductive loads... even though I still use gate resistors and the tie-across (err, crosstie?) resistors. The 12V power supply is a split bobbin -> full bridge -> 3300/.33 -> 317 -> 10/.1 (with protection diodes)

It's me, I have no idea what I'm doing. And I can't find much about using timers to control a load on/off at specific intervals... you can find examples using triacs for motor speed control and lighting control... but I need to control precisely at least 6 on/offs, with zero fu*kups. triggering the pump twice w/o triggering the motor would be like shifting without the clutch while coasting downhill... (do youngins get that analogy?

Also, does anyone know what's the deal with "low leakage" electrolytics? I mean even with these, at timing intervals of 3-4 minutes... if you need exactly 3:44.50 good luck right? So how do you get precise for longer times that make film caps impossible to use? Cascade more also? ehahe I could end up with 30 555s which is why I'm wondering if I should just bail in general and go uC where it can just sink for the opto as programmed.

Thx for the links rick, those were fun reading anyway... even though not really on target.

 

[thomas]

hmm, never seen anyone cascade 555 timers but what do i know

(but wait, i'm not that young!

)

why don't you cascade j/k's instead... connect j +k inputs to Vcc, and the 555 to the clock input. I don't know exactly what's the lowest frequency you can get without using electrolytics (or how about a $20 black gate to control your motor

) but i don't think you;d need more than 10. And since there's 2 per package, it won't be too bad.

To control the motors in the correct sequence, try looking at the design of a data selector (also used in some multipexers). Basically, it uses a multiple-input AND gate, and an inverter at some of the inputs- for example, if you want it to trigger on 0101, you'd use a 4-input AND gate, and an inverter at the first and third bits. That way, the AND gate will only get an input of 1111,(only thing that will cause it to trigger) when the circuit is really getting 0101...

If you need 6 controls, just use 6 of these data selectors, each one configured so it will trigger with the right input...

and if you use the J/k's, you don't even need the inverters, as there are already inverting and non-inverting outputs. for really high precision at long times, simply use a huge data selector connected to each jk's output...

Hmm, a microcontroller could also be used, do you have the stuff to program them? if you do, you might as well use one, it'll probably be cheaper than using tons logic gates. But its kinda overkill to use one for a timer!

 

[Joobu]

For the power up issue, why not have some sort of brown-out detection. When the power is too low, the system is put in a safe reset state. Only when the power has been high and stable for a while does the system start up again.

For precise timing (milliseconds) I'd probably start with a stable reference (crystal) and use some sort of digital counter from there on. This could either involve logic counters or as others have alluded to a microcontroller.

I'm not sure about how to implement this in logic but a microcontroller would be relatively simple and predictable. Either brute force loops or using timer interrupts. The Atmel series of microcontrollers are pretty nice for this scale of project.

If you are going to do MCUs I suggest an evaluation kit like the STK500. For ~100$ you get a board for programming and testing MCUs before plugging them into your application.

 

[no_iq]

Apheared,

I'll put in my $.02 if you don't mind.

I've worked on medical equipment for 25 years now. Here's one solution I've seen, it's a primitive digital filter.

See if you can measure the length of the 555 shot when it misfires. Build another 555 1 shot that has a period of at least that length. Put both outputs into an AND gate (maybe a NAND, depends on your polarities). The output of the gate should never fire UNTIL you press the button for a period which is LONGER than the misfire.

Mye explanation may be weak, but I think you can get the idea.

no_iq

 

[aos]

For timing circuits use the tantalum capacitors. They do have very low leakage. People may not like to use them for audio, but this is not audio.