Help me solve this design problem PLEEEZ

There is an electromagnet at my job that has a 50% duty cycle. That is it may be on for up to 5 minutes and then off for 5 before using it again. I want to design a circuit that acts as a time integrator. As long as a relay is pulled in, the timer is counting up. When the relay is out, the timer counts down to zero and stays there. If the operator isn't paying attention, and he goes over 5 minutes, an alarm sounds to alert him that he is operating the electromagnet outside its design parameters. How can this be accomplished? Any suggestions would be appreciated.

Reply to
funster
Loading thread data ...

You need an up-down counter and a slow clock - a 32,768 Hz source as used in watches might serve as the clock, and counting five minutes of that needs a 24-bit counter, for which you could use the LSI LS7166

24-bit multimode counter (available ex-stock from Farnell). Multi-mode includes up-down ....

Of course, any single chip microcontroller could handle the job - the Microchip Technology PIC series comes to mind (and Farnell have pages of variants for you to chose from) though the Atmel AVR parts are reputed to be nicer to work with. For this application you'd presumably clock the microcontroller at 32.768kHz.

------------ Bill Sloman, Nijmegen

Reply to
bill.sloman

Five minutes is too long for anything but a CMOS 555, and even then the impedance levels are so high that you'd be wise to mount all the timing components on Teflon/PTFE stand-offs.

Assembly language programming isn't all that difficult, anyway.

---------- Bill Sloman, Nijmegen

Reply to
bill.sloman

Get yourself a microcontroller and learn how to use it? Or take an osc with a chain of up/down counters, alarm connected to the overflow? Or take an old PC, use the serial or parallel port as input, and write the application any way you like?

Wouter van Ooijen

-- ------------------------------------

formatting link
Webshop for PICs and other electronics
formatting link
Teacher electronics and informatics

Reply to
Wouter van Ooijen (www.voti.nl

How about a bank of 74HC190 up/down counters? CLock it at one Hz, which you could get from a 4060 and watch xtal. You'd have to strap the first two to count 0-59 and reset, for minutes, then just decode five minutes and do the alarm thing. When they deenergize the magnet, the D/U' input gets switched, and it counts down. Add a little logic to stop at 0.

Good Luck! Rich

Reply to
Rich Grise

--
5 minutes is a little long for a 555, and 24 bits is a little much,
IMO, but a successful marriage might be a single 555 astable with a 20
 Click to see the full signature
Reply to
John Fields

Or, of course he could do the dual 555 thing. One timer for "on time" and one for "off time", which added with some logic gates allows the first timer to work and produce an error condition. Sonalert module for audio alert. No 3 months required to learn PICs.

Reply to
Mark Jones

Thanks all... Wow... the talent out there is amazing. And to think that my company was quoted from an engineering company to do the job for $30,000. The circuit I named was just the most basic one possible though. They would have added a recorder and a signal sent up to supervision when the operator violated the %50 duty cycle requirement. Keep the suggestions coming though and you may have earned a promotion for me. LOL

Reply to
funster

Now you really could use a PIC. External 512K eeprom for recording and serial-rs232 out to the supervisor. All in a PIC12F629 - 1 dollar in small quantities.

Reply to
Luhan Monat

OK, let's put on our thinking caps. Why would anybody want to limit the duty cycle???

If it's heat, put a thermistor on the magnet. When it gets too hot, turn on the "coffee break" sign.

If you can't get at the magnet, make a heater/heatsink/thermistor with the same time constant. I'd suggest a snap-action thermal switch, but make sure it's rated for lots of operations. The little round jobbers probably aren't.

Never infer a condition when you can (easily) measure it. mike

--
Return address is VALID but some sites block emails
with links.  Delete this sig when replying.
 Click to see the full signature
Reply to
mike

John Fields wrote: snip

snip

snip Hmmmmm Sounds like somebody's not getting enough fiber... mike

--
Return address is VALID but some sites block emails
with links.  Delete this sig when replying.
 Click to see the full signature
Reply to
mike

Bwahahahahahaha!

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

Google "timing relay." Like

formatting link
Look for the one you need, it might be the "delay dropout" type but that's up to you to figure out. Buy, connect; garner praise for achieving inexpensive, effective solution.

Reply to
kell

--
"make a heater/heatsink/thermistor with the same time constant"? Good
luck...  Another self-styled expert raises his ugly head and proposes
 Click to see the full signature
Reply to
John Fields

And a spare input to monitor if someone flushed the toilet, and sent that information to the supervisor as well ;)

--
Thanks, Frank.
(remove 'q' and 'invalid' when replying by email)
Reply to
Frank Bemelman

What's the problen with 24 bits? You can buy a 24-bit counter in a single package off the shelf, or program it in a medium-sized PLD. Itt would need triple precision arithmetic in a an 8-bit micro, if most of them didn't have 16-bit counters built in.

It's not like the bad old days, when 24-bits was 6-four bit ripple counters strung together ......

-------------------- Bill Sloman, Nijmegen

Reply to
bill.sloman

This is not really a "design" problem yet. What voltage is being switched onto that solenoid and what kind of access do you have to it? And you have several things going on here like that 50% maximum limit applies for on times of 5 minutes but can be higher for lesser on times? Or do you want a timer that locks out the solenoid for a duration equal to the preceding on time? Or is the timer to alarm when the operator exceeds previous off time or 5 minutes whichever comes first? If duty cycle is the main thing then a simple timer will not be enough.

Reply to
Fred Bloggs

The counter idea as mentioned in a number of the above posts is OK and can easily be implemented in a PIC or other micro or in IC logic as several have stated. However, here is another idea that you might consider: Instead of measuring time, why not measure temperature and run the coil for how ever long it takes to get to a max temp, maybe 60deg C. and then keep it off until it returns to a lower temp, maybe 30 or 40 deg C. That way you would maximize its duty cycle and get more out of when in a cool environment and not overstress it when hot. Controlling time does not give that flexibility. I would use the copper winding as the temperature sensing element by measureing both the voltage and the current in the coil. The resistance of the copper would be calculated every second or so in a microprocessor and the temperature calculated or looked up in a table from the resistance and its TC which I believe changes 0.39% per deg C. A PIC with two A to D converters built in could easily do this. The A to D's are hooked up with appropriate scaling resistors to measure the current and voltage. The output of the micro would be the same as in the other schemes. It's just a thought, probably worth about what you're paying for it. Bob

Reply to
Bob Eldred
5 minutes is a little long for a 555, and 24 bits is a little much, IMO, but a successful marriage might be a single 555 astable with a 20 second period and a 4 bit up-down counter.

Then, >What's the problen with 24 bits? You can buy a 24-bit counter in a

--
No problem, just pricey.  $49.78 from Farnell, (if they had any) plus
the crystal and the oscillator and the glue logic.
 Click to see the full signature
Reply to
John Fields

And what about the learning curve? Does the OP have any experience with PICs? How long is it going to take to get that chip programmed and debugged successfully? Sure PICs are handy, but they are not simple "jellybean" parts. We all know that you know everything John, but the rest of us have to work hard at it.

Reply to
Mark Jones

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.