Looking for a Low Cost On Delay Timer

On Tue, 20 Aug 2013 14:12:28 -0700, "Guv Bob" wrote:

--- For way under $20, this'll work, and if you need NC contacts, all you'll have to do is put either a solid state relay or a mechanical relay with a catch diode across the LED.

Version 4 SHEET 1 880 680 WIRE 0 -48 -128 -48 WIRE 160 -48 0 -48 WIRE 256 -48 160 -48 WIRE 352 -48 256 -48 WIRE 400 -48 352 -48 WIRE 528 -48 480 -48 WIRE 528 -16 528 -48 WIRE 0 0 0 -48 WIRE 160 0 160 -48 WIRE 352 0 352 -48 WIRE 528 80 528 48 WIRE 256 96 256 -48 WIRE 0 112 0 80 WIRE 224 112 0 112 WIRE -128 128 -128 -48 WIRE -96 128 -128 128 WIRE 0 128 0 112 WIRE 0 128 -32 128 WIRE 352 128 352 80 WIRE 352 128 288 128 WIRE 464 128 352 128 WIRE 160 144 160 80 WIRE 224 144 160 144 WIRE -128 176 -128 128 WIRE 0 176 0 128 WIRE 160 176 160 144 WIRE 528 256 528 176 WIRE -128 288 -128 256 WIRE 0 288 0 240 WIRE 0 288 -128 288 WIRE 160 288 160 256 WIRE 160 288 0 288 WIRE 256 288 256 160 WIRE 256 288 160 288 WIRE -128 320 -128 288 FLAG 528 256 0 FLAG -128 320 0 SYMBOL Comparators\\LT1017 256 128 R0 WINDOW 0 27 24 Left 2 WINDOW 3 12 52 Left 2 SYMATTR InstName U1 SYMBOL npn 464 80 R0 SYMATTR InstName Q1 SYMATTR Value 2N3904 SYMBOL res 336 -16 R0 SYMATTR InstName R1 SYMATTR Value 10k SYMBOL res 496 -64 R90 WINDOW 0 0 56 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName R2 SYMATTR Value 680 SYMBOL LED 512 -16 R0 SYMATTR InstName D3 SYMATTR Value NSCW100 SYMBOL res 144 -16 R0 SYMATTR InstName R3 SYMATTR Value 100k SYMBOL res 144 160 R0 SYMATTR InstName R4 SYMATTR Value 150k SYMBOL cap -16 176 R0 SYMATTR InstName C1 SYMATTR Value 0.47µ SYMBOL res -16 -16 R0 SYMATTR InstName R5 SYMATTR Value 10meg SYMBOL voltage -128 160 R0 WINDOW 3 24 96 Invisible 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V1 SYMATTR Value PULSE(0 12 1 10ms 10ms 5s 10s) SYMBOL diode -32 112 R90 WINDOW 0 -38 31 VBottom 2 WINDOW 3 -33 37 VTop 2 SYMATTR InstName D2 SYMATTR Value 1N4148 TEXT -162 344 Left 2 !.tran 20 startup

-- JF

seconds. Then output will go back off and stay off until the input goes to zero. Then cycle starts again.

with NC contacts. Tons of stuff $100+, but for this application, it has to be under $20.

Thanks John, I'm not following you. Can you sketch the circuit?

seconds. Then output will go back off and stay off until the input goes to zero. Then cycle starts again.

one with NC contacts. Tons of stuff $100+, but for this application, it has to be under $20.

Thanks again! Would this work?

NTE955M Timer, (555 Type), 200 mA Iout, CMOS 8 DIP 1.20

snipped-for-privacy@4ax.com...

snipped-for-privacy@4ax.com...

onds. Then output will go back off and stay off until the input goes to ze ro. Then cycle starts again.

ne with NC contacts. Tons of stuff $100+, but for this application, it has to be under $20.

I was going to say this sounded like a perfect fit for a 555 timer. (So what's the best RC combo to get 5 seconds from a 555?)

George H.

seconds. Then output will go back off and stay off until the input goes to zero. Then cycle starts again.

finding one with NC contacts. Tons of stuff $100+, but for this application, it has to be under $20.

mechanical

Vcc)

I was going to say this sounded like a perfect fit for a 555 timer. (So what's the best RC combo to get 5 seconds from a 555?)

George H.

I used to know all that stuff by heart, but I've been swapping modules too long. This not the usual rush job, so it's a nice change to be able to spend some time on it. Now I'm off to re-learn all the RC stuff.

Maybe like this:

in------+-------------+ | | | . . . .|. . . . | . VCC(8) . | . . +-----+-------RES(4) OUT(3)------ out _|_ | . 555 . D1 A [R1] +---TH(6) DIS(7)-- 'T` | | . . +-----+---+---TR(2) CV(5)-- | . . C1 ===== . GND(1) . | . . . .|. . . . | | ----+-------------+--- ground U1 LMC555 / 7555 D1 1N914 C1 10uF R1 470K

Grin.. sure, that's why I asked. I don't really know 555 timer specs.

So are Jasen's 10 uF and 500k a good choice? (I assume the 10uF could be an Al electrolytic... though I do have big 10uF film caps... spendy.)

George H.

modules too long. This not the usual rush job, so it's a nice change to be able to spend some time on it. Now I'm off to re-learn all the RC stuff.

That was my 2nd guess. LOL!

In industrial relay timer parlance, we'd call that an "interval timer". An "off delay" timer would drop out a period of time after the power was removed (a more rare beast since it would have two power supplies or a big capacitor inside). An interval timer is/was just an on delay timer with logic inversion before the relay driver, so it pulled in immediately upon connection to the power source and drops out after the delay.

The terms are used differently in different domains so I wouldn't count on them being valid in any particular instance. For example, some cretinous poltroons will refer to a recycle timer as an "interval timer".

Best regards, Spehro Pefhany

They're okay.. depends on the type of 555- CMOS or the old bipolar type. 5 seconds is short enough that almost anything will work. A low leakage type electrolytic of 10uF to 100uF would be cheapest in volume, for one-off, maybe a CMOS 555, a ceramic 1uF with a 5M resistor but voltage coefficient might come into play. A tantalum 10uF with 470K should work without problems.

There are two impediments to accuracy with this sort of thing- the tolerance/leakage of the cap, resistors and chip (mostly the cap) and how precisely the cap voltage is reset to zero. If, say, 4-6 seconds delay with a reasonable minimum off time is okay, then you don't need to worry much about this sort of thing.

In the line of industrial timers I designed a couple careers ago, I would clamp the cap to within some mV and used a relatively high voltage for the comparison (about 20V, IIRC) so the accuracy was excellent even if the reset time was relatively short. They used all discrete components, the 555 was available but I could see no advantages at all in using it (I did manage to make a PI temperature controller for bench top hydraulic presses with a 555 as the only active component).

If you need more like 1-5% accuracy, the resistor can be trimmed to compensate for the cap tolerance. If you're building a lot of them, a microcontroller with either a calibrated RC oscillator (1%-ish accuracy without adjustment) or a crystal (accuracy probably limited by the reset circuit) would be better.

Best regards, Spehro Pefhany

These used to be called "Delay Relays" and looked like a vacuum tube. Now there are solid state models. See .

You can find these on eBay, i.e.

You should be able to figure out a way to get what you need though you may need another regular relay as a latch.

Well under $20.

e an Al electrolytic... though I do have big 10uF film caps... spendy.)

Nice thanks, I've been sorta penciling tantalums out of my design brain. Though I did use a big one ~100uF for a triangle wave thingie years ago. The one 'gotcha' you left out was a clean voltage reference for the compara tor. I semi-screwed up the above triangle wave generator and used the posi tive supply rail as the reference. There was this little occasional 'hick

-up' in the frequency when a 'bang-bang' heater circuit in the same instrum ent would turn on and drop the positive rail by some ~10's of mV.

George H.

Yup.. even though the value of the supply voltage cancels out (if stable), any noise or other variations in the reference/charging voltage during a cycle will result in shorter (usually the problem) or longer time. So you usually will want to use the same regulated voltage for both RC charging and for the voltage reference.

It's tempting to try to coax a bit more time out of a given value of (expensive) film cap by pushing further out on the RC curve, but that increases noise sensitivity, so it's probably best to stick to around one time constant unless you've carefully designed around that kind of requirement.