switch debouncing

Hey F B, you are getting close to something really good! Now parallel C with an R3 to set any hysterias desired. Used this many times, works like a champ. Regards, Harry

Thumbing thru my old notes: If you tie Threshold and TriggerBar on a

555 and drive from a switch thru an RC you get tremendous bounce rejection.

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
I love to cook with wine.      Sometimes I even put it in the food.

What sort of response time do you need on a wheel switch?

With 10KV transients (where do you get them on a plane?) I'd be more worried about the resistors, R1 and R2, arcing over. That's an ESD-type issue, distinct from debouncing.

C, being ungrounded, no longer protects the semiconductors from ESD zaps.

I'd also worry about G1 latching up, especially at high temp, since G2 is going to overdrive its substrate diodes pretty hard. Some HC-family parts can be made to latch (parasitic PNPN scr, Vcc to ground) at modest currents, and that will fry the chip and possibly drag down Vcc of the whole system. Or an ESD arc across R2 might latch either gate.

A healthy high-frequency transient might well pump two or more edges through G1 before the G2 hysteresis can respond. Guaranteed, actually.

Overall, looks like more problems created, or at least needing to be carefully analyzed, than solved. I'd be surprised if this circuit were used on flight hardware.

As you add parts, the number of possible failure modes increases roughly factorially.

John

27.5? Speed kills!

The real pita is to have a pair of up/down buttons that have to jog a value (or a bunch of different variables) over a huge range, like

+-4095 or something. So a single push jogs up/down one tick, and if you hold it down it begins to count at an increasing rate. Damned nuisance to program!

John

| R2 | |/ | | +----| | |

I'd add D1 to discharge the cap when you hit S1, so you don't get a leisurely response. :-) Also, I don't think you need a whole

1 uF there - ISTR seeing these down to .001, but maybe that was with CMOS, and bigger resistors.

Cheers! Rich

I had some 4095 range button things with a Moto 68HC05. Initial button press counts 1. Hold for .5 sec and repeats at 10/sec. After 2 seconds it goes to 200/second and wasn't bad to program and more importantly, it is easy to use. 60 seconds after the last button press, the values are written to a serial EEPROM. The long delay ensures a minimum of write cycles without losing too many settings if power fails before the save. GG

I had some 4095 range button things with a Moto 68HC05. Initial button press counts 1. Hold for .5 sec and repeats at 10/sec. After 2 seconds it goes to 200/second and wasn't bad to program and more importantly, it is easy to use. 60 seconds after the last button press, the values are written to a serial EEPROM. The long delay ensures a minimum of write cycles without losing too many settings if power fails before the save. GG

Sure it does, or are you blind drunk ?:-)

...Jim Thompson

-- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona Voice:(480)460-2350 | | | E-mail Address at Website Fax:(480)460-2142 | Brass Rat | |

| 1962 | I love to cook with wine. Sometimes I even put it in the food.

There's an even simpler way. Lose R2. :-)

Cheers! Rich

Then again, all you've said is "a switch". If you can use a(an?) SPDT then you'd use an R-S flip flop. The logic should be trivial to figure out. ;-)

Cheers! Rich

...

Or 27.5. ;-) I'd go nuts trying to do an ASCIImatic at 5 or 10 CPS. Yes, I know I'm already crazy - you know what I mean. :-)

Cheers! Rich

...

Yeah - like a whole 18 ms to "close". I think I'd start noticing around

50-100 ms - that'd be an interesting experiment ... hmmmm....

Cheers! Rich

Ok John, If your switch closure is on the front wheel of a AB360 (Airbus) with 100s feet of wiring thru power cables with 10KV transients then a switch receiver with 1.0V of forward hysterias and +/-5V of transient hysterics and the flying cap acting as a low pass filter when the output is in the high or low states requires some parts. I have not seen a switch receiver more tolerant of noise. What say you JL?? Regards, Harry

>

I suppose it is possible for that arrangement to become a delay line oscillator at something like 30MHz, so this is fixed by limiting the slew rate across the gate's input capacitance to delay much greater than the composite gate delay. Your 5ms switch should work well with this circuit, same waveforms apply: View in a fixed-width font such as Courier. . . . . . . . VCC 74HC14 . | . [68K] G1 . | |\\ . +-[39K]-+-[220K]-| o---+---->

. | | |/ | . o | | . -| === | . o |0.47U G2 | . | | /| | . --- +--------o |---+ . /// \\| . . .

Fred Bloggs wrote:

It doesn't look anything like your oscillator and it uses Schmitt triggers and not HC04's.

I don't see why that can't oscillate at 1/(2Tpd) if it got started for some reason unless there is something to do with a minimum pulse width, finite transition times, and the location of VT- and VT+ in the input range.

Something to do with N*180d for any N. Odd Ns make oscillators, even Ns make latches. N is the number of inversions. This circuit has N=2 so it acts like a latch. I sometimes implement with a single stage non inverting comparator with input hysterias and rail to rail outputs. N=0, works like a charm, no oscillation. Cheers, Harry

G3 and the 150K are added to a AC-hysteretic circuit to make an oscillator.

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
I love to cook with wine.      Sometimes I even put it in the food.

Your no-clip-osc is close but that third output inverter simulates flipping a bounceless switch at the instant G1 and G2 change state: View in a fixed-width font such as Courier.

. . . NO CLIP OSC . . +---------------------------------------+-->

. | | . | G1 | . | |\\ | . +--[150K]--+---| o------------+ | . | |/ | G3 | . | | |\\ | . | +-------|----| o--+ . | | G2 | |/ . | | /| | . | +-o |---+ . | | \\| . | [2.7K] . | C | . +---||-----+ . | . [2.7K] . | . --- . /// . . .