Maker Pro
Maker Pro

Momentary Switch Circuit

I have a device that requires a momentary switch closure, however I
need to activate it using a toggle type switch that will remain closed.
I am hoping to find a circuit that will provide the momentary closure
when the toggle switch is first closed and then not again unless the
toggle switch is opened and then closed again.

I am inexperienced with electronics and need some direction. I thought
adding a resistor and cacpacitor might achieve what I need by allowing
current to flow while the capacitor is charging and then cutting off
when it is fully charged. I don't know if this will work or how to
reliably discharge the capacitor to reset for the next switch close.

I also looked at using a one-shot (monostable multivibrator) but they
appear to trigger continually if the input signal persists. Maybe some
additional logic could overcome this. I also wonder if an edge
triggered one-shot would avoid the problem.

I would appreciate any suggestions.
 
J

Joel Moore

[email protected] wrote in @g44g2000cwa.googlegroups.com:
I have a device that requires a momentary switch closure, however I
need to activate it using a toggle type switch that will remain closed.
I am hoping to find a circuit that will provide the momentary closure
when the toggle switch is first closed and then not again unless the
toggle switch is opened and then closed again.

I am inexperienced with electronics and need some direction. I thought
adding a resistor and cacpacitor might achieve what I need by allowing
current to flow while the capacitor is charging and then cutting off
when it is fully charged. I don't know if this will work or how to
reliably discharge the capacitor to reset for the next switch close.

I also looked at using a one-shot (monostable multivibrator) but they
appear to trigger continually if the input signal persists. Maybe some
additional logic could overcome this. I also wonder if an edge
triggered one-shot would avoid the problem.

I would appreciate any suggestions.

I'm sure there's a 555 timer circuit that will do exactly that. The only
reason I can't say for certain is because I've never made one.

Check out some of the circuits that come with Schematica's 555 timer
design software (http://www.schematica.com/555
_Timer_design/Timer_Comparison.htm). It's not freeware but the
evaluation will allow you to play with several circuit types.

Alternatively you could wait another 10 minutes and a helpful soul in the
group will probably post an ASCII schematic for you.

Lastly a properly contructed Google search would probably turn up many
solutions.

Joel Moore
 
D

Dan Hollands

An RC circuit can certainly generate a pulse but to know if it will do the
job more info is needed about the application. What are you trying to
trigger, how long does the pulse have to last, are the requirements for rise
or fall time of the pulse, what kinds of voltage and currents are involved?

Dan

--
Dan Hollands
1120 S Creek Dr
Webster NY 14580
[email protected]
www.QuickScoreRace.com
 
B

budgie

I have a device that requires a momentary switch closure, however I
need to activate it using a toggle type switch that will remain closed.
I am hoping to find a circuit that will provide the momentary closure
when the toggle switch is first closed and then not again unless the
toggle switch is opened and then closed again.

I am inexperienced with electronics and need some direction. I thought
adding a resistor and cacpacitor might achieve what I need by allowing
current to flow while the capacitor is charging and then cutting off
when it is fully charged. I don't know if this will work or how to
reliably discharge the capacitor to reset for the next switch close.

I also looked at using a one-shot (monostable multivibrator) but they
appear to trigger continually if the input signal persists. Maybe some
additional logic could overcome this. I also wonder if an edge
triggered one-shot would avoid the problem.

I would appreciate any suggestions.

There are a number of non-retriggerable monostables that would meet your
requirement - 74XX221 is a dual that I use often.

There is also a "standard" circuit for this with a 555, but I haven't used one
in over a decade.
 
I am triggering buttons on a device that simulates a USB gamepad on
PCs. Specifically the Ultimarc A-PAC. It is often used for building
custom controllers for arcade emulators. I am using it to interface
the stick, buttons, and switches on an RC airplane radio to a PC so it
can be used with the FMS RC flight simulator.

The A-PAC button terminals need to grounded to activate the switch. A
want the radio toggle switches like on/off and high/low rate to become
momentary button presses on the controller. I think something like
1/10 of a second should be adequate to register the button press. I
don't think the rise and fall time is critical here.

I don't have any specification about how much current will flow from
the button terminal to ground upon closure. Would grounding the button
through a multi-meter tell me this accurately?

Kevin
 
I thought a non-retriggerable was what I needed until I read it only
prevents retriggering during the output pulse. I need to prevent
retriggering after the output pulse because my input will remain pulled
low. I only want to trigger again after the input switch goes open and
then grounded again.

I see some possibilities based on the 555 in Schematica that Joel
referred to. I will try those but in the final product I want to use a
newer timer with duals or a quad because I have four switches to
connect.

The newer timers seem to use edge triggering instead of level
triggering. I am not certain edge triggering will eliminate the need
to worry about my persistant input causing additional triggering after
the output cycle, but I guess it will.
 
E

ehsjr

I have a device that requires a momentary switch closure, however I
need to activate it using a toggle type switch that will remain closed.
I am hoping to find a circuit that will provide the momentary closure
when the toggle switch is first closed and then not again unless the
toggle switch is opened and then closed again.

I am inexperienced with electronics and need some direction. I thought
adding a resistor and cacpacitor might achieve what I need by allowing
current to flow while the capacitor is charging and then cutting off
when it is fully charged. I don't know if this will work or how to
reliably discharge the capacitor to reset for the next switch close.

I also looked at using a one-shot (monostable multivibrator) but they
appear to trigger continually if the input signal persists. Maybe some
additional logic could overcome this. I also wonder if an edge
triggered one-shot would avoid the problem.

I would appreciate any suggestions.

Simple.

+Vcc ---o o---+---|<---- Vout
\ | Zener
o |
| |
[C1] [R1] 220K
| |
Gnd ------+-----+

When the toggle switch is off, the capacitor is connected
to +Vcc, and charges quickly. When the toggle switch is on,
C1 discharges slowly (relatively) through the Zener into
your circuit, and also through R1. When the charge on C1
drops below the zener voltage, it no longer has a path
to your circuit, so your circuit "sees" a pulse that slopes
from Vcc to Vzener. C1 continues to discharge through
R1 slowly.

Select the zener close to, but below Vcc.
For example, a 4.7 volt zener with a 5 volt Vcc,
an 11 volts zener with a 12 volt Vcc. A transistor
can be added for higher current Vout, if needed.

The larger the value of C1 and R1, the longer the
Vout pulse. The values depend upon your circuit,
but likely can be a broad range. Try whatever
you have on hand - just be sure the voltage rating
for the cap is well in excess of Vcc - at least 30
percent higher.

Ed
 
B

budgie

I thought a non-retriggerable was what I needed until I read it only
prevents retriggering during the output pulse. I need to prevent
retriggering after the output pulse because my input will remain pulled
low. I only want to trigger again after the input switch goes open and
then grounded again.

I see some possibilities based on the 555 in Schematica that Joel
referred to. I will try those but in the final product I want to use a
newer timer with duals or a quad because I have four switches to
connect.

The newer timers seem to use edge triggering instead of level
triggering. I am not certain edge triggering will eliminate the need
to worry about my persistant input causing additional triggering after
the output cycle, but I guess it will.

The 74XX221 is an edge-triggered non-retriggerable mono (x2). The input
transition which triggers it will not cause more than the one output pulse.
 
A

amdxjunk

I posted a simple schematic on
sci. electronic design. Check it out.
 
A

amdx

I'm sorry I got it backwards, it should have been
alt.binaries.schematics.electronics.
Mike K.
 
E

ehsjr

Here is the solution I ended up with. I found it at
http://home.cogeco.ca/~rpaisley4/LM555.html. I mocked it up and I get
an approximate 1/10th second pulse even when the input switch remains
closed. Please pardon me if the ASCII art isn't quite standard
notation. I am new at this.

+Vcc --+---------+------+---+
| | | |
| | | [R1] 1K
[R2] 100k [R3] | |
| | | +---+-----------+
+--[C2]---+ | | | |
| 0.1uF | +---+---+---+---+---+ |
| | | 8 7 6 5 | |
\ | | LM555 | [C1] 100uF
o | | 1 2 3 4 | |
| | +---+---+---+---+---+ |
| | | | | |
| +----------+ +----------------
| | | Output
Gnd --+----------------+-------------------+----

Thank you all for the help.


I'm curious. Why didn't you try this?

+Vcc ---o o---+---|<---- Vout
\ | Zener
o |
| |
[C1] [R1] 220K
| |
Gnd ------+-----+

Vzener is slightly below Vcc - eg 4.7 volts if Vcc is
5 V, 11 volts if Vcc is 12 volts etc.
Ed
 
T

Terry Pinnell

ehsjr said:
I'm curious. Why didn't you try this?

+Vcc ---o o---+---|<---- Vout
\ | Zener
o |
| |
[C1] [R1] 220K
| |
Gnd ------+-----+

Vzener is slightly below Vcc - eg 4.7 volts if Vcc is
5 V, 11 volts if Vcc is 12 volts etc.
Ed

Ed: That looked such an attractively simple circuit that I tried it
myself yesterday. I may have missed something but aren't there a
couple of downsides to it? I don't think the OP specified what sort of
pulse he wanted, but my starting assumptions were a clean, +ve, full
supply signal. Your circuit:

1) Delivers only a low amplitude spike, so needs amplifying to get a
'full' pulse. Using say a simple NPN stage that results in a -ve going
pulse, which may then need inverting. (BTW, such amplification would
presumably be simplified if the headroom voltage is rather larger than
your examples?)

2) Transmits switch noise.

I used a 14V supply and a nominal 12V 1W zener, with this circuit:
http://www.terrypin.dial.pipex.com/Images/Momentary-Ed1.gif

Here are a few screenshots of the results I actually saw:

Relatively 'clean' switch
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-1.gif

Noisy switch
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-3.gif

Noisy switch, detail
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-5.gif

Noisy switch, detail, after NPN stage
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-8.gif

BTW, I'm puzzled by the roughly triangular shape of the output. I
tried one simulation and got a squarish result:

http://www.terrypin.dial.pipex.com/Images/Momentary-Ed-SIM1.gif
Maybe it was just the choice of zener?
 
E

ehsjr

Terry said:
I'm curious. Why didn't you try this?

+Vcc ---o o---+---|<---- Vout
\ | Zener
o |
| |
[C1] [R1] 220K
| |
Gnd ------+-----+

Vzener is slightly below Vcc - eg 4.7 volts if Vcc is
5 V, 11 volts if Vcc is 12 volts etc.
Ed


Ed: That looked such an attractively simple circuit that I tried it
myself yesterday. I may have missed something but aren't there a
couple of downsides to it?

Yes, absolutely.

I don't think the OP specified what sort of
pulse he wanted, but my starting assumptions were a clean, +ve, full
supply signal.

He's replacing (or using in place of) a momentary switch.
Both the toggle and the momentary are subjecty to switch
bounce. Since whatever circuit he's driving will work
on a momentary, then switch bounce from a toggle is not
a factor.

But your observation is correct. The circuit will not
provide a clean, +ve, full signal supply. It is a definite
downside to the circuit for general use - it is good only
for use where switch bounce is irrelevant.


Your circuit:
1) Delivers only a low amplitude spike, so needs amplifying to get a
'full' pulse.

That's wierd. It should deliver close to Vcc. Your
Vcc is 14 volts and the zener is 12. It (the zener)
should conduct until Vcap drops to 12, meaning that
the pulse amplitude has to be at least 12. Maybe
you are scoping at the input to the zener? The
duration depends on the load, and to a lesser extent,
on the 220K resistor, but primarily on the zener.
It will be a spike, but it should be a spike whose
amplitude is over the zener voltage. RC is ~ 9.5 mS
(way too short) down to ~1/3 vcc - but it is collapsing
only 2 volts or about 14% (instead of ~ 63%) when the zener
shuts it off. You need a bigger cap! The circuit "looks at"
just the top of the cap discharge curve, which is steep.

We disagree on spike amplitude, but that doesn't change
the fact that the output is another downside, as you
indicated. You get what amounts to a spike instead of a
nice robust square wave.

Using say a simple NPN stage that results in a -ve going
pulse, which may then need inverting. (BTW, such amplification would
presumably be simplified if the headroom voltage is rather larger than
your examples?)

Yes - you need that with a low impedance load, or a long
duration pulse, or if the pulse must be square. And as long
as you stuff enough drive current into the base, it drives
to Vcc minus the transistor Vce. So you can have plenty of
headroom between the zener and Vcc and still get square wave
output.

I got way over 40 seconds with a 120(?) ohm relay driven
that way through a darlington with Hfe > 1000 in a delay
off circuit I made. Had to add a resitor in parallel with
the cap to get it down to the ~40 second target in that
circuit. I don't remember all the values, but I can find
them if they are of interest.

The headroom was to make the pulse voltage unambiguous.
RC is ~ 63% discharged - the circuit is "looking"
at the top ~6% of discharge (eg from 5.0 down to 4.7).
I didn't want the the circuit that is being driven
to "see" an ambiguous voltage that was not clearly
either a high or a low. By using the top of the
discharge curve, the voltage will be at a high
until the zener cuts it off. It has a wicked slope,
but all of it will be above the zener V. Also, I picked
common zener voltages. I run into lots of 11 and 4.7v
zeners, but rarely a 4.3.
2) Transmits switch noise.

Absolutely. As mentioned earlier, I don't think
that's a factor for the op, but it certainly is
in many applications.
I used a 14V supply and a nominal 12V 1W zener, with this circuit:
http://www.terrypin.dial.pipex.com/Images/Momentary-Ed1.gif

Here are a few screenshots of the results I actually saw:

Thanks for these - they're great! I have a question, inline
below.

I'm puzzled by the Vout curves. It should "fall off the cliff"
at 12V when the zener stops conducting, but the traces all show
Vout curving down below 12 volts.

I think the output (without the noise, which I can't draw)
should look like this:

+14 |\
+12 | \
| |
| |
| |
0 ---- ----

Can you determine why there is that curve down to 0
volts? I wonder if your zener doesn't "zen" :) ?
BTW, I'm puzzled by the roughly triangular shape of the output. I
tried one simulation and got a squarish result:

http://www.terrypin.dial.pipex.com/Images/Momentary-Ed-SIM1.gif
Maybe it was just the choice of zener?

Regarding the simulation - my guess is that they treat
the switch as noiseless, no bounce. But they don't
show the Vout "falling off the cliff" pattern - unless
the yellow is supposed to be Vout. The green looks
more like a capacitor discharge curve than the yellow,
but the green doesn't fall off the cliff. And if the
yellow is supposed to be Vout, the trailing edge is
correct and everything before it is wrong. So I'm
clueless.

Ed
 
T

Terry Pinnell

ehsjr said:
Terry said:
I'm curious. Why didn't you try this?

+Vcc ---o o---+---|<---- Vout
\ | Zener
o |
| |
[C1] [R1] 220K
| |
Gnd ------+-----+

Vzener is slightly below Vcc - eg 4.7 volts if Vcc is
5 V, 11 volts if Vcc is 12 volts etc.
Ed


Ed: That looked such an attractively simple circuit that I tried it
myself yesterday. I may have missed something but aren't there a
couple of downsides to it?

Yes, absolutely.

I don't think the OP specified what sort of
pulse he wanted, but my starting assumptions were a clean, +ve, full
supply signal.

He's replacing (or using in place of) a momentary switch.
Both the toggle and the momentary are subjecty to switch
bounce. Since whatever circuit he's driving will work
on a momentary, then switch bounce from a toggle is not
a factor.

But your observation is correct. The circuit will not
provide a clean, +ve, full signal supply. It is a definite
downside to the circuit for general use - it is good only
for use where switch bounce is irrelevant.


Your circuit:
1) Delivers only a low amplitude spike, so needs amplifying to get a
'full' pulse.

That's wierd. It should deliver close to Vcc. Your
Vcc is 14 volts and the zener is 12. It (the zener)
should conduct until Vcap drops to 12, meaning that
the pulse amplitude has to be at least 12. Maybe
you are scoping at the input to the zener? The
duration depends on the load, and to a lesser extent,
on the 220K resistor, but primarily on the zener.
It will be a spike, but it should be a spike whose
amplitude is over the zener voltage. RC is ~ 9.5 mS
(way too short) down to ~1/3 vcc - but it is collapsing
only 2 volts or about 14% (instead of ~ 63%) when the zener
shuts it off. You need a bigger cap! The circuit "looks at"
just the top of the cap discharge curve, which is steep.

I think I need a much *smaller* cap! If I make it 10nF instead of 1uf,
then I see Vout as an initial *very* brief spike.
We disagree on spike amplitude, but that doesn't change
the fact that the output is another downside, as you
indicated. You get what amounts to a spike instead of a
nice robust square wave.
Using say a simple NPN stage that results in a -ve going

Yes - you need that with a low impedance load, or a long
duration pulse, or if the pulse must be square. And as long
as you stuff enough drive current into the base, it drives
to Vcc minus the transistor Vce. So you can have plenty of
headroom between the zener and Vcc and still get square wave
output.

I got way over 40 seconds with a 120(?) ohm relay driven
that way through a darlington with Hfe > 1000 in a delay
off circuit I made. Had to add a resitor in parallel with
the cap to get it down to the ~40 second target in that
circuit. I don't remember all the values, but I can find
them if they are of interest.

You've lost me there. Can you draw that circuit please?
The headroom was to make the pulse voltage unambiguous.
RC is ~ 63% discharged - the circuit is "looking"
at the top ~6% of discharge (eg from 5.0 down to 4.7).
I didn't want the the circuit that is being driven
to "see" an ambiguous voltage that was not clearly
either a high or a low. By using the top of the
discharge curve, the voltage will be at a high
until the zener cuts it off. It has a wicked slope,
but all of it will be above the zener V. Also, I picked
common zener voltages. I run into lots of 11 and 4.7v
zeners, but rarely a 4.3.


Absolutely. As mentioned earlier, I don't think
that's a factor for the op, but it certainly is
in many applications.


Thanks for these - they're great! I have a question, inline
below.

I'm puzzled by the Vout curves. It should "fall off the cliff"
at 12V when the zener stops conducting, but the traces all show
Vout curving down below 12 volts.

I think the output (without the noise, which I can't draw)
should look like this:

+14 |\
+12 | \
| |
| |
| |
0 ---- ----

Can you determine why there is that curve down to 0
volts? I wonder if your zener doesn't "zen" :) ?


Regarding the simulation - my guess is that they treat
the switch as noiseless, no bounce. But they don't
show the Vout "falling off the cliff" pattern - unless
the yellow is supposed to be Vout. The green looks
more like a capacitor discharge curve than the yellow,
but the green doesn't fall off the cliff. And if the
yellow is supposed to be Vout, the trailing edge is
correct and everything before it is wrong. So I'm
clueless.

Ed

I'm out for rest of day, but I'll get back on the case tonight. I
suspect that 'triangle' I saw for Vout might have been an artifact of
some sort, due to poorly chosen PC-based 'scope settings. All the
simulations I've tried show a square signal, amplitude Vcc-Vz, *apart*
from that initial spike which they now show, following the massive
reduction in the value of C.
 
T

Terry Pinnell

Terry Pinnell said:
ehsjr said:
Terry said:
I'm curious. Why didn't you try this?

+Vcc ---o o---+---|<---- Vout
\ | Zener
o |
| |
[C1] [R1] 220K
| |
Gnd ------+-----+

Vzener is slightly below Vcc - eg 4.7 volts if Vcc is
5 V, 11 volts if Vcc is 12 volts etc.
Ed


Ed: That looked such an attractively simple circuit that I tried it
myself yesterday. I may have missed something but aren't there a
couple of downsides to it?

Yes, absolutely.

I don't think the OP specified what sort of
pulse he wanted, but my starting assumptions were a clean, +ve, full
supply signal.

He's replacing (or using in place of) a momentary switch.
Both the toggle and the momentary are subjecty to switch
bounce. Since whatever circuit he's driving will work
on a momentary, then switch bounce from a toggle is not
a factor.

But your observation is correct. The circuit will not
provide a clean, +ve, full signal supply. It is a definite
downside to the circuit for general use - it is good only
for use where switch bounce is irrelevant.


Your circuit:
1) Delivers only a low amplitude spike, so needs amplifying to get a
'full' pulse.

That's wierd. It should deliver close to Vcc. Your
Vcc is 14 volts and the zener is 12. It (the zener)
should conduct until Vcap drops to 12, meaning that
the pulse amplitude has to be at least 12. Maybe
you are scoping at the input to the zener? The
duration depends on the load, and to a lesser extent,
on the 220K resistor, but primarily on the zener.
It will be a spike, but it should be a spike whose
amplitude is over the zener voltage. RC is ~ 9.5 mS
(way too short) down to ~1/3 vcc - but it is collapsing
only 2 volts or about 14% (instead of ~ 63%) when the zener
shuts it off. You need a bigger cap! The circuit "looks at"
just the top of the cap discharge curve, which is steep.

I think I need a much *smaller* cap! If I make it 10nF instead of 1uf,
then I see Vout as an initial *very* brief spike.
We disagree on spike amplitude, but that doesn't change
the fact that the output is another downside, as you
indicated. You get what amounts to a spike instead of a
nice robust square wave.
Using say a simple NPN stage that results in a -ve going

Yes - you need that with a low impedance load, or a long
duration pulse, or if the pulse must be square. And as long
as you stuff enough drive current into the base, it drives
to Vcc minus the transistor Vce. So you can have plenty of
headroom between the zener and Vcc and still get square wave
output.

I got way over 40 seconds with a 120(?) ohm relay driven
that way through a darlington with Hfe > 1000 in a delay
off circuit I made. Had to add a resitor in parallel with
the cap to get it down to the ~40 second target in that
circuit. I don't remember all the values, but I can find
them if they are of interest.

You've lost me there. Can you draw that circuit please?
The headroom was to make the pulse voltage unambiguous.
RC is ~ 63% discharged - the circuit is "looking"
at the top ~6% of discharge (eg from 5.0 down to 4.7).
I didn't want the the circuit that is being driven
to "see" an ambiguous voltage that was not clearly
either a high or a low. By using the top of the
discharge curve, the voltage will be at a high
until the zener cuts it off. It has a wicked slope,
but all of it will be above the zener V. Also, I picked
common zener voltages. I run into lots of 11 and 4.7v
zeners, but rarely a 4.3.


Absolutely. As mentioned earlier, I don't think
that's a factor for the op, but it certainly is
in many applications.


Thanks for these - they're great! I have a question, inline
below.

I'm puzzled by the Vout curves. It should "fall off the cliff"
at 12V when the zener stops conducting, but the traces all show
Vout curving down below 12 volts.

I think the output (without the noise, which I can't draw)
should look like this:

+14 |\
+12 | \
| |
| |
| |
0 ---- ----

Can you determine why there is that curve down to 0
volts? I wonder if your zener doesn't "zen" :) ?

We aren't possibly at cross-purposes here? We're talking about the
circuit I illustrated above, with Vout taken from zener's anode, yes?

No, that's down to my choice of input to Circuitmaker's VCS. I
simulated a noiseless switch, like the relatively clean one I used in
my tests.)
I'm out for rest of day, but I'll get back on the case tonight. I
suspect that 'triangle' I saw for Vout might have been an artifact of
some sort, due to poorly chosen PC-based 'scope settings. All the
simulations I've tried show a square signal, amplitude Vcc-Vz, *apart*
from that initial spike which they now show, following the massive
reduction in the value of C.

I re-assembled the breadboarded circuit yesterday with a different 12V
zener. The waveforms now look OK to me. So the only remaining
questions are:

1) Amplitude; I still think it should be as seen here, Vcc-Vz, not
close to Vcc?

2) Simulated waveshape versus actual; must be something to do with
unrealistic Circuitmaker models. I'll study further.
 
T

Terry Pinnell

Terry Pinnell said:
Terry Pinnell said:
ehsjr said:
Terry Pinnell wrote:


I'm curious. Why didn't you try this?

+Vcc ---o o---+---|<---- Vout
\ | Zener
o |
| |
[C1] [R1] 220K
| |
Gnd ------+-----+

Vzener is slightly below Vcc - eg 4.7 volts if Vcc is
5 V, 11 volts if Vcc is 12 volts etc.
Ed


Ed: That looked such an attractively simple circuit that I tried it
myself yesterday. I may have missed something but aren't there a
couple of downsides to it?

Yes, absolutely.

I don't think the OP specified what sort of
pulse he wanted, but my starting assumptions were a clean, +ve, full
supply signal.

He's replacing (or using in place of) a momentary switch.
Both the toggle and the momentary are subjecty to switch
bounce. Since whatever circuit he's driving will work
on a momentary, then switch bounce from a toggle is not
a factor.

But your observation is correct. The circuit will not
provide a clean, +ve, full signal supply. It is a definite
downside to the circuit for general use - it is good only
for use where switch bounce is irrelevant.


Your circuit:

1) Delivers only a low amplitude spike, so needs amplifying to get a
'full' pulse.

That's wierd. It should deliver close to Vcc. Your
Vcc is 14 volts and the zener is 12. It (the zener)
should conduct until Vcap drops to 12, meaning that
the pulse amplitude has to be at least 12. Maybe
you are scoping at the input to the zener? The
duration depends on the load, and to a lesser extent,
on the 220K resistor, but primarily on the zener.
It will be a spike, but it should be a spike whose
amplitude is over the zener voltage. RC is ~ 9.5 mS
(way too short) down to ~1/3 vcc - but it is collapsing
only 2 volts or about 14% (instead of ~ 63%) when the zener
shuts it off. You need a bigger cap! The circuit "looks at"
just the top of the cap discharge curve, which is steep.

I think I need a much *smaller* cap! If I make it 10nF instead of 1uf,
then I see Vout as an initial *very* brief spike.
We disagree on spike amplitude, but that doesn't change
the fact that the output is another downside, as you
indicated. You get what amounts to a spike instead of a
nice robust square wave.
Using say a simple NPN stage that results in a -ve going
pulse, which may then need inverting. (BTW, such amplification would
presumably be simplified if the headroom voltage is rather larger than
your examples?)

Yes - you need that with a low impedance load, or a long
duration pulse, or if the pulse must be square. And as long
as you stuff enough drive current into the base, it drives
to Vcc minus the transistor Vce. So you can have plenty of
headroom between the zener and Vcc and still get square wave
output.

I got way over 40 seconds with a 120(?) ohm relay driven
that way through a darlington with Hfe > 1000 in a delay
off circuit I made. Had to add a resitor in parallel with
the cap to get it down to the ~40 second target in that
circuit. I don't remember all the values, but I can find
them if they are of interest.

You've lost me there. Can you draw that circuit please?
The headroom was to make the pulse voltage unambiguous.
RC is ~ 63% discharged - the circuit is "looking"
at the top ~6% of discharge (eg from 5.0 down to 4.7).
I didn't want the the circuit that is being driven
to "see" an ambiguous voltage that was not clearly
either a high or a low. By using the top of the
discharge curve, the voltage will be at a high
until the zener cuts it off. It has a wicked slope,
but all of it will be above the zener V. Also, I picked
common zener voltages. I run into lots of 11 and 4.7v
zeners, but rarely a 4.3.


2) Transmits switch noise.

Absolutely. As mentioned earlier, I don't think
that's a factor for the op, but it certainly is
in many applications.


I used a 14V supply and a nominal 12V 1W zener, with this circuit:
http://www.terrypin.dial.pipex.com/Images/Momentary-Ed1.gif

Here are a few screenshots of the results I actually saw:

Thanks for these - they're great! I have a question, inline
below.

Relatively 'clean' switch
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-1.gif

Noisy switch
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-3.gif

Noisy switch, detail
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-5.gif

Noisy switch, detail, after NPN stage
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-8.gif

I'm puzzled by the Vout curves. It should "fall off the cliff"
at 12V when the zener stops conducting, but the traces all show
Vout curving down below 12 volts.

I think the output (without the noise, which I can't draw)
should look like this:

+14 |\
+12 | \
| |
| |
| |
0 ---- ----

Can you determine why there is that curve down to 0
volts? I wonder if your zener doesn't "zen" :) ?

We aren't possibly at cross-purposes here? We're talking about the
circuit I illustrated above, with Vout taken from zener's anode, yes?

No, that's down to my choice of input to Circuitmaker's VCS. I
simulated a noiseless switch, like the relatively clean one I used in
my tests.)
I'm out for rest of day, but I'll get back on the case tonight. I
suspect that 'triangle' I saw for Vout might have been an artifact of
some sort, due to poorly chosen PC-based 'scope settings. All the
simulations I've tried show a square signal, amplitude Vcc-Vz, *apart*
from that initial spike which they now show, following the massive
reduction in the value of C.

I re-assembled the breadboarded circuit yesterday with a different 12V
zener. The waveforms now look OK to me. So the only remaining
questions are:

1) Amplitude; I still think it should be as seen here, Vcc-Vz, not
close to Vcc?

2) Simulated waveshape versus actual; must be something to do with
unrealistic Circuitmaker models. I'll study further.

Forgot to include the links to latest screenshots
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-10.gif
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-11.gif

BTW, note that in these tests I'm typically toggling the switch on and
off again after only a sub-second pause. In practice it doesn't matter
how long it remains in the On position (providing you remember to
switch it off before requiring a subsequent action).
 
E

ehsjr

Terry said:
Terry said:
I'm curious. Why didn't you try this?

+Vcc ---o o---+---|<---- Vout
\ | Zener
o |
| |
[C1] [R1] 220K
| |
Gnd ------+-----+

Vzener is slightly below Vcc - eg 4.7 volts if Vcc is
5 V, 11 volts if Vcc is 12 volts etc.
Ed


Ed: That looked such an attractively simple circuit that I tried it
myself yesterday. I may have missed something but aren't there a
couple of downsides to it?

Yes, absolutely.

I don't think the OP specified what sort of
pulse he wanted, but my starting assumptions were a clean, +ve, full
supply signal.

He's replacing (or using in place of) a momentary switch.
Both the toggle and the momentary are subjecty to switch
bounce. Since whatever circuit he's driving will work
on a momentary, then switch bounce from a toggle is not
a factor.

But your observation is correct. The circuit will not
provide a clean, +ve, full signal supply. It is a definite
downside to the circuit for general use - it is good only
for use where switch bounce is irrelevant.


Your circuit:
1) Delivers only a low amplitude spike, so needs amplifying to get a
'full' pulse.

That's wierd. It should deliver close to Vcc. Your
Vcc is 14 volts and the zener is 12. It (the zener)
should conduct until Vcap drops to 12, meaning that
the pulse amplitude has to be at least 12. Maybe
you are scoping at the input to the zener? The
duration depends on the load, and to a lesser extent,
on the 220K resistor, but primarily on the zener.
It will be a spike, but it should be a spike whose
amplitude is over the zener voltage. RC is ~ 9.5 mS
(way too short) down to ~1/3 vcc - but it is collapsing
only 2 volts or about 14% (instead of ~ 63%) when the zener
shuts it off. You need a bigger cap! The circuit "looks at"
just the top of the cap discharge curve, which is steep.


I think I need a much *smaller* cap! If I make it 10nF instead of 1uf,
then I see Vout as an initial *very* brief spike.

Why do you want a brief spike? The circuit is intended
to give something roughly 10 mS long.
You've lost me there. Can you draw that circuit please?


+12 ---+-----+---[RLY]---+-----+
| | | |
| +---[D2]----+ |
| |
OFF o o ON |
| /c
o---+---[Zd]---+-------| Q1
| | \e
[C1]470uf [R1]220K |
| | |
Gnd -------+----------+--------+

In essence, the darlington's gain multiplies the coil
resistance and places it in parallel with R1, which
affects RC.

I'm out for rest of day, but I'll get back on the case tonight. I
suspect that 'triangle' I saw for Vout might have been an artifact of
some sort, due to poorly chosen PC-based 'scope settings. All the
simulations I've tried show a square signal, amplitude Vcc-Vz, *apart*
from that initial spike which they now show, following the massive
reduction in the value of C.

I'm not surprized. The larger the C the less you
will see the spike. But the spike is most definitely
not the thing wanted in this application. Intended is
a positive voltage near Vcc that stays up for a
relatively long time - milliseconds. The "voltage up"
period is the time for the voltage to drop from
Vcc to the zener voltage.

Ed
 
E

ehsjr

Terry said:
Terry Pinnell wrote:




I'm curious. Why didn't you try this?

+Vcc ---o o---+---|<---- Vout
\ | Zener
o |
| |
[C1] [R1] 220K
| |
Gnd ------+-----+

Vzener is slightly below Vcc - eg 4.7 volts if Vcc is
5 V, 11 volts if Vcc is 12 volts etc.
Ed


Ed: That looked such an attractively simple circuit that I tried it
myself yesterday. I may have missed something but aren't there a
couple of downsides to it?

Yes, absolutely.

I don't think the OP specified what sort of

pulse he wanted, but my starting assumptions were a clean, +ve, full
supply signal.

He's replacing (or using in place of) a momentary switch.
Both the toggle and the momentary are subjecty to switch
bounce. Since whatever circuit he's driving will work
on a momentary, then switch bounce from a toggle is not
a factor.

But your observation is correct. The circuit will not
provide a clean, +ve, full signal supply. It is a definite
downside to the circuit for general use - it is good only
for use where switch bounce is irrelevant.


Your circuit:

1) Delivers only a low amplitude spike, so needs amplifying to get a
'full' pulse.

That's wierd. It should deliver close to Vcc. Your
Vcc is 14 volts and the zener is 12. It (the zener)
should conduct until Vcap drops to 12, meaning that
the pulse amplitude has to be at least 12. Maybe
you are scoping at the input to the zener? The
duration depends on the load, and to a lesser extent,
on the 220K resistor, but primarily on the zener.
It will be a spike, but it should be a spike whose
amplitude is over the zener voltage. RC is ~ 9.5 mS
(way too short) down to ~1/3 vcc - but it is collapsing
only 2 volts or about 14% (instead of ~ 63%) when the zener
shuts it off. You need a bigger cap! The circuit "looks at"
just the top of the cap discharge curve, which is steep.

I think I need a much *smaller* cap! If I make it 10nF instead of 1uf,
then I see Vout as an initial *very* brief spike.


We disagree on spike amplitude, but that doesn't change
the fact that the output is another downside, as you
indicated. You get what amounts to a spike instead of a
nice robust square wave.

Using say a simple NPN stage that results in a -ve going

pulse, which may then need inverting. (BTW, such amplification would
presumably be simplified if the headroom voltage is rather larger than
your examples?)

Yes - you need that with a low impedance load, or a long
duration pulse, or if the pulse must be square. And as long
as you stuff enough drive current into the base, it drives
to Vcc minus the transistor Vce. So you can have plenty of
headroom between the zener and Vcc and still get square wave
output.

I got way over 40 seconds with a 120(?) ohm relay driven
that way through a darlington with Hfe > 1000 in a delay
off circuit I made. Had to add a resitor in parallel with
the cap to get it down to the ~40 second target in that
circuit. I don't remember all the values, but I can find
them if they are of interest.

You've lost me there. Can you draw that circuit please?


The headroom was to make the pulse voltage unambiguous.
RC is ~ 63% discharged - the circuit is "looking"
at the top ~6% of discharge (eg from 5.0 down to 4.7).
I didn't want the the circuit that is being driven
to "see" an ambiguous voltage that was not clearly
either a high or a low. By using the top of the
discharge curve, the voltage will be at a high
until the zener cuts it off. It has a wicked slope,
but all of it will be above the zener V. Also, I picked
common zener voltages. I run into lots of 11 and 4.7v
zeners, but rarely a 4.3.


2) Transmits switch noise.

Absolutely. As mentioned earlier, I don't think
that's a factor for the op, but it certainly is
in many applications.


I used a 14V supply and a nominal 12V 1W zener, with this circuit:
http://www.terrypin.dial.pipex.com/Images/Momentary-Ed1.gif

Here are a few screenshots of the results I actually saw:

Thanks for these - they're great! I have a question, inline
below.

Relatively 'clean' switch
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-1.gif

Noisy switch
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-3.gif

Noisy switch, detail
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-5.gif

Noisy switch, detail, after NPN stage
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-8.gif

I'm puzzled by the Vout curves. It should "fall off the cliff"
at 12V when the zener stops conducting, but the traces all show
Vout curving down below 12 volts.

I think the output (without the noise, which I can't draw)
should look like this:

+14 |\
+12 | \
| |
| |
| |
0 ---- ----

Can you determine why there is that curve down to 0
volts? I wonder if your zener doesn't "zen" :) ?

We aren't possibly at cross-purposes here? We're talking about the
circuit I illustrated above, with Vout taken from zener's anode, yes?

BTW, I'm puzzled by the roughly triangular shape of the output. I
tried one simulation and got a squarish result:

http://www.terrypin.dial.pipex.com/Images/Momentary-Ed-SIM1.gif
Maybe it was just the choice of zener?



Regarding the simulation - my guess is that they treat
the switch as noiseless, no bounce.

No, that's down to my choice of input to Circuitmaker's VCS. I
simulated a noiseless switch, like the relatively clean one I used in
my tests.)

But they don't
show the Vout "falling off the cliff" pattern - unless
the yellow is supposed to be Vout. The green looks
more like a capacitor discharge curve than the yellow,
but the green doesn't fall off the cliff. And if the
yellow is supposed to be Vout, the trailing edge is
correct and everything before it is wrong. So I'm
clueless.

Ed

I'm out for rest of day, but I'll get back on the case tonight. I
suspect that 'triangle' I saw for Vout might have been an artifact of
some sort, due to poorly chosen PC-based 'scope settings. All the
simulations I've tried show a square signal, amplitude Vcc-Vz, *apart*
from that initial spike which they now show, following the massive
reduction in the value of C.

I re-assembled the breadboarded circuit yesterday with a different 12V
zener. The waveforms now look OK to me. So the only remaining
questions are:

1) Amplitude; I still think it should be as seen here, Vcc-Vz, not
close to Vcc?

2) Simulated waveshape versus actual; must be something to do with
unrealistic Circuitmaker models. I'll study further.


Forgot to include the links to latest screenshots
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-10.gif
http://www.terrypin.dial.pipex.com/Images/Toggle-Ed-11.gif

BTW, note that in these tests I'm typically toggling the switch on and
off again after only a sub-second pause. In practice it doesn't matter
how long it remains in the On position (providing you remember to
switch it off before requiring a subsequent action).

The second image with the 5 uf cap and noisy switch begins to
show a more desireable waveform at the beginning of the
capacitor discharge. Stuff a big cap in there and see what
you get. I still don't understand the voltage levels.
It's like your zener is shorted.

Ed
 
H

Howard Henry Schlunder

ehsjr said:
That's wierd. It should deliver close to Vcc. Your
Vcc is 14 volts and the zener is 12. It (the zener)
should conduct until Vcap drops to 12, meaning that
the pulse amplitude has to be at least 12. Maybe
you are scoping at the input to the zener? The
duration depends on the load, and to a lesser extent,
on the 220K resistor, but primarily on the zener.
It will be a spike, but it should be a spike whose
amplitude is over the zener voltage. RC is ~ 9.5 mS
(way too short) down to ~1/3 vcc - but it is collapsing
only 2 volts or about 14% (instead of ~ 63%) when the zener
shuts it off. You need a bigger cap! The circuit "looks at"
just the top of the cap discharge curve, which is steep.

We disagree on spike amplitude, but that doesn't change
the fact that the output is another downside, as you
indicated. You get what amounts to a spike instead of a
nice robust square wave.

You have a fundamental misunderstanding on how zener diodes function. I
recommend studying the (reverse) voltage vs. (reverse) current graph closely
in a zener diode data sheet. When a zener diode is reverse conducting, it
will have a voltage drop across it of appoximately the rated zener voltage.
In your circuit, immediately after the switch is moved to the right, Vout
should be close to Vcc-Vzener, not close to Vcc.
 
Top