Electronic toggle switch
Arnd Strube
the following requirements:
- Operate off a voltage of 3.3..4.5 V
- Draw no current when inactive (well, some micro Amps)
- Provide output current of 20..30mA at the above voltage
(e.g. ideally as if a physical switch was closed)
- Switch to ON when a momentary contact is closed and
switch to OFF when it is closed again
- Be inexpensive, small and light, preferably a single IC.
Thank you,
Arnd Strube
res.raytheon.com
transistor that handles the current.
Arnd Strube
and the clock input to Vcc via my push button. Then exactly how would I
connect the transistor to Q, and what type of transistor would you
recommend? The load will draw a maximum of 30mA at 4.5V. It's a long time
since I've designed any circuit and my knowledge in this field is very
basic, so if you could maybe point me to schematics somewhere that would be
really neat.
Arnd
Terry Pinnell
If you really MUST have a push-button, then you'll need an extra front
end stage to debounce it. A simple RC arrangement can work reliably
but needs careful choice of the timing constant to suit the particular
button. Normally, it's best to use another flip-flop or logic latch to
do the debouncing.
Frankly, I'd recommend a switch instead. You could then use the simple
circuit posted at
http://www.terrypin.dial.pipex.com/Images/4001FF.gif
In any event, you can use the same transistor driver I've used there.
Terry Pinnell
Hobbyist, West Sussex, UK
John Fields
I've posted a single-chip solution for you on
alt.binaries.schematics.electronic. S1 is a momentary pusbutton switch
shown in the resting state. The first half of the HC74 is used to
debounce the switch and the second half is wired as a divide-by-two
which will give you the toggle you need.
---
John Fields, Austin Instruments, Inc.
El Presidente Austin, Republic of Texas
"I speak for my company" http://www.austininstruments.com
Robert
Terry Pinnell wrote:
> If you really MUST have a push-button, then you'll need an extra front
> end stage to debounce it. A simple RC arrangement can work reliably
> but needs careful choice of the timing constant to suit the particular
> button. Normally, it's best to use another flip-flop or logic latch to
> do the debouncing.
>
> Frankly, I'd recommend a switch instead. You could then use the simple
> circuit posted at
> http://www.terrypin.dial.pipex.com/Images/4001FF.gif
>
> In any event, you can use the same transistor driver I've used there.
>
> Terry Pinnell
> Hobbyist, West Sussex, UK
Very nicely done, Terry- that's a beautiful drawing and circuit.
Spehro Pefhany
I haven't seen this method of debouncing used before, but it ought to
work:
+Vbatt
| 1/2 74HC74 +Vbatt-----> + to circ.
|
10K ------------ Q1 D
| | | |-------------> - to circ.
| | Q |---------1K--------||
x------------|> Cp | G |-------Rx--- 0V
| | | S
| o x---|D /Q |----x
| S1 | | | |
| o | ------------ |
| | 100K
0V | |
x-------------------x
|
0.1uF
|
0V
(no attempt is made to debounce the clock signal, so it triggers multiple
times on each switch press, but the voltage on the D line won't change
until the switch has stopped bouncing)
Where Q1 is a logic level MOSFET with suitable Rds(on) at Vg = 3.3V.
Bypass Vbatt with a suitable capacitor (eg. 220uF/6.3V in parallel with
0.1uF ceramic). If there is a lot of capacitance in the circuit across the
power supply rails you might have to add Rx to soft-start the circuit.
Maybe 3.3 ohms.
You can connect /PR and /CLR to Vdd if you don't care what state it goes
into when the battery is inserted, or use an R-C to set whichever state
you prefer. Ground (or connect to Vdd) all unused inputs on the 74HC74 and
leave unused outputs open.
Best regards,
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Spehro Pefhany --"it's the network..." "The Journey is the reward"
sp...@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
Contributions invited->The AVR-gcc FAQ is at: http://www.BlueCollarLinux.com
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Arnd R. Strube
push button though because that's the whole point. If a switch were
suitable, I'd just simply use it the "old-fashioned" way and be done...
"Terry Pinnell" <terr...@dial.pipex.com> wrote in message
news:ecihbtsahrv4eo68p...@4ax.com...
Arnd R. Strube
Arnd
"Spehro Pefhany" <sp...@interlog.com> wrote in message
news:hl4u6.98509$p66.27...@news3.rdc1.on.home.com...
Robert
did not post this.
Spehro Pefhany
> I don't know who you are, but fraudulent impersonation is a serious abuse. Spehro
> did not post this.
Alright, what did I miss. ;-)
Arnd R. Strube
Arnd R. Strube
alt.binaries.schematics.electronic. I just subscribed a few minutes ago.
What might have gone wrong?
Arnd
> I've posted a single-chip solution for you on
> alt.binaries.schematics.electronic
John Fields
>
> For some reason, your post doesn't show up for me under
> alt.binaries.schematics.electronic. I just subscribed a few minutes ago.
> What might have gone wrong?
---
I don't know, but I'll email you a copy. BTW, if you have a preferred
start-up state let me know (here) and I'll tell you what to do if you
don't know how to make it happen
Robert
Look in these pages for PUSHBUTTON ALTERNATE ACTION SPST keywords...
http://www.switchchannel.com/product/minipush/mpi.htm
http://www.nkkswitches.com/navfrm.html
http://www.switchchannel.com/ts707/
http://nt.headland.co.uk/rox/switch/grayhill/pushbutton.htm
http://www.adept.net/adept/bizlinks/ohbromberg/products.htm
http://www.e-switch.com/newsite/showcatalog.html?itype=4&stype=push
John Fields
...
> +Vbatt
> | 1/2 74HC74 +Vbatt-----> + to circ.
> |
> 10K ------------ Q1 D
> | | | |-------------> - to circ.
> | | Q |---------1K--------||
> x------------|> Cp | G |-------Rx--- 0V
> | | | S
> | o x---|D /Q |----x
> | S1 | | | |
> | o | ------------ |
> | | 100K
> 0V | |
> x-------------------x
> |
> 0.1uF
> |
> 0V
>
> (no attempt is made to debounce the clock signal, so it triggers multiple
> times on each switch press, but the voltage on the D line won't change
> until the switch has stopped bouncing)
...
---
Neat!
One small nitpick, if I may?
Let's say that Q is high and that the switch is open. /Q and D will
then be low and Cp will be high. If the switch is depressed nothing
will happen until it bounces for the first time, then Q will go low, \Q
will go high, and the voltage on D will start to rise immediately,
regardless of whether the switch continues to bounce or not.
Spehro Pefhany
> did not post this.
Of course you (and John) are right. Dumb dumb dumb..
Here's a fix for it:-
o +Vbat
|
1M S1
| ----
x-----o o-----x------> to Cp on FF
| |
0.1uF 1K
| |
0V 0V
I'll go slink off and get some sleep or a vacation or something...
Terry Pinnell
>Terry, your circuit is really neat and I found it helpful. I MUST have a
>push button though because that's the whole point. If a switch were
>suitable, I'd just simply use it the "old-fashioned" way and be done...
OK, try the one I just posted at
http://www.terrypin.dial.pipex.com/Images/SimpleToggle.gif
David Milne
Terry Pinnell wrote:
--
Regards
David Milne
ICQ 37590068
Terry Pinnell
>Just curious Terry, what program(s) did you use to generate the gif file?
David: I drew my schematic in CircuitMaker Pro 6.2c. (You don't read
sci.electronics.cad? <g>) Then I copied it to the Windows clipboard (I
actually use a screen capture program called Snagit, but PrtScr or
Alt+PrtSCr does a similar job). Finally I edited it in PaintShop Pro
6.
David Milne
Thanks for that.
--
Regards
David Milne
ICQ 37590068
All hail the Master of the Universe -
Wilbur the hairy green toad and his consort, the invisible pink unicorn.
Arnd R. Strube
then come back as to whether they worked for me.
Arnd
Tom Del Rosso
>
> Here's a fix for it:-
>
> o +Vbat
> |
> 1M S1
> | ----
> x-----o o-----x------> to Cp on FF
> | |
> 0.1uF 1K
> | |
> 0V 0V
What was your goal in doing it this way (with the RC on the data
input)? To get a faster risetime on the clock? What is the 1M and
the cap for? I don't see how it wouldn't work without them.
Tom Del Rosso
How is that a debouncer? First, the pot has to be adjusted to put the
input voltage above the upper threshold. Then you push the button and
the voltage goes to 0 and starts to rise, but if there is a single
bounce before the cap charges and the input is still low, the gate
will make another pulse.
Also, what's the negative feedback for? It's not enough for
oscillation, but what good does it do?
Nobody has suggested the common old circuit:
V+
|
|
| o
|
| o
|
|
x----1k-----x------->
| |
| |
10k 0.1uf
| |
| |
GND GND
Robert
It appears that he wants to debounce the *current*- maybe the LED is IR
and he is doing manual signalling of some kind-
whatever-....zzzz...yawn....ho hum...
What I want to know is what anarchist group is using the NG address as
graffiti in rest room stalls...;)
Spehro Pefhany
> What I want to know is what anarchist group is using the NG address as
> graffiti in rest room stalls...;)
Whichever it is, we must find their leader!
Spehro Pefhany
> What was your goal in doing it this way (with the RC on the data
> input)? To get a faster risetime on the clock? What is the 1M and
> the cap for? I don't see how it wouldn't work without them.
The problem with the other circuit is that it would misbehave on the
switch release, if it bounced (which is possible, though less of a
problem on release than on closure). The above fix confines
any (probably multiple) clocking to the closure of the switch, after which
the 0.1uF is discharged. The switch must be released for a fraction of a
second to allow it to recharge enough to clock the FF again (through the
1M). You need both R-C's for it to work reliably.
Tom Del Rosso
> The renowned Tom Del Rosso <no.spam.plea...@att.net>
wrote:
>
> > What was your goal in doing it this way (with the RC on the data
> > input)? To get a faster risetime on the clock? What is the 1M
and
> > the cap for? I don't see how it wouldn't work without them.
>
> The problem with the other circuit is that it would misbehave on the
> switch release, if it bounced (which is possible, though less of a
> problem on release than on closure).
Of course, I see now, it clocks on both sides. But you still have one
more C and one or two more Rs than the minimum requirement, so was I
right in assuming that you were trying to keep the clock risetime high
when no Schmidt input is available, or was there another reason?
Terry Pinnell
I found at Bill Bowden's web page,
http://members.tripod.com/~schematics/toggle.htm
It's a design by John Lundgren, and uses a couple of transistors.
Tom Del Rosso
> For yet another approach you might like to take a look at the
circuit
> I found at Bill Bowden's web page,
> http://members.tripod.com/~schematics/toggle.htm
> It's a design by John Lundgren, and uses a couple of transistors.
There's something Fascistic about that schematic. ;-)
Arnd R. Strube
Arnd
"Terry Pinnell" <terr...@dial.pipex.com> wrote in message
news:f10mbt0hgc507ipam...@4ax.com...
Terry Pinnell
>How is that a debouncer? First, the pot has to be adjusted to put the
>input voltage above the upper threshold. Then you push the button and
>the voltage goes to 0 and starts to rise, but if there is a single
>bounce before the cap charges and the input is still low, the gate
>will make another pulse.
>
>Also, what's the negative feedback for? It's not enough for
>oscillation, but what good does it do?
(From old Elektor magazine, author unknown)
Basically the output of the Schmitt trigger N1 changes (toggles) when
the switch is closed momentarily. This toggle function is achieved in
such a simple circuit by the fact that the inputs of the trigger are
held between the switching threshold levels. If we assume that the
output logic level (Q) of the trigger is at logic 1, capacitor C1 will
charge via Rl. When switch Sl is closed the input of the trigger will
be taken to logic 1 (because the capacitor is fully charged) and the Q
output will of course become logic 0. The capacitor will now discharge
but not completely because the closed switch will hold the level to
that existing at the wiper of Pl. However, this drop in voltage at the
input of the gate will not cause its output to change state again
because the input level will be above the lower switching thresholf
the Schmitt trigger.
This intermediate voltage level will remain while the switch is
closed. When the switch is eventually released, C1 will discharge
completely. The 0 V across the capacitor will not affect the trigger
since it is no longer connected to the capacitor (the switch is open).
Now, when the switch closed, the 0 V will reach the input of gate and
its output will once again change state. It is essential that Pl is
set correctly for the circuit to function, but it will be found in
practice that this presents no problem. Various types of Schmitt
trigger are suitable for this circuit: 4093, 40106, T4LSl4, T4LS132.
If TTL ICs are used, the supply voltage must be 5 V ą 0.25 V; for CMOS
ICs it should lie between 5 V and 15 V.
>Nobody has suggested the common old circuit:
>
> V+
> |
> |
> | o
> |
> | o
> |
> |
> x----1k-----x------->
> | |
> | |
> 10k 0.1uf
> | |
> | |
> GND GND
>
Not really in the same league...! Its noise immunity with most
push-buttons would be too low for reliable use with digital circuitry.
John Fields
---
What "noise immunity"? All you're trying to do is to keep the voltage
across the cap from rising/falling to the logic switching threshold
until the switch stops chattering. To that end,
V+
|
|
| o
| S1
| o
|
|
x----[R1]-----x------->VOUT
| |
| |
[R2] [C1]
| |
| |
GND GND
the time constant of R1C1 is chosen so that VOUT cannot rise to the
minimum low-to-high switching threshold of the logic until S1 has
stopped chattering after it's closed. The time constant of (R1+R2)C1
is chosen so that VOUT will not fall to the maximum high-to-low
switching threshold of the logic until S1 has stopped chattering after
it's opened. Of course R2 will also have to carry the switching current
flowing through S1, so that _reasonable_ resistance will form the basis
for the determination of the values of R1 and C1. Also, a smart thing
to do would be to connect the input of a Schmitt trigger to VOUT in
order to minimize the effects of the rise/fall times of VOUT.
I would go so far as to say that this circuit is far superior to the pot
circuit in terms of reliability and cost just by virtue of its
simplicity.
Terry Pinnell
>How much power would you expect that circuit to draw at 4.5V?
>Arnd
Assuming you mean current, about 1.4 mA. A lot more than my 4013
version of Spehro's circuit.
Tom Del Rosso
> "Tom Del Rosso" <no.spam.plea...@att.net> wrote:
>
> >How is that a debouncer? First, the pot has to be adjusted to put
the
>
> Basically the output of the Schmitt trigger N1 changes (toggles)
when
> the switch is closed momentarily. This toggle function is achieved
in
> such a simple circuit by the fact that the inputs of the trigger are
> held between the switching threshold levels.
Oh, damn, I saw a circuit that was almost the same as a debouncer,
with some modifications, and I didn't realize that it was supposed to
be a complete toggle that didn't need a flip flop. I understand how
it works now, but I don't know if you should bet on staying between
the thresholds over time. A pot can drift, and if the wiper developed
a poor contact, as so often happens with volume controls, it would
oscillate when you push the button. With a fixed voltage divider it
still might not be as reliable as a debouncer with a packaged flip
flop. It's clever, but things can be too clever, especially when
they're too simple.
> Not really in the same league...!
Of course, because that was not a toggle but a debouncer, especially
when driving a Schmidt trigger.
> Its noise immunity with most
> push-buttons would be too low for reliable use with digital
circuitry.
You mean if the switch bounces a lot more sometimes? You can make the
time constant 10 times as long as it needs to be.
Tom Del Rosso
news:3ABB7096...@austininstruments.com...
>
> the time constant of R1C1 is chosen so that VOUT cannot rise to the
> minimum low-to-high switching threshold of the logic until S1 has
> stopped chattering after it's closed.
I think it's even less critical than that. Even if the cap charges
past the threshold on the first bounce, all that matters is that it
doesn't discharge until the bouncing stops.
> I would go so far as to say that this circuit is far superior to the
pot
> circuit in terms of reliability and cost just by virtue of its
> simplicity.
Although it's not that simple because it needs a Schmidt and a flip
flop, whereas Terry's is a complete toggle.
Terry Pinnell
>What "noise immunity"? All you're trying to do is to keep the voltage
>across the cap from rising/falling to the logic switching threshold
>until the switch stops chattering. To that end,
>
> V+
> |
> |
> | o
> | S1
> | o
> |
> |
> x----[R1]-----x------->VOUT
> | |
> | |
> [R2] [C1]
> | |
> | |
> GND GND
>
>the time constant of R1C1 is chosen so that VOUT cannot rise to the
>minimum low-to-high switching threshold of the logic until S1 has
>stopped chattering after it's closed. The time constant of (R1+R2)C1
>is chosen so that VOUT will not fall to the maximum high-to-low
>switching threshold of the logic until S1 has stopped chattering after
>it's opened. Of course R2 will also have to carry the switching current
>flowing through S1, so that _reasonable_ resistance will form the basis
>for the determination of the values of R1 and C1. Also, a smart thing
>to do would be to connect the input of a Schmitt trigger to VOUT in
>order to minimize the effects of the rise/fall times of VOUT.
>
>I would go so far as to say that this circuit is far superior to the pot
>circuit in terms of reliability and cost just by virtue of its
>simplicity.
Superior on those two counts, agreed. ('Far superior'?)
And inferior on implementation time for any given push button, if you
define that as the time taken to ensure it won't misbehave - i.e is
immune to switching noise. Although that time could be improved by
first using a pot/preset!
Terry Pinnell
Yes, that's what I had in mind - thinking of some of the buttons I
have around the place! But I've since been having fun simulating it,
and agree that (for a simple debouncer) it does have a wide range of
immunity. Of course, that begs the question of how realistic the
'switch noise' piece of my simulation is. When the rain stops
bucketing down, I'll get to the shed and breadboard it...
Tom Del Rosso
>
> And inferior on implementation time for any given push button, if
you
> define that as the time taken to ensure it won't misbehave - i.e is
> immune to switching noise. Although that time could be improved by
> first using a pot/preset!
The 1ms time constant works with most switches, but there is no real
disadvantage to making it 50ms. With the 2 Rs of 10k and 1k, and a C
of 4.7uf, you would have to press the button for 5ms to make something
happen, and it could bounce for, well forever, as long as it never
breaks contact for more than 10 times as long as it makes contact. It
probably does, but it can still bounce for a lot longer than common
switches do with no effect.
John Fields
> Superior on those two counts, agreed. ('Far superior'?)
---
IMHO, yup! If you can send it out into the field and _know_ you don't
have not worry about whether it's going to come back and bite you on the
ass, that's far superior than having to worry about the pot drifting,
the wiper corroding, etc. just as you're drifting off to sleep...
---
> And inferior on implementation time for any given push button, if you
> define that as the time taken to ensure it won't misbehave - i.e is
> immune to switching noise. Although that time could be improved by
> first using a pot/preset!
---
Set the make TC for 100ms and you've got about 99% of the switches out
there covered.
Tom Del Rosso
>
> When the rain stops
> bucketing down, I'll get to the shed and breadboard it...
I hope it survives the elements. You're not the only guy with a
workbench in a shed, but I want to keep my electronics in a controlled
environment. Even indoors I just noticed some of my BNC connectors
are getting oxidized.
John Fields
---
Those are the silverplated "good" ones. Nothing to worry about if it's
silver oxide, which is conductive.
Tom Del Rosso
>
> Those are the silverplated "good" ones. Nothing to worry about if
it's
> silver oxide, which is conductive.
I've read that, but how do you tell if it's real silver?
Frank Bemelman
Tom Del Rosso <no.spam.plea...@att.net> schreef in
berichtnieuws
SLMu6.1711$9i1.1...@bgtnsc04-news.ops.worldnet.att.net...
Check if it conducts when oxidized ;-)
That was easy. Next please.
--
Met vriendelijke groeten,
Frank Bemelman
(reageren per email ? verwijder dan de 'x' uit mijn emailadres)