Best way to make a 555 retriggerable?

[barabb...@email.it]

I'm using a double NE555 (called '556'), in which both halves act as a
one-shot. I need them to be retriggerable, that is, if a trigger pulse
comes before the completion of an ongoing pulse, the ongoing pulse is
extended by an amount equal to the preset time constant.

I found several ways on the 'net to achieve this:

1. 2.
http://sam.electroastro.pagesperso-orange.fr/dossiers/555/555.htm
(see "montages" n. 2 and 3)

3.
(Starting with a 555 in the usual monostable configuration) It's
achieved by wiring the capacitor between pin 6 and GND and the
resistor between supply voltage and pin 2. Pins 2 and 6 are wired
together, so pin 7 is not used.

4.
Same as the usual 555 in monostable configuration, but reset and
trigger are wired together: at each incoming pulse, the timer is reset
even if the previous output pulse hasn't finished yet.

I'm trying to build a device that tells me if an incoming signal
(Horizontal sync from a VGA card) has a frequency that is below or
above a certain value. The first one-shot has a time constant
corresponding to the period of this fixed frequency value (54 us, 18.5
KHz). The incoming signal is wired to the trigger pin of the first
half of the chip. When the frequency is higher than 18.5 KHz, the
interval between the pulses is less than 54 us, so the first one-shot
is retriggered constantly and stays high. If the frequency drops below
18.5 KHz, the pulses come at a lower rate, thus allowing the one-shot
to basically oscillate at the incoming signal's frequency (for
instance: the incoming signal has a frequency of 15 KHz (66 us): the
one-shot will produce a 54 us pulse every 66 us). The output of the
first one-shot is fed to the second half of the 556 (second one-shot),
which is set with a time constant of 84 us: when the first half stays
high because the frequency is above 18.5 KHz, this one remains low
because it is never triggered; when the first half is oscillating at
the incoming signal's frequency, the triggering pulses will always
come at intervals lower than 84 us, so the device will be constantly
retriggered and give a constantly high output.

Thanks very much for your time.

[Phil Hobbs]

Unless this is for a contest, the best way to get a dual retriggerable
monostable is to use a 74HC123. They're designed for the job.

One-shots are fine as long as you don't care if the delay is wrong by a
factor of 2 or 3, but that doesn't fit your application. (Read the data
sheet carefully, and compute the worst-case deviations at the corner
cases of device-to-device variation, temperature, supply voltage,
capacitor tolerance, and so forth, and then add a safety factor, and
you'll see what I mean. One-shots stink.)

If this is a one-off hobby project that only has to sort-of-work for a
short time, you can use tweaked one-shots. If you want it to actually
work properly many times, or over temperature, or after rattling round
your workbench drawer for a year, you'll need something a lot more
accurate than a one-shot.

If this is for a real-life application, you could use a small
microprocessor with a crystal clock and a timer/counter unit. (All
micros that I've ever seen have a timer/counter that's good enough for
this job.)

If you don't want to learn to use micros, but still want accuracy and
repeatability, you'd be very much better off using a crystal oscillator
to generate the limiting frequency (18.5 kHz) and then use a CD4046
(phase detector II) to detect whether the frequency is below or above that.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

[John Fields]

---
You've described what you want the circuit to do if the input
frequency is higher than, or lower than 18.5kHz, but what do you want
the circuit to do if the input frequency is precisely 18.5kHz?

--
JF

[John Larkin]

HC123 has a +-15% pulse width tolerance from -55 to 125C. That ain't
bad.

I like one-shots. They have lots of uses... resetting things, flashing
LEDs, decoding Manchester and 1-wire data streams, demodulating
wideband FM, all sorts of stuff.

Here's my dual tach:

ftp://jjlarkin.lmi.net/DoubleTach.jpg

John

[TTman]

>
> ---
> You've described what you want the circuit to do if the input
> frequency is higher than, or lower than 18.5kHz, but what do you want
> the circuit to do if the input frequency is precisely 18.5kHz?
>
> --
> JF

It will blow up ? :)

[Fred Bartoli]

barabb...@email.it a écrit :

Divide your frequency by two with a 74HC74 in order to have an accurate
50% square wave.
Use that to drive a small 2N7002 MOS that you'll use to reset a small RC
circuit. That'll generate a sawtooth ramp, whose amplitude is (almost)
proportional to your signal period. Then compare this to a reference
level you purposely define.

Above your reference frequency you'll have no output pulse. Below your
comparator will output pulses at F/2.

Maybe you can convince the second half of your 74HC74 to latch the
pulses so that your detector outputs constant levels, but it's too late
for me here. Someone else?


--
Thanks,
Fred.

[John Larkin]

On Wed, 9 Nov 2011 22:56:35 -0000, "TTman" <pcw1...@ntlworld.com>
wrote:

The world will end.

John

[John Fields]

---
Cute, but seemingly irrelevant since it only indicates the magnitude,
but not the sign of the error.

Consequently, it seems to have little to with the subject at hand,
which is how to identify whether the signal is below, above, or at the
reference frequency.

--
JF

[John Fields]

---
Well, probably not, since a singularity won't have been generated
since there's some R in the circuit.

--
JF

[John Larkin]

What does that mean? It converts frequency to voltage. It doesn't
indicate anything.

>
>Consequently, it seems to have little to with the subject at hand,
>which is how to identify whether the signal is below, above, or at the
>reference frequency.

We were discussing one-shots. I posted a one-shot circuit.

Actually, my tach could be used with an analog comparator to indicate
whether the input was above or below some frequency threshold. I'm not
sure how one could ever determine that a frequency is exactly at some
reference... you'd have to wait forever to find out. The OP sensibly
asked for "below or above."

John

[John Larkin]

Since it would take forever to determine if a frequency was precisely
18.5 KHz, the world would have ended!

John

[mike]

Some design goals/specs might be in order.

Define two frequency thresholds, A and B.
Below A, the output should be low.
Above B, the output should be high.
Between A and B is a region where you can't tell
if the frequency is higher or lower than desired.
What should the output do in that region?
Is there a problem if the output changes randomly
even for a fixed input frequency? This might require further
increase in the distance between A and B.
What do you do when noise or mode change glitches
the signal you're measuring?

Compare the allowable width of the uncertainty region
with the tolerance and stability of the resistors, caps, 555's that you
can afford.
That will determine if it's even possible to do what you want with
a 555.

I like the micro controller solution described elsewhere in this thread.

[Don Lancaster]

On 11/9/2011 3:04 PM, barabb...@email.it wrote:
> I'm using a double NE555 (called '556'), in which both halves act as a
> one-shot. I need them to be retriggerable, that is, if a trigger pulse
> comes before the completion of an ongoing pulse, the ongoing pulse is
> extended by an amount equal to the preset time constant.
>

Replace it with a PIC, of course.

--
Many thanks,

Don Lancaster voice phone: (928)428-4073
Synergetics 3860 West First Street Box 809 Thatcher, AZ 85552
rss: http://www.tinaja.com/whtnu.xml email: d...@tinaja.com

Please visit my GURU's LAIR web site at http://www.tinaja.com

[barabb...@email.it]

On 10 Nov, 05:25, Don Lancaster <d...@tinaja.com> wrote:
> Replace it with a PIC, of course.

That's what I wanted to do in the first place. The project I was
trying to do is described in this page, with source and all:
http://members.optusnet.com.au/eviltim/scart.htm
But my PIC programmer doesn't see the 12C508, no matter what software
or operating system I use. I can't code for PICs, otherwise I would
adapt the source to one of the other PICs I have, which my programmer
can perfectly read / write (PIC16LF84A, PIC16F628A, PIC18F2550).

[Okkim Atnarivik]

John Larkin <jjla...@highnotlandthistechnologypart.com> wrote:
: Since it would take forever to determine if a frequency was precisely

: 18.5 KHz, the world would have ended!

Ahh, another twist with 'practical' vs. 'pedantic' definition of
what counts as 'infinite', in latching relays or (ohmygod) 555's .
This is going to be a long thread...

Regards,
Mikko

[John Devereux]

Clearly JF has now come around to the practical definition. So all is
well, peace and harmony all around. :)


--

John Devereux

[Rich Webb]

On Wed, 9 Nov 2011 14:04:08 -0800 (PST), barabb...@email.it wrote:

>I'm using a double NE555 (called '556'), in which both halves act as a
>one-shot. I need them to be retriggerable, that is, if a trigger pulse
>comes before the completion of an ongoing pulse, the ongoing pulse is
>extended by an amount equal to the preset time constant.

The poor ol' 555; it just can't get any respect. ;-)

One tool that's available which may help is the 555 Designer from
Schematica <http://www.schematica.com/index.htm>. It's free to download
although some capabilities are masked in un-registered "demo" mode.
probably at least worth checking it out. There are a couple of options
shown for setting up a retriggerable input to a monostable layout.

#disclaimer <just a customer>

--
Rich Webb Norfolk, VA

[John Fields]

On Wed, 9 Nov 2011 14:04:08 -0800 (PST), barabb...@email.it wrote:

>I'm using a double NE555 (called '556'), in which both halves act as a
>one-shot. I need them to be retriggerable, that is, if a trigger pulse
>comes before the completion of an ongoing pulse, the ongoing pulse is
>extended by an amount equal to the preset time constant.
>
>I found several ways on the 'net to achieve this:
>
>1. 2.
>http://sam.electroastro.pagesperso-orange.fr/dossiers/555/555.htm
>(see "montages" n. 2 and 3)
>
>3.
>(Starting with a 555 in the usual monostable configuration) It's
>achieved by wiring the capacitor between pin 6 and GND and the
>resistor between supply voltage and pin 2. Pins 2 and 6 are wired
>together, so pin 7 is not used.
>
>4.
>Same as the usual 555 in monostable configuration, but reset and
>trigger are wired together: at each incoming pulse, the timer is reset
>even if the previous output pulse hasn't finished yet.

---
That won't retrigger because the output will go low every time RESET\
is asserted, regardless of the states of pins 2 and 6, and will stay
low for as long as the trigger pulse is active.

Here's the way I do it, which is the same as Montage n°2.

Version 4
SHEET 1 960 772
WIRE 224 144 -96 144
WIRE -192 208 -288 208
WIRE -96 208 -96 144
WIRE -96 208 -192 208
WIRE -64 208 -96 208
WIRE 256 208 160 208
WIRE -192 240 -192 208
WIRE -64 272 -96 272
WIRE 192 272 160 272
WIRE -96 336 -96 272
WIRE -64 336 -96 336
WIRE 320 336 160 336
WIRE 352 336 320 336
WIRE -192 368 -192 320
WIRE -96 368 -96 336
WIRE -96 368 -192 368
WIRE -192 400 -192 368
WIRE 224 400 224 144
WIRE 224 400 160 400
WIRE 320 464 320 336
WIRE -96 480 -96 368
WIRE 16 528 -32 528
WIRE 192 528 192 272
WIRE 192 528 96 528
WIRE -288 560 -288 208
WIRE 192 560 192 528
WIRE -288 688 -288 640
WIRE -192 688 -192 464
WIRE -192 688 -288 688
WIRE -96 688 -96 576
WIRE -96 688 -192 688
WIRE 192 688 192 640
WIRE 192 688 -96 688
WIRE 256 688 256 208
WIRE 256 688 192 688
WIRE 320 688 320 544
WIRE 320 688 256 688
WIRE -288 752 -288 688
FLAG -288 752 0
FLAG 352 336 OUT
SYMBOL Misc\\NE555 48 304 M0
SYMATTR InstName U1
SYMBOL voltage 192 544 M0
WINDOW 3 24 104 Invisible 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
WINDOW 0 33 19 Left 2
SYMATTR Value PULSE(5 0 1 .1U .1U .1 .5 10)
SYMATTR InstName V1
SYMBOL voltage -288 544 M0
WINDOW 123 0 0 Left 2
SYMATTR SpiceLine Rser=0
SYMATTR InstName V2
SYMATTR Value 5
SYMBOL res 336 448 M0
WINDOW 0 -42 36 Left 2
WINDOW 3 -57 70 Left 2
SYMATTR InstName R3
SYMATTR Value 1000
SYMBOL res -176 224 M0
SYMATTR InstName R1
SYMATTR Value 2meg
SYMBOL cap -176 400 M0
SYMATTR InstName C1
SYMATTR Value 1µ
SYMBOL pnp -32 576 R180
WINDOW 0 58 35 Left 2
WINDOW 3 52 64 Left 2
SYMATTR InstName Q1
SYMATTR Value 2N4403
SYMBOL res 0 512 M90
WINDOW 0 65 59 VBottom 2
WINDOW 3 68 59 VTop 2
SYMATTR InstName R2
SYMATTR Value 1000
TEXT 40 720 Right 2 !.tran 0 10s 0 .001s startup uic

--
JF

[John Larkin]

How long would it take you to decide if a signal was exactly 18.5 KHz?

John

[John Larkin]

On Thu, 10 Nov 2011 08:27:31 -0500, Rich Webb
<bbe...@mapson.nozirev.ten> wrote:

>On Wed, 9 Nov 2011 14:04:08 -0800 (PST), barabb...@email.it wrote:
>
>>I'm using a double NE555 (called '556'), in which both halves act as a
>>one-shot. I need them to be retriggerable, that is, if a trigger pulse
>>comes before the completion of an ongoing pulse, the ongoing pulse is
>>extended by an amount equal to the preset time constant.
>
>The poor ol' 555; it just can't get any respect. ;-)

The uA709 has the same problem.

John

[Fred Bloggs]

On Nov 9, 5:04 pm, barabbadr...@email.it wrote:
> I'm using a double NE555 (called '556'), in which both halves act as a
> one-shot. I need them to be retriggerable, that is, if a trigger pulse
> comes before the completion of an ongoing pulse, the ongoing pulse is
> extended by an amount equal to the preset time constant.
>
> I found several ways on the 'net to achieve this:
>

> 1. 2.http://sam.electroastro.pagesperso-orange.fr/dossiers/555/555.htm

There is such a thing as frequency-to-voltage conversion
http://www.national.com/ds/LM/LM231.pdf

[Okkim Atnarivik]

John Larkin <jjla...@highnotlandthistechnologypart.com> wrote:
> How long would it take you to decide if a signal was exactly 18.5 KHz?

As a practical man I'd say that one gets more and more accurate estimates
for the frequency when one measures longer and longer. Just like the
(dc) power gain estimate of a latching relay increases without bound
(which could be practically called 'approaching infinity') when measured
longer and longer. The dc-SQUID coincidentally has the same property:
power gain approaching (practically) infinity when signal frequency
approaches dc. MOS transistors would behave similarly if their gates
did not leak, if I'm not mistaken?

In the pedantic viewpoint the 18.5 kHz measurement accuracy could
be brought down to femtohertz level, maybe, before the sun expands into
a red giant. Even then one can imagine fleeing in a spaceship with the
measuring apparatus onboard. Reaching atto-Hz accuracy then depends on
whether the universe is cosmologically open or closed.

Actually, I'd
expect the convergence towards the 'exact' frequency to be slower than
linear in time, once the reciprocal of the 1/f noise corner of the
apparatus is exceeded. I'm too tired to think now what consequences
the Heisenberg uncertainty would bring in - it works to the other
direction I figure?

Regards,
Mikko

[Rich Grise]

Don Lancaster wrote:
> On 11/9/2011 3:04 PM, barabb...@email.it wrote:
>> I'm using a double NE555 (called '556'), in which both halves act as a
>> one-shot. I need them to be retriggerable, that is, if a trigger pulse
>> comes before the completion of an ongoing pulse, the ongoing pulse is
>> extended by an amount equal to the preset time constant.
>
> Replace it with a PIC, of course.
>

No "use a pic" posts are allowed without a schematic, software (firmware)
listing, and programming instructions.

Thanks,
Rich

[Rich Grise]

To the ordinary roob in the street, "infinity" is "how much electricity
is in the light socket." ;-)

Cheers!
Rich

[Rich Grise]

John Larkin wrote:
> On Thu, 10 Nov 2011 12:12:23 +0200 (EET), Okkim Atnarivik

>>John Larkin <jjla...@highnotlandthistechnologypart.com> wrote:
>>: Since it would take forever to determine if a frequency was precisely
>>: 18.5 KHz, the world would have ended!
>>
>> Ahh, another twist with 'practical' vs. 'pedantic' definition of
>>what counts as 'infinite', in latching relays or (ohmygod) 555's .
>>This is going to be a long thread...
>

> How long would it take you to decide if a signal was exactly 18.5 KHz?
>

I guess that depends on the parameters of "exactly." To the nth degree,
obviously forever, and that _still_ wouldn't get it, because before the
signal started, its frequency was zero.

;-)
Rich

[Jim Thompson]

Does the OP really need that accuracy, or is he attempting to know
H-sync so he can switch to the appropriate synchronization?

...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

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

[ehsjr]

Probably the only thing you need to change in the source code is:
processor 12c509
__config 0xfea

If using a 12c508 you should use processor 12c508 instead of
processor 12c509

The _config 0xfea line gives the PIC the configuration word.
That word configures the PIC for how you need it to operate
in general, before it attempts to execute any program instructions.
The configuration word in the source code tells the 12c508 to:
1) use the internal RC oscillator in the PIC
2) disables the ^MCLR pin (and tie it to +Vdd internally)
3) turn code protection off and
4) disable the watch dog timer.

I mention this, because you might need to change the configuration
word if you use a different PIC, and you might need to change the
hardware, too.

For example, AIRC the PIC16F84 doesn't have an internal RC timer, so
you would need to wire the R and C externally to that chip. That
information is seen in the configuration word definition. The
configuration word is defined in the datasheet. It all sounds very
complicated, and it is the first few times. But you can get help here.

As to why you can't get the 12c508 to work, assuming you changed
the processor 12c509 in the source above to processor 12c508:
I'm guessing your MPASM suite has nothing for the 12c509 (or 12c508
if you changed it), so you could look for that at microchip. Here's
what I have for the 12c508 in my MPASM Suite\LKR directory:
12c508.lkr
12c508a.lkr
12c508a_g.lkr
12c508_g.lkr

And in my MPASM Suite directory I have:
P12C508.INC
P12C508A.INC

I haven't tried the code - don't have a 12c508 - so what I posted
is just for reference and not a tested solution. :-( But if
you're missing the link & include files for the '508 you need to
add them to get it to work.

Ed

[Okkim Atnarivik]

Jim Thompson <To-Email-Use-Th...@on-my-web-site.com> wrote:
: On Thu, 10 Nov 2011 19:33:16 +0200 (EET), Okkim Atnarivik
: <Okkim.A...@twentyfour.fi.invalid> wrote:
: > In the pedantic viewpoint the 18.5 kHz measurement accuracy could

: >be brought down to femtohertz level, maybe, before the sun expands into
: >a red giant. Even then one can imagine fleeing in a spaceship with the
: >measuring apparatus onboard. Reaching atto-Hz accuracy then depends on
: >whether the universe is cosmologically open or closed.

: Does the OP really need that accuracy, or is he attempting to know

: H-sync so he can switch to the appropriate synchronization?

True, the OP doesn't need the femto-Hz accuracy.

Regards,
Mikko

[John Fields]

---
All he's trying to do is determine whether the H-sync freq is less
than or greater than 18.5kHz.
--
JF

[Jim Thompson]

Me, I'd run a XTAL (or tight R/C) controlled count-down to 18.5kHz,
use a digitally phase-jerkable oscillator and compare at the end of
each cycle. (Basically a frequency-only comparison... accuracy =
1/counter-length)

I've done this at much higher frequencies... as in satellite
down-links.

[John Larkin]

On Thu, 10 Nov 2011 19:33:16 +0200 (EET), Okkim Atnarivik
<Okkim.A...@twentyfour.fi.invalid> wrote:

Life's easier if you remove the requirement to measure exact
equality... of anything physical.

It's fun to trigger a scope from one atomic clock and look at a second
one on a vertical trace. Zoom up to, say, 1 ns/cm. It looks triggered.
Come back in half an hour or so, and it's drifted a little.

I'm not sure how people compare atomic clock frequencies in the big
leagues. They are talking stabilities of 1 part in 1e18 these days.

John

[Mike]

John Larkin <jjla...@highNOTlandTHIStechnologyPART.com> wrote:

> I'm not sure how people compare atomic clock frequencies in the big
> leagues. They are talking stabilities of 1 part in 1e18 these days.

> John

ISTR some of the ultra stable devices are optical, like aluminum or mercury
ions. They have a technique to count down to rf frequencies, so they can do
the comparison at high GHz frequencies. This makes it much easier to get
down to 1e-18 in a reasonable time frame.

Mike

[Jim Thompson]

Larkin follows Popular Science Magazine closely ;-)

[John Larkin]

On Thu, 10 Nov 2011 15:08:00 -0700, Jim Thompson
<To-Email-Use-Th...@On-My-Web-Site.com> wrote:

>On Thu, 10 Nov 2011 21:29:50 GMT, Mike <sp...@me.not> wrote:
>
>>John Larkin <jjla...@highNOTlandTHIStechnologyPART.com> wrote:
>>
>>> I'm not sure how people compare atomic clock frequencies in the big
>>> leagues. They are talking stabilities of 1 part in 1e18 these days.
>>
>>> John
>>
>>ISTR some of the ultra stable devices are optical, like aluminum or mercury
>>ions. They have a technique to count down to rf frequencies, so they can do
>>the comparison at high GHz frequencies. This makes it much easier to get
>>down to 1e-18 in a reasonable time frame.
>>
>>Mike
>
>Larkin follows Popular Science Magazine closely ;-)
>
> ...Jim Thompson

Have I mentioned lately what an idiot Thompson is?

John

[Jim Thompson]

On Wed, 9 Nov 2011 14:04:08 -0800 (PST), barabb...@email.it wrote:

>I'm using a double NE555 (called '556'), in which both halves act as a
>one-shot. I need them to be retriggerable, that is, if a trigger pulse
>comes before the completion of an ongoing pulse, the ongoing pulse is
>extended by an amount equal to the preset time constant.
>

I've been busy on a project, so I'm late to this discussion.
Retriggering a 555 one-shot _accurately_ is non-trivial.

But here's a way to do an accurate frequency compare...

http://www.analog-innovations.com/SED/FrequencyCompare_SED.pdf

which is a variation of a floppy data extractor I did in 1983...

http://www.analog-innovations.com/SED/FloppyDataExtractor.pdf

at GenRad, to handle extreme floppy-drive speed variations in portable
(briefcase) testers.

It boils down to being a "jerkable" VCO... in the frequency-compare
case, clocked by 1MHz... use a 555 if you must ;-)

I've also used this scheme on a satellite down-link for clock
restoration.