Simple pulse stretcher

[George Herold]

Hi guys, I need a little digital pulse stretcher. I've got pulses
that are 5ns to 25ns in width, and I'd like to bump them all up to all
be at least 10ns. I've got a few spare 74HC14 inverters. I was
thinking of just feeding one through a diode followed by an R/C to
ground. Any other ways? There might be an extra AND gate too.
(not my circuit so I'm not quite sure.)

Thanks

George H.
(and who ordered the second t in stretcher)

 

[Joerg]

Hi guys, I need a little digital pulse stretcher. I've got pulses
that are 5ns to 25ns in width, and I'd like to bump them all up to all
be at least 10ns. I've got a few spare 74HC14 inverters. I was
thinking of just feeding one through a diode followed by an R/C to
ground. Any other ways? There might be an extra AND gate too.
(not my circuit so I'm not quite sure.)

The diode thing is what I usually do. But 10nsec through a 74HC14? Now
that's a strrrrretch.

 

[George Herold]

Isn't 10ns width just a wee-bit tight for 74HC parts?

                                        ...Jim Thompson
--
| James E.Thompson                                 |    mens     |
| Analog Innovations                               |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon athttp://www.analog-innovations.com|    1962     |

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

OK, maybe I should stretch to 20 ns.

George H.

 

[George Herold]

The diode thing is what I usually do. But 10nsec through a 74HC14? Now
that's a strrrrretch.

OK thanks, I'll do 20ns then.

George H.

 

[Joerg]

OK thanks, I'll do 20ns then.

Still a bit of a stretch but if you need pulses this short there are
faster versions of the HC14.

 

[George Herold]

Something like this:

http://www.abvolt.com/misc/pulse_stretch.jpg

You got the idea.

/*
Placeholder for banal comments
*/

Vladimir Vassilevsky
DSP and Mixed Signal Designswww.abvolt.com

Yeah I was thinking of something like that too. The diode thing was
an easy 'hack' and seems to be working just fine. (I stuck in 10k
and 10pF and have about a 50ns minimum pulse.) Life is good.

George H.

 

[Joerg]

Yeah I was thinking of something like that too. The diode thing was
an easy 'hack' and seems to be working just fine. (I stuck in 10k
and 10pF and have about a 50ns minimum pulse.) Life is good.

Diode plus Schmitt is the better solution. The main reason is that doing
it with piped-in gates relies on their prop delays. I have seen that
blow up in production when people did it. All it takes is one or more
manufacturers improving a semiconductor process, things become faster,
suddenly the fix ain't working no more. Or only sometimes. Datasheets
typically do not specify a guaranteed lower limit on prop delay. For
example, there is no entry for it in the "min" column for the 74HC14
from TI:

http://www.ti.com/lit/ds/symlink/sn74hc14.pdf

 

[Jamie]

Hi guys, I need a little digital pulse stretcher. I've got pulses
that are 5ns to 25ns in width, and I'd like to bump them all up to all
be at least 10ns. I've got a few spare 74HC14 inverters. I was
thinking of just feeding one through a diode followed by an R/C to
ground. Any other ways? There might be an extra AND gate too.
(not my circuit so I'm not quite sure.)

Thanks

George H.
(and who ordered the second t in stretcher)

I think you're going to have a little problem with that inverter in
regards to the speed you are referring to.

Jamie

 

[[email protected]]

Hi guys, I need a little digital pulse stretcher. I've got pulses

that are 5ns to 25ns in width, and I'd like to bump them all up to all

be at least 10ns. I've got a few spare 74HC14 inverters. I was

thinking of just feeding one through a diode followed by an R/C to

ground. Any other ways? There might be an extra AND gate too.

(not my circuit so I'm not quite sure.)

The 5ns is way out of line for HC. You can apply the pulse directly to an emitter follower loaded with parallel RC and then to HC to stretch it to anywidth you want. The BE junction is your diode. HC will not work reliably across VCC and temperature with anything less than about 25ns.

 

[George Herold]

---
"Pulse stretcher" implies that the leading edge of the source pulse
should be presented at the input of the load simultaneously with its
generation, and that its fall should occur after the desired delay.

The former is, of course, unrealizable, while the latter is easy.

How much delay can you tolerate between the generation of the leading
edge of the source pulse and its presentation to the load?

Yeah there's already ~20ns of latency between the input pulse and the
output so a bit extra ain't too bad. I'll knock the 10pF down to 4.7

George H.

 

[George Herold]

The diode+RC thing is perfectly feasible with an HC14, likely at 20 ns out and
certainly at 30.

A bit of inductance in series with the R would be interesting.

Here's a variation on Vlad's circuit:

https://dl.dropbox.com/u/53724080/SED/Stretcher.JPG

--

John Larkin                  Highland Technology Incwww..highlandtechnology.com  jlarkin at highlandtechnology dot com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators- Hide quoted text -

- Show quoted text -

Whee... fun!

George H.

 

[George Herold]

Yeah in this case there's certainly already some delay. But it
doesn't really matter.
OK it matters.. it all adds to the built in latency of the device.
But at the moment there's some 'software' latency that's ~x10's as
bad.
So I can burn 10's of nanoseconds with abandon.

George H.

 

[George Herold]

That's pretty much unavoidable. I doubt that it matters to George.

 and the


How do you know that? I didn't specify the parts, other than max PD. Theycould
be implemented as a tapped delay line, passive or silicon. Or one could trim Vcc
to tweak prop delay. There are lots of possibilities.

Post something better.

--

John Larkin                  Highland Technology Incwww..highlandtechnology.com  jlarkin at highlandtechnology dot com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators- Hide quoted text -

- Show quoted text -

In this case I'm timing a pulse stream. But absolute time doesn't
matter as much as relative time. So as long as the prop. delay
doesn't stop more than one pulse from traveling down the signal chain,
then it may not matter.

George H.

 

[[email protected]]

The 5...6 ns per gate is what you can practically expect from today's HC.



VLV

Please link to a datasheet- all I'm seeing is 20ns nominal performance on both Tpd's and minimum pulse widths into edge activated inputs like clk's, R's, and S's.

 

[George Herold]

Another one of JL's non-working "designs".  A quick check of 74HC
proves my supposition that a 5ns wide pulse goes nowhere... at least
reliably... something JL never worries about... but that's what you
get when you're pompously ignorant.

                                        ...Jim Thompson
--
| James E.Thompson                                 |    mens     |
| Analog Innovations                               |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon athttp://www.analog-innovations.com|    1962     |

I love to cook with wine.     Sometimes I even put it in the food.- Hide quoted text -

- Show quoted text -

Yeah Jim the 5ns pulses weren't getting counted....
Sorry I missed a reply to you. There's a comparator on the input and
I put the diode R/C thing between it and the first 74HC14 gate...

George H.

 

[George Herold]

The 5ns is way out of line for HC. You can apply the pulse directly to anemitter follower loaded with parallel RC and then to HC to stretch it to any width you want. The BE junction is your diode. HC will not work reliablyacross VCC and temperature with anything less than about 25ns.

Cool 25 ns is just what I'd expect if I cut my C to 4.7pF

Thanks Fred.

George H.

 

[[email protected]]

truth? Take a good scope, pulse generator and see for yourself.

Okay, you know what you measured, but you need to be real careful with layout using that parts family. It used to be a point of advantage for a logic family that it was unresponsive to full logic level transitions of durationless than the fastest edge rates in the system. Interwiring capacitive coupling into a nice high impedance input converts the transition into a nice square pulse there. I think Lenin covered that one.

 

[[email protected]]

A *triggered* pulse generator.

 

[George Herold]

Okay, you know what you measured, but you need to be real careful with layout using that parts family. It used to be a point of advantage for a logic family that it was unresponsive to full logic level transitions of duration less than the fastest edge rates in the system. Interwiring capacitive coupling into a nice high impedance input converts the transition into a nice square pulse there. I think Lenin covered that one.

Hi Fred, As Vlad said the HC series was really doing just fine with
10ns pulses and perhaps only 'missing' 1% of the 5 or 6 ns ones. I
don't have a commercial pulser that will do anything less than ~20ns,
so it's hard to test. I had a pulse swallower circuit on a piece of
copper clad, but I might have ripped it up to make something else...
Anyway I don’t see any reason not to play it safe and make the minimum
pulse 25ns.

My ‘scope is only 200MHz, so I’m not really sure I even believe the
5ns... what’s that rule of thumb for ‘scope bandwidth/ rise time? tau=
1/(3*BW) ?

George H.

 

[Spehro Pefhany]

Hi Fred, As Vlad said the HC series was really doing just fine with
10ns pulses and perhaps only 'missing' 1% of the 5 or 6 ns ones. I

The tinylogic one-shot that John mentioned is guaranteed to trigger
off 2.5ns pulses (with a 5V 'Vcc'). Available in a nice friendly
0.65mm pitch 8-pin package (also a 0.5mm pitch or 1 x 2mm BGA if you
are in need of pain).