1. Home
  2. Electronics Design
  3. Schmitt trigger pulse stretcher, How big an R?

Schmitt trigger pulse stretcher, How big an R?

I'd like a to make a pulse stretcher.. with a pulse output of ~100ms (I might need less.) I've got a few unused gates, 74HC14 and Nand gate. So the "standard" diode, R and C.

formatting link
(I was too lazy for ascii art.)

For long times (and short input pulse times) I need a small C and big R. I need the input pulse long enough to discharge the cap, given the max current of the IC (~10-20 mA) I need C of 10 nF and R = 10 meg ohm.

Am I asking for problems with such a big R? (Or what's the input bias current of a 74HC14?) This will be in room temp conditions.

For fun I tried 100 meg and 1 nF. It worked fine on the bench.

George H.

Reply to
George Herold
Loading thread data ...

For all intents and purposes the input Z is just capacitive (~4pF) until you exceed the rails then the ESD structures kick in.

So the limit will be the strays on your board. I have never in my lifetime used a 100Meg resistor >:-} ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142   Skype: Contacts Only  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 

             I'm looking for work... see my website.
Reply to
Jim Thompson

Grin, Well I grabbed the 100 meg to make a x100 'scope probe... (otherwise the probe loaded it down) And then I thought it would make a good check.... OK I'm off to try 1G and 100pF, if that works I'll feel pretty comfortable with 10 Meg.

George H.

Reply to
George Herold

1 Gig gave too short a pulse... I'm getting leakage through the diode. (I stuck my soldering iron tip on the diode body.. smaller pulse width, more leakage.)

George H.

Reply to
George Herold

Reasons you don't want to go too high on the R:

1: The diode (as you found). I think a transistor with B&E shorted makes a better diode. Junction diodes leak less than Schottky diodes. A buffer into an open-collector transistor _might_ leak less -- but I won't guarantee that!

2: Take a perfect diode with no leakage whatsoever. Wipe it with your greasy, salty fingers. Leave flux on the board. Put the assembly in a humid environment. Poof -- leakage. Ditto the cap.

I generally try not to use anything higher than 1M, and then I keep an eye out for inaccuracies caused by using it. This is probably overly suspicious -- but it's what I tend to do (never mind that I'm working on a board right now that has an 0603, 1.24M resistor on it).

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com 

I'm looking for work -- see my website!
Reply to
Tim Wescott

Please understand that working gigaohm circuits need special construction techniques, like selected insulation materials and guard rings. Also, the proper things to probe the voltage are low-current operational amplifiers (or just FETs of course).

For your original problem, consider a shorter time pulse stretcher triggering a 555 for creating the 100 ms pulse.

--

-TV
Reply to
Tauno Voipio

Yeah I tend to think of 1 Meg as a max. The 1G was just to "explore" the edges. If it works with 1G then 10 Meg should be fine... (at least that's the story I'm telling myself.)

George H.

Reply to
George Herold

You could use a PNP emitter follower, or an FDV301, to discharge the cap harder. That improves the tradeoff.

The HC14 should leak picoamps. 10M should be fine.

You can cascade stretchers to improve the ratio.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

Transistor c-b junctions are generally much better diodes than diodes. BFT25 leaks fA.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

A properly clean board will be OK at Gohm levels without exotic techniques. Tohms are more difficult.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

I made a current source (voltage reference->R-> opamp to common of Vref.) that at the low end makes 10nA with the 1 Gig resistor. (Which is why I have 1G in stock.) I've not had any problems with it, (knock on wood). I test it with (1%) 100 Meg and 10 Meg R's it's good to 1%, so leakage has to be below 1nA... 100 pA range.

Yeah.. If I added another IC I can make it a flip-flop and then I wouldn't need the pulse stretcher... that may be a better solution.

George H.

Reply to
George Herold

Can you wire up a CD4040 or similar as a pulse stretcher? I'm thinking it'd at least take three chips once you added necessary glue logic, which wouldn't be good. Then you could have a good fast count (for small-value caps and reasonable-value resistors), yet still have a long delay.

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com 

I'm looking for work -- see my website!
Reply to
Tim Wescott

No, wait -- CD4060. Hook up an RC oscillator per the National data sheet, and disable the oscillator when pin 3 (Q14) is high, with an open- collector NPN transistor. Reset to start.

On power up, you get a "stretched" pulse, then the count stops when Q14 goes high. Your pulse to be stretched goes into the reset input.

You'd need an inverter on Q14 to get the right sense, but George is using an inverter anyway, so that's "free".

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com 

I'm looking for work -- see my website!
Reply to
Tim Wescott

Yeah I'm coming to that conclusion. I read some spec sheet ('HC14) that said bias current was 0.1uA max and 1uA over the temperature range... But with my 1G R I heated the IC (with soldering iron) and saw no change in the pulse width. (1n4148 repalced with 2n3904 BE) Sorta 1 ms out of 100 ms, so change in leakage with temperature is less than ~50 pA (5nA/100) ... measuring leakage with a pulse width.

George H.

Reply to
George Herold

Oh right the c-b leaks less.. oh well the b-e was good enough.

George H.

Reply to
George Herold

Use a 555? IIRC, the incoming pulse just sets a flip-flop rather than discharging the cap.

Or, if you're a 555 hater, use a '123 (yea, yea -- there's '123 haters, too).

Or a set of NOR gates to make a SR flip-flop (for fast acquisition) with your Schmitt-trigger timer for "off".

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com 

I'm looking for work -- see my website!
Reply to
Tim Wescott

If you use rosin flux (No-Clean in Sn63 paste is good old RMA) there aren't many issues of that sort below about 100M. I usually start using guard rings above 50M or so, because I'm generally after high accuracy. I've used SMT resistors as large as 50G with no problems.

You do need to use guarding, and put a slot under the resistor so you can get the flux out.

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 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

1N914 types can be very leaky. They may be non-lithography chips, just diced from a big wafer, with edge damage. And photosensitive. 
--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

This gives you a substantial (10:1?) advantage... |\ .-[R]-+---+----| >o-- | | | |/ |\ | |o---+---| --- C |/ |\ | 'HC14 | | === ===

How long is your input pulse, the one being stretched?

Cheers, James Arthur

Reply to
dagmargoodboat

Oh that's cute...(I think that is what JL suggested) I'll mash it up. (Does it matter if the R goes to the supply rail instead?)

The input pulse is an external trigger. I thought requiring a ~100 ms input pulse was too much... 10 us seems about right to me.

George H.

Reply to
George Herold

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.