Halving Pulse Width/ Pulse Width Division

Flip it over so that "on" time is controlled by the Discharge Pin, then invert the output. ...Jim Thompson

Sounds like a job for a transistor and an R/C circuit at the base.

The likely reason for the duty-cycle limit is the current limit of the discharge transistor. So, it'll get shorter if you use a smaller timing capacitor (and larger current source resistance). There's limits to that, so a second approach would be to use two '555s, one for period-setting, and a second wired for pulse-time-setting, which is the 'monostable' function. A dual version of the NE555 is NE556, so this can still be a one-chip solution.

On Saturday, April 2, 2016 at 4:16:18 PM UTC-4, Chris wrote: However, I still need to

One of the simplest ways is to use an arduino. If you buy them from China , they are pretty cheap. And the first example of programming is to blink a light. You can select delay times to the millisecond. So if you want t o have a delay for say 90.123 seconds , no problem. Want the on time to be .025 seconds , no problem.

Yes more expensive than a 555 for one signal, but cheaper for three or more signals.

eBay item number:

Dan

On AliExpress I found 5 digispark modules for $1.47 including shipping. So that is 30 cents a piece.

Dan

Amazing! When Pis are down to that price, I might buy one. ;-)

OK, I'm going to go with whit3rd's suggestion, since I have a massive pile of 555s and need to use them up anyway.

Next question is (different app altogether) with a 555 in monostable mode, I need to generate a on-off pulse of 150ms duration, but triggering it with a typical momentary push-button switch takes more than 150ms, so it's not working as it should. I need to somehow trigger it with a much shorter pulse in order to reliably get a 150ms one out. Any ideas?

I'm having a pretty good time doing Pi programming at the moment. Once you figure out the various Linux incantations required to get control of the hardware, it's pretty slick. Running the compiler on the target system is a good way to make sure your toolchain is archived along with your code. ;)

The UART has a tty connected to it by default, and also gets kernel messages. That can make things a bit strange when you first try it, but it isn't hard to turn off.

Use ten! Chains of jittery monostables and tables covered with BNC cords...the physics grad student ethos of the early '80s, in a nutshell. ;)

Well, getting into the spirit of the thing, you can AC-couple the first

Cheers

Phil Hobbs

I was just kidding. I've been using a Pi for about 4 years now in an embedded application as a timelock and -touch wood- it's been fine. As they continue to proliferate I'm sure they'll end up finding all sorts of obscure uses the developers could never imagine.

Can you throw us a bone here? How does that work?

Series RC between output and trigger, resistor to ground from trigger. Pick the time constant (R_shunt C) to be something adequately short such as 10 ms, and put 1k or something in series to limit the current through the protection diodes. (R_series

Like this...

Not elegant... just takes advantage of Discharge Pin being Open-Drain (or collector, depending on vintage). ...Jim Thompson

Hello, hello? What's going on here? I thought you'd ditched PSpice some years ago in favour of the vastly more "cost effective" LTS, Jim?

I don't like LTspice's schematic entry... as klutzy as using Paint

But I regularly test circuits generated in PSpice with LTspice... just click on the .CIR file and select "Open with" LTspice.

I also do that now with TopSpice (customer request that Spice models run on LTspice and TopSpice).

Which presents problems...

TopSpice didn't, until recently, cope with pins named +IN, -IN, etc (math expressions). TopSpice also doesn't know what ATANH means, and can't do a .DC LIN PARAM... :-(

LTspice overly linearizes many mathematical expressions and regularly barfs on anything that is really fast with a G-source.

(Besides PSpice is quite cheap NOW... I'm stopped "tithing" maintenance when the Cadence/OrCAD cretins stopped supporting PSpice Schematics and required using Crapture... gag me with serving spoon ;-) ...Jim Thompson

I'm impressed! Thanks for that, Jim.

You're obviously not using the std ckt:

Timing is approx ln2 x C x R so you need an ON time R of 5% of the sum of ON and OFF time R's to get 5% duty and 1.25-2.5% for your wishlist duty cycle. Using 1.25% as example ends up like so:

Please view in a fixed-width font such as Courier.

. . . . VCC . | . | . [R] . | 555 . 1n4148| . --|DIS . | | . | [80xR] . | | THRESH . ------+------>TRIG . | . C --- . --- . | . | . | . --- . GND . . . . T=ln(2) x C x 81R=90s T ON neglects 1N4148 drop . . . 80R~100K makes C~1500uF . . R~ 1.2k 80R~100k . . . VCC . | . | . [1.2k] . | 555 . 1n4148| . --|DIS . | | . | [100k] . | | THRESH . ------+------>TRIG . | + . C --- . 1500uF --- . | . | . | . --- . GND . .

You are quite welcome! ...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

Simplest fix for 2.5/1.25 % nominal duty select:

Please view in a fixed-width font such as Courier.

. . VCC . | . | . [1.2k] . | . +-----o . | | . [1.2k] | JMPR . | | . +-----o 555 . 1n4148| . --|DIS . | | . | [100k] . | | THRESH . ------+------>TRIG . | + . C --- . 1500uF --- . | . | . | . --- . GND . .

Also is most flexible.

I'm not seeing that from here sometimes they mix lot and piece prices, (on the same page) are you sure that's not $1.47/piece