Propeller Schmitt Trigger Inputs

Phil Pilgrim (PhiPi)

In another thread the Propeller's lack of a Schmitt trigger input option was mentioned. With one spare pin and a few resistors, however, it's possible to convert as many inputs (to be read individually) to Schmitt trigger inputs as you want. In the following schematic, A0 to A2 are Schmitt triggered inputs, and A3 is the feedback pin. Before reading any pin, set A3's output state to the state last observed on the pin being read. This provides the positive-feedback biasing required to create a hysteresis band.



Using 4.7K input resistors and 470K feedback resistors, you will get about a 1% hysteresis band. Lower-valued feedback resistors will increase the hysteresis.

-Phil

deSilva

Quite ingenious! An idea of your own? The corresponding hardware solution would require a flipflop for each port..

It works to nearly upto Rin = Rfeedback...
But note that bandwidth - as always - is constrained by your sampling intervalls; it will not work if the signal changes faster.

Phil Pilgrim (PhiPi)

deSilva,

Right. It's only effective if you sample at the Nyquist rate or faster.

This is a technique I came up with years ago to sample phototransistor outputs from a wind instrument. There were four fan blades arrayed in the cardinal directions, each with a 50/50 black/white disk on the hub and an IR emitter/receiver pair. The "virtual Schmitt triggering" used five pins on a PIC: four for input and one for feedback.

-Phil

Graham Stabler

very clever

Don D

I nominate this as a worthy entry for the good thread index.

Maybe a one-source compilation of ingenious virtual circuitry would be useful for the prop community?

deSilva

ANY compilation of the 1001 things from this forum would be useful

Fred Hawkins

(Well duh) Or more properly, when stuff goes over my (yours too?) head, fall back on Wiki for an explanation.

In electronics, a Schmitt trigger is a comparator circuit that incorporates positive feedback.
 
When the input is higher than a certain chosen threshold, the output is high; when the input is below another (lower) chosen threshold, the output is low; when the input is between the two, the output retains its value. The trigger is so named because the output retains its value until the input changes sufficiently to trigger a change. This dual threshold action is called [b]hysteresis[/b], and implies that the Schmitt trigger has some memory. 

The benefit of a Schmitt trigger over a circuit with only a single input threshold is greater stability (noise immunity). With only one input threshold, a noisy input signal near that threshold could cause the output to switch rapidly back and forth from noise alone. A noisy Schmitt Trigger input signal near one threshold can cause only one switch in output value, after which it would have to move to the other threshold in order to cause another switch.

http://en.wikipedia.org/wiki/Schmitt_trigger

OzStamp

Hi Fellow Proppers



Single gate schmitt trigger input for Propeller or other.
The TLC555 low volts 555 timer makes an excellent real schmitt trigger also.

Big note * the input to output is INVERTED ..

Added the image as well in jpeg· ..
Thnx for the pointer John T

Will be in touch with you later.. just got back from Sydney.

Cheers
Ron·· Melbourne Oz
Home of the PropBus ...





<!--EndFragment -->

Post Edited (OzStamp) : 11/11/2007 4:59:02 AM GMT

QuattroRS4

Ron,
Seems to be an issue with 'Snippy0001.jpg'

John T

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
'Necessity is the mother of invention'

Beanie2k

What's wrong with a 4093?

OzStamp

Hi Beanie

Nothing wrong with a 4093..

Also chips like 40106· + 4584··are fine also.

Maxim even has some dedicated pushbutton schmitt/debounce chips..

Philips and I believe TI also have some single gate Schmitt chips..
From memory they are called PICO gates ...or similar.

PHiPi circuit idea is really cool and does work.

We did something similar with the SX chip and the comparator in/output
we sampled an analog signal via a resistor divider and had a trip point.
Some feedback and it created a handy signal for a "battery charge or no charge.."

cheers

Ron Melbourne Oz

deSilva

This has become a little bit confusing now..
ANY chip able of bi-stable states (which generally means containing a kind of flip-flop) can be utilized as a Schmitt trigger. What Phil showed you was, how a Schmitt trigger worked! The standard circuit is this one below.

He especially pointed out that using the speed of the Propeller (to fulfill Nyquist criteria) and its internal memory, just ONE feed back line suffices to emulate MANY Schmitt Trigger inputs.

If someone is interested I can post a list of 300 chips with Schmitt-Trigger input and two or three circuits of how to build a Schmitt trigger from transistors...

Post Edited (deSilva) : 11/11/2007 9:36:13 AM GMT

deSilva

OzStamp said...
..We did something similar with the SX chip and the comparator in/output
we sampled an analog signal via a resistor divider and had a trip point.
Some feedback and it created a handy signal for a "battery charge or no charge.."

I cannot see the connection at the moment... What is "similar"? The use of a resistor? Handlng of an analogue signal?

Fred Hawkins

OzStamp said...

Philips and I believe TI also have some single gate Schmitt chips..
From memory they are called PICO gates ...or similar.

TI has many (1 gate, 2 gate and 3 gate versions, inverting or not...) This one is their little logic chip for 3.3v and·has 3 gates.

http://focus.ti.com/docs/prod/folders/print/sn74lvc3g17.html

To poke around in their little logic chips, try this finder tool

http://focus.ti.com/paramsearch/docs/parametricsearch.tsp?familyId=339&family=logic&uiTemplateId=SZVI_T

Beanie2k

Cool. I was a bit mystified because some of the proposed solutions seemed to me a bit over-complicated. I guess it's another case of the "green eyeshades syndrome" caused by doing electronics design at for-profit companies for too long.

deSilva

@Beanie2k: You were mystified by what? And what "de-mystified" you?

Post Edited (deSilva) : 11/11/2007 9:34:52 AM GMT

OzStamp

Hi De Silva

My last comment was in reference to my posting months ago.
Suggestion similar to Phils suggestion.
I added this principle to a SX 28 circuit see post
http://forums.parallax.com/forums/default.aspx?f=7&m=204932

cheers Ron

deSilva

@Ron: I see. You created a 1.5 bit DAC to generate 1.5, 2.5, or 3.5 V and used it as a comparator reference input, something not available at the Prop.

The basic idea Phil used for his Schmitt-trigger was using the - single - trigger point of a Prop input by biasing the signal (which in fact is not reactionless, but high value resistors will do their part....)

So your batterie charge monitor could look something like this on the Prop:

            .---.
            |   |  1M
            |   |  ___
            |   >-|___|-. 470k
            |   |       | ___
            |   <-------'|___|-<
            |   |
            |   |
            '---'
(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)

Post Edited (deSilva) : 11/11/2007 1:04:08 PM GMT

bambino

Very nice Phil,
I have been thinking of Schnitt-Triggers this week. I was toying with the idea power supplies. Using a zener to bias an op amp running the gate of a Mosfett in a buck topography. The constant swithing is getting the fett to hot because it doesn't really let it shut off. I was hopeing to get a comparator and keep it anolog, but it's nice to know the Prop could fill in the gap.

I am not fimiliar with Nyquist criteria however, Is that speed compatable with Buck switching?

Mike Green

bambino,
Do a web search on "wiki nyquist" and "wiki nyquist frequency" and you'll get a much better explanation than I can provide.

Tracy Allen

Clever idea, Phil, and good for counter inputs where the edges are slow or noisy.

I was thinking about how that could be applied to the COG counters, but came up short. The EDGE detect feedback modes of the counters have negative feedback, B := !A1. Even if there were a positive feedback mode, B := A1, or if one were to put a fast inverter on the B pin, there would still be a runt pulse, due to the one cycle delay, and that would cause the counter to increment when that runt pulse is tested as A1 & !A2. No go. It would take a counter mode that does not exist. (or just an external Schmitt trigger.)

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

deSilva

That was also my first idea: Counters! But even if it worked you would need one for each channel... Phil's main idea is not a Schmitt trigger, but MANY Schmitt triggers.

bambino

Mike,
I don't know if it could work well or not. I am thinking you would need 50khz to 500khz for the PWM out. Is the prop nyquist capable of 25khz. At any rate I would still like to make it analog.
A cool circuit none the less!!

deSilva

@Bambino: If you are talking of the Schmitt algorithm there is no PWM.

(a) You set the feedback to the old value.
(b) You read the input
(c) You update the memory

That's all.

The issue is, when you do not come back often enough to this piece of code.
Consider a drop of the line from 1 to 0 , and than going up to 1/2. You SHOULD read 0, but you do 1, if you do not get this drop. You will get it when you sample faster than twice the freqiency you are interested in (and to avoid aliasing, just use a hardware low-pass, best of second order. You see: Doing Schmitt triggering right in software is not exactly simple

But I think you can survive by spending a COG for updating the inputs. This could work @ 2 MHz/number_of lines.

Post Edited (deSilva) : 11/11/2007 11:18:34 PM GMT

bambino

No, PWM is a poor choice of words. I was just refering to the output signal controlling the mosfett signal being read by the Schmitt trigger. If hysterisis could be used to help buffer the switching frequency then the Fett might be able to run cooler.
I don't know, It's just food for thought.