Unused outputs... short em?

[simmisj]

Hei.
I wanted to ask if it would be wise to short the outputs that I am not using.
For an example. If I take a 08M and only use one output and one input. Instead of pulling the unused pins high or low can't I just short them, to destroy them, and then I don't need the pull up/down resistors.

I'm thinking this in the way to use as little power as posible for my application and also to make it as small as possible.

The main questions are: is there a high posibility to destroy some other function of the chip and also if I short the output is it the same as if the chip did not have the output in the first place (when looking at the fluctuation it can couse when not using pull up/down resistor).

Thanks

Simmi

 

[BeanieBots]

Instead of pulling the unused pins high or low can't I just short them, to destroy them, and then I don't need the pull up/down resistors.

Of course you can just short the unused pins to destroy them. In fact, shorting output pins is a great way of destroying most chip, not just PICAXE. (once destroyed, you won't need the resistors either!).

Input pin on the other are another matter. It's a good idea (but not essential) to pull unused input pins to 0v via a resistor or by direct link. For pins which can be either input or output, just make them output, then they can't float. Or, pull down via a resistor just in case they ever become outputs and then get destroyed.

 

[Jeremy Leach]

I expect you'll get a few answers on this one !


Shorting to destroy pins - I can't see why you'd want to even contemplate this ! Imagine all the microscopic components packed together on the silicon wafer on the chip. You come along and short outputs and destroy output transistors. It's like setting a house on fire on a road and hoping it won't affect neighbouring houses.

Maybe you're confusing inputs and outputs? It's generally a good idea to have a pullup or pull down resistor on unused inputs, but it doesn't matter on outputs.

 

[MartinM57]

You're mad...but I like the way you think. I'm sure that sort of thinking will come in useful some time. Just not this time

 

[BeanieBots]

OMG, I just re-read and simmisj IS serious. I thought I was being sarcastic at somebody who had mistaken outputs for inputs, but no, I missed the point that simmisj is totally barking

I like the analogy Jez, especially if you consider terrace housing.

 

[geoff07]

Take a look at this: http://ww1.microchip.com/downloads/en/DeviceDoc/01146B.pdf

 

[simmisj]

Martin: lol thanks

So having a floating outputpin uses no current at all? I know there needs to be a circle for current to flow but I thought maybe internally it would take some current even when not in use.

But about the input pins. The idea about shorting it was to physically remove the connection. What is the difference in current draw with it grounded or shorted? Ofcourse I am talking about AFTER you short it
Does someone have a number from some test he/she made about keeping an input pin floated versus pulling it down to 0V?

Simmi

 

[BeanieBots]

This has been covered many many times.
A floating input does just that. It floats.
Every time it floats between one state and the next, all the internal transistors switch and hence draw a little power. The difference is quite small but in a very low power battery powered circuit it can make a difference.

 

[westaust55]

Hey simmisj

Why not beat Rev Ed to create the PICAXE 05M

get out the hacksaw and cut off the bottom end at a slight angle to remove physical legs ,4,5 and 6)

Then you have the supply the serial in, pin4 for input and pin0 for output.

 

[hippy]

Destroying the input or output drivers is a novel approach but it's impossible to judge how that could be done cleanly, without damaging anything else or how it would even achieve the desired goal.

The idea of blowing internal electronics up is not new; it's the basis of OTP and EPROM devices and security fuses inside chips; "fuse" being the obvious connection here. Willy-nilly blowing things up which are not designed to be blown up is a very different path, not recommended other than to see what effect that has before throwing the chip in the bin.

I recall a company involved in rail signalling designed a majority voting system where the loser was electrically destroyed to guarantee it played no part in further decision making, but have no details of how that was implemented.

 

[Dippy]

A floating 'Output' pin?
When a PIC port.pin is set to 'output' the driver is high or low.

I don't know the difference when floating or tied with an input pin.

Why don't you measure it yourself?
You could do it far more quickly than waiting hours for someone to reply.

My only comment would be that when Microchip do their specifications for low power consumptions they tie their input pins.
Tieing them provides some stability but as to the difference in power consumption I don't know.

All I do know is that if you are considering destroying I/O stages and/or PIC peripherals in an attempt to save power then you should call a Psychiatrist first.

If all this is an attempt to get minimal power consumption then go to Microchip's website and find the Application/Design Note on minimising power.
All the faffing around has already been done by engineers.

 

[simmisj]

All the faffing around has already been done by engineers.

Yeah that I know. But those chips are designed to be used in many types of applications not a specific one and thus providing some number of output pins and some number of input pins wether you use it or not.

I was wondering if by shorting them out and eliminating them I could make the chip like it would be custom made for use in applications with only one input and one output.

I guess people don't like this idea since now I am considered a crazy m***** f****

But experimenting is what I like. I have some 08M's around, not as many as I would like, I can try to short out and measure the current draw. And ofcourse see if they still work after shorting them

And yes my only goal is to take the power consumption to the absolute minimum.

You can think of my application like a temperature sensor with a LED output.

Pseudo code being something like:
Do
Sleep for a minute.
Wake up.
Turn on temperature sensor.
Measure temperature.
Turn off temperature sensor.
If TEMP > 25° LED = 1
Else LED = 0
Loop

And the LED would be blinking to save more power.

Simmi

 

[Hugh McCormack]

....I recall a company involved in rail signalling designed a majority voting system where the loser was electrically destroyed to guarantee it played no part in further decision making...

Oh the applications I see for that in Canadian politics at the present time!

 

[Dippy]

Simmi,

All this has been done before... apart from destroying parts of the PICAXE.

May I suggest getting onto the Microchip support service before wasting your time and money.
They may be able to change your mind or, at least, supply you with a comfortable straight-jacket.

If you start trying to pop I/Os and peripherals you may pop something you didn't plan to.
I don't know how predicatable it would be.

If you actually read PIC Data Sheets you will see that many can operate at very, very low currents with a little care from you.

And so many people get obsessed with faffing about with the PICAXE and forget about their other electronics.

Flashing an LED is an obvious step to saving power - they've been doing that since LEDs were invented.
But YOUR choice of LED could make significant differences to power consumption.
Replacing bipolar technology with MOS (where practical) can make huge differences.
But newbies don't realise this.
There are SO MANY techniques depending on your application.

Anyway, feel free to blow up your PICAXEs.
Let me know when you are going to start and then I can buy shares in Rev-Ed
As a famous detective said; "You'll thank me later."

 

[BeanieBots]

.....And yes my only goal is to take the power consumption to the absolute minimum.

You can think of my application like a temperature sensor with a LED output.
....And the LED would be blinking to save more power.

This just gets funnier and funnier

So you want to reduce power by a few fractions of a uA and you have an LED
Halve the LED resistor value and turn it on for a few milliseconds less and you will save more power than you could if you had a custom chip designed without the outputs you don't want.

As mentioned, read the PIC datasheet and app notes to get a feel for the numbers we are talking about here.
Then consider the current consumption of your ENTIRE circuit.
Then look at the spec of your battery.
Then ask yourself how much longer it would run even the PIC drew ZERO current.

 

[Dippy]

I'm using a peltier chip to power my PIC and display the temperature.

 

[hippy]

And yes my only goal is to take the power consumption to the absolute minimum.

There's an impressive 'absolute minimum' demonstration of an 08M flashing a LED on a set of batteries which has been running for quite some time now on the forum ( couldn't find the post with a quick search nor recall the author ).

The quest for an absolute minimum may ultimately be fruitless in requiring more effort and cost than it would actually ever save so beyond an academic perspective could be rather pointless -- if the energy used in destroying the internal drivers is more than the removal of them saves it seems you've incurred a loss which would be hard to claw back except over a very, very long time.

It's also often not worth worrying about absolute minimum when there are other ways in which consumption could be reduced, battery life extended, and particularly where length of life meets with end-user expectations. Any circuit is a compromise if it's to do anything at all and absolute minimum is actually subjective; 'absolute minimum while doing this' is the only real absolute. If a simple change to LED flashing or other operation saves more than blowing a chip part out of circuit then it seems obvious to me the best way to go. You'd have to exhaust all other paths before that becomes sensible.

No one will stand in the way of your experimenting, and I'm sure we'd like to hear the results if you do. Other than it being a goal in itself; what are you hoping to achieve with 'absolute minimum' ?

 

[simmisj]

My application is not that of what I mentioned. I just said you could think of it like that.
I know how to calculate the avarage power consumption and I am aware of the fact that the chip uses very low power compared to the devices that it usually controls.
In my application the chip is always powered but the "LED" is almost never on. So lowering the power consumption of the chip is a big factor.
I just wanted some feedback from the people here if they have done this before or not and also some comments fromt he proffesionals if it is theoreticly feasable.

Thank you

Simmi

 

[geoff07]

If you read the Microchip document that I posted a link to earlier you will see what the manufacturers think. Not all is relevant to Picaxe as we don't have access to all the instruction set but Chapter 2 explains how to deal with I/O pins, and also how to use I/O to power other devices such as sensors.

And I shan't be voting in THAT kind of election!

 

[simmisj]

Hippy: Didn't see your post before I posted my last post. Thanks for this nice information. I think I have seen that thread before and I will try to find it.

Well I can give an example of one of my projects that I have made. It's a circuit that turns on when I open the refrigerator and if it is not closed after one minute it starts to warn me that the refrigerator is open. The main power saving function of this circuit is ofcourse that the circuit is only on when the refrigerator is open.

I made it because my refrigerator tends to not close properly sometimes.
My main goal was to make it compact and to keep the power consumption in the minimum.

 

[BeanieBots]

I just wanted some feedback from the people here if they have done this before or not and also some comments fromt he proffesionals if it is theoreticly feasable.

Thank you

Simmi

Have I ever tried this myself?
Not exactly, though I have blown the odd ouput unintensionally.
You might be interested to know that the common way of telling when a chip is partially damaged is by its uncharacteristically HIGH current draw. This is because 'blown' involves a meltdown of the semiconductor device which results in lots of partial shorts within the chip.

Is it theoretically feasable?
To blow an output open circuit without damaging the rest of the chip. Yes.
To save current as a result. Very unlikely. (I won't say impossible).
Would it work if you tried it. NO CHANCE WHATSOEVER.

Several people HAVE won the lottery and that has odds of 14 million to one. Personally, I'd stick with the lottery as being a safer bet.

 

[hippy]

I just wanted some feedback from the people here if they have done this before or not and also some comments fromt he proffesionals if it is theoreticly feasable.

I doubt anyone's done it, not deliberately. And if they have done it accidentally would not have thought to see if it reduces power consumption.

Is it theoretically feasible ? I honestly don't know. I'd guess that it could work. Which I imagine puts you at the forefront of this exciting new research area.

I suspect you'll fare better with a scanning electron microscope, a de-capped chip and a laser to cut into silicon substrate than using a car battery and a 'quick zap'. It probably would count as interesting and valid research if done properly. It may even revolutionise the way ultra low-power chips are designed in future.

 

[eclectic]

Battery power

http://www.picaxeforum.co.uk/showthread.php?t=8353&highlight=current+year*

e

 

[Dippy]

If anyone could find it then eclectic is the man for the job....
I haven't looked - was it the right one?

Tip for simmi:-
Did you know that when you get down to low limits the PCB design can make measurable differnences?
I haven't tried burning out some of the tracks on PCBs... maybe that'll improve things

 

[John West]

The problem with blowing up part of a microchip is that - unless it was designed for it - any shorting out of an output puts the entire silicon chip at risk.

I used to hook the mains in random places to old transistor radio circuit bds. then plug them into the wall and watch the fun. I never knew if I'd get a slow glow or a nasty bang - or both - out of it.

What you're curious about is much the same. Doing to a silicon chip that for which it was not designed will have erratic results. It will indeed do something. But there's no way of telling how much. From it doing just what you want to killing the entire PIC - it's a not well advised gamble.

These are tiny devices with microscopic distances between active components. Even the tiniest of heat or spark internally is - on a larger scale - much like a random hand grenade or a lightning bolt striking a building. Who's to say how much damage will be done - or to what?

Have fun - and be sure to buy lots of spares. Rev. Ed. will thank you for it.

 

[lbenson]

... demonstration of an 08M flashing a LED on a set of batteries which has been running for quite some time now ...

"Low-Power Battery Backup Reference Design" -- http://www.picaxeforum.co.uk/showthread.php?t=8353

Two years, 10 1/2 months now, and still blinking every 7 seconds on 3 AAs.

[Yes, same as eclectic found--I hadn't yet read to the end of the thread.]

 

[hippy]

Yay! That's a very good result and it's approaching its third anniversary. Fingers crossed and sorry for forgetting it was your work. Plus thanks to eclectic who always does a great job as search engine.

I also enjoyed John West's analogy to lobbing a hand grenade into a building. That's pretty close to the mark as witnessed by the magic smoke released and bits of chips flying off when done accidentally. A controlled and directed detonation is a different thing and it wasn't sarcasm in post #22.

 

[geoff07]

I think that Ibenson's led will blink for many years yet.

I have an HP-14B calculator that HP gave me on their 50th anniversary in 1989 (I used to buy a lot of minicomputers from them), that I use frequently, and that is still running on the original three little hearing-aid style cells after 21 years!

And my hen-house door is opened and closed each day by a pic with 4xAA and they last typically about 3 years. The pic controls a motor that winds the plywood door up on a thin cord, so real energy is used.