Hi Olin Latrop

In addition having an input held at 1/2 Logic the voltage between 0 and Vdd will cause the CMOS Device's internal gate to heat up.  Floating inputs is certainly the worse thing anyone can do with CMOS.

I recall the early days of CMOS where many engineers that had little practices in design using this new family.  I seen many products that actually got placed on the Market, only to come back and have to be reworked.  

Most had experience with TTL.  In TTL it was OK to float an input and get away with it.  But CMOS Device Inputs being Hi Z, allow them to charge up from pick up of the surrounding field (noise).  

In a CMOS INVERTER as an example, once the input gate charges enough to reach UNKNOWN LOGIC a voltage of 1.5 v above Vss (with a 5 V supply), the output of this gate starts to conduct between the two internal output FET.  Once your input reaches Mid Voltage of 2.5 Volts,  Both these FET's are on. allowing  current is flowing between the Source and Ground, using the two internal FET as a load.  After some time they might heat up enough to be damaged.

Well I am glad you asked this question.  Many people getting involved with CMOS today need to understand this major point.  


Good Luck, you just took another step up.

Dan Marchesani


Olin Lathrop wrote:

Brendan, I simply ignore unused pins.

Bad idea.

CMOS inputs are very high impedence and can pick up noise if not driven.
Noise on an input will increase overall current draw from the extra clock
edges, and can be worse if the input is held in the mid region.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu

-- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email listserv@mitvma.mit.edu with SET PICList DIGEST in the body

Wouter van Ooijen wrote:

Well if you have your unused Pins set as OUTPUTS, and floating, a few
things can happen.

A basic course on logic learned me (long ago) that when a particular
line of reasoning states that a contradiction is true, you are allowed
to take *any* proposition as beging true. So once you have achieved the
above effect (both floating and outputs) I guess a lot more than just 'a
few things' can happen..... (like the S/N ratio on the PICList suddenly
reaching 1.0).

Wouter van Ooijen

Hi Wouter van Ooijen
-------------------------------------------------------------------------
In the following I (Dan Marchesani)wrote

Well if you have your unused Pins set as OUTPUTS, and floating, a few
things can happen.  Accidental shorts to these pins can happen.  You'll
have to check the data sheets on what kind of protection there is
against this.  Shorts from a voltage source higher then the device's Vdd
may damage them.  Out side of shorts there is no real problem setting
them to OUTPUTS.  If I may add, as OUTPUT, the programming if it sends
logic changes to this unused pin, current during transit still normally
occur internally to the output gate.  Also if your supply source is from
a batt, then you might take this current into consideration, and not set
your unused pins to output.
-------------------------------------------------------------------

It is tough to try to determine what is the best way to handle so many possibilities.
Like many others here on the list, we are trying to sort the best method.  

Mr. Olin Lathrop wrote some really good points on the matter.  In my thinking of the many possibilities, I had to many sticking points.  Olin made a very good point of one thing I gave little thought about.  What will happen if we decided to tie our unused pins to ground and set them to INPUTS.

Well taking the possibility such as a processor going crazy.  Changing the I/O configuration from INPUT to OUTPUT.  Gee we are going to be  left with a short to ground.  

Yes there are many things to consider, but once the path is reached, there is a final solution that will best fit all possibilities.  

UNUSED PINS - should be set as OUTPUTS, and not connected to anything.  At some point in time later in the development of a project, these unused pins can simply be set as needed in F/W and hardware can be attached to the floating pin(s) without having to undo connections made if tied to something.

If  in the final development you may still have unused pins, there is no extra cost in the production model finally being  released.
 
Dan Marchesani
-- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email listserv@mitvma.mit.edu with SET PICList DIGEST in the body

Hi All

In trying to decide on a configuration, we take into consideration, what can happens that is not in our control.  Generally, once a design that goes out into the field.

We can make a list of these problems we know that can happen.  (This is a bit out of scope of unused pins per say.)  Make a list of these problem.  Look at it and then select either IN or OUT.  We'll find there is no perfect way.  In all our attempt to come up with a solution, we'll find, none that can fit all problems.  But some are better then others.  

======================================================
Wouter van Ooijen wrote to me and stated the following: 

A basic course on logic learned me (long ago) that when a particular
line of reasoning states that a contradiction is true, you are allowed
to take *any* proposition as beging true. So once you have achieved the
above effect (both floating and outputs) I guess a lot more than just 'a
few things' can happen..... (like the S/N ratio on the PICList suddenly
reaching 1.0). 

======================================================

For the NEWBIES, there is no doubt in regard to selecting unused pins as INPUT that floats.  Its a NO NO.  This is really basic.

Outside the scope where other things can and do go wrong it is a different story.  If they go wrong it is not the intention in general of the design to address all problems. We should try to address them, but there are times it can not be address until it has been field tested.

It is the intention of the project to work as designed, and if it does then that is good.  We as engineers should address as much protection into our designs as best we can.  Mainly to "protect" our self from the flack later....  But something will go wrong no matter what.  Even the best minds got flack in their faces......  You just have to take it the best you can, and go on.

Dan Marchesani
=====================
Olin Lathrop wrote:

In the real world, unused and used OUTPUTS are prone to taking a short,
and can be damaged, no more differently.

In the real world, your unused pins won't get shorted because they don't
go anywhere near the customer interface. (usually not connected at all)

Not to mention that you've got much bigger problems if something is shorting

-- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email listserv@mitvma.mit.edu with SET PICList DIGEST in the body