Peter wrote:

>>> Potential and voltage are one and the same in theory but in practice it
>>> is more complicated.
>>
>> I'm not sure this is on firm ground. I'm no ace in physics terminology, but
>> I think "voltage" is a term for the value of an electric field potential,
> 
> The definition of potential is 'the energy required to move a punctiform 
> charge unit from the point being measured to infinity'. All normally 
> measured voltages are *relative* differences of potential, not absolute. 

I still don't see how you can make a difference between 'voltage' and
'potential' -- at least not any difference that would be relevant to the
Seebeck effect. (At the very least, you should be more careful with the
papers you cite in a given context: the one you cited uses 'potential
difference' and 'voltage difference' as synonyms :)

Every voltage is relative to something. Every potential is relative to
something. No difference here. (Like in a normal circuit: voltages are
usually given relative to ground. They represent potentials relative to the
generally not known -- in absolute terms -- ground potential.)

Absolute potential is relative to infinity, and has a voltage that is given
relative to infinity. Difficult to measure, as you state, but still
existent as concept. No difference either. (You could say, for example,
that a certain potential has 1 kV -- absolute, i.e. relative to infinity.
That 1 kV would then be its voltage, no?)

> In this sense a voltage is a potetnial *difference*, 

A voltage is a difference just as a potential is a difference. Either must
be referenced to something (this 'something' includes infinity) to make any
sense.

> but if only a potential exists then a voltage must not also exist 

Describe a situation where a potential exists but no voltage... For every
point that has a known potential, I can give you its voltage (relative to
infinity); just tell me its potential :)  You seem to not take into account
that the reference that is implicit in the concept of an absolute potential
(infinity) necessarily always exists, and is used as reference point for
expressing the value of this absolute potential (in volt). 


> But the definition of the t. voltage is absolute, not relative. 

Oops... Didn't you just base your argument regarding the difference between
the concepts of 'voltage' and 'potential' on the statement that potential
is absolute whereas voltage is relative? And here you introduce the concept
of an 'absolute voltage'?

This inconsistency left aside (as it is not really relevant for the
thermoelectric voltage)... is it really absolute? I thought it was
relative. The Seebeck effect seems to talk about the /difference/ of
potential that is caused by a /difference/ in temperature along a
(semi)conductor. I don't see anything absolute here. 

Again citing from the paper you mentioned: "Seebeck effect: A temperature
difference between two points in a conductor or semiconductor results in a
voltage difference between these two points." And later: "A voltage is
therefore developed between the hot and cold ends with the hot end at
positive potential." Just in case my explanation is not scientific enough
:)

This is confirmed by the practical ways to measure that potential
difference (in a typical thermocouple setup): you measure a difference of
two Seebeck voltages, both a function of the temperature /difference/
between two points. Every thermocouple measurement needs to reference the
temperature difference you gain from the measured voltage to some absolute
reference (the so called cold junction temperature) -- and this can't be
measured with a thermocouple (alone), never ever. This is not because we
measure the difference of two Seebeck voltages; this is because the Seebeck
voltages themselves are voltage differences based on the temperature
difference.


> So while the t. potential should be there, 

Where is it, as an absolute potential? As I see it, it is always,
necessarily, relative to another point -- at least what they call the
Seebeck effect. Of course, you can always reference both the temperature
and the potential of both ends of the rod under test to infinity (absolute
zero of both temperature and potential). But that doesn't seem to add much
value -- compared to looking at it as one point (with a temperature and a
potential) relative to another point (with a temperature and a potential). 

I also don't think you can speak of a Seebeck effect if you have only one
single point (with its temperature and its potential). You need two points,
with a temperature difference between them, connected by a conductor, for
the Seebeck effect -- which causes a potential (or voltage) difference
between them. I don't think that the electric potential of a single point
(with uniform temperature) changes while changing its temperature. The
Seebeck effect (aka thermoelectric voltage) needs a temperature
/difference/ between two points on the conductor. Nothing absolute in this.


> ... it must be measured with reference to something else. This implies
> the need for at least *one* contact or quasi-contact. Even the pin/bar
> proposed by Olin has one contact. 

No. The "contact" could be (and should be) of the same material as the rod
under test. This may put some restrictions on the range of suitable
materials, but not on the principle itself.

I think you may be mixing up the fact that in order to measure the Seebeck
effect in a closed current loop we need necessarily two different materials
(and therefore measure the difference of the Seebeck effect between the two
materials) with some other things.

A few affirmations (all IMO, of course :)

1- In order to create a closed loop where current can flow (which is the
case of all "normal" ways to measure voltages), it is necessary to use two
different materials in order to see a Seebeck effect. If the loop consisted
of the same material, the Seebeck effects on both sides of the hot spot
would cancel themselves out. That's the thermocouple: the resulting voltage
is a function of the difference of the Seebeck effects in both materials
and the difference in temperature between the hot and the cold end. (This
implies a certain configuration of what is hot and cold, but can be
reformulated for any other configuration.)

2- The Seebeck effect itself does not need a second material; it happens
within one material. It creates an electric field, which implies a
potential difference (aka voltage difference) between points of different
temperature along the conductor. These different potentials can be measured
(without introducing a second, different material) in a number of ways --
but all these methods have one thing in common: they don't rely on a
continuous current flow, therefore they don't need a closed loop, which
makes them conceptually different from the way we commonly measure
thermocouple voltages. However, measuring electric potentials in such ways
is nothing particularly strange (in laboratories that deal with electric
fields, at least).

3- While it may be that a common use of 'voltage' and 'potential' exists
that implies that 'potential' is absolute and 'voltage' is relative, the
author of this distinction doesn't adhere to this use (he earlier suggested
an 'absolute voltage'). I agree with the author's practical sense (here and
in general <g>) and say that both voltage and potential can be relative
(and in most applications are) and absolute. (For example, mechanic
potential energy is pretty much always used in a relative form, and it's
conceptually pretty similar to an electric potential.) I also think that
the value of an absolute electric potential is given in volt, and probably
can therefore be called a 'voltage' (an absolute one, maybe, but still).

If there is a difference between the terms, I'd say the voltage is the
value of the potential (in the sense of "we have here an electric potential
with an unknown voltage" or so)... but then, as long as we talk about
'bigger' and 'smaller' (as in the Seebeck effect), we always talk about the
(relative) values.


Not sure this is of interest to many, but for me it is surprising that we
need to discuss this (and important that we do, at least for me -- I could
be wrong on any or all questions :).

Gerhard

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist