[Re: thermal tips re expo converters:]

[Ian Fritz]

Juergen -- 

Thank you for the further thoughts. I'm afraid, though, that my main
point still may not have gotten through. Please indulge me while I try
again. 

A proportional controller at equilibrium has to have a finite error
signal to operate. In other words, the final temperature has to be below
the set-point temperature in order to produce a difference (error)
signal to be multiplied by the loop gain to hold the heater current at
its equilibrium value. [Imagine an opamp with an open loop gain of 10
for an analogy].

The lower the loop gain, the larger this error is. For a given loop
gain, the size of the error signal will depend on how fast heat is being
transported away from the system, which in turn is determined by the
ambient temperature. Thus the equilibrium temperature of the system will
vary with the ambient temperature. In other words, lower gain leads to
decreased regulation. 

Conversely, good regulation requires high gain, which can only be used
when there is tight coupling between the object whose temperature is
being controlled and *both* the heating and cooling sources.

Another approach would be to use "proportional plus integral" control,
where the integral of the error signal is also fed back, in order to
eliminate the equilibrium error signal. 

If the arguments about not needing a fast cooling rate were true,
designers would not be heating their chips to the 60-80 deg C range.
This is way, way above any ambient temperature that would ever be
encountered, and the higher temperature clearly must degrade the
converter performance.

Hope this is clearer now.

  Ian



Haible Juergen wrote:
> 
> In correction to my earlier post, I have to agree with Martin and Ian.
> The thermal capacity would not be changed by isolation, but the
> thermal resistance is increased. Thus the "release" time constant
> of the system is increased. In theory, you could still get a stable
> system by adjusting the "attack time constant" (heater current)
> and the gain & time constant of the regulation loop to get it stable,
> but in the end you'd have a much slower regulation. You save energy,
> but your regulation becomes slow.
> I still have to think about whether slow regulation and higher impedance
> to the "error source" (ambient temperature) would somehow cancel
> for short fluctuations of ambient temperature. (?) But the warm up time
> would be much increased for sure.
> I think temperature *compensation* + isolation is the best solution.
> 
> JH.