[time-nuts] SMPS or conventional?

[Gerhard Hoffmann]

Am 22.12.20 um 00:43 schrieb Hal Murray:
> [Old mail, context is TECs] bruce.griffiths at xtra.co.nz said:
>> If the drive current ripple is too high fatigue failure from cyclic 
>> thermomechanical stress can be significant. 
> Do good data sheets say anything about that? Is there a frequency term 
> in there? Can I use PWM, which is as much ripple as you can get, as 
> long as the frequency is high enough? If yes, ballpark of how high? 
> Physically, where is the heat/cold generated? Is it mostly at the 
> junction? 

The g**gl search for "TE coolers ripple current" delivers the first 2 
hits of 3.65e6:


< 
https://www.google.de/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwiQ4YK0tuDtAhWKCewKHckIALoQFjAAegQIBRAC&url=https%3A%2F%2Fwww.ii-vi.com%2Fwp-content%2Fuploads%2F2020%2F07%2FMRLW_Working_with_Thermoelectric_Materials.pdf&usg=AOvVaw3pbUbFp4SsfdKsBzRNf-Wi 
    >

(that looks dangerous  :-)  and


< https://www.ferrotec-nord.com/technology-of-thermoelectric-module/    >


Both agree that ripple > 1 KHz is harmless. But then the current delivered
to the TEC is quite large usually -  and It's a different question if 
you want such
large and dirty currents on your table or in your device.

The thermal resistance between the hot and cold side is not very large
and that explains part of the suboptimum efficiency.

You do not want such a thing to regulate the temperature of a crystal etc
because if you externally heat or cool one side, that will soon propagate
to the other side; delta T stays constant until your regulator takes note
and corrects it.


Oh, and the google search shows that everybody and their dogs make
TE controller chips.


Gerhard