[time-nuts] Frequency referenced temperature regulator

[jimlux]

Javier Herrero wrote:
> Hi,
> 
> Why not use a more conventional approach using a good thermistor and a 
> PWM IC? I built some in the past with Peltiers and an LM3524 for cooling 
> a photomultiplier in a nitrogen oxyde analyzer, and the result was quite 
> good. I also used the same approach (but not with Peltier) for heating. 
> Of course, the frequency stability was not time-nuts grade... not even 
> near :)
> 
> NTCs like the YSI 44031 have very good repeatabilty, and the resistance 
> vs. temperature curve is quite steep, so you really don't need too 
> stringent voltage stability needs.
> 


or for more precision/linearity what about a platinum resistance 
thermometer (PRT)...