Hi:

It depends on both your setup and on your molecular system. For example, in a Peltier controlled cuvette holder in our Aviv UV or CD instrument one can change temperature much faster than the cuvette and contents can equilibrate. Using a thermocouple in a 1mm path length cuvette (no stirring), one can be sure of cuvette content temperature equilibration with rates of about 1 deg/min. However, whether the sample in the cuvette will equilibrate on this timescale is another issue. I have much more experience with proteins. An example melt is 0.2 C steps with 5 min equilibration once the cell holder reaches the set temperature. Oligos will equilibrate faster than longer sequences. Stirred microcuvettes will be slower but can ensure better mixing.

Another issue is photo-decomposition during the experiment. Programmable systems can close the slit during temperature change and equilibration, opening during a constant temperature segment while data is collected. Ideally one has enough sample to try a few protocols and compare them.

I don't think the overall rate of change is as important as accuracy and stability of temperature and ability to program experiments. If you are interested in temperature jump experiments, Peltier is not the technology to use.

Hope this helps. Report back on what you decide to buy. PM me if you want to discuss particulars.

I only have experience of DNA melting curve analysis for real-time PCR. For that I have noticed that the ramp rate / time per temperature step affects the Tm (at least on a RotorGene)

I think the answer of   Kurt D. Berndt is OK