Re: [Ntp] Antw: [EXT] Re: Temperature Compensation for NTP?

[Ulrich Windl <[e-mail]>]

>>> Kurt Roeckx <[e-mail]> schrieb am 10.12.2020 um 21:09 in Nachricht
<[e-mail]>:
> On Thu, Dec 10, 2020 at 08:56:27AM +0100, Ulrich Windl wrote:

...
>> I'm not deep in that subject, but couldn't a non-overshooting response be
>> compensated by feeding samples at a higher rate (smaller "poll")?
> 
> I'm not sure what you mean with compensated.

I mean the time until the offset is below a specific threshold.
In my simple memory there exists one parameter that is the highest that does not cause overshoot; I'd consider that to be the optimal parameter.

> 
> If you have a higher sample frequency with the same impluse resonse,
> you will also faster change your phase to match. The same impulse

Yes, Intended.

> response at different sample frequency means that it takes the
> same amount of samples.

Guessed so.

> 
> Note that if you want the same impulse response at different
> sample frequencies, you need different feedback parameters.
> 
> So with a damped / undershooting response and a higher sample
> frequency you can have a faster response.

...without overshoot. So what are the drawbacks?

Regards,
Ulrich