# Student's question about sampling freq

T

#### thinh phan huu

Deer ListMembers,

Can You give me some help on these :
(My questions are on sampling in digital control)

1. How to choose an optimal sampling rate ?
(error rate and stablibility of system)

2. Why people say when rising sampling frequency, the
digital control system may come unstable ??

If could, pls tell me what site to find info on

Expecting

Thinh Phan

J

#### Jerry

> Deer ListMembers,
> Can You give me some help on these:
> My questions are on sampling in digital control
> 1. How to choose an optimal sampling rate?
> (error rate and stablibility of system)

One can do complicated diagrams and computer
simulations, but a rule of thumb would be to
sample at a rate at least four times as fast as the highest frequency input

> 2. Why people say when rising sampling
> frequency, the digital control system may
> come unstable ??

It seems wrong. Who says this? Is it in a
book?
. . .
Maybe if scan time cannot keep up? If the scan time is not as fast as the sample time, faster sampling would be a waste.

J

#### JimU

> 2. Why people say when rising sampling frequency, the
> digital control system may come unstable ??

Karl Astrom published an article about twenty years ago that speaks to this issue. It's a theoretical account of this behavior given certain noise model assumptions.

N

#### Nassir Abedi

If you want to handle in digital like analog you should choose the sampling freq. more than 10 times the highest dynamic in your plant. If you want to use digital way it could be less.
Nassir Abedi

T

#### Tom

It is ten times the highest frequency - normally the bandwidth. The zero order hold contributes significant phase lag at half sampling freq and you must back off from this. 1/10 is a good figure. As for sampling too high - the higher you sample the better as it becomes more like an analogue system. Digital systems have an inherent one step time delay - more phase lag.

P

#### pankaj

here is the way u find the sampling rate>>

THE SAMPLING RATE IS GREATER THAN OR TWICE OF UR MAXIMUM FREQUENCY. OR ELSE U GET THE ALIASING EFFECT.
F(samp)>= 2*F(max)