What does resonace peak mean when it is represented with dB

justdoit wrote:

It sounds like you are working with an optical drive actuator? The
specs are usually rather vague, but you can assume that zeta is about
0.0325 to give 24dB peak. The mm/V spec is often specified at a few
frequencies, but 0.85mm/V is probably the low frequency gain spec
before the resonant peak.

I hope that helps.

fred

Hi Fred,

Yes it is an optical drive actuator. Are you also doing optical disk
servo? My msn is: wt70707@hotmail.com.

I think 0.0325 is square of zeta and zeta is 0.18. I have much
confusion on reading spec of the actuator. From an article, I read the
resonance frequency as of the spring suports the OPU and the 1st mode
frequency as the resonace frequency of the VCM motor. Am I right?
Another confusion is about the quality factor Q(e.g. equals 16): what
information does it convey? I think the resonance frequecy and
resonance frequency peak can tell what the 2-order system is. If
quality factor is used to determine resonance peak, what is the
relation between Q value and resonance peak?

Best Regards,
Thomas

justdoit wrote:

Q is the ratio of the energy stored in the resonant system(1) to the
energy dissipated in one cycle of oscillation. Damping lowers the Q. Q
is roughly the ratio of resonant amplitude to amplitude at a frequency
remote from resonance. Q ans zeta are different ways to specify or
measure the same thing.

Jerry
___________________________________
1) for a mass and spring, the energy sloshes back and forth between
kinetic in the mass and potential in the spring.
--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

Hi Thomas,

I worked on optical drive servo until last year. I think zeta should be
0.035  - you can check by drawing a Bode plot. You should get around
24dB peak. Zeta=0.18 is too high. The device behaves like a spring and
a mass with negligible viscous damping which is why zeta is so low.
Zeta and Q are related in reciprocal  fashion with a factor of 2
included.

fred

Fred Stevens wrote:

Hey Fred:

Is this one of those systems where you bring the resonance into the loop
and control it, or is it one where you put a big ol' notch in your
controller and hope the resonance doesn't move too much in production?

--
-------------------------------------------
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Hi Tim,

I have seen both approaches used and  I have heard that gain scheduled
methods are also used. Optical pickup actuators are quite precision
devices, so their characteristics are reasonably well controlled. The
main challenge is the nonlinear S-curve transfer characteristics of the
laser array and the requirement that modern optical drives must be able
to play all types of media; CD, DVD, RMS etc. with different LASER
wavelengths. "Blue Ray" should be very interesting, but I'm out of the
loop with this technology now.

fred

The S-curve can be searched open loop. We are meeting with how to
stabilize the focus in the beginning. We want to tune it with PID, but
the critical P is hard to find.

How doya think about "blue ray" and "hd-dvd", both is difficult to
control and they are very different. But the boss want to make
compatibility of these two and also DVD and CD.

Thomas

Hi Tim:

Notch filter is used to suppress the higher frequency resonance of the
actuater. What doya mean with "bring the resonance into the loop and
control it"?

Thomas

justdoit wrote:


If the system is well behaved at the resonant frequency of the plant you
can damp out the resonance with differential control.  For example, I
have a client/former employer who's systems generally have a poorly
defined, low-Q resonance around 0.1 to 1Hz, and another, much higher
frequency one, that is rather high-Q and well-defined for any given
product line.  The loop is closed well above the 1Hz point but below the
higher resonance.  To keep the higher resonance from interfering with
stability we notch it out.

Here is an extremely ugly and rough root-locus plot of what you'll see
if you wrap that resonance:


                                        |
                                        |
                                        |
                                        |
                                        |
                                        |
                                  ---   |
                                /     \ |
                               /       x|
                              |         |
                              |         |
                              |         |
                              |         |
===<========x---------o<=====|======>o-+---
                              |         |
                              |         |
                              |         |
                              |         |
                               \       x|
                                \     / |
                                  ---   |
                                        |
                                        |
                                        |
                                        |
                                        |
                                        |
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de


--
-------------------------------------------
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Tim Wescott wrote:

(from email)
 > Tim:

 > I have simulated 2-resanance(4-order) system with high-Q second
 > resonance. I injected measure noise with the same frequency of the
 > second resonance. The notch works well.

 > Thanks for your "rough root-locus plot". But I can not discern it.
 > Would you mind beat my brain on how to "see if I wrap that resonance"
 > on root locus.

The root locus is for a controller where the resonance is brought within
the active bandwidth of the controller.  By placing a zero close to f=0
(with a differentiator) you add damping to the resonance, which acts to
increase the effective damping coefficient of the system.

--
-------------------------------------------
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com