NAS/MEDIA PLAYER線性電源pcb

billymud

johnme 發表於 2014-4-5 00:33
Sounds great, I will ask you for the polyester caps later.

Here is a modified version.

yup, i'll free u some Panasonic polyester caps when I'm home.

johnme

billymud 發表於 2014-4-5 01:04
yup, i'll free u some Panasonic polyester caps when I'm home.

what value do you have?

johnme

billymud 發表於 2014-4-5 00:46
remote sensing for a better job.

isn't the remote sense better connected to the load side?

if it is connected to the output jacks, I doubt the effectiveness.

billymud

johnme 發表於 2014-4-5 01:05
what value do you have?

103~ 334, ..... Likely.....

billymud

johnme 發表於 2014-4-5 01:12
isn't the remote sense better connected to the load side?

if it is connected to the output jacks, ...

That depends on the length of DC  output cables..

johnme

billymud 發表於 2014-4-5 01:59
103~ 334, ..... Likely.....

Perfect!!!

johnme

billymud 發表於 2014-4-5 02:01
That depends on the length of DC  output cables..

Agree.

Usually, I make DC cable about 8 inches for external PSU.

Besides remote sensing, do you have any idea of choosing output capacitors ( both values and types)
making the transient response better?

billymud

johnme 發表於 2014-4-5 09:55
Agree.

Usually, I make DC cable about 8 inches for external PSU.

i used to acquire Panasonic & Nichicon caps from RS-components Ltd.  
Sometimes, I've Electrolytic and Tantalum caps populated in mix.  

royccw

johnme 發表於 2014-4-3 21:34
手痕做多個sigma11 modified版,
用tmc compensation,
但係可以還原做原本sigma11,

C Hing, please confirm your adjustment pin cap typical value around Cadj=22uF, too big value would not help for output ripple or may cause side effect.

johnme

royccw 發表於 2014-4-7 17:48
C Hing, please confirm your adjustment pin cap typical value around Cadj=22uF, too big value would ...

Thank you for the reminder.

The choice of capacitor size depends on what frequency you are working at.

The LT1084-NAS PCB provides the user the choice of using an electrolyte cap C7 and/or a film cap C8.

The 5mm pitch of C7 enable the user using as large as 220u, although I think (not tested)
that 100u is more than enough.

In my prototype, I only use C8 = 0.47u without using C7, it sounds good to me.

Have a look at the attached figure,
you may realize why I provide the option of using C7 and/or C8.
The LT108x family's rippple rejection at both ends of the frequency spectrum are relatively weak.
i.e. work best around 60Hz to 1kHz.

Also, note that in the figure, Cadj is 220uF, depending on what frequency range you are aiming at.

In the datasheet, it suggests 0.22u, if you aiming at optimizing ripple rejection around 10kHz.

Hope these information help.