Re: [CAD_CAM_EDM_DRO] Re: What size capacitor should I use? resistors....

turbulatordude wrote:

>--- In CAD_CAM_EDM_DRO@yahoogroups.com, "cnc_4_me" <cnc4me@g...>
>wrote:
>
>
>>Attn CNC power supply people...Please DO NOT put a resistor between
>>your bridge rectifer and your capacitor...Because of the time
>>constant increase (T=RxC)x5, you will render the recovery of
>>capacitor voltage to be to long in time to be of any use to PWM
>>drives...
>>
>>T = time in seconds
>>R = resistance in ohms
>>C = capacitance in farads
>>X 5 = 6 time constants for a full charge
>>
>>example... Power supply has 30,000Uf and you add a .5 ohm resistor.
>>C = 30,000Uf is 10,000 to the minus 6 farad.
>>R = .5 ohm
>>T = C x R
>>T = .03farad x .5 ohms = .015 seconds for 1 time constant.
>>It takes 5 time constants to charge a capacitor...So..
>>5 x .015 seconds = .075 seconds for capacitor to recover...
>>.075 seconds = 75mS = 75000uS
>>
>>We are PWMing around 22khz or 22,000 cps..At this frequency each
>>chopped wave can be as small as 1/22,000 = .0000454 seconds...Or..
>>.0454mS....Or 45.4uS
>>
>>With resistor in circuit it takes 75mS to recover and we need a
>>
>>
>fully
>
>
>>charged cap in .045mS....Adding a .5 ohm resistor makes the system
>>
>>75mS/.0454mS = 1670 times to slow.
>>
>>And now you can see why many power supplys use buss bar or very
>>
>>
>heavy
>
>
>>wire....
>>
>>
>>Wally
>>
>>
>
>I understand the desire to limit the starting inrush current as well
>as the need for the smallest time constant.
>
>would it not be prudent to charge the cap slowly up to some value,
>and at that point, switch over to the raw voltage directly ?
>
>It would seem a simple electromechanical relay would be a possibliity
>to eleminate the hum of initial power on.
>
>
>

This is a common technique, a slow-start circuit. I have a big VFD that
does that.
It has a couple of problems, though. One is that momentary line power
dips can
cause a massive surge when the power comes back on after a few low or missed
cycles. Without something that continuously limits peak current, the
rectifier has to
deliver massive short current pulses to the capacitor. On larger power
supplies in
the KW range, this can mean pulling 100 Amp peaks every half cycle of
the power
line. A 30 A rectifier can only take this abuse so long, and the 100 A
ripple
current will cook the caps, too. A carefully thought out series
resistor can reduce
the peak current to a reasonable level with very little increase in
ripple or
worsening of load current regulation.

Jon