LM7805 smoothening capacitor

[Spectacular Butter]

How do i determine the value of the capacitor used to smoothen supply voltage from the LM7805?

I read through some they used 220uF. some 1uF, some 100pF. Does it matter?

 

[Jay.slovak]

Basically, all 78xx stabilizers require 100nF cappacitors on both input and output. You may need extra cappacitor on input if your recitifer is far from 78xx. Adding extra 10uF cappacitor to output will reduce noise&ripple.

 

[Nigel Goodwin]

Basically, all 78xx stabilizers require 100nF cappacitors on both input and output. You may need extra cappacitor on input if your recitifer is far from 78xx. Adding extra 10uF cappacitor to output will reduce noise&ripple.

The 100nF isn't critical, anything upto 10uF seems to work perfectly well, but you MUST have some capacitance, as close as possible to the chip, to prevent instability (which could be intermittent).

 

[audioguru]

The output capacitor doesn't reduce ripple, the extremely high gain of the regulator IC does it.
The output cap also improves the regulator's transient response to abrupt input voltage changes and abrupt output load current changes, since the IC is fairly slow to respond and has overshoots.

 

[guzzimac]

If you do put a capacitor on the output don't forget the diode (1N4148 or equivalent) from output to input - anode connected to output and cathode to input. Leave this off and you run the chance of zapping the regulator every time you switch off.

 

[bonxer]

I've never put a diode in that way. I've never known I should. Have I just been lucky :?: Why is it that a diode should be in there?

 

[audioguru]

The 78xx datasheets show a 1N4002 high-current diode (not a flimsy 1N4148) across the regulator to protect it if the input is suddenly shorted to ground!
An input would short if it has an over-voltage crowbar SCR circuit activate. Normally the input wouldn't short to ground but instead would gradually drop its voltage as the main filter cap discharges.
The datasheets also show that there is nothing to gain by having an output cap larger than 10uF, and the regulator doesn't need a protection diode with a 10uF or less output cap even if the input was suddenly shorted to ground.

 

[lord loh.]

The LM 317 datasheet says that the difference in capacitor values is due to the construction of the capacitor....

It said 1µF of solid tantalum capacitor is equivalent to 25µF aluminium electrolytic...

Also it says that the input capacitance is needed only id your IC is 6inches or further away from your filter (pi or shunt capacitor filter)...

The output capacitor improves ripple rejection...the datasheet say it...but the improvement saturates beyond a certain value of capacitance used.

Ofcourse you need not bother about it if you have to just light a few leds or something... It begins to matter when you use level sensitive Digital circuits...The TTL moise margin shall normally reject any ripples after an 7805. but it may be crucial in ECL circuits...

And I just wondered if I could call that diode reverse biased between input and output as a freewheeling diode ?

 

[audioguru]

The LM 317 datasheet says ....
The output capacitor improves ripple rejection...the datasheet say it...but the improvement saturates beyond a certain value of capacitance used.

The LM317 datasheet says that a capacitor of up to 10uF on the reference terminal improves ripple rejection of frequencies above 120Hz. The 78xx series don't have a reference terminal.
The graphs show that an output cap up to 10uF improves transient response to abrupt input voltage changes or abrupt load current changes. Maybe you are confusing the resulting output voltage overshoots with rectifier ripple.