PS Pin Capacitance LM3886

[Puffin]

It seems that the PS pin capacitance is suggested to be between 1000 and 2200uf for a single 3886. I have built a bridged LM3886 and after re-reading the notes it seems that they suggest 4700uf for bridged.

Is the reccommendation an absolute i.e is there any detriment to using higher than 2200uf (or 4700). The reason I ask is that I have some good quality 10,000uf caps. I will give them a go and see how they sound.

I would however be intersted to know if the extra capacitance will just be wasted or whether it will have any other detrimental effect ?

[jerome69]

No problem, it will work better

[AndrewT]

Hi,
when suggesting the smoothing capacitance requirement for a bridged amp I argued that bridged needed 4times as much capacitance as a single amp into the same load. Most came back and said "rubbish".
Try four or more times the local decoupling. it might help.
But remember the fast caps are the ones that go on the pins. The electrolytics can be a little further away.

[Puffin]

jerome69 & AndrewT. Thanks for your responses. Andrew I am going to show my ignorance now and ask what denotes a "fast cap" ?

[AndrewT]

fast cap=low value polyfilm or ceramic or silvered mica.

[tangmonster]

Originally posted by AndrewT
Hi,
when suggesting the smoothing capacitance requirement for a bridged amp I argued that bridged needed 4times as much capacitance as a single amp into the same load.

I agree , It's simple math


1 Chipamp into 8 ohm = xAmps
1 Chipamp into 4 ohm = xAmps x 2

2 chipamps into 8ohms bridged(each chipamp "sees" 4ohms) =

xAmps x 4

So your current has multiplied by 4

C = 0.7(I)/ΔE(f)

where C = capacitance in farads, I = dc load current in amperes, ΔE = peak-to-peak ripple voltage, f = ripple frequency (generally 120 Hz for full-wave or 60 Hz for half-wave), and 0.7 is the complement of the rectifier-current duty cycle, which is assumed to be 0.3


http://electronicdesign.com/Articles...ArticleID=9809

IT all depends on the ripple voltage you are happy with at high loads (high volume)

You do have SOME play with music because you don't have a contstant current pull , but If you want to quote an RMS voltage these calculations are important.

am I overlooking something?

[AndrewT]

Originally posted by tangmonster
I agree , It's simple math

but we are in the chipamp section.
Nobody here wants to admit they learned anything at school.
In the car audio section, peer pressure alone forces them to claim they never went to school

[gootee]

Originally posted by AndrewT
but we are in the chipamp section.
Nobody here wants to admit they learned anything at school.
In the car audio section, peer pressure alone forces them to claim they never went to school

<GRIN!>

[digi01]

Originally posted by AndrewT
fast cap=low value polyfilm or ceramic or silvered mica.

i have use the beacon cap of camera,350v/120uf.they work great.guys can try it

[weissi]

without all those theoretical aspects aside (poor ripple rejection of LM3886 in the >500kHz region) just put 220µF || ~100n caps for every chip you use. But them as near as you can to the PS pins, bulk capacitors can stay more far away then.