[sdiy] Shutdown on chip-amp

[Johannes Öberg]

Sorry if I'm silly, but how do I pull the base of a PNP to ground from
a PIC chip? Can I really do that so that the PNP draws no power when
the PIC is in power down mode?

I also thought about using a small mechanical relay (if only because
they are so easy to understand!), and connecting a diode in series
with the coil. Would this be a good solution (if there's something I'm
not thinking about), and is the diode really necessary? (my idea is
that it will prevent any leakage from the PIC in powerdown mode).

Many questions, but I'm learning! :-) Thanks everybody for your answers so far.

/J

On 5/2/05, Roy J. Tellason <rtellason at blazenet.net> wrote:
> On Monday 02 May 2005 01:45 pm, Johannes Öberg wrote:
> > Short version: I need to make a small mono power amplifier (probably a
> > LM386) shut down into some sort of low-power mode when not used, and
> > this will be controlled by a PIC. I'm a total newbie.
> >
> > Longer:
> >
> > I'm building a small batteryoperated door-bell out of an ISD2560
> > (lofi-sampler-on-a-chip) and a PIC. I guess I could have saved on
> > parts, but it's my own "design" so it's a bit of an overkill to get it
> > working.
> >
> > Anyway, I'm trying to reduce the power consumption as much as
> > possible. I'm using the PIC in shutdown mode with interrupt-on-change
> > so it only draws power when you press the door bell button. The ISD
> > has a built-in power saving mode.
> >
> > All this works so far.
> >
> > However, while the ISD can drive a speaker directly, it is way too
> > quiet for my needs. Therefor I need to use an external amplifier. The
> > only one with built-in power down mode I can find readily available is
> > surface mount :-(
> >
> > I looked at the LM386, but how do I make it power down? Can I just put
> > some kind of PIC operated switch on the voltage supply for the LM386?
> 
> I would.  Except that it being a single-ended amplifier which has a capacitor
> at its output you may get a bit of a "thump".  Maybe not,  or it may not be
> much of a problem,  depending on what you use for a speaker.  You could
> always tailor the response a bit.
> 
> > Which one would you recommend? It doesn't matter if it's expensive
> > because this is going to be one-time quick-and-dirty fix.
> >
> > Also, unfortunately, I'm short on time, and only have access to parts
> > from www.elfa.se (the only local electronics supply), aparts from the
> > random stuff I have at home of course.
> 
> I would suggest looking at what Don Lancaster shows in his CMOS Cookbook, if
> you have a copy or can get access to one.  Some sort of a "sleep" circuit
> which he constructs out of a couple of gates configured as a flip-flop (not
> really needed if you have a chip pin that'll toggle for you) and driving a
> power FET.
> 
> If you want to use a bipolar transistor instead of a FET,  use a PNP,  and
> pull the base toward ground to turn it on,  it'll be much easier to drive it
> into hard saturation that way,  compared to having to supply some particular
> voltage and current to drive an NPN into saturation when an output is high.
> 
> > BTW, the LM386 only gives out 325 mW. Is that enough? I have little
> > idea about these things, I have a 1W radio somewhere that is awfully
> > loud even at low pot settings, so that's why I figure 325mW would be
> > enough. Am I way off?
> 
> Depends on which variant of that chip you end up with,  and what you're using
> for power.  I have a couple in one of those little boards that radio shack
> used to sell as "headphone amplifier" kits,  and power the thing with a 9V
> wall wart style power supply that's capable of kicking out an amp or so of
> current,  and it gets pretty loud.  I think I figured once that they were
> putting out about a watt per channel.  For your application that should be
> plenty of power.
> 
>