Ben Jackson wrote:
> This is half question/half cautionary tale.  When I set up my MPLAB
> project for the 16F628, either I chose to turn on brownout detection
> (BOD) or it defaulted to on.  It's battery powered, using 3x AA for
> a net VDD of about 4.5V.
>
> When I switched from using one surplus 7-segment display at low power
> (for testing) to my new MAN8610 (very nice .8" digit displays) at full
> power with multiple segments, I could not get the PIC to stay on.  In
> addition to the normal decoupling (.1u and 1.5u) that I had been using
> all along, I needed almost 100uF to smooth the power out enough to
> drive the displays and keep the PIC on.
>
> When I finally realized I was hitting BOD, I turned it off, and it
> works without the 100uF.  So my question is:  Is this an acceptable
> solution, or is the fact that I *would* be triggering the BOD a sign
> that I can expect problems anyway?

Ben, the problem is related to the 3xAA cells (probably carbon cells,
right?).

They have a not so low internal impedance, and the peak currents when
multiplexing segments was high enough to drop the voltage of those cells to
a BOD point.  The 200uF capacitor in parallel with the cells accumulated
enough energy to supply it fast to the multiplex current pig. As the
capacitor has an impedance much lower than any regular carbon cell, the
capacitor was the power supply for the multiplex and kept the cells with a
voltage more stable.

If you think about it, your circuit probably have some small caps around.
The LED multiplex was trying to suck current from the AA cells and it "was
not there" (due the cells impedance), so, the circuit ended up sucking
current from the small caps you have in the board, and this can cause all
sort of problems.  A larger cap in parallel with the cells is a solution.
Some other 47 to 100uF cap in parallel to the (anode) + side of the mux
leds can also help the mux to have energy very close, not needing to cause
any kind of noise or disorder in the current.

If later you will use a regular power supply, it will have some sort of
voltage filter, as a capacitor or a voltage regulator that will take care
of this issue.

If instead of carbon AA you used NiCad or NiMH cells, even with a lower
voltage (3.6V for 3 cells or 4.8V for 4) it would never happens, since
those cells have a very low internal impedance and would feed the led mux
without any significative voltage dropout.

You need to do some lab tests, your cells versus power drain.  Measure
everything individually, what consumes how much current, and draw a table
about it.  You can then try to optimize power consume to meet your wishes.
If you intend to keep using carbon cells, you probably will need to
substitute some power pig units for more skinny ones.

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads