Michael Grass wrote:
> Fred,
>
> I appreciate your input but think again:
>
> 1. A large (dimensional 12mm * 25mm) cap is not a problem. Electrically
> 1000uF reduces the effects on the supply and stiffens the supply for
> regulating purpose. True it is electrically large and I have no problem if
> you reduce its size.
Not my input ultimately, but National Semiconductor. ;-)
The larger the cap the bigger the inductor, and they best filter 60Hz AC line
hum.
At high freqenies (alternator) and fast-rise spikes (Rotax regulator), a large
cap will do nothing if not make things worse. In looking at various avionics
schematics, I suspect this is one reason one often finds a small value cap
on the 14V input, as little as .01uF, some a hundred. Some use small caps and
a choke, the effective way if needed to condition high-freq, noisy input.
> 3. The cap on the output is not harmful to the 7805 since it is completely
> self protected.
The subtle issue is the storage capability of the cap, the larger the worse.
When
14V power is removed from the circuit, and if the input lead goes to ground
(a master relay can briefly do that), the 7805 regulator can be destroyed.
Plus maybe pop a fuse. The diode I suggested prevents that and is recommended
by National Semi if any input grounding can happen.
> The [output] cap helps to stabilize the relatively high inrush
> current to the motor from the servo.
It sure will, but the inrush current (caused by a large output cap) when you
merely
switch on the circuit is many times the current jolt to the regulator than
motor startup should cause. IOW, a net loss in abuse to the regulator IC, the
more microfarads the worse.
> 3. Input and output caps can be omitted but are beneficial. Rule here is
> that the input cap should be 10 times larger as the output cap to have a
> stable regulating result. Anything between 100uF and 1000uF for an input and
> 10 to 100uF for the output would be my recommendation.
Actually my texts cite no 10/1 rule, but rather National says the output cap is
mandatory and should be selected (value and ESR) for phase shift relative to
input noise frequency. Else it can oscillate, and that you could hear in the
avionics (plus a DC servo wouldn't work too well)! As practical matter, the
noise
products on the input should be small enough, due to what the battery does,
so this shouldn't happen. Further effective is a tantalum cap, due to its
low ESR, they say. 1 uF typical.
Regards,
Fred F.
|