Page 176 - Troubleshooting Analog Circuits

Letters to Bob I63

GROUND

-1 5V
(a) (b)
Figure 13.4. Depending on your applications, you may want to connect antirevenal diodes across (a) or in
a series with (b) a power supply’s pins. Or maybe both.

pected signals to their source. Again, any conducted interference should be at least 20
dB below requirements.
7. No matter how you measure the near-field radiation from your design, such radiation
should be at least 20 dB below the 3-meter radiation limit. To make these measure-
ments, you’ll need a shielded, screened test room.
Thomas L. Fischer
Pacific West Electronics
Costa Mesa, CA

Dear Bob:

You recommend using “antireversal” diodes across a power supply’s input to
protect circuitry against reversal of the power supply’s lead (see Figure 13.4).
However, if a power supply does get hooked up backwards, high currents will flow
through the diodes, which might degrade or ruin them. Remember, the diodes are a
part of the circuitry, too. Instead, I recommend connecting the diodes in series with
the input pins (Figure 13.4b).
Now the board has protection but with virtually no diode current.

Marvin Smith
Harbor City. CA

Dear Mr. Smith:
You are correct in some cases, and I guess I was delinquent in not mentioning
them. For example, if you have a battery, putting the diode in series with the correct
path may be appropriate. Then, if reversal happens, the battery won’t be crowbarred
and you will avoid damage to the battery and it environs. However, if you have a 5-V
bus, a diode in series with the supply would both waste a large fraction of the total
voltage and possibly spoil the supply’s regulation.
Even with a 15-V supply, where the wasted power might be acceptable, the bounce
and poor regulation of the supply might hurt the accuracy of the circuits that the
supply powers. The diode’s impedance may cause poor regulation. So, in cases
where a regulator drives the power busses, the shunt antireversal diodes are a good
idea. The solution presumes, of course, that the regulators are short-circuit proof.
The worst part of your circuit, Mr. Smith, comes to light when one of the power-
supply wires falls off or becomes disconnected. Then, the -14-V bus could get pulled
to +5 or +10 V, depending on what loads are between the +I4 and -14-V busses.
Many linear circuits can get very unhappy if fate pulls their negative-supply pin