45 years ago in the LM340 regulator days it was conventional to always put a 1N400x in series
with the regulator input.? The problem of shorted regulators passing full voltage through them was
well known and accounted for with the series diode.? Somehow we seem to have forgotten that

knowledge over the intervening 40+ years.? Doing this means that you have to account for the

additional 0.65 volts of diode voltage drop, but it is a small price to pay for the added security.

?

The datasheet also shows the shorted-input protection diode across the regulator device (Fig. 15)
and a protection diode across the output (Fig. 26).

Arv K7HKL
_._

On Sun, Jun 11, 2017 at 11:08 AM, Jerry Gaffke via Groups.Io <jgaffke@...> wrote:

My curiosity overcame me.? Pulled the LM7805 from my Raduino (so I was sure I was working with the same manufacturer that hfsigs uses) and verified proper operation with 12v in, saw 5v out.? Reversed the 12v leads into the LM7805, put on the safety goggles, and powered up.? Was powering from a 2.5 amp 12v supply. ? The reversed LM7805 was sucking enough current to pull down my 12v switcher to -5v going into the LM7805, saw -2v going out of the LM805.? Could well be more than -2v going out if I had used a beefier 12v supply.? Regardless, -2v is sufficient to damage some parts behind the LM7805.? After a few seconds the LM7805 got hot enough to vent and stink up the room.

So, the LM78** parts offer little to no protection from reversed power supply leads.? Not surprising, they were designed 40 years ago when such niceties would have been too expensive to include.? A reversed power supply to the rig could easily blow the Nano on the Raduino.? The Nano has an internal 3.3v regulator that may or may not save the Si5351, but no guarantees.? The LM78L09 powering the analog VFO and LM78L05 for the IRF510 bias likely have the same issue.

The LM2940* family of regulators does offer reverse polarity protection and has a lower dropout voltage.? However it is twice as expensive ($1.00 vs $0.50 from mouser for 10), they recommend at least 22uf on the LM2940* output to maintain stability. ? As mentioned previously, having an LM2940-12 at the main Bitx40 supply going into the board would protect us from reversed power leads, and also avoid trouble from supplies that exceed 12v such as a stack of 4 lithium batteries.? The IRF510 should be powered separately, as we may want considerably higher voltages and currents there and the IRF510 is plenty cheap.

I am left very curious if Matt's Raduino is still functional. ? Does the display come up if you apply 12v to it?

Jerry, KE7ER


On Fri, Jun 9, 2017 at 08:27 pm, Jerry Gaffke wrote:

If the LM7805 and LM78L09 regulators avoid passing negative supplies somehow,
might not be just too much damage from reversed power.
But I don't see anything in the LM78** datasheet suggesting that is the case.
Would be interesting to see what they do.

?