I've submitted this project to the Component Abuse Challenge.

I had some further experiments and measurements I wanted to do, but life has a way of changing our plans.

In this case, changed plans means it'll be at least one month, more likely two before I can get back to my electronics work bench.

Some of the things I wanted to try:

1. Remove the pull-up on the JK inputs.  My memory says the original circuit didn't have it.  Both inputs should float high, and with them connected together, the toggle function should be enabled without the pull up.  The sketch I found shows the pull-up, but it was drawn at least 20 years after my dad built the first one.

2. More accurate measurements of the clock signal.  The clock signal is very low, and doesn't come close to meeting the input specifications of the 7473. I want to be sure what it is really doing.

3. Restore the blinking "power on" LED.  The original had the LED on the 3909.  It let you know the circuit was on and that the battery wasn't dead. You needed that to be sure of an open "diode under test" (DUT) when neither of the test LEDs lit up.  Power on LED lit, DUT connected, no test LEDs lit = diode failed open.

4. Test on 5V, test on 9V from a power supply.  I have assumed that the lower current from the carbon zinc battery protected the 7473 somewhat.  That may not, however, be true.  I built one diode tester combined with a 9V power supply.  The diode tester was powered from the 9V supply.  As a school project, I built a 9V DC power supply to use for my projects. 110VAC to 9VDC. Completely linear - transformer, diode bridge, 7809, filter capacitor.  The benchtop power supply had a diode tester built in.  My memory swears that the diode tester was powered by the 9V output.

---------

The idea to use the 3909 came later.  Radio Shack sold blinking LEDs at the time.  The first version of the diode tester had a blinking LED connected to the clock input of the 7473.  Since the LED blinked at like once per second, the test LEDs also alternated slowly.  My dad wanted to have the appearance of both LEDs lighting simulataneously, so he needed a faster oscillator.

Radio Shack's "documentation" (description on the back of the package) for the blinking LEDs said that they had an integrated LM3909 for the blink function.  I don't know if that's true, but my dad took them at their word.  If the blinking LED could trigger the 7473 clock, then an LM3909 ought to do it as well - and it did.

At least one of the diode testers my dad built used a dual red/green LED.  Both in one housing, just two pins with the two LEDs in antiparallel.  When driven with a (relatively) fast AC signal, it would glow yellow.  With that version of the diode tester, you had "yellow" to indicate a shorted diode.