So, what does happen when the local oscillator frequency matches the carrier frequency? The mixer output, after a low-pass filter would be at 0 Hz. Would that be a DC voltage? It makes sense. Since the carrier is alternating between lower-level and higher-level, the voltage should do the same. Low-level, high-level? You mean logic levels? I need to look in to this.

Digging through the parts heap, I located a 3 MHz crystal. Divide by 50 to reach 60kHz, sounds easy enough. Let's go! Oscillate. Come on, oscillate! Ah, here we go. Time to divide. The 7490 has independent divide by 5 and divide by two stages, a pair of these chips will make short work of divide by 50, and there will be a left over flip-flop. Divide by 5, divide again by 5, and finally divide by two and we have a 50% duty-cycle square wave. What is this, 59980 Hz? Hours of tweaks later, and it's only up to 59900 Hz. I know, use a 74HC chip. Adjust the load capacitor, and woo-hoo 59.9994 kHz! Close enough.

If the mixer's output is DC, I'll have to remove the coupling capacitors between the mixer outputs and low pass filters. I guess that will work. Then it was time to add the low-pass filter between the mixer and first stage differential amplifier, but where should the corner frequency be? Most of the previous circuits had corner frequencies somewhere in the 5 kHz to 20 kHz range. At  0 Hz there is no real need for bandwidth, right? Let's set the corner at 5 Hz anyhow. At that frequency, the filter is going to take a nice bite out of 60 Hz power-line noise, too. Fascinating.

Taking a first look at the output of the diff. amp shows it to be stuck at V+. Just a wiring error, right? Connections checked, all are correct. The voltages all look right, the problem must be the differential voltage and gain causing the problem. Checking the inputs reveals a difference of about 40 mV. Where is that coming from? The resistors are acceptably matched, the diodes are similarly close, that leaves the filter capacitors and TTL outputs. Sure 'nuf, looking at the outputs on the scope shows a very small difference in output low voltages. Adding a 33 ohm resistor in series with one of the outputs moves the diff. amp's output off of the rail. Having to trim each output individually won't scale well, perhaps this circuit requires MOSFET switches.

Looking at the output demonstrates the need for a chart recorder of some sort, possibly a hacked thermal receipt printer. Until that project is completed, a comparator and LED will have to suffice. It looks like the blinking matches WWVB's signal, but it also appears to be drifting in an out of sync with it. Will bringing back the Q signal (quadrature) keep things in phase?