Adding 1 PPS and 10MHz Outputs
to the Z3801A
I wanted to make my Z3801A GPS-disciplined oscillators more useful in my shack by making 1 PPS and additional 10MHz signals available on the rear panel, and providing a 1 PPS signal on the DCD line of the serial connector so the unit could serve as a reference clock for an NTP server. Although I ultimately want to add a distribution amplifier/splitter for the 10MHz sine wave signal, I thought it would be easier to initially add several 10MHz TTL square wave outputs. This page describes how I did that.
Jeff Mock has published a modification that provides a nice RS-232 level 1 PPS signal on the DCD line, so I used that as the starting point. (Note: I've learned that Jeff's web page is no longer there. A mirror is at http://support.ntp.org/bin/view/Support/Z3801AReceiverModifications.) Jeff taps a 1 PPS TTL level signal off of U5, pin 2 (a Lucent 1141LL, whatever that is). I used that signal to drive my 1 PPS circuit.
After probing around on the board, I was unable to discover a 10MHz TTL level square wave anywhere. However, TP5 in the front portion of the main board has a nice 10MHZ sine wave available, which unfortunately is riding on top of a DC bias voltage. I used a piece of RG-174 cable to bring the signal from TP5 back to my circuit. The front pin of the adjacent TP4 complex is at ground, so I connected the RG-174 shield there.
I built some buffer circuitry on a piece of perf board. Here's the schematic; some description and pictures of the board and its installation follow:
The 1 PPS circuit is very simple. I used a 74AC04 hex inverter, with the first section providing an initial inversion so that the end result remains a positive-going pulse. The output is a pair of inverters parallelled together through 51 ohm resistors; this combination provides a nice low-impedance signal that can drive a reasonable length of coax. The unused sections of the 74AC04 must be tied to something; if they float, the chip may oscillate or otherwise find itself in a state of high current draw. I tied the unused inputs to the output of the first inverter, so they are available as additional 1 PPS signals if needed.
Since the 10MHz signal is a sine wave and I want to generate TTL square wave outputs, I struggled a bit with the best way to couple the signal into the digital circuitry. Finally, I settled on an idea used by Brooks Shera in his GPS-disciplined oscillator controller. Brooks used the input circuit of a 74HCT4046 PLL chip as an input conditioner. I wasn't able to find the HCT version (and it's not recommended for new designs, anyway), so I used a 74AC4046 instead. The AC series part is faster and has lower jitter, so hopefully it will work at least as well in this application as the HCT version. My circuit has an added capacitor at the input that Brooks didn't use; this keeps the input circuit from loading down the DC voltage on TP5.
Oops... I just noticed that the schematic doesn't show a 0.1uF bypass capacitor at the power input (pin 16) of the 74AC4046. Sorry about that.
The output of the 74AC4046 is fed to another 74AC04, with the sections parallelled into three pairs of outputs via 51 ohm resistors.
I built the board on a piece of perf-board and mounted four SMA connectors on the rear panel for the new outputs. The board is supported by the short wires leading to the SMA connectors; this wouldn't be wise if the unit were being banged around, but it should suffice for my application. The board draws only a few milliamps at 5 volts, and I tapped power for it from the +5V test point on the power supply board. Ground is via a solder lug attached to one of the SMA chassis connectors.
Here are some pictures of the installation:
Tapping the 1 PPS signal.
By the way -- thanks to Mike, WB8GXB, who did the modifications to the main board; his skill dealing with tiny surface mount components was a huge help.
The 10MHz tap.
The board (the bottom isn't as pretty).
Note the 5 volt power connection point on the power supply board.
Connectors added on the rear panel.