I've got a prototype run of boards heading back. They replace the ATTiny841 with an ATMega328P, and add the Venus838 module. The processor upgrade quadruples the flash space, necessary because of the constant threat of running out of space in the 841. The Venus838 module on an electrical basis is almost a drop-in replacement for the PA6H. It takes 3.3 volt power, an external 50Ω antenna, serial I/O and gives an active-high PPS and status LED outputs. There is no internal antenna, like there is in the PA6H, so the external antenna is mandatory (in reality it always was. Placing the PA6H in the extruded aluminum case pretty much negated its ability to get reception, and the case makes the oscillator more stable by isolating it from ambient airflow).

Unlike the PA6H, there is no internal provision for antenna power, but the antenna input has a blocking cap already, so there's no need for a separate one to make up a bias-T. All that's required is a 0.033µH inductor on the power line. I've also decided to add an AP2331 current limiter to protect from shorts and tame any inrush from hot-connecting the antenna. There's also a UHF rated TVS diode to provide some ESD protection on the antenna input. There's also a solder jumper to allow switching between 5 and 3.3 volt antenna power. 5 volts is the better choice, as it comes directly from the input rather than through an LDO.

The phase discriminator circuit will have a slight change. The 680pF cap is going to replaced by a 1000pF one. That increases the time constant so that the full scale voltage is nearer to 1.1 volts. That allows us to select the 1.1 volt reference in the ATMega328P, since it lacks a 4.096 volt one. That lets us maximize the use of the ADC resolution, and also results in an almost perfectly linear response.

The new firmware will parse the quantization error message and correct the phase discriminator output accordingly.

The board is still designed to fit in the current aluminum case, so it will retain the two status LEDs of the current design rather than switching to the LCD.

From a user perspective, the potential gains to be had from this design are likely to be had from the GPS module's ability to give excellent timing results with less than perfect reception. In addition, the hope is that the module will be more stable than the PA6H. The PA6H occasionally shows signs of GPS phase steps and micro-outages, even with the excellent reception I get from my outdoor marine antenna.

We'll see how that all goes.