NEW

The GPS receiver is now a quad-constellation module that supports reception of GPS, GLONASS, Galileo and BeiDou satellite systems.

The power input is now a USB C jack. The power requirements are 5W at 5V. The power supply option is now also USB C and includes a 6 foot cable. The output is now a single SMA jack with a 10 dBm sine output.

What is it?

This is a GPS disciplined high precision oven controlled crystal oscillator. It has a short-term Allan deviation specification of < 1e-11 @ 𝜏 1s, which means that most of the time the frequency will be within 0.1 ppb. GPS is used to continuously discipline the frequency to maintain accuracy.

The output frequency of 10 MHz is provided as a +10 dBm sine wave. The output channel is supplied on an SMA jack.

The unit optionally comes with a 5 VDC power supply (90-240 VAC 50/60 Hz with a North American plug - a passive prong adapter is all that would be required for worldwide use). If you wish to power it yourself, it requires 5V@1A supplied over a USB C connector.

The board has three LEDs, labeled FIX, 0 and 1. The FIX LED comes directly from the GPS module and is its indication of a GPS fix. The 0 and 1 LEDs indicate the operating mode of the firmware. If they're blinking back and forth, then there is no GPS fix. If they're not, then they represent a binary value 0-3, which mean coarse (0 - both LEDs out), fine (1 - 0 on, 1 off), finer (2 - 0 off, 1 on) and finest (3 - both on) mode.

The first time it is powered up, it will take several days before the unit will completely stabilize and the phase range will settle, but the frequency will be well disciplined in the meantime. Restarts after that should be able to obtain a lock faster (depending on how long it's been powered off). It is recommended that the unit be (more or less) permanently powered, as the stability of the crystal will improve with age and use.

There is an SMA connector for an external GPS antenna (required and not included). The oscillator will hold-over when GPS is unavailable (however, it will not attempt to "guess" at the aging over time while holding over. Good, constant GPS reception is required for maximal accuracy). There is also a mini-DIN 6 diagnostic connector that supplies a separately buffered PPS output, diagnostic serial output from the controller and serial NMEA I/O from the GPS module.

The board has a 6 pin ISP header for firmware updates. Using it requires an AVR PDI programmer with a 6 pin cable and a pogo pin programming adapter. The firmware requires only standard AVR GCC and AVR libc to build, and is open source and available on GitHub.

How good is it?

The design is intended to produce an Allan variance of around 1E-11 or better (assuming good GPS reception and good mechanical and temperature stability). The free-running OH300 oscillator has a short-term spec of no worse than 1E-11 @ 𝜏 10^0, but usually achieves around 8E-12. The pink graphs are comparisons between a unit running against a Thunderbolt after several days of warming up. They are reasonably characteristic.

Why did you make it?

I had a project where I needed to perform particularly accurate frequency measurements and wanted those measurements traceable to some respected standard. This was the cheapest way I could find to do it. It is not intended to compete with commercial GPSDOs (which are far costlier, but you get what you pay for), but rather offer an entry level frequency standard for a reasonable price.

What makes it special?

In terms of accuracy and value, this is the best unit I know how to make. And for most hobbyist and light commercial uses, it should be more than adequate.