First Pilot Version of External PLL for GPS frequency reference of VHF WSPR Transceivers and Transverters
The external Pll project block diagram is as shown below.
An OEM, single board GPS receiver (RoyalTek REB21R) is used with an external active antenna to provide one pulse/second GPS timing as in input to a PIC processor running VE2ZAZ GPS/DO code. The PIC counts the frequency of the local OCXO, compares it to the 1 PPS signal and steers (disciplines) the OCXO to be phase coherent with it. This digital PLL has configurable sample size, loop bandwidth, holdover and a host of other features that can be accessed with a stand-alone Windows executable that manages it all. The PIC code that accomplishes it is from VE2ZAZ's project.
Once the GPS receiver has acquired satellites and is providing the 1 PPS signal, the PIC steers the OCXO to the correct frequency. For the OCXO we are presently using, approximately .1 ppb accuracy, 1 part in ten billion, is achievable. This 10 MHz reference frequency is available as an output from the pilot external PLL box which can be seen in the photographs.
In addition to this GPS/DO, the PIC also runs code that configures a TI CDCM7005 so that an external VCXO, ranging from low- MHz to 1500 MHz, can be phase locked to the 10 MHz OCXO signal. This is accomplished with 16 bit M/N integer dividers so that the VCXO is coherent and a rational multiple of 10 MHz. The PLL loop parameters as well as several derivative outputs of the VCXO are configurable.
There are two versions of the external PLL box, one version the GPS10P is designed to phase lock a VCXO inside an amateur transceiver or transverter. The main locked output from the box has two signals on it. One is a sample of the transceiver or transverter master oscillator that is being compared to the 10 MHz reference. The other is a VCXO DC tuning voltage that will steer that same oscillator to be phase coherent with the 10 MHz reference. Thus, on an Icom IC-706 MKIIG, a tap of the 30 MHz master oscillator signal at low level, -20 dBm comes into the external PLL and a DC voltage produced by the PLL rides on the same coax going back to the Icom transceiver. In order to use this VCXO-in-the-radio approach, the master oscillator of the radio must be modified in order to become a VCXO. This VCXO only needs enough tuning range to be able to correct the time & temperature variations of the particular oscillator.
The second version of the box, the GPS10V, has a built-in VCXO. This oscillator has been added into the design and is locked by the PLL. RF from it is injected into the radio via the coax - seen in the upper right of unit. Only this connection to the radio, the GPS antenna and a power connection to the external PLL box are necessary. However, a 10 MHz output is also available for SDR or hamshack test equipment use. In addition, an external 10 MHz input allows a reference other than the internal OCXO to be used to lock the radio. This could allow, for example, a rubidium standard to be used for frequency reference instead of the GPS constellation.
As of the end of January 2013, I have completed assembly and testing of the basic GPS10P. This first unit has been configured for the IC-706 and can be seen in operation with that transceiver in the photo above. The LEDs associated with the various functional blocks of the unit indicate status. Those LEDs are, GPS antenna present, OCXO status (same functionality as described for the VE2ZAZ design), frequency reference source (internal OCXO or external), PLL lock, VCXO signal present. Together they allow quick assesment of functionality of the external PLL. Here are two pictures of the pilot unit GPS10P with a rather beat-up label, it's been taken apart and put back together a lot of times in the course of debugging operation on this first pilot unit!
There are two RS232 connections. One is the GPS receiver connection, which can either run the SiRF chipset diagnostic code or just deliver normal GPS sentences. The second is the GPSDO/PIC output as described in the VE2ZAZ documentation.
All components except the power connector, a linear regulator and the GPS antenna connector are on the PC board. In the picture below you can see the GPS receiver mounted above the main board. There is a lithium backup battery installed on it which allows hot-start of the GPS in about 100 ms. after loss of power and restart.
The first time the unit is powered up, the GPS receiver must acquire satellites and figure out where it is. This can take considerable time but after doing it once, future starts are very fast. Also when the unit is first powered up after a long time, the OCXO takes a day or two to settle to full stability. I think this is typical for OCXOs. Some more effort can be spent in tweaking the loop parameters and some additional styrafoam put around the (bare metal) OCXO case to reduce convection currents in the box but even without this I am seeing on the order .1 ppb stability. This is probably enough to allow WSPR operation all the way to 10 GHz. Here is a picture of the inside of the lid:
The unit seen above is the GPS10P and does not have the on-board VCXO running. Most of the parts are loaded and I intend to verify the design and operation of the GPS10V now that the GPS10P is properly phaselocking the Icom. I have crystals on-hand to validate the design for the Yaesu transceivers, which require a 22.625 MHz crystal and the Icom IC-910 which needs 30.200 MHz. I hope to order a 116.000 MHz crystal to verify Elecraft transverter operation as well.
My next tasks are to complete and verify the VCXO and GPS10V design for the various radio types and John, K6PZB and I will probably build up a few more units of each type. It turns out that this is a *very* time consuming process. Apart from the development and debugging effort, each unit takes on the order of 8 man-hours of box machining, part loading and soldering. There are a lot of parts and they are being hand-loaded since at low volumes it isn't cost effective to have them machine loaded and soldered by a board house. If/when we do the first "production run" of units, we may decide to have this done for us. Please feel free to contact us with any questions or suggestions!
Glenn n6gn January 30 2013