GPS Timing and Frequency Reference Hardware

Our GPS clocks that we will be using typically produce three outputs:
1 - 10 MHz Reference signal
2 - 1 Pulse Per Second (PPS) Reference Signal
3 - NMEA or TSIP Data which, among other things, gives us our position and Time of Day.

10 MHz
This signal is used for making sure the carrier frequency of the transmitter(s) is held within 1 Hz of each other. Most commercial repeaters that can accept an external reference input will use 10 MHz. There are exceptions but they are few and far between. This signal can also be used to drive the 9.6 MHz PLL board as well. This board being used to provide the 9.6 MHz GPS locked reference for the RTCM or Voter board to allow the simulcast using PL/CTCSS tones. 

1 PPS and NMEA\TSIP Data
These are provided directly to the RTCM or Voter Board. These allow the voter board to synchronize and time stamp the packets for examination by the host running the Allstar software with the Voter Module. 
The same goes for the TX audio sent from the voter module to the transmitter sites, which allows the TX audio to be synchronized among all the sites so it can be transmitted at the same time (or with the correct offset if needed). You may be wondering about TSIP, this is the data format the old Trimble Thunderbolt GPS receivers used. In my opinion, unless you can get the old Trimble Thunderbolts for very cheap, I would avoid them for a few reasons, one being that they are quite old and another being the power supplies require to run them are complicated. All the equipment I will be mentioning here will use the NMEA format. 

GPS Disciplined Oscillators (GPSDO) Devices we can use

- Trimble Thunderbolt 
These units used to be a dime a dozen as some say. Now they are rare, expensive and worst of all old. 
I do not recommend these for to be used anymore.


This unit has been around for awhile and is a great, cost effective GPSDO that provides all our outputs. 
I did a video review of one at one point but then a ham friend of mine did a MUCH better one. 
Here is a link to the video: BG7TBL GPSDO Review by VK7HH.
These units are primarily what I am using for my sites. I have found them to be reliable and have more than enough accuracy and precision to get the job done. They run about $150.00 on Ebay and they can sometimes also be found on Amazon. Something like this would be required for a TX and RX site. 
They provide 1 PPS, 10 MHz and RS232 9600 Baud NMEA outputs. Here is a wiring diagram drawn up by K9MMQ to connect this unit to an RTCM, wiring it to a Voter Board is similar.

- Leo Bodnar Mini GPSDO

This unit will only output 400 Hz to 810 MHz so it is a bit of a one trick pony but it does it very well for a good price. This can be programmed to give you your 9.6 MHz needed for the RTCM or Voter board or it can be used as a reference for just about anything else. The down side is that there is no 1PPS or NMEA data coming from it to use. Leo Bodnar Mini GPSDO

- Ublox GPS Module (Goes by various names)

This is a little self contained module that has a Ublox GPS receiver on it. This unit can be powered by +5Vdc or by Micro-USB. They typically come packaged with a small magnet mount active GPS antenna. This unit gives you 1 PPS and NMEA data out at 9600 baud. Here is a link to one on amazon:
Amazon Gowoops GPS Module with 3m Active Antenna
These units are perfect for a voting receiver site since a 10 MHz or a 9.6 MHz reference are not needed. 
Their output is at TTL levels, this isn't an issue with the RTCM since the NMEA data levels are selectable between TTL and RS232 levels. The Voter Board however is only designed for RS232 Levels so a level shifter would need to be employed between this. These are cheap an easily obtained from Ebay or Amazon. Here is an Amazon link to the one I have used successfully. 
Amazon TTL to RS232 Level Shifter

Garmin 18x LVC Puck

These have been used in the past and I have used one as well. The first thing I will say is that I do not recommend them for use. They have some quirky issues that I do not like. They are not all that sensitive of a GPS receiver. These things will see no satellites where other receivers will see at least 3 or 4. 
They have an issue where they sometimes can shift off by a whole second on their time which causes issues with the system. Jim Dixon experienced this issue himself and devised a sort of work-around called "puck-it" which was meant to compensate for this one second shift issue.
These units provide RS232 NMEA data at either 4800 or 9600 Baud and 1 PPS. The other issue I personally have encountered is the 1PPS duty cycle on time of the pulse is short. I have noticed this to be an issue with the RTCM I was using it on at the time causing intermittent problems. As I stated earlier, I would not recommend this for use, I added it here for reference or in case anyone had a bunch of them already and wished to use them. 

The Multiple GPS Antenna Issue
When using multiple GPS devices you run into the issue of have to use multiple GPS antennas which can get annoying. As you may already know, typically GPS antennas are active devices that contain a Low Noise amplifier to send the received GPS signal down the feedline with relatively little loss. The antenna is usually powered by either a +3.3 or +5 VDC bias voltage to power that low noise amplifier. There are specific GPS splitters that perform this task properly. These splitters when in a two-way split configuration have one port that allows this DC bias to pass and the other one that blocks the DC but contains a resistor load so the the GPS receiver its connected to does not go into alarm for a possible bad GPS antenna. The third port goes to the antenna and provides the antenna with bias. These splitters go from a two-way split on up to as many as 10. Here is an example of one made by Time Machines.

Some of these splitters are passive which are usually the 2 way split ones, splitters with more ports have a tendency to be separately powered devices that contain distribution amplifiers to overcome the splitting loss to reliably feed multiple GPS receivers. The one pictured below is a 4 way splitter that is powered by an external power supply and does not use the bias provided by any of the receivers. 

I have a tendency to prefer the units made by Time Machines because they work and they are reasonably priced new. Ones made by GPS Networking or Symmetricom are excellent as well but are very expensive new but can sometimes be had on Ebay for cheap.