[time-nuts] To use or not to use transmission line splitters for GPS receivers

[Magnus Danielson]

Dear Edgardo,

On 10/09/2012 04:31 AM, Edgardo Molina wrote:

> Now to the point if you kindly allow. I got involved in a round table
> discussion around the use of GPS antennas for time and frequency GPS
> receivers. I tried to make some points from my personal perspective.
> I got resistance from the audience and the topic went hot very quickly
> and didn't set at the end. Honestly I would like to share my doubts
> and opinions with you in order to enhance my experience about the topic.
> In the end it could also be beneficial to close this debate tomorrow
> while attending to CENAM's time and frequency forum.
>
> Facts and thoughts:
>
> 1. The time and frequency attendees at CENAM`s time and frequency forum
> is reluctant to use GPS antenna splitters for a number of reasons I
> couldn't conceive.
> 	1.1 They argued that cross talk could happen among ports. I doubt
> it with the newer models. I have experience with HP and Symmetricom
> units and they state their products cancel cross talk.

Cross-talk can occur, indeed. The main issue should be the 
connected/unconnected reflections, only very secondary would be leakage 
of signals from one GPS receiver to another. Both is being handled by 
measuring and require sufficient port isolation.

> 	1.2 They argued that there could be problems from the power
> feeding of the antenna and mismatches at the receiver antenna port
> voltages. Again I doubt it if one uses receivers in the same voltage range

There can be. Therefore you should select to use antenna splitters that can
1) Select which port to be feeded from.
2) Back-signal the loss of antenna to the GPS receiver.

There are double frequency GPS splitters able to do that. I will soon 
get some for my private lab. I will measure one I borrowed tomorrow.

A third issue would be that passive splitters will give you loss of 
signal level. Therefore is active splitters recommended.

> 2. I have been experimenting with GPS constellation coverage with
> different brands of antennas. I have found different gains, different
> radiation patterns and as a result different satellites in view for
> identical GPS receivers.
 >
> 	2.1 I have found that using a single antenna and a two port
> HP splitter I get the same radiation pattern, gain and identical
> satellites in view for a set of identical receivers. For
> comparison purposes I feel this is an adequate scenario.
> 	It is easier for me to take care of the transmission line
> length and errors caused by phase differences, attenuation and delays.
> 	2.2 Two identical GPS receivers each one with it's own antenna,
> could eventually cause spatial diversity reception for a system of
> two receivers conceptually set as one for comparison purposes.
> Different satellites being tracked on each receiver if not connected
> to a common antenna. Even if antennas and transmission lines are identical.
>
> Question is: Am I wrong doing the above mentioned assumptions while
> considering the use of GPS transmission line splitters? I which
> scenarios are the splitters recommended? In which cases they are
> rather not to be considered?

Just using a transmission line style Wilkinson power splitter alone is 
not recommended. The isolation effect between ports is terrible. Using 
three-port couplers allows for the ports reflection to be loaded into a 
separate resistor. I think I recall seeing that on the splitters that I 
have openend up. Thus, it is fairly trivial to isolate the splitter.

The Agilent L1 splitters uses resistive power-dividers, which has better 
isolation at the cost of higher loss, but then they include an amplifier 
to overcome that, but the amplification is at the input side.

If you require isolation between ports, it can be designed in and if you 
get the devices I have seen, it should be fair isolation.

Similarly, the power require similar isolational effects, and again 
commercial devices addresses these issues.

When in doubt, measure the effect. Network analyzers for 1.2-1.7 GHz 
isn't hard to come by these days. I'll toy around a little with the one 
at work.

> Better ask as to start buying more antennas or feel comfortable with
> my original RF distribution design using splitters.

The joy of using multiple antennas is one thing, but if you have a good 
antenna up, then good cabling in combination with good splitters should 
work well.

> Your kind comments and expert advise is always welcome. Thank you!

My 2 öre contribution to the discussion.

Cheers,
Magnus