[time-nuts] Rooftop antenna and splitter

Ben Hall kd5byb at gmail.com
Sun Jan 27 13:58:06 UTC 2019


Hi Denny and list,

Some thoughts from me...worth about what you paid for them considering 
my Time-Nuts membership card was revoked the week before last for 
heresy, hahaha!  ;)

> PCTEL GNSS1-TMG-40N 
> (https://www.pctel.com/antenna-product/global-gnss-timing-antenna-gnss1-tmg-40n/)

I've got the GPS-only version of this antenna as my main antenna bought 
off the e-place - example:  <https://ebay.us/pbDgU2>  (no connection to 
the seller, not even sure if this is the place where I got mine)

It's a very nice antenna and has a nice mounting kit with it.  The 
filters on the GPS-only version I've got are tight enough to block out 
GLONASS.  IMHO, this is nice, as it shows they paid attention to the 
design.  On the other hand, now that I do want to receive GLONASS...I've 
got one of the eBay shipped-from-China "CORS RTK GNSS Survey Antenna 
high gain measurement GNSS GPS GLONASS BDS" antennas on the way that 
folks were discussing in the uBlox F9 thread here.

> And the following splitters:

I have no experience with commercial active splitters.  I've used either 
commercial passive units, homebrew active units, or modified commeriical 
passive units.

 >The run from the antenna to the splitter will be 30-35 feet, and
>from the splitter to the units will be 3-5 feet. I’m wondering about 
>the need for the 40dB vs the 26dB. I haven’t looked at any passive 
>splitters, but even with the 40dB I’m thinking won’t offer enough to 
>support even a 1x6 splitter.

Your proposed setup is pretty close to mine.  From my PCTEL to the 
Polyphaser lighting arrester is about 30 feet, and from the arrester to 
my splitter is about 10 feet.  The 30 feet is Cable Experts CXP-1318, 
the 10 feet is RG-58.  Depending on what receiver is on the splitter, 
add a couple of feet of RG-316.  (some are farther, some are closer, but 
it's about two feet from the splitter to the Z3801 or the TruePosition.)

My splitter is a modified Mini-Circuits ZC6PD-1900W.  This is a six way 
splitter, theoretical loss 7.8 dB.  I've modified it by adding a DC 
block capacitor to 5 of the output ports so that the PCTEL is fed +5VDC 
from the Z3801, and none of the other receivers see this DC voltage. 
This was simple - cut the traces at the five output ports I wanted to 
block, soldered on a SMD capacitor of value I cannot remember.

The VNA at work says I've got from 9 to 11 dB loss from 1.0 GHz to 2.5 
GHz.  It's actually a little less, but those are easy numbers to use.

Why the passive splitter?  I was able to get these off of the e-place 
for like $10 each...while an active 6-way splitter was at least $100 used.

This setup, while sub-optimal in so many ways, works pretty well! 
Here's a screen dump from Lady Heather:

<http://www.kd5byb.net/heather/trueposn.gif>

Lowest signal is 40 dBc...and it's pretty much *always* has the full 8 
satellites it can simultaneously track.

The Z3801 shows the lowest signal strength right now as SS = 118 for PRN 
20, and it's not tracking PRN 14 if you want to compare it's signal 
strengths versus the TruePosition.

Doing some quick loss calcs:  (warning:  I've not had my coffee yet, so 
very prone to error!)

30 feet of CXP-1318 = 2 dB loss (had to assume similar to Belden 9913 as 
Cable Experts doesn't have loss above 400 MHz)

arrester = 1 dB loss (a guess)

10 feet of RG-58 = 2.5 dB loss

splitter = 10 dB loss

2 feet of RG-316 = 1 dB loss

Sum of the loss = 16.5 dB.  Probably should toss a few more dB in there 
because of connectors, etc...

So even with my crappy, far-less-than-optimal setup, I get pretty good 
results.

Like I said earlier, I haven't had my coffee yet so all of the above may 
be totally worthless.  ;)

thanks much and 73,
ben, KD5BYB




More information about the Time-nuts_lists.febo.com mailing list