[ASK] Simple way to check GPS active antenna ?

[BravoV]

Purchased an used GPS outdoor active antenna for my upcoming GPSDO installation, for biasing it needs 5 VDC and <28mA, datasheet attached below.

Without having to use a proper bias tee, just a quick & dirty way, can I just bias it directly with DC power, and then coupled with a RF DC block adapter that is connected to my spectrum analyzer to check if its working as its used, and also to measure the amplified signal ? (1575.42 ± 10 MHz)

That frequency is in my RF DC block adapter freq range, and also I will be using a low noise/ripple bench PS that will be set accordingly with current limit enabled.

It this method adequate for my purpose just to verify and measure it, just once before I install it high up there ?

Any other considerations or suggestions will be appreciated.

[Ian.M]

Well, you'll have to build or buy a bias T. How else are you going to get power onto the coax and still tap its RF output without exceeding its rated <2 VSWR?

I've tested used active GPS antennae by basically the same method you propose  except I hand-built an inline 78L05 based 5V regulator  which was basically a pair of bias Ts back to back with a shared DC block capacitor + the regulator itself, as the spectrum analyser I was using was intended for analog satellite installation work, and outputted a switchable 13V or 18V to the LNB.

[BravoV]

Well, you'll have to build or buy a bias T. How else are you going to get power onto the coax and still tap its RF output without exceeding its rated <2 VSWR?

Well, the idea is to solder directly the pos & neg wires from the power supply at the middle core and shield like this below example photo.

Can this be done ? I don't need the power measurement accuracy, what I need is just to know if the antenna is working.


Random image from Google

I've tested used active GPS antennae by basically the same method you propose  except I hand-built an inline 78L05 based 5V regulator  which was basically a pair of bias Ts back to back with a shared DC block capacitor + the regulator itself, as the spectrum analyser I was using was intended for analog satellite installation work, and outputted a switchable 13V or 18V to the LNB.

Do you have any photo ?

[Ian.M]

Sorry, no photos - it was many years ago and I didn't even have a digital camera back then.   

Up at 1575 MHz, assuming cable with a velocity factor of 80%, a quarterwave is slightly under 48mm.  As a rough rule of thumb, stubs or discontinuities less than 1/10 of that - approx 5mm long will only have a minor effect on the signal.  Unfortunately BNC T connectors have side arms considerably longer than 5mm, then you have an uncontrolled transition into whatever the PSU cable is, so, if there is a very strong signal within the antenna's passband, there is a small risk of damage to it due to excessive VSWR.

If you were to join two BNC chassis sockets back to back, with a 4.2mm diameter center conductor between them and a 9.65mm I.D. outer conductor, the result will have a fairly small impedance discontinuity, mostly due to the actual BNC connectors, as air spaced hardline with those dimensions has an impedance very close to 50 ohms.     For the very short length of a bias T directly built between the center pins of BNC chassis sockets, the center conductor difference from 4.2mm dia isn't too important. Bulk up the center pins with fine wire if you are feeling fussy!  Therefore if you simply cut a short length of 10mm bore copper pipe into two semicircular half pipes, solder one between the threaded exteriors of the back of the BNC connectors leaving just enough space between them for a small SMD capacitor between their center pins, and wind a RF choke out of fine magnet wire round the shank of a small lightly waxed drill bit as a former, then dope the choke with a drop of superglue so it holds its shape, solder one end to one BNC center pin, and bring the other out through a pinhole drilled in the half pipe, and take it immediately to a SMD cap soldered on end to the exterior of the half pipe, you've almost completed a bias T.  All that's left to do is to solder on the DC wires and the other half pipe to compete its shielding.   That's more or less how I built my inine 5V regulator except I had two chokes, two caps on the exterior and a 78L05 regulator between them.

I remember I got a reasonably clear spike at the right frequency on the spectrum analyser that was easily distinguishable from the noise floor.  It vanished if I switched off LNB power or when I took it indoors under a concrete roof that pretty much totally blocked the sky view, so I was fairly confident it was picking up the GPS signals.

[edpalmer42]

The easiest way to do this is with a splitter.  You can use either a real GPS splitter or a satellite TV splitter that has power passing on at least 1 port.  These are commonly used with satellite receivers to provide power and control for the LNB.  I don't have any links handy, but I've seen reports that they work fine.  The GPS splitters will have the advantage of terminating the antenna ports back to the receiver so you don't get an antenna alarm.  You might have to add that feature to a satellite splitter.

Having a splitter in the link is nice because it allows you to add receivers for testing without disconnecting your main receiver.

The 'standard' GPS splitters are the ones from HP/Agilent/Symmetricom.  You might have to wait a while to find one at a reasonable price.  There are other good brands.  Do some research.

Be careful.  Remember that some spectrum analyzers will fry if 5V gets fed back into the input.

[BravoV]

Just use some coax connectors or coax cable and build a bias T.
You can just wind some blocking coils that should be good enough for a quick test.  Go for Z=500 or Z=1000 and it'll work for a quick test.  Or maybe lambda/4 good coax into an AC short.

Please apologize, I'm just an electronics hobbyist, what inductance of that coil will be ?

The worse problem may be seeing anything above the noise floor of your SA.  Even through a LNA your "signal" will just be wideband "noise" over a couple MHz span.

Even of clear sky , clear 180 view and no obstruction at all, what can I expect the "typical" signal strength on the GPS frequency that come out from my antenna ?

Usually it is easier just to plug the antenna into a real GPS receiver unit and look at the unit's reported signal strength and visibility of various SVs over the hemisphere to see if your antenna solution has good signal reception capability vs. noise and angle.  Would give a more meaningful answer than a SA by far.

Yes, that is the plan originally, its just I have this complicated timing issue, that in order to install my antenna high up there, I have to borrow a long ladder from a friend that lives few blocks away, and thats also has time constraint too as its not always available. And to make it worst, my GPSDO board is not arrived yet. 

[BravoV]

Sorry, no photos - it was many years ago and I didn't even have a digital camera back then.   

Up at 1575 MHz, assuming cable with a velocity factor of 80%, a quarterwave is slightly under 48mm.  As a rough rule of thumb, stubs or discontinuities less than 1/10 of that - approx 5mm long will only have a minor effect on the signal.  Unfortunately BNC T connectors have side arms considerably longer than 5mm, then you have an uncontrolled transition into whatever the PSU cable is, so, if there is a very strong signal within the antenna's passband, there is a small risk of damage to it due to excessive VSWR.

If you were to join two BNC chassis sockets back to back, with a 4.2mm diameter center conductor between them and a 9.65mm I.D. outer conductor, the result will have a fairly small impedance discontinuity, mostly due to the actual BNC connectors, as air spaced hardline with those dimensions has an impedance very close to 50 ohms.     For the very short length of a bias T directly built between the center pins of BNC chassis sockets, the center conductor difference from 4.2mm dia isn't too important. Bulk up the center pins with fine wire if you are feeling fussy!  Therefore if you simply cut a short length of 10mm bore copper pipe into two semicircular half pipes, solder one between the threaded exteriors of the back of the BNC connectors leaving just enough space between them for a small SMD capacitor between their center pins, and wind a RF choke out of fine magnet wire round the shank of a small lightly waxed drill bit as a former, then dope the choke with a drop of superglue so it holds its shape, solder one end to one BNC center pin, and bring the other out through a pinhole drilled in the half pipe, and take it immediately to a SMD cap soldered on end to the exterior of the half pipe, you've almost completed a bias T.  All that's left to do is to solder on the DC wires and the other half pipe to compete its shielding.   That's more or less how I built my inine 5V regulator except I had two chokes, two caps on the exterior and a 78L05 regulator between them.

Ian.M, thank you for the concise build instruction, really appreciate your help here. 

I will try to find if I can source cheap & easily these parts, as probably it will be used only once.

I remember I got a reasonably clear spike at the right frequency on the spectrum analyser that was easily distinguishable from the noise floor.  It vanished if I switched off LNB power or when I took it indoors under a concrete roof that pretty much totally blocked the sky view, so I was fairly confident it was picking up the GPS signals.

Actually this is exactly what I'm expecting, using quick & dirty hack biasing the antenna, while temporary placed probably in front of my house, or even in the middle of the road  to get wide view of the sky, and then see if it does "something around" the 1575.42 MHz at my spectrum analyzer. Of course with minimal risk damaging the LNB, but yet quick & easy way.

[BravoV]

The easiest way to do this is with a splitter.  You can use either a real GPS splitter or a satellite TV splitter that has power passing on at least 1 port.  These are commonly used with satellite receivers to provide power and control for the LNB.  I don't have any links handy, but I've seen reports that they work fine.  The GPS splitters will have the advantage of terminating the antenna ports back to the receiver so you don't get an antenna alarm.  You might have to add that feature to a satellite splitter.

Having a splitter in the link is nice because it allows you to add receivers for testing without disconnecting your main receiver.

The 'standard' GPS splitters are the ones from HP/Agilent/Symmetricom.  You might have to wait a while to find one at a reasonable price.  There are other good brands.  Do some research.

Thanks for the tip, its just I worry I can't afford that at this moment, but I will keep looking forward into this matter.

Be careful.  Remember that some spectrum analyzers will fry if 5V gets fed back into the input.

Thanks for reminding, yep, any connections to my spectrum analyzer will have to go through the external DC Block 1st, and probably with an attenuator too if I'm not too sure.

[edpalmer42]

The easiest way to do this is with a splitter.  You can use either a real GPS splitter or a satellite TV splitter that has power passing on at least 1 port.  These are commonly used with satellite receivers to provide power and control for the LNB.  I don't have any links handy, but I've seen reports that they work fine.  The GPS splitters will have the advantage of terminating the antenna ports back to the receiver so you don't get an antenna alarm.  You might have to add that feature to a satellite splitter.

Having a splitter in the link is nice because it allows you to add receivers for testing without disconnecting your main receiver.

The 'standard' GPS splitters are the ones from HP/Agilent/Symmetricom.  You might have to wait a while to find one at a reasonable price.  There are other good brands.  Do some research.

Thanks for the tip, its just I worry I can't afford that at this moment, but I will keep looking forward into this matter.

That's why I mentioned the satellite splitter.  You can get them from places like Home Depot or Walmart for less than $10.

Here's one description of how to do it.  The amplifier is optional, but remember that a 2-way splitter will drop the signal level by 3 dB, and a 4-way splitter drops it by 6 dB.

http://www.prc68.com/I/4GPS.shtml

Ed

[texaspyro]

Don't waste your time with all that tedious mucking around in hyperspace.   Just wait for your GPS to come in and see how well the antenna works with it.  It's the only way to be sure.

[dmills]

The GPS signal is actually below the noise floor! You will see nothing on a SA as only the processing gain lifts the signal above the noise and your analyser does not know how to do that.

This is why the initial acquisition phase for a totally cold booted GPS rx takes so long, it does not know roughly what it expects to see, so has a much lower processing gain. Once the thing has an accurate time and an almanac downloaded, the receiver knows what most of the signal should contain, so it can do extensive correlation to effectively narrow the 'bandwidth'.

About all you can really do is power the thing and see if you get an amplified noise floor on the SA, a bias T and DC block, job done, you will not see the GPS signal itself.

Regards, Dan.

[BravoV]

Don't waste your time with all that tedious mucking around in hyperspace.   Just wait for your GPS to come in and see how well the antenna works with it.  It's the only way to be sure.

God of GPSDO boards has spoken, aye Sir. 

There is a new development, my board arrived and also access to ladder is granted for long time. 

Currently pursuing the coax cable terminations, as N type is new for me. 

Again, many thanks to all contributors, lets assured, even I cancel this test, I've gained quite some knowledge from you all. 

[BravoV]

The GPS signal is actually below the noise floor! You will see nothing on a SA as only the processing gain lifts the signal above the noise and your analyser does not know how to do that.

This is why the initial acquisition phase for a totally cold booted GPS rx takes so long, it does not know roughly what it expects to see, so has a much lower processing gain. Once the thing has an accurate time and an almanac downloaded, the receiver knows what most of the signal should contain, so it can do extensive correlation to effectively narrow the 'bandwidth'.

About all you can really do is power the thing and see if you get an amplified noise floor on the SA, a bias T and DC block, job done, you will not see the GPS signal itself.

Regards, Dan.

Thanks Dan, really, an eyes opener, I didn't aware of this. 

[Ian.M]

Well I remember seeing something on the spectrum analyser, enough to convince me the antenna was working.   However, in the light of Dan's comments above, it must have been a spurious signal, possibly being radiated from a nearby malfunctioning active TV antennae. (See http://gpsworld.com/the-hunt-rfi/ )

[ConKbot]

The GPS signal is actually below the noise floor! You will see nothing on a SA as only the processing gain lifts the signal above the noise and your analyser does not know how to do that.

Partially true. The signal is below the thermal noise floor for a bandwidth needed to receive the data. (~1-2 MHz for the 1.023 Mchips/s ) Buf if you choke the RW/VBW way down to drop the noise by a lot and have enough gain between the antenna element and SA (a 30dB LNA in the active antenna may not be enough, depending on the SA)  you can get a spectrum.  But if you have a poor or damaged LNA you may end up getting just a response graph of the filter after the LNA, if present.

That being said, the easy way of figuring out if an active antenna is working is to hook it to a known good RX. A lot easier than setting up a bias tee, SA, etc.

[BravoV]

Partially true. The signal is below the thermal noise floor for a bandwidth needed to receive the data. (~1-2 MHz for the 1.023 Mchips/s ) Buf if you choke the RW/VBW way down to drop the noise by a lot and have enough gain between the antenna element and SA (a 30dB LNA in the active antenna may not be enough, depending on the SA)  you can get a spectrum.  But if you have a poor or damaged LNA you may end up getting just a response graph of the filter after the LNA, if present.

That being said, the easy way of figuring out if an active antenna is working is to hook it to a known good RX. A lot easier than setting up a bias tee, SA, etc.

Noted, as I was naively thought I can see the bulge of the GPS frequency in my spectrum analyzer.

I will test it directly at the GPSDO receiver board directly, no more fussing around.

[BravoV]

Closure post, problem solved, bought an used Symmetricom 58535A distribution amplifier. 

[cdev]

Just for future reference, its relatively easy to make a single-frequency bias-tee that works okay, its making a broadband one with an acceptably flat response that's hard.

[TheSteve]

Just noticed this thread. If I compare a completely blocked antenna to one mounted on a tower with full view of the sky I do see a difference. In both cases you can see a rise in noise floor which peaks around the proper frequency. So am I seeing some GPS signals or is it purely atmospheric noise, to me it kind of looks like it is the GPS signals.

edit - added third pic with a really long sweep (0 Hz RBW) to lower the noise floor.

[cdev]

Doppler shift on LEO satellites is the give away. They describe sort of an S shape as they fade in and fade out. Although the degree of shift varies based on how close they are to you at their closest point and how fast they are going the track as shown on a frequency waterfall should start high and fade out low in a distinctive curve that you will only see with space signals.

[hamster_nz]

Doppler shift on LEO satellites is the give away. They describe sort of an S shape as they fade in and fade out. Although the degree of shift varies based on how close they are to you at their closest point and how fast they are going the track as shown on a frequency waterfall should start high and fade out low in a distinctive curve that you will only see with space signals.

The GPS Space Vehicles are not in low earth orbit, but semi-synchronous orbit (each orbit takes ~12 hours). They only have +/- 5kHz of Doppler shift, and the change in frequency is very slow.

Because the C/A signal uses direct synthesis spread spectrum signal you won't see any particular signal or carrier - it looks like noise. Also, the transmitted power is ~25W at 20,000km, covering the best part of half the earth so they are very weak - they are actually below the noise floor, so GPS relies on the conversion gain of "unspreading" the signal the signal to allow it to be received.

What is seen is pretty much all you can expect to see - an increase in signal over about +/- a few MHz of the L1 frequency.