Author Topic: Antenna selection on generic GNSS modules (Ublox etc.)  (Read 591 times)

0 Members and 1 Guest are viewing this topic.

Offline MarkMLl

  • Regular Contributor
  • *
  • Posts: 201
  • Country: gb
Antenna selection on generic GNSS modules (Ublox etc.)
« on: June 13, 2021, 04:33:53 pm »
I have a number of GPS etc. modules which I'm using for experimental purposes, with (genuine) Ublox neo-xx cans onboard. In general I select these to have provision for an external antenna, very often they support both an external and an onboard antenna.

I've seen various mutterings that a capacitor should be removed so that there is only a single RF source. Can anybody provide definitive or at least informed information on this?

In more detail, the attached photo shows part of a GPS module with SMA and IPEX connectors paralleled at the bottom and left, and the centre connection of the onboard (ceramic patch) antenna towards the top right.

The external antenna is connected to C8 and L1 in parallel, with C8 connected to U3 which is an unmarked six-pin device.

The internal antenna is connected to C4 and L2 in series, hence to U3.

The fragmentary instructions I've seen suggest that C8 should be removed if the external antenna is being used... which is of course rubbish since that's the in-path device. However the seller who advised that must have got the notion from somewhere.

If U3 is a passive combiner, e.g. from Mini-Circuits, then is is reasonable that an unused RF source should be isolated since an out-of-phase signal would mess things up.

If U3 is an active switch, e.g. from Peregrine as mentioned in https://www.u-blox.com/sites/default/files/products/documents/GNSS-Antennas_AppNote_%28UBX-15030289%29.pdf , then it's not reasonable to expect to have to remove components.

I lean towards U3 being active, since the component layout on the module is approximately as described in the Ublox appnote- even if the open-circuit detection has been omitted or simplified. However I'd be interested in the community's opinion or experience.

MarkMLl
 

Online evb149

  • Super Contributor
  • ***
  • Posts: 1901
  • Country: us
Re: Antenna selection on generic GNSS modules (Ublox etc.)
« Reply #1 on: June 13, 2021, 04:53:41 pm »
Hard to tell from the picture but my guess is:
1: SMA ----> C8A ---> L1A -----> 2nd module pin is the RF microstrip connection from the UFL & SMA into the module pin.

2: L1 <-----<---R2----< 5th module pin is a series circuit which probably enables the module to feed DC bias voltage as a bias Tee  along to the SMA and UFL antennae for their LNAs if they need DC bias on the feed for such.

Looks like: module #2 RF RX <--- C8A <-- C8B <-- microstrip RF 50Z trace <--- U3-6.
Therefore maybe we have:
U3-3 <--L2A <--- L2B <--- C4B <-- C4A <--- patch ant.
Therefore U3 should perhaps be a LNA and the patch antenna signal comes through the PCB via into C4A and U4 is not relevant to antenna switching.  If the patch ant has a built in LNA then I do not know what U3 would be useful for.

So if the UFL and the SMA antennae are not connected then all is well.  But to connect either optional antenna, C8 should be removed if U3-6 is an RF LNA output feeding through C8.

If u3 is a switch, where would its control pin be? I see perhaps:
U3-1: ground plane connection?
U3-2: ground plane connection?
U3-3 RF trace from L2 --- C4 --- via to patch antenna?
U3-4 C5A in parallel with a thick trace.  I'm guessing this is a decoupled VCC since C5B is maybe grounded and the trace seems to go to other possibly decoupling capacitors elsewhere.
U3-5: No connect?    Or maybe tied to U3-4 under the chip to tie it high?  Maybe this is the LNA enable pin.
U3-6 RF trace.

Looks possible to be LNA, not possible to be a switch unless there is a control signal

I don't know why L2 would be in series with C4 though that is a little odd forming a series resonator if that's their connection.

And the connection from L2 to U3 is not so much like a 50 ohm microstrip as is on other traces.



« Last Edit: June 13, 2021, 05:17:56 pm by evb149 »
 

Offline MarkMLl

  • Regular Contributor
  • *
  • Posts: 201
  • Country: gb
Re: Antenna selection on generic GNSS modules (Ublox etc.)
« Reply #2 on: June 13, 2021, 05:06:23 pm »
The five-pin U4 is a voltage regulator, marked 4A2D.

I suppose that one thing I could usefully try is looking at the six-pin chip (U3) and finding whether it has CMOS-style power on it, rather than being a passive splitter/combiner.

MarkMLl
 

Online evb149

  • Super Contributor
  • ***
  • Posts: 1901
  • Country: us
Re: Antenna selection on generic GNSS modules (Ublox etc.)
« Reply #3 on: June 13, 2021, 05:16:13 pm »
Is pin U3-5 high?
Does U3 look like it could be in the following package, for example?

https://www.digikey.com/en/products/detail/nxp-usa-inc/BGU7005-115/13513598
q.v. that data sheet.  Looks like the physical connections I'm inferring match the pinout if the enable pin is high.

 

Online evb149

  • Super Contributor
  • ***
  • Posts: 1901
  • Country: us
Re: Antenna selection on generic GNSS modules (Ublox etc.)
« Reply #4 on: June 13, 2021, 05:28:25 pm »
It couldn't be combining / splitting anything if it has only one RF input and one RF output.
Though the lack of consistent transmission line use is a bit confusing it seems clearly to be a 1-in 1-out device
with a Vcc, Ground, and a couple other pins.

I'm fairly convinced it's a LNA similar to the guesstimated one I mentioned.

The RF layout of the board isn't great though with the large stubs everywhere depending on what antenna is being used but I guess it works.

The five-pin U4 is a voltage regulator, marked 4A2D.

I suppose that one thing I could usefully try is looking at the six-pin chip (U3) and finding whether it has CMOS-style power on it, rather than being a passive splitter/combiner.

MarkMLl
 

Offline MarkMLl

  • Regular Contributor
  • *
  • Posts: 201
  • Country: gb
Re: Antenna selection on generic GNSS modules (Ublox etc.)
« Reply #5 on: June 13, 2021, 05:56:30 pm »
U3 is marked "LY". I noticed this when it started catching oblique sunlight, but now find that there's one very specific direction that LED illumination has to be applied which also shows it clearly... same applies to three modules branded differently so this looks like somebody's reference design.

That means it might be a MAX2659ELT+ with gnd on 1 and 2, RF in on 3, Vcc on 4, shutdown/ on 5 and out on 6. Input is shown via series L and C.

If that's correct it suggests that it's only in the "ceramic patch" antenna circuit, that the SMA/IPEX antenna is in parallel with its pin 6,

In which case if the antenna on the SMA connector is being used the internal antenna should be disabled, and that if this isn't done by the Ublox module telling the amp to shutdown removing C8 is the obvious way.

HOWEVER: component numbering isn't consistent, and on other modules that cap is identified as C2 etc... the important thing is to identify the SMA -> can line and look for a C and L in parallel.

MarkMLl
« Last Edit: June 14, 2021, 06:51:43 am by MarkMLl »
 

Online evb149

  • Super Contributor
  • ***
  • Posts: 1901
  • Country: us
Re: Antenna selection on generic GNSS modules (Ublox etc.)
« Reply #6 on: June 13, 2021, 06:33:14 pm »
I don't think it matters much if the amp is shut down or not.
There's only one perfect possibility which is that at the signal frequency the stub RF line "terminated" by the impedance looking into the disabled output of the LNA is essentially an RF open circuit at the patch feed line junction location with the SMA/IPEX feed line.
Though that's just barely possible, it wouldn't have been a design requirement, and I doubt the LNA data sheet even guarantees what its small signal S22 output impedance is in the disabled state, and anyway they would have needed to tune
the line length and geometry on the patch feed stub to make that happen while simultaneously making it a good 50 ohm feed
for the powered state.  Not really practical / reasonable.

So therefore the only right thing to do is remove C8 or the equivalent capacitor for the different modules, just like the "conventional wisdom" says.  Because whether or not the LNA output is 50R or something else when the LNA is off, it's going to disturb the signal line from the antenna you do want to use.

The parallel L is nothing to worry about, that's almost certainly the bias-t feed in to the SMA/IPEX.  So unless you don't
want DC bias to possibly be applied there, leave the L alone.

U3 is marked "LY". I noticed this when it started catching oblique sunlight, but now find that there's one very specific direction that LED illumination has to be applied which also shows it clearly... same applies to two modules.

That means it might be a MAX2659ELT+ with gnd on 1 and 2, RF in on 3, Vcc on 4, shutdown/ on 5 and out on 6. Input is shown via series L and C.

If that's correct it suggests that it's only in the "ceramic patch" antenna circuit, that the SMA/IPEX antenna is in parallel with its pin 6,

In which case if the antenna on the SMA connector is being used the internal antenna should be disabled, and that if this isn't done by the Ublox module telling the amp to shutdown removing C8 is the obvious way.

HOWEVER: component numbering isn't consistent, and on other modules that cap is identified as C2 etc... the important thing is to identify the SMA -> can line and look for a C and L in parallel.

MarkMLl
 
The following users thanked this post: MarkMLl

Offline MarkMLl

  • Regular Contributor
  • *
  • Posts: 201
  • Country: gb
Re: Antenna selection on generic GNSS modules (Ublox etc.)
« Reply #7 on: June 13, 2021, 07:49:21 pm »
So therefore the only right thing to do is remove C8 or the equivalent capacitor for the different modules, just like the "conventional wisdom" says.

I suspect that by now "conventional wisdom" is going to be this thread since the situation has not been adequately explained elsewhere :-)

I'd add that my workshop is thoroughly screened and even without sorting out that cap, I've been getting fairly good results with a cheap external antenna sold as https://www.banggood.com/GPS-Active-Marine-Navigation-Antenna-10-Meters-With-BNC-Male-Plug-Connector-New-p-1019248.html However the thing I want to "get into the record" is that the thread described there as "G3/4" actually matches standard conduit fittings... at least in the UK.

MarkMLl
 

Online evb149

  • Super Contributor
  • ***
  • Posts: 1901
  • Country: us
Re: Antenna selection on generic GNSS modules (Ublox etc.)
« Reply #8 on: June 13, 2021, 08:14:58 pm »
Yeah I figured that it might just still work anyway even with C8 in place.

It really depends on how good your antenna is and how much gain the external antenna LNA has.
If the LNA has say 12dB of gain, but the loss from the parallel terminated LNA is something like 6dB then
you've still got a reasonably good signal strength coming in just maybe several dB less than you could have.

You could quantify it roughly if you want -- use the same positioning of the same external antenna with C8 still present
and wait until you collect enough satellite signal strength measurements to see what your repeatable peak
signal strength is for a satellite in a given angular spot maybe around 20-45 degrees below the zenith or whatever.

Then leave everything the same but remove C8 and remeasure and see if your signal strength is a few dB better as expected for similarly located SVs as before.

In good conditions even an antenna without a LNA can work so there's usually a few dB of margin with good geometry and a decent size antenna with a short enough feed line that there's not too many dB feed loss.

EDIT: wow that's a lot of LNA gain in that antenna you linked --
"Lna gain(without cable):42±2dB" -- though that is also a very long cable on it.  I wonder what the cable loss is.
Anyway that's certainly a reason (hopefully decent antenna with lots of LNA gain added) why the unit will see the signal with or without C8 installed.


I suspect that by now "conventional wisdom" is going to be this thread since the situation has not been adequately explained elsewhere :-)

I'd add that my workshop is thoroughly screened and even without sorting out that cap, I've been getting fairly good results with a cheap external antenna sold as https://www.banggood.com/GPS-Active-Marine-Navigation-Antenna-10-Meters-With-BNC-Male-Plug-Connector-New-p-1019248.html However the thing I want to "get into the record" is that the thread described there as "G3/4" actually matches standard conduit fittings... at least in the UK.

MarkMLl
« Last Edit: June 13, 2021, 08:18:55 pm by evb149 »
 
The following users thanked this post: MarkMLl

Offline MarkMLl

  • Regular Contributor
  • *
  • Posts: 201
  • Country: gb
Re: Antenna selection on generic GNSS modules (Ublox etc.)
« Reply #9 on: June 13, 2021, 08:46:47 pm »
EDIT: wow that's a lot of LNA gain in that antenna you linked --
"Lna gain(without cable):42±2dB" -- though that is also a very long cable on it.  I wonder what the cable loss is.

I suspect that its quoted bandwidth- 1575MHz+/-5 is fairly accurate: it's good for GPS but I can't see any of the other constellations at all... and would prefer to. https://gssc.esa.int/navipedia/index.php/File:GNSS_navigational_frequency_bands.png

I'd normally rip it apart to see how well it's made, but my mast is in an extremely exposed position and right now I need it to stay waterproof. It might possibly resemble the one discussed at https://www.reddit.com/r/RTLSDR/comments/fvycl5/how_can_i_modify_this_type_of_gps_antenna_for/

I'm losing quite a bit since I'm putting it through one of these splitters https://www.aliexpress.com/item/1005001614629367.html ... they appear to "do what it says on the can" quite nicely.

MarkMLl
« Last Edit: June 14, 2021, 10:27:43 am by MarkMLl »
 

Offline Miti

  • Super Contributor
  • ***
  • Posts: 1086
  • Country: ca
Re: Antenna selection on generic GNSS modules (Ublox etc.)
« Reply #10 on: June 14, 2021, 01:06:11 am »
I have two of those modules, U3 is LNA for the passive ceramic patch antenna installed on the other side. Remove C8 to use it with external antenna.
Fear does not stop death, it stops life.
 
The following users thanked this post: MarkMLl

Offline MarkMLl

  • Regular Contributor
  • *
  • Posts: 201
  • Country: gb
Re: Antenna selection on generic GNSS modules (Ublox etc.)
« Reply #11 on: June 14, 2021, 06:48:09 am »
I have two of those modules, U3 is LNA for the passive ceramic patch antenna installed on the other side. Remove C8 to use it with external antenna.

Except where it's marked U2 (as on two out of three of the modules I've just done), and we've thrashed out why it's only that cap that's significant.

MarkMLl
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf