Reducing GPS antenna voltage

[DaJMasta]

I've got a GPS receiver that's designed for building installation GPS runs and provides 24V to the GPS antenna connector.  Most GPS antennae run on 3.3-5V, and the ones that have the power handling to run at 24V are typically pretty expensive... I don't need the long run or outdoor capability, so I'd like to be able to use the real cheap antennae with the receiver.

What is the best way to drop the voltage?

It seems like the antenna itself needs 10-20mA optimally, so while I can try something like a resistor divider... it's fairly inefficient and may not be great for those kinds of currents.  I've tried tracing the antenna input path in the receiver to find where it drops to logic level, but it heads into the field of chips on a multilayer board after considerable input protection (gas discharge tube, huge diode, 2W+ resistors, etc), and I haven't been able to figure out where it makes the transition to just try dropping 5V on from the receiver.

Is there another method that would be simple and viable?  Something like a linear regulator in a module, AC coupling and dropping in the new DC with an inductor seems reasonable... but I don't have much experience with that sort of RF design and it would involve making a whole module.  If there were a way to either modify the antenna (one of the cheap, square ones with a ceramic element and a couple of transistors), or modify the receiver (if I can find where, no service manual is available) to do this without making another inline module and all the expense and design associated with that?

[Vgkid]

Generally the antenna voltage is passed through an inductor(serves as an AC block), find that inductor, and follow the ttrace where the 24v is input. either modify the trace(cut it) and solder in a wire going to the lower voltage. Or follow the trace to a connector on the board, where the 24v is present, and just desolder that pin, and feed in 5v.

I took an example from one of the Ublox integration manuals that I had.

[Ian.M]

Plan B:
Do the same hack at the antennae end.  Again there will be an inline coupling cap with an inductor tapping the DC on the cable side of it.

Plan C:
Two LNB power injectors back to back with a 5V regulator (+ decoupling caps as per datasheet) connected between their DC in ports, common to the cable shield.

[DaJMasta]

And that's basically what I see, the trick is that I don't know where the signal is being tapped off, so I'm not sure where is alright to cut (I would assume DC side of the inductor would be enough, but the trace still goes to a lot of stuff and disappears on a connected board).

Here's the board:


The connector is in the upper left, with the center pin connected to GT1 and C9.  L1 is the inductor I think is used to inject the DC, but that would imply that C9 is for data, but then there's a bunch of filtering on that trace, and it goes down to pin 3 on the top of the card edge connector and basically disappears into a via.  On the other side of L1, the power goes through R1 and R2, the cathode of CR1 and then to pin 4 on the card edge connector, which takes a very short trip (with a very thin trace) and disappears into a via next to another inductor.  The board pictured is multilayer.

Breaking the trace below CR1 should cut out the +24V supply, but because of the way it trails off and because the other side doesn't seem too obvious as being a signal path, I'm hesitant to just go ahead and do it.


The LNB injector could be a good option, if nothing else because I could leave all the parts intact... will look into pricing.

[Ian.M]

It looks like removing the big green resistors R1 and R2 and patching in a 5V regulator there would do the job. Lift one end of them and check if the track going to L1 is hi-Z with respect to the rest of the board.  It would alsao be easy to revert if unsuccessful.

[DaJMasta]

Poked around a bit more and verified that pin 3 on the card edge was signal, as it goes straight to some RF transformers that while pin 4's trace gets close... it does not connect.

So I made the mod, ended up cutting the trace below CR1 because I couldn't bring myself to just lift the resistor legs and leave that beefy reverse protection diode hanging, and while there wasn't a 5V source on the backplane, I found it on an unused header on the main board and flew a wire over to the antenna.  Seems to work fine with the new antenna, but apparently it also monitors the output current of the 24V to the antenna... and its now throwing a minor error because there isn't any being used.  The tough part is that nominal is 110-170mA for the antenna.... which means 2.5-3W of resistor and wasted power if I wanted to convince it that everything was fine...


Anyone have a good idea of how to spoof the current sensor when I'm not even sure which chip is doing the sensing? 

Maybe there's a way I can disable it in software instead...

[ajm8127]

It might be sensing the current across a resistor and monitoring it with an op-amp/comparator or ADC. If this was the case you could use a voltage divider to just give it the voltage it is expecting or feed the 5V through a properly sized resistor to allow the receiver to see the actual current of the antenna.

Like you said in your first post most active GPS antennas use 3.3 to 5 volts at like 50 mA or less. I am surprised this receiver is expecting 150 mA at 24V to flow into the active antenna. As you said that is multiple watts.