Hi,

I got two cellular mobiles.
I want to connect them to an external antenna, via RF switch.
I'd like to operate the switch via PC USB Port.

Can you please advice on RF switches for cellular applications - for both WB and GSM bands?

Thanks in advance.

 

Hi

What are you looking to spend? Can you make a PCB?

The reason I ask is because you could go the route of connectorized switches, where you would have 3 SMA ports for the RF and probably two terminals to control which way the switch is thrown (where they must be inverse of each other). This is the easiest route since you won't have to worry about impedance matching a microstrip, just wire it all up with coax. Your looking at ~$50 - $80 for one of these, depends on frequency range, isolation, insertion loss, etc. Look here to start:

http://www.minicircuits.com/products/Switches.shtml

If this is out of your budget, you can get a switch IC but you need to spin a board for this to really work, you can't breadboard GHz frequencies. Here are some RF switches for a few dollars or less.

http://www.skyworksinc.com/products.aspx?CategoryID=79

You also may be able to get away with a 2-way RF power splitter. This way you wouldn't have to flip the switch, both would work. This is assuming that the isolation is great enough that one cell phone doesn't damage the other's receiver, and that whatever multiple access method the two use works as well on a piece of coax as it does in the air. If they were at different frequency bands it would definitely work.

Cheers,

Steve

 

Hi Steve,
Thank you so much for your help.

I reviewed Mini-Circuits RF Swtiches.
This one is actually what I'm looking for:

http://www.minicircuits.com/pdfs/MSP2TA-18-12.pdf

I already got a 12V Power Supply which I can control via a PC USB port, and that way I'll be able to switch the RF switch (added a pic below).

The only problem is its high cost - 190$ per unit.
It has a very wide frequency range - DC to 18GHz - but I only need DC to 3GHz frequency range.
I was hoping for the product's cost to be below 100$.

You got a recommendation on such product?

Thanks.
I'll share here additional info I find

 

Hi

I think there are a couple of reason's for the high price. First it works to 18GHz, and second it's a mechanical relay, not solid state. Unfortunately solid state won't work if your control can only supply 12V or be off (open circuit). What is this USB controlled supply you're using? I'm not sure you'll find something relay based for the cost you want (but I haven't looked extensively!)

I was thinking something like this, for about $60

http://www.minicircuits.com/pdfs/ZX80-DR230+.pdf

You're going to need to drive one control line with 5V/0V and the other with 0V/5V respectively so you'll always need power applied. You could use a FTDI USB-Serial with a small MCU to take serial commands and toggle lines. Or more simply, use a FTDI-Serial but toggle the RTS or CTS lines. Or use a FTDI to parallel chip, FT245R. Sparkfun makes breakout boards for all of these but they are very overpriced. Just make sure these chips can source over 1mA.

Let me know more about the supply and we can work out a control solution.

 

Hi Steve,
Thank you very much again.

The ZX80-DR230+ is indeed something I can afford.

The PSU I got is this:
I got a USB-4761 - this is an 8-channel relay controlled via USB port (http://www.advantech.com/products/USB-4761/mod_C1E301AB-CDC8-45C0-B610-6AEA44B544AE.aspx).
To one of its channel, I connect any PSU I want (i.e. 5V / 12V), and I connect the PSU to the desired product that I'd like to switch.

You think it'd be possible to integrate this USB-4761 into the control section somehow efficiently?

Thank you again Steve.

 

Sure, that should work nicely. You'll need a power supply at 2.7V - 5V (ZX80-DR230+ Supply Voltage operating range), that can supply say > 10mA. It's too bad the USB controlled relay doesn't bring out the 5V from the USB cable to a terminal. If your not comfortable breaking it open to tap off VBUS, you could run a second USB cable, cut it open, and pull off VBUS and ground. It should draw so little current that the PC will be fine with it. Or use some independent power supply.

Now wherever it comes from, attach VCC to COM, and attach resistors between NO and GND and between NC and ground. The resistor will pull the open side to ground. Choose a resistor for say 1mA current draw through the closed circuit path. Then wire the NO and NC terminals to the two control terminals on the RF switch. And don't forget to wire up all the VCC and GND terminals of the switch as well.

I'm a bit curious at this point, do you mind sharing some more details about the purpose of this switching?

 

Hi Steve,
Thanks a lot again!

You seem to have worked it out really great and efficiently.
If I got it right, the below drawing describes how it should work.
What do you think?

As for the purpose of this switching:
I'm testing a cellular chip in front of live network.
So what I do is connecting the cellular chip to an external antenna which is located on the roof of the building.
The cellular chip is also connected to a PC, so I can see how it communicates with live network in different scenarios - Idle, Voice Call, Internet connection, etc.

Sometimes, when I get a weird behavior (for example, the cellular chip doesn’t manage to register on a cell and therefore remains in No Service), I want to see how a Reference cellular mobile behaves, to determine whether the strange behavior is a live network issue, or an issue of the cellular chip that I test.

You see the point?

 

Hi Again,

The diagram looks right to me. Thanks for the background on the application, makes perfect sense. Good luck.

Steve

 

Thanks for the feedback

It is said in the datasheet that the device's linearity degrades below 1MHz.
However, they don't provide a graph which describes this degradation in linearity.

What do you think of it?

 

You don't care about below 1MHz, I think the lowest cell phone carrier frequency is something in the 800MHz. Glad to see your going over the datasheet though! Things you might care about:

* Input 1dB compression is min 28dBm. Verify your cell phones don't put out more power than this (I highly doubt it but worth checking).
* Isolation from RF1 - RF2 is min 37dB. Verify one transmitters output minus this isolation is below maximum allowable power on the other transmitters input. Again I doubt this is a problem but worth checking.
* Switch specifies 0VDC on input ports, if this is not true, you need a DC blocking cap. 10pF works well for 2.4GHz. If you have a schematic for the output stage of the cell phone you can check to see if there is a DC block in the path, a lot of the times there is as it is part of a harmonic filter. If you don't have this info, you could tie a 1k resistor from rf output to ground and measure it with a voltmeter.
* Switch has absolute max of 24dBm on an input which is not switched to the common port (the switch has to absorb this power). Unless you can guarantee the transmitter which is not switched to the antenna is off, verify it's maximum output power is below this level.

 

Hi Steve.

Thank you very much again

I checked it and the conditions are satisfied, beside the 0V DC input condition.

Thanks for bringing it up.
I read datasheet of course, but missed that line.

I think it might be an issue because I cannot get the schematics of reference cellular mobiles, and it isn't that easy to connect a resistor to the output port of the transmitter, as its connector isn't so "user friendly" for that purpose.

Are you familiar with a way of connecting a DC block capacitor without needing to solder it or cut the wires? i.e. a connector with a built-in DC block capacitor.
I mean, I'd like to easily connect a RF cable from the cellular mobile to the RF switch, using connectors only.

--
Actually, I found PIN Diode RF switches, which have built-in DC block capacitors.
You see any reason why not using these?
I found this one, which is a little more expensive (90$ per unit): ZSDR-230+ http://www.minicircuits.com/pdfs/ZSDR-230+.pdf
The only problem with it is that its 1dB compression point is 19dBm, while a cellular mobile can reach 27dBm.
You know where else I can find low cost RF switches (connectors)?

You don't care about below 1MHz, I think the lowest cell phone carrier frequency is something in the 800MHz. Glad to see your going over the datasheet though! Things you might care about:

* Input 1dB compression is min 28dBm. Verify your cell phones don't put out more power than this (I highly doubt it but worth checking).
* Isolation from RF1 - RF2 is min 37dB. Verify one transmitters output minus this isolation is below maximum allowable power on the other transmitters input. Again I doubt this is a problem but worth checking.
* Switch specifies 0VDC on input ports, if this is not true, you need a DC blocking cap. 10pF works well for 2.4GHz. If you have a schematic for the output stage of the cell phone you can check to see if there is a DC block in the path, a lot of the times there is as it is part of a harmonic filter. If you don't have this info, you could tie a 1k resistor from rf output to ground and measure it with a voltmeter.
* Switch has absolute max of 24dBm on an input which is not switched to the common port (the switch has to absorb this power). Unless you can guarantee the transmitter which is not switched to the antenna is off, verify it's maximum output power is below this level.

 

I didn't know there would be such a thing until I googled for "DC block sma cable". They seem to exist but are outrageously overpriced, for what is essentially an SMA cable with a $0.10 capacitor inline. Here's the most affordable one I found in the first 10 or so results:

http://www.ecmstockroom.com/process...ibute3=&itemAttribute4=&originSearch=DCBlocks

I really would bet there is a DC block on your reference cellular mobile. Is any part of the PCB exposed? Do you have a picture of it? Often most of the circuit will be under a shield but the RF will come out, there will be some antenna match or filter components, then an antenna or RF port. Can you see any of this stuff? Can you ohm across any of it?

 

Hi Steve,

Wow I didnt know such things exist.
It is indeed over priced.
Thanks for sharing it!

I test my cellular chip vs many different references, it changes all the time according to the test requirements.
For example, Samsung Galaxy-S I/II / Galaxy-ACE, Ericsson, etc.
Their antenna port (where I connect my RF Cable) is very small and it's hard measuring through it.

What do you think on PIN Diode RF Switches?