1
\$\begingroup\$

Considering that the SDR board has one TX/RX output/input and one RX input, I am wondering how could I use one antenna for a half duplex communications on the same frequency.

The idea is to use the TX/RX output for transmitting and the RX input for receiving. The problem is that the required output power is around 47dBm (which is achieved after a HPA) and if I use a fairly good circulator (for joining Rx with TX) as this one, it gives a 20dB isolation and that is not enough as it would burn the whole thing with the reflected power (max. input power of the SDR is -15dBm after the LNA!). Scheme: scheme

As an alternative, I have thought of using a electronically controllable switch that could open one path while closing the other. The problem with this option is that so far I have only found very expensive solutions and my budget is limited.

Can anyone give me some guides on building a PC controllable switch for this purpose (should withstand around 50W of power while offering the typical 50\Ohms of impedance)?

Otherwise, any different ideas/approaches on how could I solve the problem?

\$\endgroup\$
  • \$\begingroup\$ Most standalone RF modules are expensive unfortunately, especially for high power handling. Depending what your link budget looks like you might consider adding an attenuator on the RX side before the LNA to absorb some of the reflected TX power. You could also look into a reflected power cancellation circuits. \$\endgroup\$ – Captainj2001 Nov 13 '16 at 1:21
1
\$\begingroup\$

As mentioned by Sam, if you're doing half-duplex only, your problem is a lot easier, and you don't actually need a circulator; just a set of switches.

You don't mention which SDR device you're using, but a lot of devices, including the USRP series of SDR peripherals, have GPIOs for exactly that purpose: external switch Notice that at least in the case of USRPs (and if I judge by your nomenclature, you might have one of these), the SDR device itself supports switching the GPIO state based on the transmit/receive settings – i.e. you can automatically switch externally when disabling the transmit stream.

Please make sure that post-amplifier crosstalk doesn't fry your RX electronics – for example, the frontends available for USRPs are rather sensitive (ie. NF <<10dB, sensitivities below thermal noise level), and thus can only withstand free of damage maximum inputs of typically -15dBm. Hence, switching off the HPA before turning on the RX LNA might be a good idea – again, GPIOs might be helpful here; the modern USRPs support timed commands for addressing such operations in a hardware-timed manner.

\$\endgroup\$
  • \$\begingroup\$ Thanks for the fast answer @Marcus Müller! This is exactly my case. I am using the USRP B210. And I was thinking in using this HPA. I believe that it is a passive component. When you suggested to turn it off before switching to receive mode, did you mean, from the URSP TX channel? I think that I do not fully get it. So if you are just in transmit mode, can't the crosstalk fry the RX channel anyway? Alos, do you have any exampe of a switch being able to handle around 50W, and also offering the GPIO control? (I have not found one yet!) \$\endgroup\$ – B.N Nov 21 '16 at 20:21
  • \$\begingroup\$ A HPA (==high power amplifier) is by definition of the word "amplifier" not a passive component! \$\endgroup\$ – Marcus Müller Nov 21 '16 at 20:22
  • \$\begingroup\$ ohh right! Obviously, I am not an expert. I thought that an active element is one that can be controlled. My bad! \$\endgroup\$ – B.N Nov 21 '16 at 20:25
1
\$\begingroup\$

You might want to change your title as it implies that you want simultaneous transmission and reception, half-duplex is way easier to achieve.

Now, even for half-duplex, 47dBm TX power with a -15dBm RX limit is a difference of >60dB (or one million to one power ratio) so what ever solution you use needs good isolation.

With a good PIN diode switch and with the LNA powered down while the transmitter is operational (and real good shielding, but that goes without saying), that'll give you a reasonable amount of isolation between TX and RX.

Failing that, a good RF relay, but all this assumes half-duplex, if you decide you want full-duplex then that's a whole other can-o-worms!

\$\endgroup\$
0
\$\begingroup\$

Well, RF switches for 50+ watts are $200 and up. How fast do you need to switch from receive to transmit? Relays to switch this power will take ~ 40 to 80 mSec to switch, and you need to switch them without RF power, so you'll need to switch off the amplifier first. PiN switchs aren't out of the question, depending on the frequency of operation. and could be homebrewed with some work.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.