Receiver Sensitivity improvement

Hello,

For 315 MHz operation I am using CC1100 Transceiver. This transceiver has -111 dBm sensitivity while using 1.2 kbaud and 58 kHz receive filter. I want to improve sensitivity of this receiver by placing an external LNA. Up to which levels can I improve this sensitivity? By the way I am planning to use SGL0363Z as external LNA which has 20 dB gain and 1.1 NF. (If it is correct, according to my calculations noise floor is -125 dBm for 58 kHz rx filter)

Many Thanks,

You have run into a common problem now-a-day. The problem is most of these modern communications chips no longer allow the user to access the IF frequency before demodulation. The newer ICs either use soleley digital filtering of the IF, or use a zero IF scheme. So...you are sending 1.2 kbaud, and should have at most a 5 khz IF bandpass filter, but are forced to use 58 KHz.

So what can you do?

Yes, you can put a LNA in front of the chip. Noise floor is -174 dBm/Hz +1.1 dB NF + 10 log 58,000 = -125.3 dBm. So you are not going to do any better than that, and probably not even get that due to other noise sources in the chip. But you will probably get better than -111 dBm. I would personally do the LNA and then a bandpass filter centered at 315 MHz between the LNA output and the chip....to limit out of band noise causing trouble. Also, you might have trouble if the transmitter gets too close, due to all that system gain.

You could change ICs. Melexis has some IC's that allow access to the IF frequency at 10.7 MHz. Throw a narrowband ceramic or crystal filter in there and it will help a lot (although you will need accurate transmit and receive LO frequencies!). Microchip has a reciever like that also.

Anybody know of other "modern" Reciever ICs that allow access to a 10.7 MHz IF for filtering?

Thanks,

Also I have another question,

I am also planning to use an RF switch to switch connection to single antenna from transmitter or receiver. Do you have any suggestions or idea about using a switch before LNA? Is it practical or are there better solutions? Circulators or duplexers are seem to be impractical.

Any switch (or loss) you put in front of the LNA (between the LNA and antenna) translates into noise figure. There's a number of switch manufacturers out there. Mechanical relays and PIN switches have the lowest loss. Fet switches have higher loss but are easier to work with. PIN switches have issues when working in the MHz. Mechanical switches are slow. Choose your poison.

As it is generally straightforward to achieve a noise figure below 5-dB, I have my doubts that your modification will effect more than about 6-dB improvement under the conditions of use envisaged by the chip designers - indeed, it is more than likely that the performance under those specific conditions will be degraded. Nevertheless, depending on where the chip designers have made their compromises, you may see improvement under your specific operating conditions; this depends largely on what interfering signals are received at your antenna and on the available recovery time (between transmission and reception).

ge's comment about switching loss is well-made - and you will have to redesign the matching arrangements.

In practice you may need to place a roofing filter both before and after the additional amplifier.