CC1190: CC1190EM evaluation module - impedance match LNA / PA modes

- The LNA has a optimal source impedance that covers a fairly large area of the smith chart. The optimal load impedance for the PA covers a relative small area in the smith chart.

- What is important is the load impedance (TX) and the source impedance (RX). Even for an LNA it's not given that the optimal source impedance is a conjugated match of the impedance seen into the LNA input. Put differently, even if you see a given impedance in the match is designed to give optimal performance with 50 ohm in/ out on the antenna connector/ interface to a RF chip (for the EM PA_IN/ LNA_OUT).

- The input impedance will be dependent on the output impedance and the other way around due to how the design is done. Therefore the performance/ BOM is to some extent dependent on what you use in combination with CC1190. Therefore I would also look a the reference designs we have with CC1190 + CC1190 and CC1310 + CC1190.  

Do you get a different sensitivity than you expect since you ask? 

Hi TER,

So essentially I'm not looking at re-tuning the source impedance of the LNA, but the end result that I would like to see is that in PA output mode (regardless of the load impedance, but let's say it's close to 50 Ohms) the output power that I'm getting is over 26,5dBm ouput power when switching on a carrier at full power (that is with about 7-8dBm on the PA input). This I can confirm easily as I can measure (and have done so) with a spectrum analyser.

The source impedance on the LNA, well as I guessed it's a little more tricky, but again I'd prefer it to be closer to 50 Ohms when I look into the SMA. From what it sound like, is that you say the end goal is best to take as the sensitivity. So do you suggest that I should take the sensitivity measured as an optimising goal rather ... This will be tricky to tune it this way as it's not directly measurable with an instrument like a VNA etc... whereas I can measure the output power easily to see if I get expected output power close to 27dBm.

The SMA load will be mainly a dipole antenna like the ones from Nearson C463AM-860 and the C463AM-925 which are specified as 50 Ohms nominal with SWR < 2:1. Then there are some 1/4-wave monopoles from Linx  and/or SamWoo and some specific internal antennas for which we may require a specific LC tuning circuit to get closer to a 50 Ohms than it would be otherwise with nothing ....

Regards, Mike

Ideally all designs that use CC1190 should be measured using source- and load pull where the sensitivity and output power + harmonics are measured as a function of loads that cover the wanted section of the smith chart. For designs that follows the reference design this is not required but since you deviate from the stack-up used in the reference design this applied to you.

Would you be able to get hold of a line stretcher for a few days and see how the performance change depending on where you are in the smith chart? 

For RX, I would use the sensitivity as an indication. We have typically seen 3 dBm improvement (+ 3 dB loss in the SAW filter) compared to the datasheet numbers.

For TX: I would also check the harmonics and the power consumption since these change depending on where you are in the smith chart. 

Hi,

Unfortunately we won't have the time to do these type of measurements and accessibility to a suitable variable delay is not guaranteed in any case.

Given the situation, the tuning aspect is really optional for the moment and its going to be more practical to check harmonics for PA mode and RSSI level for Rx mode which are more readily measurable, and then make relative comparisons with previous PCBs or an evaluation card. Problem of having these non 50 Ohms optimal impedance for Rx and Tx modes, is that adding an extra length of coaxial to the SMA connector on the CC1190EM will move these points around the SC quite a bit... which was my initial motivation for looking at getting both modes closer to a 50 Ohm even though this isn't the most optimal from a datasheet performance point of view … We have some antennas with pig-tailed 50 Ohm coaxials with a pre-cut length already planned for use ...

BTW is the CC1190EM that is mentioned in the datasheet still an active "part" or is it no longer available ?

Regards,

MM

The CC1190EM was used when CC1190 was new to verify the part but it has not been available for years due to some of the reasons listed in my last post. 

As said, I believe that you should treat the input and output (antenna connector) as 50 ohm point. At least this is possible on the reference design. Since you had issues with the SMA connector impedance it's something strange going on with your board since the antenna output should be a 50 ohm connection and if you add a 50 ohm connection to this it should not affect the performance. When we do testing on CC1190 based boards we add 1m + coax to do measurements without issues.