Tuning an Existing Antenna

Doug had an ISM transceiver whose performance he wanted to improve. The analysis started with simulations of the current design and baseline measurements in order to see if there was anything reasonable that we could do. (Sometimes it just isn’t feasible to try and revise an existing product due to form factor or power requirements, and sometimes the desired performance is simply not reasonable.) We saw a couple of ways the product might be improved and decided on a minimally invasive strategy to try to optimize the existing product’s performance at a minimum cost. In other words, we were going to try to optimize the componentry and avoid a new “board spin.”

This was an interesting project because Doug had followed the transceiver chip manufacturer’s reference design pretty closely, and yet the transceiver did not perform as well as it should have in the field. Several units of this particular product had already been installed and most of them worked OK; but some installations didn’t link up reliably. Doug was hoping for a solution that would enable him to swap out the problematic units without triggering customer acceptance issues of the non-problematic ones. This presented a two-fold problem: Could we sweeten the performance significantly but without changing the external appearance of the equipment?

Simulation data showed that we could…

pattern

It’s not often the case that one can simply change some board components and realize an improvement in the antenna pattern, but it worked out in our favor here. Even if we were not able to improve the input match (the thing we were initially hoping to do) this transceiver will definitely have less “dead spots” than it originally had. In addition, we were able to dramatically improve the input match to the antenna, thus allowing more of the received signal to be transformed into information. We are now making the RF do what is expected of it…

rl

Measured data showed that we did!

ctl_dn

Above is a spectrum analyzer screenshot of the signal strength before optimization…

match_dn2

…and here is a spectrum analyzer screen shot of the signal strength after optimization.  On average, there was about 8 dB more available signal after optimization!

A month or so later, we got some feedback from Doug, who wrote:

Mark,

Today we did actual field distance testing [using] a modified and un-modified [base unit].

The test was free-field and nearly ideal. The transmitter was held at height above ground of about 6 feet. The base units were then moved away from the transmitter in increments. The [Semtech] XE1203 transceiver chip in the base unit has very coarse resolution (2 to 9) and has significant receive capability well below an RSSI reading of 2. Both base units were held in the upright orientation facing the transmitter.

Distance un-mod signal new-mod signal
0.0 Miles ok 9/9 ok 9/9
0.1 Miles ok 3/9 ok 5/9
0.2 Miles ok 2/9 ok 4/9
0.3 Miles ok 2/9 ok 3/9
0.4 Miles No Packet ok 3/9
0.5 Miles No Packet ok 3/9
0.6 Miles No Packet ok 2/9
0.7 Miles No Packet ok 2/9
0.8 Miles No Packet ok 2/9
0.9 Miles No Packet No Packet

Just as expected, with the antenna performance improved by 8.2dB, the reception range a little more than doubled.
The eating has proven the pudding!

Based on the success we have had I consider the base unit antenna improvement project complete. Once again, nice work!

Doug

Share Button
Comments are now closed for this article.