I compared some Skyworks FM evaluation tuners that I recently obtained to some TEF6686 radios that I have. That prompted me to obtain a TEF6687 radio that some claim improve the RDS sensitivity over the 6686. You can identify the IC by the codes on them: F8602 for the TEF6686, F8605 – TEF6686A and F8705 for the TEF6687. The difference can be seen below. RSPduo and Airspy HF+ SDR radios were also evaluated.
I removed a shield covering the TEF6686. The TEF6687 looked like it was installed on a board where a 6686 was removed. Excess flux had to be removed from on top of that IC before its code could be seen.
The measured spec comparison is below. I started with just measuring the minimum dBuV input level required for RDS decoding at different injection levels. Since that depends on the radio's sensitivity as well as its decoding ability, I added the minimum RDS injection level required for decoding a moderate 23 dBuV signal.
Whether the TEF6687 was better or not depends on which TEF6686 you are comparing it to. Based on the radio that you have you could call it a little worse, about the same or a bit better. The SDR radios had impressive figures over their narrow dynamic range. The manufacturers make the dynamic range seem large by moving that narrow window around a much larger claimed bandwidth. Those suffer when trying to receive a weak station close to a strong signal. That will be shown in the coming adjacent frequency selectivity figures that I need to complete.
While the NXP specification sheets do not claim increased RDS sensitivity for the TEF6687, they do promote improved stereo separation as part of the 6687's Stereo Improvement algorithm. I did not attempt to document that improvement.
I removed a shield covering the TEF6686. The TEF6687 looked like it was installed on a board where a 6686 was removed. Excess flux had to be removed from on top of that IC before its code could be seen.
The measured spec comparison is below. I started with just measuring the minimum dBuV input level required for RDS decoding at different injection levels. Since that depends on the radio's sensitivity as well as its decoding ability, I added the minimum RDS injection level required for decoding a moderate 23 dBuV signal.
Whether the TEF6687 was better or not depends on which TEF6686 you are comparing it to. Based on the radio that you have you could call it a little worse, about the same or a bit better. The SDR radios had impressive figures over their narrow dynamic range. The manufacturers make the dynamic range seem large by moving that narrow window around a much larger claimed bandwidth. Those suffer when trying to receive a weak station close to a strong signal. That will be shown in the coming adjacent frequency selectivity figures that I need to complete.
While the NXP specification sheets do not claim increased RDS sensitivity for the TEF6687, they do promote improved stereo separation as part of the 6687's Stereo Improvement algorithm. I did not attempt to document that improvement.