CC3100: How to interpret the calibration file, phy.cal. What parameters, and their significance, are associated with XMT and RCV?

Hi Mike,

The calibration file is for use by the NWP only. Do not write to it. We don't publish how to interpret this since it is not useful to the developer

-Aaron

Aaron,

I asked about the cal file because of the problem I've already posted.

CC3100: CC3100: Successful Radio Tool CW XMT Testing / Failure for Radio Tool RX Testing

I thought the calibration file may shed light on what's going on with the extremely low, almost no signal detected, RSS readings reported by the Radio Tool.

We connected the CC3100BOOST board to a spectrum analyzer and observed a transmitted packet.

We repeated using our two prototype boards and observed similar wave forms.

It appears that the transmit function is working correctly.

I reviewed the prototype board artwork and compared it against the CC3100BOOST board. The RF trace CPW, including the GND via stitching, is a duplicate of the CC3100BOOST artwork, including stack up.


The one difference I found is the number of GND vias used in the DC-DC area near the L3, L4, and L5. The CC3100BOOST board has about 30 GND vias in the area just to the right and below silk screen marking 012.

We do not have that density of GND vias.

In hindsight I should have added more vias. I will do so on the next board spin.

Any theories on the impact to RSS for not having a sufficient number of GND vias in the DC-DC area?

I would have thought that an insufficient number of GND vias would impact the transmit function more than the receive function.

I cannot come up with a reason why transmit works but receive does not.

For the next board spin, what is the protocol for requesting TI to do an artwork check?

Thanks.

Mike

If you haven't already please review the design checklist: processors.wiki.ti.com/.../CC31xx_&_CC32xx_Hardware_Design_Review

I believe I found the problem, the 40 MHz clock source.

Even though I read the CC3100 data sheet requirement for using a sine wave or clipped sine wave source (1.2 Vpp), out of habit I used a 40 MHz square wave HCMOS clock oscillator.

I will change to a 40 MHz crystal.


RTC source is a 32.768 kHz square wave HCMOS clock oscillator.


I assume the CC3100 internal synthesizer / PLL uses the 40 MHz to produce the CW signals for the XMT/RCV mixers.

When using the radio tool we see CW tones and Continuous results on the spectrum analyzer.

The CW tones are at the correct frequency, as set by the radio tool, and the dBm levels agree with the CC3100BOOST board.

Same observations for Continuous mode.

Assuming the 40 MHz is the problem, I do not understand why the signals look good on the spectrum analyzer but fail to go out over the air.

Same with respect to RX mode.

Can you provide some details on the synthesizer / PLL function and how it would react to the 40 MHz square wave harmonics?


Thank you.