CC430 and CC1190 not enough power with SimpliciTI

The following appnote is also valid for CC430: http://www.ti.com/lit/an/swra151a/swra151a.pdf. The PA on CC430 is the same as for CC1101. The output power is dependent on if wirebound or multilayer inductors are used.

Have you measured output power conducted on your CC430+CC1101 board? If you want us to try to find out why the output power is poor on your board you need to upload at least a schematic. The layout of the RF part would also be good. If you don't want to make these public send me a friend request and you will be able to send them to me directly.

8666.PDF TCR1_V01_1.rar

Hi TER,

You'll find board and schematics in the rar file. I've tried to copy and merged some EVBoard from Texas : CC1101+CC1190 and CC430.

Inductors are multilayer Ceramic, excepted L2 (11nH) and L24 (4,1nH) which I find only in wirebound. Is it a problem?

I'll be back with measured output power, but my first test (difference between RFstudio exported settings for SimpliciTI) seems to show that different output power settings on RFstudio make same output power from my board.. ?? Even when exporting PATABLE.

Thanks for your help.

Hubert

5023.130826_01 - tests de puissance d'emission TCR - V01_2.xls

Here are my test result.

Some explanations :

Software : sending 2 message, first with CC1190 OFF (PA=HGM=0), second with CC1190 ON (PA=HGM=1). It's why there is 2 power measurement by line. The most important one is when CC1190 is ON.

Test : Measurement is done by a CC1110 EVBoard as a sniffing device (Smart RF Packet sniffer). My sending device is in the next room, same place and same orientation of the antenna, each test.

Observation :

1) Output power is effectively better with my prototypes (MASP430+CC1101+CC1190, yellow line 4 to 6).

2) Change of PA_TABLE0 (in register view on the right) with RFStudio seems to change nothing on measured output power, even when exporting PA table (line 7 to 15).

3) Change of PA_TABLE0 (with selecting TX Power) with RFStudio... there's some little difference, but not sure.

4) Change seems to be done when modifying PA table directly in mrfi_radio.c (line 16, 17 and 28 to 31)

I don't really understand what is done by RFStudio...
only export register file (so TX_power is not exported)?
And when export PA Table, is the #define PA_TABLE used by SimpliciTI?
(note : I found a bug is RFStudio, when checking PA Table option, the exported file doesn't have the #define PA_TABLE. I had to had it to the file myself).

Regards,

Hubert

Do you have access to a spectrum analyzer to measure conducted output power? It is difficult to debug from random radiated tests. To see if one board is better than the other the same PA_TABLE values have to be used. Are you able to control the boards with Studio to avoid issues with SimpliciTI settings?

I'm not an expert on SimpliciTI but:

Looking in SimpliciTI Developers Notes, table 5: It is some limitations for exporting this value.

It looks to me that the PA_TABLE is set by the following:

static const uint8_t mrfiRFPowerTable[] =
{
  0x0D,
  0x34,
  0x8E
};

in the mrfi_radio.c in the family5 folder. I don't have a proper text editor on this PC so I was not able to compare line numbers.

I have forwarded the #define PA_TABLE in the exported file question to the Studio programmer.

Hi Hubert,

The PA_TABLE registers in SmartRF Studio is given an "*" at the end of the name. This is to indicate that it is not a normal register. In SmartRF Studio we call them virtual registers. The reason is, as you probably have figured out, that programming of the PA_TABLE is done in a different way. The "#define PA_TABLE" you can get if selecting that option is only to make it a bit easier. The "define" must be put on the right place and used by the code. I'm not familiar with the implementation of SimpliciTI, but the code snippet from TER indicate that it is different. The values exported by SmartRF Studio in the define statement must be copied and put in the definition of mrfiRFPowerTable[].
If you select 10 dBm and "ramping" in SmartRF Studio the exported PA_TABLE must be copied like this:

static const uint8_t mrfiRFPowerTable[] =
{
0x00,
0x03,
0x0f,
0x27,
0x8d,
0xcd,
0xc8,
0xc3
}

Problem is that the exported PA_TABLE can be used by several implementations and for the moment the exported "#define PA_TABLE" statement doesn't fit them all. Som copy and paste will be needed. A future enhancment of the export function would be to enable modifications of the template used for the PA_TABLE.

Regards,
Øyvind

Hi Øyvind,

It's what the yesterday and today tests make my realize...

I also realize why there was so much misunderstanding of test result all along my development.

A good improvement from TI would be to indicate more clearly that settings from RF Studio does only partially sets SimpliciTI settings. I've lost so much time without knowing this, and perhaps that my 433MHz board (without CC1190) with good TX power,  would be enough range for my application...

So, with good TX power I've found a few better range.

Then I found that there was a mistake in nomenclature (RF inductor L132 was 12nH instead of 18nH), so I've found another few better range.

But the range is still not as good as the prortype (based on the CC1101CC1190EMK868).

I also notice that my board has multilayer inductor and CC1101CC1190EMK868 board has wirebound inductor (L22, L121, L131, L132, L24, L23, L25, L21, L122). Do you think that changing them to wirebound would improve range ?

Regards and thanks for help,

Hubert

You could get about 2dB higher output power with wirebound inductors