CC2538-CC2592EM-RD: CC2538+CC2592 Custom Board does not transmit with modified sampleswitch code.

Its gets power from VDD_EM pin available at rf06 EB right now. I have tried powering it from the same pin when EB is powered from a wall charger rated at 1.5amps,5v. The system behaves same in bothcases

Following is the screenshot of what I am testing . While testing for range, I plug 2530 based end point connected via ccdebugger to one laptop and cc2538 based coordinator connected via rf06 on other laptop. I then move away with 2530(and 1st laptop) from 2nd laptop(2538+2592). And PER and BER starts increasing beyond 8 mtrs. I hope I have made myself clear now. 

Please note that I follow processors.wiki.ti.com/.../LNA_with_Z-Stack-Home link to set transmit power by using following command 

// Set the transmit power level
ZMacSetTransmitPower(TX_PWR_PLUS_19); 

just below following section of code:

void zclSampleSw_Init( byte task_id )
{
  zclSampleSw_TaskID = task_id;

..
..

  
#if defined (OTA_CLIENT) && (OTA_CLIENT == TRUE)
  // Register for callback events from the ZCL OTA
  zclOTA_Register(zclSampleSw_TaskID);
#endif

I note that this is valid for Z-Stack-Home-1.2.x release . My folder name is Z-Stack Home 1.2.2a.44539 so I believe this is the right way of setting tx power.

No. Its not a dk it is also a custom board with cc2530 and without any cc259x part. Can you please share your screenshot and tell distance between the CB and EP? Does your system has 2530+259x and 2538+2592? Is the receiver in hgm enable mode?

CC2530 doesn't connect to CC259x so there is no HGM enable on receiver. The distance is about 30cm.

Oh . Sorry , EP means End Point. it is running SampleLight application. CB means coordinator board which is running SampleSwitch Application. Anyway, following is the result of 2530EM+rf05 (replaced my custom 2530 board with cc debugger) :

Does it mean the 2538+2592 HW is at fault FOR SURE? (Please note that PA_EN and LNA_EN toggle perfectly fine (seen on DSO) when used with SmartRF Studio.)

Following is the test result with custom 2538+2592 and custom 2530 board:

Following is the test result with custom 2538+2592 board and 2530EM+rf05:

What can be inferred by the difference in RSSI values in the two scenarios?

I just got reminded that the custom boards actually worked few days ago. The modified codes were running on the custom boards and when the 2538 side sent the data to end point, I could see the data at end point (2530) by breaking the debug instance. If I remember right, the communication was happening flawlessly during that test. My memory is bit weak. Luckily i captured both debug sessions that day. I will share the pictures if you want.

I get avg RSSI of -39dbm at 2530EM at 30cms , tx power 7dbm at cc2538EM . What can be the issue here? I have replicated the test you did yesterday. Following is the screen shot:

<- this is my custom 2538+2592 board.

And this is my test setup identical to yours->  .

Please share your test setup picture.

I do so . Anyway, now my application code is able to transmit and toggle lna and pa pins properly. (The issue was incorrect values of OBSSEL and CCTEST registers)

What is weird is that LNA pin stays at 0 and goes to -3.24v at the time of transmission. PA pin stays 0 and goes to +3.24 volts while transmission is happening. Can you explain why it could be happening?

Okay. So the problem was with DSO configuration, even though it was set to work in DC coupling mode, it was still in AC coupling mode . Changing setting again and again resolved the issue.

Coming to the RF performance issue there are few things that i noted in the AN130 and cc2592 datasheet:

1. The track length between RF pins of 2538 and 2592 should be as per the reference design of 2538+2592. In our case, the length is not same and tracks are not straight. Can this be a problem?

2. When we cut the green wires, that are LNA and PA enable pins, the system still works but at a very small distance (less than a meter) when these wires are soldered back to the corresponding pins of 2592, the range increases upto 3 meters. What does this mean suggests?\

We did not replicate the reference design of 2538+2592 emk. We dont have the same antenna design too. Have you designed any custom boards with these two devices?How did you do it ?

And lastly, What according to you would be a better idea?:

1. To edit the reference design and re-use as many components as we can and get it fabricated ? or

2. To tweak our present design and try resolving issues in the hardware?