CC2540: CC2540 initial freq drift

Hank Zhang

Expert 2040 points

Part Number: CC2540

I found the following problems in the CMW270 test：

initial freq drift=30kHz,It is a bit high.What might be the reason.What happens if this parameter fails

over 7 years ago

0 leo chen over 7 years ago

Prodigy 235 points

hi，hank
I think it's related to the baseband algorithm of chips.

0 Albin Zhang over 7 years ago

TI__Expert 6110 points

Hank,

fine tune the frequency accuray might improve it. (tune the CL)

Layout can cause the problem. you can post you snapshot here, I can help to check.

it will cuase package loss.

BR. Albin

0 Hank Zhang over 7 years ago in reply to Albin Zhang

Expert 2040 points

PCB-20171010.pdf

Hi Albin

As the attachment

0 Albin Zhang over 7 years ago in reply to Hank Zhang

TI__Expert 6110 points

looks the de-cap is not close to the chip and crystal lines are also long.
the noise might be inject here.

please tune the frequency accuracy in 20k, to see of the drift can be better.
BR. AZ

0 Hank Zhang over 7 years ago in reply to Albin Zhang

Expert 2040 points

We tune the CL, the frequency accuray improve,but "initial freq drift" did not improve

0 Albin Zhang over 7 years ago in reply to Hank Zhang

TI__Expert 6110 points

what's the command you used?

0 Hank Zhang over 7 years ago in reply to Albin Zhang

Expert 2040 points

PTM Mode, download the CC2540_USBdongle_HostTestRelease_All.hex

0 Albin Zhang over 7 years ago in reply to Hank Zhang

TI__Expert 6110 points

what's the packet payload type and length you used?

0 Hank Zhang over 7 years ago in reply to Albin Zhang

Expert 2040 points

payload:10101010
length:37

0 Sverre over 7 years ago in reply to Hank Zhang

TI__Mastermind 39860 points

The issues is most likely related to VCO pulling when switching on the PA. The frequency will then be "off" before it settles towards the programmed RF frequency. Not worked with the CC2540, but I know that for newer BLE products we have a pilot tone prior to the preamble to mitigate the initial drift. Will check if this is the case also for CC2540 and if this tone (if present) can be extended.

0 Hank Zhang over 7 years ago in reply to Sverre

Expert 2040 points

1.What's the effect of the initial drift too high?This is the most important question
2.What was the result of the initial drift test at CC2540EM

0 Sverre over 7 years ago in reply to Sverre

TI__Mastermind 39860 points

Soy, but there is no pilot tone on the CC2540. Agree with Albin's suggestion that you should look a the routing from the crystal to the CC2540 crystal pins and the crystal decoupling capacitors.

0 Fredrik K over 7 years ago in reply to Hank Zhang

TI__Guru** 106705 points

Hank Zhang said:
1.What's the effect of the initial drift too high?This is the most important question

You will not be able to pass Bluetooth qualification. Violating the PHY requirements may also lead to packet drop and connection problems.

Hank Zhang said:
2.What was the result of the initial drift test at CC2540EM

All CC254x EVMs pass the BT PHY requirements.

Cheers,
Fredrik

0 Hank Zhang over 7 years ago in reply to Fredrik K

Expert 2040 points

Are there any relevant test reports？I didn't see the results of "initial Freq Drift " in the datasheet

0 Fredrik K over 7 years ago in reply to Hank Zhang

TI__Guru** 106705 points

Hi Hank,

As you can see here: processors.wiki.ti.com/.../How_to_Certify_your_Bluetooth_product we have a QDID for the Controller subsystem which includes PHY qualification.

Cheers,
Fredrik

0 Hank Zhang over 7 years ago in reply to Fredrik K

Expert 2040 points

Hi Fredrik
I changed another manufacturer's crystal after testing the phenomenon as follows：
initial freq drift
At first, “the initial freq drift ” was out of standard. After waiting for 3 minutes,"the initial freq drift" returned to normal

0 Fredrik K over 7 years ago in reply to Hank Zhang

TI__Guru** 106705 points

Hi Hank,

As discussed, there are some layout issues in your design which you should address.

Cheers,
Fredrik

0 leo chen over 7 years ago in reply to Sverre

Prodigy 235 points

I would like to ask you to understand the general debugging scheme of CC2540. This problem has not been solved up to now. Thank you

0 Fredrik K over 7 years ago in reply to leo chen

TI__Guru** 106705 points

Hi Leo,

As has been discussed already, your PCB design has weaknesses which is most likely causing frequency pulling of the internal VCO. I would recommend you to redesign the PCB. Pay close attention to the decoupling caps making sure these are close to the supply pins as well as having short and direct ground return paths.

Cheers,
Fredrik

Bluetooth®︎

Bluetooth forum

CC2540: CC2540 initial freq drift