range of custom pcb

[karthikeyan]

Hi,

  My nrf 52840 pdk and my custom pcb with nrf 52840 soc. both are running with same application ble_proximity example from sdk 14.1 and softdevice s140 5.0. when i set the Tx_power to 4dbm. i am expecting the range of custom pcb should be equal to my pdk. but the range of custom pcb is very low when compared to pdk. how can i achieve the range of custom board equal to pdk?

thanks and regards,

karthikeyan

[FormerMember]

A typical mistake that leads to shorter range is wrong value of the load capacitors for the 32 MHz crystal. The value of the load capacitors should be calculated according to the formula in the "Crystal and load capacitors" section in this tutorial.

Wrong values of the load capacitors leads to an offset of the crystal frequency, and an offset of the carrier frequency.

Could you check if this is the problem on your board?

[karthikeyan]

hi kristin,

  after several reviews of custom pcb and BOM from the help of nordic team. 32 MHz crystal is 

ABM11-32.000MHZ-D2X-T3 CRYSTAL, 32MHZ, 10PF, 2 X 1.6MM

according to 10pf crystal. the load capacitor is choosed

15pf. so i dont think the load capacitor will be problem.

regards

karthikeyan

[karthikeyan]

hi kristin,

    already there was a private question with the heading "custom board programming fails"  which i closed on last friday. there i have attached the gerber file and BOM. if you cant check it. i will upload gerber. but how can i make the case private?

thanks and regards,

karthikeyan

[karthikeyan]

hi kristin,

  any update?

regards,

karthikeyan

[FormerMember]

What is the case number on the other case?

[karthikeyan]

hi kristin,

   this is the Case ID: 206489

[FormerMember]

I suspect the range problem could be related to the 32 MHz crystal, because C8 is placed far from the crystal.

Do you have a spectrum analyzer available? If so, could you use the spectrum analyzer to measure the output frequency? (The radio_test example can be used for testing.)

J1: I assume that a switch is placed on J1, so that the output of the radio is connected to the antenna?

[karthikeyan]

hi kristin,

   i assume that you read all the comments in that previous ques. so i cant go for a another pcb design and prototype. want to make changes in this itself rather than going for a new design.

Unknown said:

Do you have a spectrum analyzer available? If so, could you use the spectrum analyzer to measure the output frequency? (The radio_test example can be used for testing.)

sorry i am not having the spectrum analyzer.

Unknown said:

J1: I assume that a switch is placed on J1, so that the output of the radio is connected to the antenna?

thanks and regards

karthikeyan

[FormerMember]

After looking at the PCB, I think the problem is related to J1. J1 is a connector like this?

As you can see on p. 3 in that datasheet, there is no connection through the connector, it means that the antenna is not connected to the radio. Could you remove J1, scratch away the connections to ground where the the connector is placed, and solder so that the antenna is connected to the radio? And then test again?

[AmbystomaLabs]

As Kristin pointed out, you placed the wrong connector at J1.  Your micro-coax connector has a ground connection where your antenna is hooked up. I believe you thought you had bought the testport connector.  They look similar.

You can either pull J1 and bypass or just place the correct test port connector. Or, hook up an antenna via micro-coax.

Here is a link to the MuRata testport connector that most people use.  You have to use a special probe with it.  You cannot use a standard micro-coax connector.

https://www.digikey.com/product-detail/en/murata-electronics-north-america/MM8130-2600RA2/490-4981-1-ND/1775925

[karthikeyan]

thanks both kristin and ambystomalabs i will test with both the ways and let u know.

regards

karthikeyan

[AmbystomaLabs]

Assuming the artwork under J1 supports it, a good bypass will be a 6-20pF chip cap probably 0805 or 1206 package size. It won't be perfect but it should work good enough and let you continue work if you don't have the testport connector at hand.

Or a short piece of wire will work too.