nr51822: No radio with custom PCB

RE: nr51822: No radio with custom PCB

Asbjørn — Mon, 16 Feb 2015 10:14:58 GMT

This look like a good improvement with regard to ground. It look like a really small board, but to me it seems like you have added the amount of ground plane on both top and bottom layer that you could. Because of its small size it wouldn't be the best performing board, but you should be able to get some OK range with this one nevertheless.

You should move the trace that runs under the L1 component. It seems to be a button (BTN2 or 3?), but it could be something else that's powered. Avoid routing signal traces underneath the trace from the balun towards the antenna. Would it be possible to use a gpio on the opposite side of the chip to avoid this conflict?

RE: nr51822: No radio with custom PCB

Locky — Sun, 15 Feb 2015 23:12:59 GMT

You still will want to keep the capacitor to tune the Antenna to your PCB. Its best to done with a full pi-network.

RE: nr51822: No radio with custom PCB

Henry — Sun, 15 Feb 2015 10:28:06 GMT

Dear Asbjørn,

Thanks for your help. I redesigned the PCB, please see the updated version above. Do you think that this should be sufficient?

I am still not so confident due to the fact that even on very close distance, I was not able to get any signal. But as you said, maybe the lack of GND plane did create so much interference that the signal was killed completely.

Do you see at first sight anything wrong in my design?

In regards to the balun, I use the BALF-NRF01D3 version (instead of BAL-...). This should be compliant with the NRF51822 QFAA G2 chip I use. In the Balun datasheet (page 5) it says that no capacitor is needed (in contrast to the BAL-.. version). Does this mean I can leave out the capacitor from the PI network?

Thanks Henry

PS: Dont think my oscilloscope is capable of measuring 16MHz

RE: nr51822: No radio with custom PCB

Asbjørn — Wed, 11 Feb 2015 17:15:16 GMT

The lack of ground plane may be the reason for why you don't see any RF output as the antenna would suffer. Normally you would be able to see at least a very weak signal if you have the phone and board touching each other, but not necessarily. It's good that you see the 1.8 V, that indicates that the chip at least is trying to transmit. Regarding the startup of the external crystal and measurement, is the oscilloscope able to measure up to 16 MHz?

RE: nr51822: No radio with custom PCB

Henry — Sat, 07 Feb 2015 09:45:02 GMT

Dear Asbjørn,

Thanks for your comment.

You are right, the GND pane is indeed missing and I will add this to my next design for sure. Could this mean that the antenna does not work at all? I tried to get as close as possible with both, phone and sniffer probe but did not receive any signal.

I have check with the oscilloscope and was able to identify the 1.8 V peeks. So the chip does seem to transmit.

In regards to the crystals: I was able to see the LF 32kHz crystal signal. But the 16MHz crystal I was not able to see anything on my oscilloscope. But also on the working Nordic Smart Beacon Board I was not able to measure the HF crystal... Can this mean that my pocket-oscilloscope is not sufficient?

RE: nr51822: No radio with custom PCB

Asbjørn — Thu, 05 Feb 2015 11:15:19 GMT

It's difficult to tell from your screen shot, but it looks like the ground plane is missing, at least on the top top layer. Do you have a ground plane incorporated into your board? The antenna will not work as intended if there's no ground plane no the board.

Secondly, if the ground plane is insufficient, you should still get a couple of inches of range. Could you test by holding the phone very close/on top/under the beacon and see if you get any readings at all out of it.

I've recommended this as a debug option for many customer, it's to check the VDD_PA pin on the board. It will go up to 1.8 V when the chip is actually transmitting. So you should see that VDD_PA goes up to 1.8 V with an oscilloscope for every time you expect the beacon to send a package.

You should also verify that the external 32/16MHz crystal is actually starting up.

RE: nr51822: No radio with custom PCB

Henry — Wed, 04 Feb 2015 20:26:58 GMT

Hi Audun, Thanks for your comment. Sorry, my bad. I have the mandatory X1: 16MHz and the optional LF X2 with 32.768kHz. In order to see wether the soldering of X2 might be the problem, I did change the FW to use the internal clock and the result was the same, although you could see with the oscilloscope that X2 was not active in that case. So the LF crystal seems to operate fine.

RE: nr51822: No radio with custom PCB

Audun — Wed, 04 Feb 2015 18:56:37 GMT

Hi Henry. To clarify: did you mean to write 32 kHz crystal or 32 MHz crystal? The nRF51822 requires a 16 MHz or 32 MHz crystal to operate the radio. A 32 kHz crystal is optional. In the case you're using a 32 MHz crystal this needs to be configured in firmware, because the default behavior is to assume the high frequency crystal is 16 MHz.

RE: nr51822: No radio with custom PCB

Henry — Wed, 04 Feb 2015 13:44:38 GMT

Here a picture of how the balun is placed.

Please let me know your comments

Thanks

EDIT: Attached please find a new version, this time with GND plane. Please let me know your comments, if this should be sufficient.

Thanks, Henry