ANT Forum

Greetings

I have an NRF51422 QFAAEO running S210 Soft Device version 3.0.

I have written a Bootloader application that uses Ant-FS as the upload mechanism. During testing we noticed a strange behavior reproducible across multiple hosts / client combinations that we cannot explain. Hopefully you guys can explain if this is bug or feature.

We are using the Ant-FS reference Host project for development. When I run this application, It reports itself as "SW Version: AML1.300, Library version: AKX0.309 ". We are using Ant USB sticks of 3 different manufacture so don't believe that is a factor.

On the host application, when I select "Options" and then select the "Ant Channel" tab, I have a group of options labeled "RF Frequency" with text box fields for "Search Frequency" and "Connect Frequency" as well as a check box labeled "Random Connect Frequency". The value of these options plays a deterministic role in whether the host can successfully upload to the client.

If the checkbox is unchecked and the search frequency is the same as the connect frequency (the default if it has never been changed), our upload test of uploading a 42KB bin file in bursts of 1KB fails roughly 90% of the time.

If those two frequencies are far apart, for example 55 and 99, the exact same test succeeds 95% of the time. If I bring the two frequencies closer and closer together, the failure rate gets proportionately higher.

Examining the logs from the host for the failed attempts shows the client is simply not responding to multiple Ant-FS messages from the host.

Why would sharing a frequency between the "connect" and "search" fields cause interference? There are no other Ant devices running when we conduct these tests and the range of the client to host is just a few inches.

Is this bug or feature?

I am cross posting this to the Nordic Semiconductor Devzone pages at
Link to same post on Nordic Devzone

Any help or insight appreciated.
David
     

Hi David,

The ANT-FS spec states that the host *shall* request a change in frequency in the Link Command. If this change is not requested then I would not be surprised if the client simply ignores the link command.

Have you graphed your results? Is it a linear proportional relationship over the whole range you've tested or do you see a step or flat line anywhere?

Thanks,

Kat      

Hi again!

A few other suggestions from the team here:

You could try testing in a different environment and see if the results change. There could be interference from wifi, or other ANT devices. Generally you should avoid 2457MHz, as this is the ANT+ frequency, and wifi channels 1, 6, 11 may also present an issue.

Also, not sure about your hardware setup, but your antenna may need tuning. Nordic would be the ones to support you with this.

Thanks,

Kat
     

Have not tried the detailed testing on graphing failure rate vs. frequency. We had been using 55 for a frequency so it is possible we are getting wifi interference.

You've given me some more steps to try but now that we have a stable workaround I don't know if the customer will pay for more investigation. I will try to find time to check out if it is frequency separation that is the trick or absolute frequency near wifi that was the root cause. That should be an easy test.

Thank you for the reply.

David