Missing (some) wiimote connection attempts - is it improper page scan configuration?

I have a basic Bluetooth stack that is able to interface pretty well with a wiimote. I've noted that (after pairing, that is the wiimote is bonded to my host controller), a connection is not established every time a wiimote button is pushed. Monitoring my host controller, it looks like the page scan is not "seeing" the wiimote paging to establish a connection (there is nothing at all sent from the host controller to my local host). A connection is eventually established by pushing a wiimote button another time (or two). It seems to be about a 50% connect success rate, with no particular pattern - which wiimote button is pushed doesn't make a difference.

I'm wondering if this is due to improper setting of Page Scan Interval, Page Scan Window, or Page Scan Mode on my host controller? I'm using PSM R1, which is what I saw the wiimote announcing in a request. I could try to set the host controller page scan to continuous, but wouldn't that leave no bandwidth for the data communication?

Thanks.



Edited 1 time(s). Last edit at 07/16/2013 08:53PM by ags000.

I have confirmed that my host BT controller is never sending an event to my host when a wiimote connect attempt is missed. I put a scope on the HCI TX line to be sure there wasn't something wrong with my UART sampling that line. I don't have any way to monitor the actual radio, but this does indicate that my BT controller is never reporting any activity (which I extrapolate to mean it is never hearing the wiimote page)

My local BT controller was using default page scan settings. I modified the page scan window, doubling the default (from 11.25msec to 22.50msec) and no longer observe any "missed" wiimote connection attempts. Unfortunately, without being certain of the cause, this is an empirical solution only - until I see another missed connection.

Still hoping someone else has seen this or can explain or point to other ways to confirm my suspicion and/or that my solution is robust.

The wii uses page scan type 1 (interlaced scan) which performs two scans (each of the duration of the scan window) back to back.

1) My problem was missing connect attempts initiated by the Wiimote. So if I understand correctly, in that scenario the Wiimote is paging, and my local controller is page scanning. Does the page scan type used by the Wiimote (specifying how the Wiimote behaves when it is page scanning, i.e. when it is made discoverable by pushing the sync button under the battery cover) have implications on how it pages, and how my local controller's page scan type should be configured?

2) Likewise, is the correct approach to double the default page scan window (making it now as long as the interlaced (two back-to-back) scans in page scan type 1) or is the proper approach to use the default page scan window but specify page scan type 1?

3) Also: I'm far from an expert, but have been studying the BT spec quite a bit. I may have missed it, but don't see any way to directly query a remote controller, or any event resulting from a remote command that includes the remote controller page scan type. Is this something that requires specialized diagnostic equipment to sniff from the radio transmission itself?

Thank you.