get rssi of peripheral device when DCENTRAL_LINK_COUNT=2

[enaon]

I got some of those cheap itags off aliexpress, less than a euro each, and they work very nice with espruino, they can be used as a remote button.
https://www.aliexpress.com/item/1005009223485666.html

my aim is to use them on the cats, I want to know if they are on the garden or they are out for a walk, so I am making an an app for that. My problem is, I have two central connections enabled, and I think this may be the reason I cannot get the rssi when connected to one of the tags. It this a fact? Is there a way to get the rssi of a peripheral, if I have build the firmware with DCENTRAL_LINK_COUNT=2 ? firmware version is 2.27.

this does not work
function BluetoothRemoteGATTServer.setRSSIHandler(callback)

and i have no rssi field on the bluetooth object.

=BluetoothRemoteGATTServer: {
  device: BluetoothDevice: {
    id: "ff:ff:11:9b:xx:xx public",
    gatt:  ... ,
    "#ongattserverdisconnected": [
      function (reason) { ... }
     ]
   },
  connected: true, handle: 1 }

[enaon]

DCENTRAL_LINK_COUNT=2 seems not be the problem, I tried on the bangle.js2, 2.27, it is the same, I cannot get the rssi of the peripheral. Has something changed, or am I doing something really wrong?

sorry for posting too soon, it seems that the itag does not send the rssi when connected, I thought this was on the connection level, it seems that rssi is device realated? I can get the rssi from another device I tested, one that is running esruino, If one has some time to explain please do.

[gfwilliams]

I thought this was on the connection level, it seems that rssi is device related?

I'm surprised too! I'm glad it works on something else though...

But you say you're actually running Espruino on the tags? I guess it's possible the BLE stack is expecting a 'transmit RSSI' entry in the BLE advertising data it received so it can scale somehow?

... but if you're not connected and you just scan for advertisements, you should get an RSSI?

[enaon]

But you say you're actually running Espruino on the tags?

No they are generic itags off aliexress, the cheap ones, they work fine with espruino in the sense that they are unlocked, I can get the button press, send the alarm etc.

... but if you're not connected and you just scan for advertisements, you should get an RSSI?

Corect, if I am not connected, I do get the rsssi, this is why I thought the two connections firmmware build was the issue, but it is not, I tested another device as peripheral, one running espruino, and I do get the rssi when connected from the watch to the peripheral.

so it seems that the itags do not send the rssi when the watch is connected to them, which is a real shame, I wanted to use the watch to find the itags whithout making them sound their alarm, and a fast rssi reading could be super, but ok, the scan is fast enough.

it actually works amazingly, anyone with cats can feel me I think, they like to hide inside boxes etc, the watch now scans them, I can understand if they are inside the room, no more searching under the bed etc :)

i will upload a bangle.js 2 version, using my script for the UI, just so you can test it if you like, as a generic scanner maybe, you can change it to scan whatever you like, the UI is what I would like for you to test if you care.

[enaon]

I'm glad it works on something else though...

I tested some more, and I can now fully appreciate the sarcasm on this comment :)

so I take it that

// Start listening for RSSI value updates
gattServer.setRSSIHandler(function(rssi) {
  console.log(rssi); // prints -85 (or similar)
});

is not working any more, do you maybe remember the last firmware version where it was working?

[gfwilliams]

I wasn't intending to be sarcastic - I thought you'd said that setRSSIHandler had worked on some devices but not others.

I just looked into this and I put a fix in, so cutting edge builds should work now.

It would have worked but only on the second device if you had more than one active connection! It looks like this got broken when support for >1 central connection was added to the firmware

[enaon]

I wasn't intending to be sarcastic

Why not, I had my tests messed up, again, I think I deserved it, I was laughing when I read what you said after I realized. :)

I just looked into this and I put a fix in, so cutting edge builds should work now.

thank you so very much

It would have worked but only on the second device if you had more than one active connection! It looks like this got broken when support for >1 central connection was added to the firmware

ok, thanks for that, I am sure I saw it working at some point, but could not replicate.

[enaon]

edit: ok sorry again, I saw this espruino/Espruino#2465, I upgraded the Bangle's softdevice to 6.1.1 and scanning on phy:"both" works fine.

@gfwilliams , I think I tested correct this time, not important so I am reporting it here:

when on 2v27.51 ( or 2v27 )and DFU is 2v25, on the BangleJS2, if I do an NRF.setScan with phy:"both", I get:

Uncaught Error: ERR 0x7 (INVALID_PARAM) (:2291)

When on the magic3 on 2v27.51 o(r 2v27 or 2v27.47), it works as expected.

I used Clean Boot on the Bangle to be sure. Scanning for coded or 1mbit works.
bangleJS2

>NRF.setScan(function(d) {
:  print (d.id) ;
:}, {  phy:"both" });
Uncaught Error: ERR 0x7 (INVALID_PARAM) (:2166)
    at REPL (:3:19)
}, {  phy:"both" });
                  ^
>NRF.setScan()
Uncaught Error: ERR 0x8 (INVALID_STATE) (:2166)
    at REPL (:1:13)

Magic3

>NRF.setScan(function(d) {
:  print (d.id) ;
:}, {  phy:"both" });
=undefined
7c:e7:12:52:xx:xx public
dd:3b:8b:d3:xx:xx random
7c:e7:12:52:xx:xx public
dd:da:ab:ef:xx:xx random
c6:d5:34:52:xx:xx random
>NRF.setScan()
=undefined

[gfwilliams]

Well, I just tried it on my Bangle.js, both with 2v27 and 2v27.49, and it works great.

Also I don't believe anything has changed in the bluetooth scanning functionality between 2v27 and the latest build. Is it possible your Bangle with the issue was using an old softdevice?

edit: ahh, I see it was - so with the latest softdevice everything is ok?

[enaon]

Well, I just tried it on my Bangle.js, both with 2v27 and 2v27.49, and it works great.

sorry about your time, I should have searched better.

so with the latest softdevice everything is ok?

yes, super ok. I have placed a lot of different tags on a lot of cats, they help with the tests, "coded" on scanning makes an important difference.

Out of curiosity, I used Daflasher and fanoush's "magic3" softdevice, is there a more Espruino way?

[gfwilliams]

Ok, great! On the latest nRF52480 firmwares I just added process.env.SOFTDEVICE which tries to work out what the softdevice is.

is there a more Espruino way?

You mean updating the softdevice? Personally I'd use Nordic's DFU app, but other than that the magic3 firmware file is best (it's the same as I'd end up making).

It should be possible to make something that would work with the Firmware Update tool in the app loader, but I haven't tried it. I think if you care about the 6.1.1 softdevice you're probably proficient enough to use the DFU app :)

[enaon]

I think if you care about the 6.1.1 softdevice you're probably proficient enough to use the DFU app

true, and atc1441's Daflasher works fine on the Bangle too, I am more than happy with everything, just wondering it there was a build in system that I overlooked. Thanks again.