nrf connect pc pca10059 crashes on receiving notifications (sometimes)

[Peter Kenyon]

I am using the nrf connect desktop application on windows with a PCA10059, and have found the application to be buggy when receiving a flood of notifications, this works fine using the pca10031 dongle but I want to use data length extensions and large att mtu

[Edvin]

Hello,

Are you sure it is not the other device, the device that you are connected to, that receives an error? Does it says that it is disconnected, or is it actually nRF Connect that hangs?

[Peter Kenyon]

nope its nrf connect that crashes, and only with the pca10059 dongle,I think its related to the att mtu settings somehow

[Edvin]

Can you please specify what version of nRF Connect you are using? (Version number), and what you mean by "application hangs"? Do you get a white screen, and then it closes, or does it become slow?  Have you tried using it with other devices than dongles? E.g. the nRF52832 DK or the nRF52840 DK?

BR,

Edvin

[Peter Kenyon]

I am using nrf Connect version v2.6.1.

I have successfully been using the pca10031 dongle V1.1.0 and everything is OK.

I changed to using pca10059 dongle 1.0.0 and I am having problems when receiving notifications,

I have just tried the pca10056 DK and that works OK,

so there is a problem with the PCA10059 dongle when receiving floods of notifications

The nrfConnect app becomes slow but can be recovered by reselecting the device

[Peter Kenyon]

I just tried another brand new NRF52840 dongle, and let nrf connect update bootlaoder and firmware, and the same problem exists, when receiving notifications it hangs

[Edvin]

Hello,

Thank you for reporting this, and for the details.

I have replicated this, and I see the same behavior. Also, I see that this does not happen while using a DK, just as you say.

I have reported this internally, and they will have a look at this.

Since it works on the PCA10040/10056 (I tested), the issue is not the large MTU, but I believe it is rather the throughput on the USB that is capped. At least that is my theory.

Best regards,

Edvin

[Peter Kenyon]

I had a quick look using a sniffer (from Nordic - great job guys), and it looks like the slave device starts to receive a seq no error in the LL and continually resends the notification, so the two devices are stuck sending and non acking, until the slave timesout 

[Edvin]

Hello,

Thanks for the update. I added this note to the internal ticket.

BR,

Edvin