Reset BlueNRG-MS by connecting multiple devices simultaneously

I use BlueNRG-MS stack 7.2c, stackmode = 3

When connected only ONE device (mobile phone) to BlueNRG-MS everything works fine, but when  startup

or

for additional second connections after some time (3 - 10 minutes) BlueNRG-MS reset by the watchdog, before it come to the event EVT_HARDWARE_ERROR=0x10 (event = 0x10, subevt_code = 0x01) and (event = 0x10, subevt_code = 0x02)

Tell me, please, what should be done to avoid reset BlueNRG-MS?

log:

[05728777] event = 0x3E (EVT_LE_META_EVENT), subevent = 0x0001 (EVT_LE_CONN_COMPLETE)

[05728779] EVT_LE_CONN_COMPLETE status = 0 ; handle = 0x0801 ; role = 1 ; interval = 30.000000, latency = 0 , supervision_timeout = 72, master_clock_accuracy = 5

[05728897] [HCI_Event_CB] event = 0xFF (EVT_VENDOR), ecode = 0x0800 (EVT_BLUE_L2CAP_CONN_UPD_RESP)

[05728897] conn_handle = 0x0801 ; code = 0x13 ; identifier = 1, result = 0

[05729257] [HCI_Event_CB] event = 0x3E (EVT_LE_META_EVENT), subevent = 0x0003 (EVT_LE_CONN_UPDATE_COMPLETE)

[05729257] EVT_LE_CONN_UPDATE_COMPLETE status = 00 ; handle = 0801, interval = 168.750000, latency = 0, supervision_timeout = 100

....

[05732794] master_devices = 1 ; slave_devices = 0

[05732800] ENTER_DISCOVERY_MODE

[05732802] aci_gap_set_discoverable() OK

....

[06198908] [HCI_Event_CB] EVT_HARDWARE_ERROR event = 0x10, evt_code = 0x01

[06202624] [HCI_Event_CB] EVT_HARDWARE_ERROR event = 0x10, evt_code = 0x01

[06204480] [HCI_Event_CB] EVT_HARDWARE_ERROR event = 0x10, evt_code = 0x01

[06204820] [HCI_Event_CB] EVT_HARDWARE_ERROR event = 0x10, evt_code = 0x01

[06205674] [HCI_Event_CB] EVT_HARDWARE_ERROR event = 0x10, evt_code = 0x02

[06220667] [HCI_Event_CB] event = 0xFF (EVT_VENDOR), ecode = 0x0001 (Firmware started properly)

[06220667] EVT_BLUE_INITIALIZED Reset caused by watchdog