Lost RF-Frames on high datarates with more than one sender

Hello!

We are using a bunch of XBee-Pro modems for sending IP-traffic over a network. When two stations send a 600 byte IP-packet to the same station at the same time it occurs that some RF-frames may be lost.

Configuration:
XBee Pro series 1 with 802.15.4 mode with Ack (MM=02), datarate set to 111111 baud (also tested lower/standard ones), API-mode, FW-version 1.0.C.8 (also tested 1.0.A.4)
USB-serial converters are FTDI FT232BM

The IP-packet is fragmented properly and send to the modem. We get an ACK that the everything was fine. Furthermore, by using a ZigBee-Sniffer on an additional station we can see that the RF-frame was send correctly as well as we can see the ACK-frame.

Monitoring the RTS-line of the receiver shows that the FTDI never requests the modem to stop sending data. Manually pulling RTS to ground while RTS is enabled (ATD6=01) shows that the modem will stop sending data - so this is working.

There is always an entire RF-frame missing. Not a single byte more or less. Therefore it looks like the frame is discarded at some entity which is aware of the frame-structure. Since the FTDI has no knowledge about the structure of an RF-frame we assume that the frame is discarded at the modem. The question now is why that happens although the RTS line stays high (logical 0) at all times as well as we receive Acks for all frames.

Is there any way to check if there was a frame discarded at the modem because of a buffer overflow? It would be very nice if there would be a command for that.

Any help is highly appreciated!

Best regards,
Robert

You must use 2 stop bits with that high of a data rate (or some other means of throttling the data to avoid overruns).

Hi!

Thanks for your hint! Anyway, the sending is fine as the frame is transmitted correctly (so no overrun at the Tx side). At the receiver side the effective data rate is controlled by the XBee-module, so there should be no problem? In fact, requiring 2 stop bits would make the comm even slower as therefore things even worse?

Best regards,
Robert