Latency and serial buffer clogging

[Electrumpet]

Dear all,

I don't know where I should ask this question but this place could be
as good as any.

Thus far it seems that directly assessing sensors with the XBEE inputs
seems faster and with less latency then a micro controller / XBEE
combination.

I do not use the XBEE API library but I use MAX/MSP directly to access
the data.

What I see in MAX is that two to five API packets are sent
simultaneously with breaks in between from 16 to 40 ms. That does not
really help latency.

(computer is a MacBook Pro 2011 Intel core i7, 2,0 GHz.) (baud rate =
115200, sensors on all the analog inputs of the XBEE)

The speed in itself is good. When I put the sensor sample rate (IR) at
6 I get 1000 samples in in 6429 ms. Since this is a constant number I
suspect that there are no losses (unless I move the receiver away). So
the 'clogging' has probably to do with a serial protocol somewhere on
the line and not with some kind of interference.

Do people experience the same kind of 'clogging' with the XBEE-api in
processing?

I was wondering whether this has to do with the serial object in MAX/
MSP or that it has to do with the FTDI serial driver. Maybe using the
D2XX driver is better then using the virtual com port driver that I
use at the moment? I have no clue though what the difference is
between them (a pointer there is also welcome).

Can somebody here help me out or at least point me to the right forum
(or person)?

Best, Hans.