MCP2155 Performance

Do you have a question? Post it now! No Registration Necessary

Translate This Thread From English to

Threaded View
Hello!

I'm developing a board based on the MCP2155 to transfer data via
IrCOMM between a PDA (Compaq Ipaq - Pocket PC) and a device connected
to the board. I've already set it up to work, but i can only achieve a
transfer rate of about 12000bps. The MCP2155 datasheet asserts it can
achieve "typical" rates of up to 38000-50000bps. I've repeated the
test with a WinXP-laptop instead of the PDA but the results were the
same.

I've verified that the actual infrared bitrate (on the air link) is
115200 and i transfer data also at 115200 on each side. Since I was
quite disappointed with the effective bitrate, I connected an
oscilloscope to the MCP2155 and i observed the following repetitive
pattern during the transmision of a file from the MCP side to the PDA:

(approximate time measurements)

1) The primary (PDA) transmits for about 12'5 msec (144 bytes @
115kbps)
2) Less than 1 msec later, the PDA activates its CTS output. Now is
when i feed the MCP with the file data, until CTS deactivates. CTS
stays active for about 5 msecs (~64 bytes @ 115kbps) as i expected.

3) Now comes the big problem! ~25 MSECS of innactivity LATER the MCP
starts to transmit, for about 7 msec (~80bytes @ 115kbps).
4) After a brief turnaround of 2 msec, the PDA transmits again and it
starts over from 1)

Summarizing: the MCP has transmitted effective data for 7msec of a
period of 50msec, so it is working at 14% of the IR bitrate, which
renders the ~12000bps bitrate i get. The delay in 3) (which i don't
understand what is due to) and the long transmission time of the PDA
in 1) (which surely has to do with the Irda Stack in the PDA) spoil
the effective bitrate.

Has anybody an explanation for that strange delay in 3)?   or simply,
Has anybody got higher bitrates between the MCP and a commercial Irda
Stack?

Many thanks in advance!!
 Victor

Re: MCP2155 Performance
Caveat: I have no experience at the low level of IrDA.

Given that caveat, a suggestion: Test with the tx and rx at least 30cm
apart. I've had unusual problems with IR transfers when the separation
was too low.

Site Timeline