xAPmcsRFID IO

[Uncle Michael]

Michael,
Sorry for the dumb questions, but just getting my feet wet on xAP. I've installed xAPmcsRFID and it appears to be working, but the input stream from the receiver appears to be segmented in unexpected ways. The result is that the tags do not register as present. I've included a screen shot of the IO window that I hope will be of use. Did I miss a step in the setup?

[Michael McSharry]

The IO window only reflect the delivery of characters from the COM port UART. It does not reflect the decoding. The data is buffered and processed when the expected protocol is satisfied. The IO window is primarily a debugging aide.

I'm on the road now so I cannot look at what I have running and what a normal stream looks like. When there are breaks in the data it is likely due to events delivered from the UART that have no data such as Ring, Overrun, etc. These usually make no difference.

[Uncle Michael]

Thanks for the reply.
I'm certainly not comfortable speculating on what may be going on under the surface, but what I see in the browser is a very long list of candidate tags that look like the fragments I see in the IO window. It also appears that tag detection for the complete tags has a very low success rate because they are listed as absent most of the time, even though they are coming in frequently (albeit broken into multiple lines).

Are you thinking that the problem may be in the RFID receiver or in the serial port? I'm using a serial to USB converter. Could that be a causing the problem?

[Michael McSharry]

It has been awhile since I worked with the RFID, but I think it is not unusual for the system to produce erroneous tag id information. These were designed to be as very cheap so there is not much intolerance at the hardware level and this task is left to software to make it usable.

The tag protocol looks to be 4 characters followed by a space delimiter and it looks as if xapmcsRFID is using the first of these two as an identifier. The stream that you showed before looks to be reasonable for two receptions. All the space in between is likely the serial port delivering events without data.

xapmcsRFID allows you to accept and reject messages. I think you will find a limited number of erroneous patterns. Once you have determined one to not be a real tag then mark is as rejected so the software will automatically filter it out and never show it to you again. Over time you may accumulate a few others than can be given the same treatment. What is important is you Accept the good IDs and set the criteria to give the desired tolerance to periods of no tag receptions.

[Uncle Michael]

Michael,
Thanks for the explanation. The problem I'm seeing is that the time between recognition of good IDs is quite long. The software seems to generate an erroneous ID well over 99% of the time (possibly closer to 99.9% - or more). I'd really like to improve the ratio substantially. Is that possible?