TRF7970A： TRF7970A

Dear Li - 

Assuming you are 100% sure that one unit on its own has stable software functions and you are satisfied with positive and negative test results, I would recommend that you check (with logic analyzer) the details of each unit when they "crash". I would suspect that you are getting a "collision"error that cannot be resolved and so therefore the system "crashes". You would be able to see that in bit1 of register 0x0C of the TRF7970A and possibly, since this is most likely over the air collision, between units, you may also see CRC or framing error bits set in that register. 

Obvious solution is to separate the readers by proximity, so their magnetic fields are not interacting with each other - if the end application just has the one reader installed. If this test is real life and you have many readers close together, then you would need to synchronize (actually this would be more dithersync'ing)  the reads periods between readers, so only ones which are not close to each other are active. You could either do that by MCU time control of the individual readers or leverage the RSSI register values to make decisions about the external magnetic environment. 

You can also check the Q of your antennas - if it is too high, this will cause problems like this, too - and easy solution would be to reduce the Q (for ISO15693 is should be around 16-20)

if your end equipment has the need for many reader antennas in close proximity (like some kind of item level smart inventory system), then you would be better off using just one TRF7970A and multiplexing the antennas to have 100% control of the environment. 

See ==> http://www.ti.com/lit/an/sloa167/sloa167.pdf and http://www.ti.com/lit/an/sloa231/sloa231.pdf