Hello!

We have been getting some help from Mark and John down here at the sidewalk to hammer down our TAB->L2/L3 BER.  We have been making a lot of progress:

1) We have new VRB testing firmware that calculates the parity of data from the TAB and compares it with the expected parity.  It also monitors and latches glink ready, glink word error, and checks for words outside of event markers.  As there is no data transfer (just the XOR calculation) this test program can handle a very high input rate.

2) We have new TAB transmitter testing firmware that sends 128 16-bit words counting from 0 to 127.  The transmitter firmware has only one clock domain (cp53) and does not require external L1As.

3) Tonights' late night discovery:  changing the values of the phase-lock loop capacitor on the VTM glink receiver chips cures our transmision woes!

With these changes, the TAB is happily hammering L3 at 100kHz.  After about 40 minutes, we arrive at 250 million events transferred with no errors.  This puts our bit error rate at less than 8E-12.

We have tested similar (XOR) code at L2, but it cannot handle such a high rate.  We have sent 10 million events to level 2 resulting in a slighly less stringent BER.

Once we have our ORC, we can do overnight tests to make these limits even stronger.  However, it seems clear that our TAB->L2/L3 link is now reasonably robust (a solid green line!)

cheers,
mike