So we were aware of the errata for BRK1900 when we purchased the board
( * Voltage translation for control signals to the QSFP cages is non-functional. While transceiver routing is functional, the lack of usable control signals may render some QSFP modules inoperable on the BRK1900. A fix for this is expected in revision Bxx.).
We figured we could still do some testing as we haven’t needed to use these control signals on the Fiber QSPF devices we have used in the past. We even purchased the FormericaOE QSFP specified in the BRK1900 transceiver performance section of the (ECM1900 - Opal Kelly Documentation Portal). However we can’t get light out of either QSFP fiber modules. The only exception is at boot up with a different transceiver before the VIO4 voltage is set by Petalinux.
What IBERT settings are you using? I was going to try in system IBERT with the IP core we were using but can’t get there because VIO4 turns off the*transceiver when it gets set to 1.2V or 1.8V (if I place the DAC SYGYZY in PORT A).
What are you using as your reference clock? Refclk 0 or 1 on which transceiver bank (226, 227)? I am using QSPF 1, bank 227 ref clock 0
How are things are connected? I put the fiber transcier into QSPF1 and am using a scintillator to look at the light which goes off when VIO[4] is set
Are you using JTAG to program and use IBERT? To program I have generated new PL design LED clock heartbeat and our fiber interface and packaged with petalinux to a sdcard image. I have JTAG to see the In-system IBERT
What is your bring up procedure starting from powering on the board?
**Put in transceiver into QSPF1, set up scintillator in beam path, load sdcard, power on and let boot. See light till the VIO[4] setting is set on boot up. Write bits to take the card out of reset (QSFP1_RESET_B, QSFP1_MODPRS_B, QSFP1_MODPRS_B and QSFP1_MODSEL_B set high. Stop working without light no need to do IBERT test with loopback fiber **