Forum Discussion

Altera_Forum's avatar
Altera_Forum
Icon for Honored Contributor rankHonored Contributor
15 years ago

LVPECL VCCIO 3.3V instead of 2.5V

Hi, I am facing an issue for my PCB which uses Arria2GX FPGA. Due to pin constraints, I am forced to use an LVPECL dedicated clock on a bank which has LVCMOS33 I/Os(Bank 4A for EP2AGX125E device to be exact). I have provided VCCIO as 3.3V but LVPECL (which shares the same VCCIO) requires VCCIO not greater than 2.6V. will this be an issue and cause the clock receiver to fail? I cannot use any other banks because all other clock input pins are used.

3 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    15 years ago

    Apart from the question, if the differential receiver possibly can work with VCCIO of 3.3V and which performance degradation must be expected, Quartus will require a VCCIO setting of 2.5V for the respective bank to allow an differential I/O standard in it. So you would have to declare all LVCMOS I/Os of bank 4A wrongly as LVCMOS25 in addition.

    But you can hardly expect official clearance for unspecified operation of Altera devices, I think.
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    15 years ago

    FvM, thanks for your reply.

    Yes, the person who compiles the FPGA build has decided to falsely give LVCMOS25 and generate the build. For LVCMOS this won't be an issue I guess, since noise margins are almost compatible for LVCMOS33 and LVCMOS25. But my worry is whether the 3.3V would cause the LVPECL receiver to malfunction. I am using AC coupling and the differential swing is within the range of Altera spec. Besides, the common mode bias which I have provided is 1.25V(assuming 2.5V)..

    since the supply is now 3.3V, should I increase the bias? I still think 1.25V is pretty safe even with 3.3V.
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    15 years ago

    Personally, I expect the receiver to work at 3.3V, the maximum speed most likely reduced. Because Altera won't give a specification for differential receiver operation at 3.3V, you have to acquire the typical data yourself, provide a sufficient margin and hope, that it isn't ruined by device parameter variations.