Forum Discussion

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

stratix II Dual-Piece Lid FBGA getting very hot

Hi,

I have developed a test board that is approx 6" x 8" with a stratix II gx running some code. It runs at a bout 40 degrees C (maybe slightly less).

I then took components off the design, added a small connector to put the large interface connectors on a duaghter card and made a board that is essentially the same from the fpga point of view, but is only about 2" by 6".

When we got the small board stuffed it came back with the dual piece lid as opposed to the old style lid whixh is on the large board.

The new small board with the dual piece lid is running at about 60 degrees C. This is way too hot. The small board works (as long as it does not get too hot) and has plls and serdes in it so it must have all the voltages it wants.

The current draws are -

Large - Good Board -

Vcc25 = 0.118A

VCC15 = 0.194A

VCC1.2 = 0.795A (0.52 core, 0.275 pll)

Small - Hot Board-

2.5V = 0.111A

1.5V = 0.148A

1.2V = 1.388A (0.77 core, 0.618 pll)

I do not know what would make the core and peripheral 1.2V be so affected.

I can find no shorts between the supplies and no supplies pulled to ground, the impedance (static, board off) between a supply and ground is approx the same on both the good and hot board.

The place I keep looking is the new package, is that causing the excess heat? Is there some pin difference that is causing the issue? I am really not sure where else to look now.

Martin

5 Replies

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

    You didn't clarify whether this condition exists with or without the FPGA programmed. Is it the exact same code as before?

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

    The problem is the board gets hot even if it is not programmed (about 50C).

    If I put in the same code as before it gets hotter around 65 C.

    The peripheral/PLL 1.2V current stays fairly constant with temperature, but the core 1.2V current climbs with temperature, quite predicatably.
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    15 years ago

    Again Altera has exceeded my expectations about how bad their support could be.

    Not once did they ask me for any relevant info to debuig the issue and we were working all channels for support.

    The issue basically came down to the 1.2V regualtors being too close to the fpga, the regulator got the fpga warm so it drew more current which made it get wamrer which made it draw more current........

    The flat top and the bump top operate at the same temperature, we got two identiacal boards one with the bump top fpga and one flat top fpga and ran the same code.

    The temperatures were almost identical.

    P.S. don't tell altera let them figure it out on their own.
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    15 years ago

    By taking an interest-

    Asking for boards, looking at layout making suggestions sending in someone who might have seen something like this.

    Basically putting in the time that other companies do with their clients who are industries that require technical turnaround in hours instead of days or weeks.

    Now support is provided by outside personnel who handle multiple lines. These personel cannot focus on the multitude of prts offered by Altera.