Forum Discussion

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

Preparing environment for production department

When we released our first board with Cyclone FPGA we had two JTAG connectors populated on the board: one for directly programing serial eeprom and the second for directly accessing FPGA chip (for debugging). We could install Quartus Programmer as a standalone program on the production pc to load the content of the serial eeprom.

Now I am preparing second generation board with Cyclone FPGA and I am tempted to leave only one JTAG header on the board, directly connected to FPGA, leaving the task of programming serial eeprom to the Nios Flash Programmer ported to our custom board.

What will be the production environment for such purpose? I use NIOS IDE to connect to FPGA, program it using JTAG with Nios Flash Programmer content and using this Flash Programmer to load serial eeprom with real system configuration content and CFI flash with my software files.

How can I make these tasks as easy as possible for production people? Can I install somehow not the whole Quartus not the whole NIOS IDE but a simpler program on a production computer, preferably only Flash Programmer to require the operator to follow very simple procedure to program new boards and know if the board was correctly programmed?

3 Replies

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

    Hi Pszemol,

    There is a "Serial FlashLoader Whitepaper" that describes how to program an Active Serial Device over a JTAG chain connected to Cyclone (and now Cyclone II) devices. The Cyclone II handbook makes a reference to the Serial FlashLoader, but I cannot find it on Altera.com.

    Basically, the Serial FlashLoader bridges the JTAG interface to the Active Serial interface with a special sof file. Once the Cyclone device is loaded with this file, you can program the EPCSx device via JTAG (connection to Cyclone). Since this is a special design, the user design you have will be erased.

    This could work well in a production environment, but in an embedded environment you may run into some trouble.

    I can send you the whitepaper if you are interested.

    -Terry
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    21 years ago

    Knowing how production environments are always hard pressed for time, I wouldn't do it. Instead I'd leave some pads on the bottom of the board to get direct access to the serial eeprom. I'd guess you'd want a pin jig of some sort to automatically test the boards anyway. Then during test, you can load the serial eeprom with whatever sof you want reliably and quickly. The other way takes a fair bit more time, and has more links in the chain to go wrong. The extra board cost of a pad layout is small, and generally the real-estate cost should make up for the testability you gain.

    Bart