Forum Discussion

vlad's avatar
vlad
Icon for New Contributor rankNew Contributor
5 years ago

How to constrain the delays from ports to combinational block?

Hi! I'm trying to transmit data from one FPGA to another (DE1-SoC Cyclone V boards).

The method I'm using is to have two wires, such that a transition on one wire means that a 0 is transmitted, and a transition on another wire means a 1 is transmitted.

Now by passing this 2 wires through an XOR gate I can essentially generate a clock (200 MHz) (so data is transferred at both rising and falling edge of that clock).

Then I feed that generated clock into a PLL that doubles the frequency (400 MHz) and phase shift it (180 dg) so I can sample the rx0 and rx1 wires to see which one transitioned so I can put a 0 or a 1 in my shift register.

The problem is when I check the path timing, from rx0 port to XOR gate is around 8 ns, and for rx1 to XOR is around 6 ns, how do I constrain them to be the same delay? And from rx0 and rx1 to the registers that samples those the delay also must be the same.

I tried set_min/max_delay, net_delay etc. in SDC, but they seem to be ignored.

Thank you!

5 Replies

  • sstrell's avatar
    sstrell
    Icon for Super Contributor rankSuper Contributor
    5 years ago

    Can you post your .sdc? Also make sure that your .sdc file has been added to the Timing Analysis settings. Those are huge delays so something's fishy with either your constraints or your design.

    #iwork4intel

    • vlad's avatar
      vlad
      Icon for New Contributor rankNew Contributor
      5 years ago

      Hi, Thank you for the reply.

      Here is my SDC file:

      create_clock -name input_clock -period 20.000 [get_ports {CLOCK_50}]
create_clock -name rx_xor      -period 4.0    [get_pins {rx_gen_clock|combout}]
 
derive_pll_clocks
 
set_false_path -from [get_clocks {input_clock}] -to [get_clocks {u_tx_pll|tx_pll_inst|altera_pll_i|general[0].gpll~PLL_OUTPUT_COUNTER|divclk u_tx_pll|tx_pll_inst|altera_pll_i|general[0].gpll~FRACTIONAL_PLL|vcoph[0]}]
 
set_false_path -from [get_registers *async_rx*] -to [get_registers *error_counter*]
set_net_delay -from [get_registers *async_rx*]  -to [get_registers *error_counter*] -max -get_value_from_clock_period dst_clock_period -value_multiplier 0.8
 
set_max_delay -from [get_ports {RX[0] RX[1]}] -to [get_pins {rx_gen_clock|combout}]  8
set_min_delay -from [get_ports {RX[0] RX[1]}] -to [get_pins {rx_gen_clock|combout}]  7.5

      Report data path:

      report_path -from [get_ports {RX[1] RX[0]}] -to [get_pins {rx_gen_clock|combout}] -npaths 100 -panel_name {Report Path} -multi_corner
 
Delay       From      To 
7.512	RX[0]	rx_gen_clock|combout
5.627	RX[1]	rx_gen_clock|combout

      And the paths in chip planner:

      That essentially what I want to do (I could make the tx_pll clock 200 Mhz by using ddo registers for the outputs)

      Btw, I have synthesized it once and it kinda worked, so I could transmit data at 400 Mbps between two FPGAs (sending just an incrementing counter and checking on the other side that I got the right value),

      but it's too unreliable especially if I slightly change the code then everything changes. Thanks

  • EngWei_O_Intel's avatar
    EngWei_O_Intel
    Icon for Frequent Contributor rankFrequent Contributor
    5 years ago

    Hi Vladislav

    Thanks for your inquiry and sharing of your design constraint. Let me take a look and get back to you.

    Thanks.

    Eng Wei

    • KhaiChein_Y_Intel's avatar
      KhaiChein_Y_Intel
      Icon for Regular Contributor rankRegular Contributor
      5 years ago

      Hi Vladislav,

      We do not receive any response from you to the previous question/reply/answer that I have provided. This thread will be transitioned to community support. If you have a new question, feel free to open a new thread to get the support from Intel experts. Otherwise, the community users will continue to help you on this thread. Thank you

      Best regards,

      KhaiY