cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Highlighted
Visitor
Visitor
1,399 Views
Registered: ‎03-07-2018

KCU105 DDR4 Pinout

Jump to solution

I have consulted the KCU105 User Guide [UG917 (v1.8) July 26, 2017], Appendix D "Master Constraints File Listing".  I have also generated a DDR4 core (using Vivado 2016.4) and its example design, including a constraints file example_design.xdc.  Unfortunately the information in these two sources seems contradictory.  To give just two examples,

   Signal DQ[0]

      In ug917 DQ[0] is AE23  (AE23 is ADR[7] in the example design).

      In the example design DQ[0] is AM22 (AM22 is DQ[31] in ug917).

   The 300MHz clock (sys_clk_p or SYSCLK_300_P)

      In ug917 SYSCLK_300_P is AK17 (AK17 is DQ[17] in the example design)

      In the example design sys_clk_p is AJ23 (AJ23 is DQ[10] in ug917).

These two sources are both specifically about the KCU105 board (not the generic defaults).  Which should I believe?

 

Thanks

 

joe

Tags (1)
0 Kudos
1 Solution

Accepted Solutions
Highlighted
Observer
Observer
1,688 Views
Registered: ‎01-15-2015

Did you set the IP settings as shown below?

ddr4_kcu105.png

View solution in original post

4 Replies
Highlighted
Observer
Observer
1,358 Views
Registered: ‎07-07-2015

If you set the project and IP settings to KCU105, I think that it will be the same pin arrangement as the document.

project_kcu105.pngddr4_kcu105.png

0 Kudos
Highlighted
Visitor
Visitor
1,353 Views
Registered: ‎03-07-2018

I did do that (the example design's Vivado even displayed a picture of the board); but the pin arrangements are different.  But thanks for the message.

 

joe

0 Kudos
Highlighted
Observer
Observer
1,689 Views
Registered: ‎01-15-2015

Did you set the IP settings as shown below?

ddr4_kcu105.png

View solution in original post

Highlighted
Visitor
Visitor
1,208 Views
Registered: ‎03-07-2018

Sorry for the delay in replying.  Yes, the last suggestion made the problem disappear -- literally.  All the explicit pin locations vanished from the constraints file, and were replaced, for each pin, by a constraint saying in effect that they should go wherever the board said they should go.

 

The example design still didn't synthesize for me though.  Vivado said that two pins (the calibration-complete pin from the core and the pass/fail pin from the testbench) didn't have their IOSTANDARD specified, and I couldn't find a way to get past that successfully.  So I'm afraid I gave up, and went to my own design instead, which I got successfully working (though I have another question about it, which I'll ask on a separate thread),

 

Thanks for the help.

 

joe

0 Kudos