Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

[SOLVED] Debugging PCIe interface

Status
Not open for further replies.

nickagian

Member level 4
Joined
Mar 19, 2009
Messages
71
Helped
2
Reputation
4
Reaction score
2
Trophy points
1,288
Location
Zurich, Switzerland
Activity points
2,001
Hi,

In a system of ours, we have a CPU that is connected over PCIe with an on-board FPGA.

CPU is root complex, compliant with 2.0 and 1.0 and the FPGA is the endpoint (only 1.0 compliant).

Unfortunately the CPU tells me that the link is down (it's a new design, I have never seen it come up yet).

Attached you can see what I measure at the two data lines.

The 100MHz is there and running. It is created from the CPU and sent directly to the FPGA.

It seems to me as if the RC sends some kind of beacon to detect possible EP. But the question is why my EP doesn't respond or why the RC doesn't recognize the response.

Is there anyone with experience on PCIe that could help me debug? I am almost without any ideas on how to proceed.

I have thought the following:
- That the Quality of the clock is not good enough and that's why the FPGA doesn't respond, but the truth is I cannot check this with much detail.
- Wrong signal levels. If I am not mistaken, this is not the case.
- The schematics are correct. I have the necessary AC blocking capacitors on all lines, TX are connected with RX.
- The length of the Signal lines is not more than 25cm (Maximum recommended is around 50cm)

Nikos
 

Attachments

  • PCIe_signal_from_EP.PNG
    PCIe_signal_from_EP.PNG
    19.8 KB · Views: 7
  • PCIe_signal_from_RC.PNG
    PCIe_signal_from_RC.PNG
    18.7 KB · Views: 6
  • PCIe_signal_from_RC_zoom.PNG
    PCIe_signal_from_RC_zoom.PNG
    16.5 KB · Views: 6

nickagian

Member level 4
Joined
Mar 19, 2009
Messages
71
Helped
2
Reputation
4
Reaction score
2
Trophy points
1,288
Location
Zurich, Switzerland
Activity points
2,001
Well, after some days the answer came to me: I had used a wrong value of the series AC-blocking capacitors. Specifications say min. 75nF and I had 10nF on the direction endpoint -> root complex. I changed them and everything was ok after that...!
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top