Arria 10 GX , Cyclone 10 GX transceiver phy compatibility

Hi Manolis,

As I understand it, you're observing some issues in the datapath between the C10 transmitter and the A10 receiver. Regarding your specific question on the electrical compatibility between C10 TX and A10 RX, I’ve reviewed their datasheets and, based on the specifications, there should be no issue interfacing the two devices electrically.

To help facilitate further debugging, I’d like to check on a few points:

1. Status Signals:

Could you please share the status signals from both the C10 TX and the A10 RX? This information would help us assess the current state of the link and determine the next steps in debugging.

2. Signal Integrity Check:

Have you had a chance to probe the C10 TX SMA output with an oscilloscope to inspect the eye diagram? This would help rule out any signal integrity issues at the transmitter side.

3. Loopback Test:

You might also consider performing an SMA loopback from the C10 TX to the C10 RX. This can help verify whether the C10 TX is functioning correctly in isolation.

Please let me know if you have any concerns. Looking forward to your update.

Hi,

Regarding your inquiry on the XCVR electrical specs, you may refer to the following tables in the Intel® Cyclone® 10 GX Device Datasheet and Intel® Arria® 10 Device Datasheet:

1. Receiver Specifications

2. Transmitter Specifications

These tables include key parameters such as VOCM, VICM, VID, and others.

Please feel free to reach out if you have any questions or need further clarification.

Thank you!

Hi,

Regarding the status signals, feel free to share the following A10 RX signals for further review:

rx_cal_busy

rx_is_lockedtoref

rx_is_lockedtodata

These will help in assessing the RX calibration and lock status.

As for rx_std_signaldetect, it typically works in conjunction with the signal detect threshold set in the QSF file. Just to confirm, do you observe this signal de-asserting when the SMA cable is disconnected?

Thank you.

Hi,

Regarding the specification, the following is an example for the RX specifications for VID and VICM:

Hi,

Thanks for your update regarding the A10 rx_is_lockedtodata = 1 when the SMA is connected. Based on this, I presume that rx_cal_busy = 0 and rx_is_lockedtoref may be toggling, which suggests that the A10 RX CDR has reached the LTD (Lock to Data) state.

To assist further, I’d like to check on a few points:

Debug Suggestions:

1. Fixed Pattern Transmission

Have you had a chance to transmit a fixed pattern (e.g., 0xBC) from the C10 TX to the A10 RX? Using a fixed pattern can simplify debugging and data monitoring, and is generally less susceptible to noise compared to PRBS patterns.

2. Eye Diagram Inspection

Would you be able to use an oscilloscope to inspect the eye diagram as close as possible to the A10 RX package balls? This might help identify any signal anomalies.

Additional Checks (for completeness only, I believe you have checked these):

1. Are the reconfig_clk, refclk, and CLKUSR clocks sourced directly from free-running oscillators?

2. Please double-check the clock frequencies and signal integrity for any irregularities.

3. Also, verify the power supplies—especially those related to the transceiver—for any potential issues.

Let me know if you have any concerns or need further assistance. Thank you!

Best regards,

Chee Pin

Hi,

Just following up on my previous message—please feel free to reach out if you have any questions or concerns.

Also, I noticed that the snapshot I included earlier regarding the RX specifications might not have displayed correctly. For your reference, you can find the receiver specifications in the A10 device datasheet under Table 33: Receiver Specifications.

Let me know if you need any further clarification.

Best regards,

Chee Pin

Hi,

Thank you for your participation in this thread.

As we haven’t received a response to our previous message, we’ll now transition this case to community support. If you have a new question or need further assistance from Intel experts, please feel free to open a new thread.

Otherwise, community members are welcome to continue supporting you here.

Thank you again, and we appreciate your engagement.