Forum Discussion

CDai0's avatar
CDai0
Icon for New Contributor rankNew Contributor
6 years ago

Transceiver toolkit reference design: why is BER never 1 or close to 1

I have questions on the reference design for transceiver toolkit on cyclone 10 dev kit provided here: https://fpgacloud.intel.com/devstore/platform/17.1.1/Pro/cyclone-10-gx-xcvr-toolkit-reference-des...
Deshi_Intel's avatar
Deshi_Intel
Icon for Regular Contributor rankRegular Contributor
6 years ago
Hi ChenXi, The reference design setting is correct else the transceiver toolkit BER won't activate and respond at all. May I know which protocol that you are validating that required until BER > 10E-2 as most of the protocol is targeting BER passing rate around 10E-12 only ? I don't have C10 dev kit board that I did ported over the reference design to A10 SI dev kit board to check out. It's taking very long time to reach BER 10E-2 with the reference design but theoretically as long as "number of tested bit" and "number of error bit" are increasing then the BER will gradually increase as well. However, it will take longer time for the BER increment when the BER number is getting larger over time. Do you see the "number of tested bit" and "number of error bit" stop increasing ? Else the BER should still increase but in very slow pace as the BER number is getting larger. Thanks. Regards, dlim
CDai0's avatar
CDai0
Icon for New Contributor rankNew Contributor
6 years ago

Hi Dlim,

I don't want a BER>1.5E-2, my custom hardware is always showing a BER of 1.5E-2. I am wondering what this means since I observe that in Cyclone 10 dev kit, I never see BER above 1.5E-2, it maxes out at 1.5E-2, and I find that strange.

Yes, I see "number of bits tested" increase and "number of error bits" increase. The BER number does not seem to increment, it starts at 1.5E-2 on my hardware and stays there.

On the cyclone 10 dev kit, I tried without loopback and then reset checker to recalculate the BER, and it is still 1.5E-2. Since there is no physical connection, I would expect the result to be number of bits test=number of error bits for a BER of 1, but unexpectedly I get 1.5E-2.

Now the question is, is 1.5E-2 just the maximum BER for transceiver toolkit for some reason? Or is there maybe a problem with my quartus project when I ported it? From your response, it seems like if there was a problem during port, the toolkit would not activate BER at all.

When you test the transceiver toolkit, can you disconnect the loopback and reset the checker, does the BER go to 1 or 1.5E-2?

Thanks

Chenxi