Re: [Discuss-gnuradio] [Re: BER value]

[Marcus Müller]

Hi Antonio,

Could you tell me how can I solve the problem with the offset between the DPSK mod and demod?

Use a BER analysator with some kind of large window. In effect, I'd just write the raw send symbol stream to a file, do the same with the receive symbol stream, and calculate the error offline, if I hadn't had packets.

Likewise, as you say, I am using a packet-based transmission.

That makes BER a bit of a "secondary issue", doesn't it? Packets have checksums, and for all packets that contain bit errors, there's (hopefully) a high probability the checksum will be wrong, and the whole packet needs to be considered lost. Also, assuming symbol errors are equally likely for all symbols in a packet, you should be able to derive a PDF for the number of wrong bits per wrong packet, based on AWGN – yeah, it's that stuff from digital comm basics :)

Anyway, having packets makes the whole measuring a lot easier; you'll need to write a block that buffers both the send packets and the receive packets. I don't know your packet format, but it possibly has a packet number. Use that packet number to delete received packets from the send packet buffer. If a packet is broken, reconstruct its number from the next/previous successfully received packet. Note down the correct received packets as successfull bits, and count the missing and broken packets.

Best regards,
Marcus

On 17.07.2016 21:33, ANTONIO TAMAYO wrote:

Hello Martin,

Thank you for your response. 

Could you tell me how can I solve the problem with the offset between the DPSK mod and demod?

In the same way, how can I garantee that I am comparing the correct bits? Using another block?

Likewise, as you say, I am using a packet-based transmission. For this reason I need to perform more sophisticated comparisons to calculate the BER, what do you recommend me to calculate BER?

Thank you all,

Antonio

On 06/13/2016 09:46 AM, MARTIN BRAUN wrote:

>Antonio,
>
>I'm not sure why it's telling you 'no BER', but your setup is not
>suitable to test BER anyway. Your DPSK demod will always be offset w.r.t
>the DPSK mod input (because of sync delays etc.).
>
>Such BER measurements only work well if you can guarantee you're
>comparing the right bits.
>
>Since you're doing packet-based transmission, you'll need to come up
>with something more sophisticated comparison.
>
>Cheers,
>Martin
>
>On 06/12/2016 12:22 PM, ANTONIO TAMAYO wrote:
>> Hello all,
>>
>> I'm trying to calculate de BER value of the comunications system of the
>> image attached. I'm using a DQPSK modulator and demodulator. I think
>> that I have made mistakes in the configuration of my sistem because the
>> BER value always is "none", but I can't solve them. Somebody could help me?
>>
>> Regards,
>>
>> Antonio.
>>
>>
>> _______________________________________________
>> Discuss-gnuradio mailing list
>> address@hidden
>> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>>

_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio