Re: [Discuss-gnuradio] CPFSK mod/demod + strange behavior

[Olivier Goyette]

Excuse me for not understanding everything, but there seem to be a lack of comprehension coming from me.

1- Where is your "625 kHz" coming from ?

2- The links I provided are only the reference where I took the information for trying to build things up. I didn't include the link where it specifies the RRC and I don't have it with me now. I will feed it tomorrow.

3rd and most IMPORTANT ! Keep in mind I've never touched an USRP in my life before nor do I have skills in digital radio design. I know that it's a big project for someone as newbie as me, but we gotta learn somewhere. Now if you agree with me, instead of replying to numerous e-mails that in my opinion tend to give less useful info and more opinion on what's good or what's not good, would it be possible to start from the beginning with the parameters I have ? Does someone has ever done CPFSK and understands it ? I don't mean offense to anyone but people on forums tend to give more their opinion and it doesn't help me at all. I want to understand what's behind the wheel.

USRP N210/WBX daughtercard

baud rate = 1.041667 Mbps

sample rate = 6.25 Msps

sample/symbol for CPFSK mod = 6

modulation index (k) = 0.6

center frequency = 978MHz

delta F for CPFSK = +- 312.5 kHz

roll-off factor for the Root Raised Cosine filter = 0.5

data is encoded using Reed-Solomon

thank you

Olivier

2016-06-09 19:28 GMT-04:00 Marcus Müller <address@hidden>:

Hi Olivier!

On 10.06.2016 00:05, Olivier Goyette wrote:
>
> You are wrong !
>
That being the case very often, I think the error's on your side, this
time :)
So, I will need you to *explain* how you can transport 1.somethingMb/s
through a 625kHz wide channel if every time you use that channel to
transmit a symbol, only one bit can get sent. How do these numbers add
up for you? You've got a serious problem with your pulse shaping, there.
>
> 1 sample/symbol is when I run simulation (no transmission over the air).
>
Yes, and as said, that's hardly a FSK. FSK is "I encode data by changing
the frequency. Frequency is the derivate of phase, or in the discrete
case, the phase increase between samples"; in your 2-CPFSK, you either
increase the phase of the transmit signal, or you decrease it by exactly
the same amount (assuming you send -1 and 1 bytes).
Now, with sps=1, you change that phase increase every sample – that's
nothing but a bad differential BPSK that you've involuntarily built.
>
> When I want to transmit over the air ( cable between Tx and Rx ) I
> need to use 6 samples/symbol.
>
> This is what I've told earlier, 1 sps is the only way I could send a
> message and retrieve it at the end of the transmission chain by doing
> a simulation. If I use a sps higher than 1, I can't find a way to get
> my message back
>
Well, and it's absolutely no use to try and do that over the air if the
simulation doesn't work with 6 sps.
As said, the real world usually makes things harder, not easier, and
things that don't work in simulation rarely work out in the real world
through magic.

You cited a word document with links (that's seriously the worst format
you can transport hyperlinks in, by the way). Which one of those
contains the standard that says you have to do an RRC on the FSK
modulated signal? I presume, and Achilleas seems to agree on that, that
there might be a misunderstanding here. Are you sure the RRC goes
*after* the FSK? As I mentioned in my last mail, it's not a good idea to
low pass filter after an FSK modulation. That just seems counter-productive.

Best regards,
Marcus

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