Re: [Discuss-gnuradio] Help understanding AX.25 encoding and tagged streams

[Cinaed Simson]

On 08/26/2017 05:36 PM, Robert McGwier wrote:
> At one time rational resampling meant that the interpolation and
> decimation were each integers so the rate change is the ratio of two
> integers and thus rational.

All integers are rational. But not all rationals are integers.

1,2,3, are integers and rational.

The quotient of rationals is rational.

Hence

  1/2 = .5000000......
  1/3 = .3333333......
  2/3 = .6666666......

are all rational.

> 
> If one or more of your inputs is fractional>

I can input 1/2 in the rational resampler which satisfies the "rate
change is the ratio of two integers and thus rational" - and the machine
will reduce it to .5000000....... - which is also rational.

> I believe you need the
> arbitrary rate resampler.

I've never used an arbitrary or fractional resampler.

Maybe by an arbitrary resampler they mean it can used with either
rational or irrational resampling ratios?

And maybe a fractional resampler is to make a small adjustments to
sampling rate - like on the order of a couple of percent?

-- Cinaed

> 
> Bob
> 
> 
> On Aug 15, 2017 10:16 PM, "Cinaed Simson" <address@hidden
> <mailto:address@hidden>> wrote:
> 
>     On 08/14/2017 07:14 AM, Jack White wrote:
>     > Hello, knowledgeable ones,
>     >
>     > I need to put together testbed satellite uplink using AX.25
>     encoding and
>     > GMSK modulation. The flowgraph source is an Ettus USRP X310 with a
>     > WBX-120 daughterboard.
>     >
>     > The flowgraph is a modification of one used with a USRP B210. That
>     radio
>     > had handle different sample rates to the X310, so some level of
>     > resampling is required - in this case from the proposed 614.4ksps to
>     > 400ksps.
>     >
>     > I understand the premise of constructing the AX.25 packet, but I'm
>     > having real difficulty understand tagged streams and how the
>     tagging and
>     > the PDU format is affecting my transmission.>
>     >
>     > ​
>     > ​
>     > ​
>     > The disabled low-pass filter and the enabled and encircled rational
>     > resampler are my additions. The AX.25 line coding hierarchical
>     block is
>     > below.
>     >
>     >
>     > ​
>     > ​So, when I run the flowgraph and try to set a command from the
>     > groundstation, I get a huge load of buffer underrun errors in the GRC
>     > log:
>     >
>     
> UUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU
>     >
>     
> UUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU
>     >
>     
> UUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU
>     >
>     
> UUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU
>     >
>     > If I try sending the command again, I get a tP error, followed by more
>     > underruns.
>     >
>     > In my head, it looks like the USRP is expecting a combined sample of a
>     > certain length, but it doesn't get it. The buffer is more than half
>     > full. Then the next packet comes in, of the same length, and you
>     get one
>     > buffer filled with half of one packet and half of another.
>     >
>     > The USRP /is/ transmitting something, because I checked the
>     transmitting
>     > antenn400k/614.4a on a spectrum analyser.
>     >
>     > Can anyone help me understand the flowgraph, especially the tagged
>     > streams and PDU input, and get this working?
>     >
>     > Cheers,
> 
>     I don't really understand what you're doing but I have a couple of drive
>     by comments.
> 
>     The produce of 2 rationals is always rational - and the quotient of two
>     rationals is always rational.
> 
>     I would replace the 2 serial rational resamplers with one, namely,
> 
>       (8/1)*(400000/614400)=(8/1)*400/614.4) = 5.208333333333333/1
> 
>     or
> 
>       15.625/3
> 
>     That is, interpolate by 15.625 decimate by 3.
> 
>     Since the sampling rate of the sink is 400000 Hz, then the sampling rate
>     coming out of the GFSK Mod needs to be 76800 Hz.
> 
>     And the baud rate out of the GFSK Mod would then be 76800 Hz/8 sps =
>     9600.
> 
>     If reducing the 2 serial rational resamplers to one doesn't help,  I
>     would poke the front side of GFSK Mod - add a throttle in front of the
>     GFSK and set it to 76800 Hz - see what happens- not recommend for
>     general use.
> 
>     And I would add a Message Debug to Tagged Stream Multiply Length Tag and
>     set the Length scalar to 1 - 512 times the packet_length seems high -
>     but then I don't know the packet_length either.
> 
>     -- CInaed
> 
>     >
>     > --
>     > Jack White
>     > address@hidden <mailto:address@hidden>
>     <mailto:address@hidden
>     <mailto:address@hidden>>
>     > 07875 813 745
>     >
>     >
>     > _______________________________________________
>     > Discuss-gnuradio mailing list
>     > address@hidden <mailto:address@hidden>
>     > https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>     <https://lists.gnu.org/mailman/listinfo/discuss-gnuradio>
>     >
> 
> 
>     _______________________________________________
>     Discuss-gnuradio mailing list
>     address@hidden <mailto:address@hidden>
>     https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>     <https://lists.gnu.org/mailman/listinfo/discuss-gnuradio>
>