Topics

Forward error correction - experiment


"g7ork" <del_swt@...>
 

FEC can provide significant improvements in overall data speeds in a mobile environment - but it is only one factor that needs to be considered. Modulation mode, symbol rate, interleaving etc also need to be taken into account, and the "best" solution for one set of circumstances could turn out to be absolutely useless in others. In the end it will always be a compromise, however I am firmly of the opinion that we can do a lot better than what we have today.

It would be worth your time reading some of the later ETSI standards for 4FSK systems - TS 102 490 for instance. The relatively low symbol rate makes it better in multi-path conditions, but still keeps a half decent data rate by using 2 bits per symbol. In this particular case, some improvement in rx sensitivity is also gained by having a 6.25kHz channel spacing and hence a smaller noise bandwidth. TETRA's 4piDQPSK scheme is even better in some respects, but suffers in a ham market by requiring a linear PA and Rx system. 8FSK with trellis coding could be interesting as well....

Lots to think about.....

--- In UniversalDigitalRadio@..., Marshall Denny <MarshallDenny@...> wrote:

I will have a look at FX.25.

But, with the flexibility that the UDR will give us. Lets keep all good
ideas on the table until we have a good reason to remove it.

The potential of FEC is not will known in the ham community.

For example, compare the potential for operating in noisy conditions
between psk31 and olivia.
Olivia will operate below the noise floor.
I have worked Olivia on 80m with thunderstorms in the region with 50 over
S9 static crashes with 100% copy.

If well done, this could allow very poor signals to get through at low
speed.
And, also allow 56k packet under conditions that would not work otherwise.

Respectfully,
Marshall
AI4CM

On Wed, Mar 27, 2013 at 4:18 AM, zl2wrw <ross@...> wrote:

**




--- In UniversalDigitalRadio@..., "marshalldennyai4cm"
<MarshallDenny@> wrote:

Testing with variable levels of FEC added to an otherwise standard ax.25
protocol.

That sounds very much like the FX.25 protocol:
http://en.wikipedia.org/wiki/FX.25_Forward_Error_Correction

It's even backwards compatable with plain old non-FEC AX.25 equipment.
Though it does not rectify some of the other historical defects of the
AX.25 protocol that are due to the adapting of wireline X.25 to a wireless
physical layer...

73 ZL2WRW
Ross Whenmouth




--
Respectfully,
W. Marshall Denny II
Software Development Engineer
206 734 9242 cell

For if you altogether hold your peace at this time,
then shall relief and deliverance arise to the Jews from another place;
but you and your father's house shall be destroyed:
and who knows whether you are come to the kingdom for such a time as this?
Esther 4:14 KJV


Marshall Denny <MarshallDenny@...>
 

I will have a look at FX.25.

But, with the flexibility that the UDR will give us.  Lets keep all good ideas on the table until we have a good reason to remove it.  

The potential of FEC is not will known in the ham community.

For example, compare the potential for operating in noisy conditions between psk31 and olivia.  
Olivia will operate below the noise floor.
I have worked Olivia on 80m with thunderstorms in the region with 50 over S9 static crashes with 100% copy.

If well done, this could allow very poor signals to get through at low speed.
And, also allow 56k packet under conditions that would not work otherwise.

Respectfully,
Marshall
AI4CM 

On Wed, Mar 27, 2013 at 4:18 AM, zl2wrw <ross@...> wrote:
 



--- In UniversalDigitalRadio@..., "marshalldennyai4cm" wrote:

> Testing with variable levels of FEC added to an otherwise standard ax.25 protocol.

That sounds very much like the FX.25 protocol:
http://en.wikipedia.org/wiki/FX.25_Forward_Error_Correction

It's even backwards compatable with plain old non-FEC AX.25 equipment. Though it does not rectify some of the other historical defects of the AX.25 protocol that are due to the adapting of wireline X.25 to a wireless physical layer...

73 ZL2WRW
Ross Whenmouth




--
Respectfully,
W. Marshall Denny II
Software Development Engineer
206 734 9242 cell

For if you altogether hold your peace at this time,
then shall relief and deliverance arise to the Jews from another place;
but you and your father's house shall be destroyed:
and who knows whether you are come to the kingdom for such a time as this?  Esther 4:14 KJV


"k7udr" <bhhoyer@...>
 

The problem with FX.25 is that in order to be backwards compatible they bundle the FEC bits at the end of the packet. As no-one to my knowledge has actually deployed FX25, I prefer a clean sheet approach so we can look at not only FEC, but other choices of modulation and encoding as well.

Backwards compatibility can be assured by connecting on either AX25 or D-STAR, then negotiating upwards amongst capable radios.

After we have the UDR shipping, we'll be starting a new independent group, sponsored by TAPR, to investigate open methodologies for advancing amateur radio.

Exciting times ahead!
Bryan K7UDR

--- In UniversalDigitalRadio@..., "zl2wrw" <ross@...> wrote:



--- In UniversalDigitalRadio@..., "marshalldennyai4cm" <MarshallDenny@> wrote:

Testing with variable levels of FEC added to an otherwise standard ax.25 protocol.

That sounds very much like the FX.25 protocol:
http://en.wikipedia.org/wiki/FX.25_Forward_Error_Correction

It's even backwards compatable with plain old non-FEC AX.25 equipment. Though it does not rectify some of the other historical defects of the AX.25 protocol that are due to the adapting of wireline X.25 to a wireless physical layer...


73 ZL2WRW
Ross Whenmouth


"zl2wrw" <ross@...>
 

--- In UniversalDigitalRadio@..., "marshalldennyai4cm" <MarshallDenny@...> wrote:

Testing with variable levels of FEC added to an otherwise standard ax.25 protocol.

That sounds very much like the FX.25 protocol:
http://en.wikipedia.org/wiki/FX.25_Forward_Error_Correction

It's even backwards compatable with plain old non-FEC AX.25 equipment. Though it does not rectify some of the other historical defects of the AX.25 protocol that are due to the adapting of wireline X.25 to a wireless physical layer...


73 ZL2WRW
Ross Whenmouth


"marshalldennyai4cm" <MarshallDenny@...>
 

I am proposing a useful experiment for this very flexible hardware for the group to consider.

Testing with variable levels of FEC added to an otherwise standard ax.25 protocol.

This might allow high speed 56k packet under conditions where it could not be used otherwise.

This might also allow higher speeds mobile that would otherwise be practical.

Respectfully,
Marshall