ARQ HF Path Simulations; Success!


Tony
 

All:

After much experimentation, I am now able to run HF path simulations with Vara, ARDOP and other ARQ modes. The setup consists of 2 PC's along with VAC to route the audio between the path simulator and various digital mode apps. Transformers on all cables eliminated a mild ground loop issue which made for an very clean signal.

There are some limitations to this setup due to VAC latency: I managed to work around it for now. I'm currently gathering some preliminary data which I'll post ASAP.

Tony -K2MO


Andrew O'Brien
 

Great news .. my predictions are that under tough conditions ARDOP fairs slightly better . Under good conditions VARA wins easily .

Andy K3UK.

On Mar 13, 2018, at 8:29 PM, Tony <DXDX@optonline.net> wrote:

All:

After much experimentation, I am now able to run HF path simulations with Vara, ARDOP and other ARQ modes. The setup consists of 2 PC's along with VAC to route the audio between the path simulator and various digital mode apps. Transformers on all cables eliminated a mild ground loop issue which made for an very clean signal.

There are some limitations to this setup due to VAC latency: I managed to work around it for now. I'm currently gathering some preliminary data which I'll post ASAP.

Tony -K2MO





Rick Muething
 

Great work Tony.

Are you using the HF simulator Ionospheric Simulator by Michael DL6iak?  He recently made an update for me (version 1.8) that allows you to save up to 4 commonly used setups which is helpful in running over 4 standard paths.

Could you send me or post a simple sketch of your setup?  I can supply you one that I have been working with using the HF simulator above.

Thanks,

Rick Muething, KN6KB  rmuething@cfl.rr.com

On 3/13/2018 8:29 PM, Tony wrote:
All:

After much experimentation, I am now able to run HF path simulations with Vara, ARDOP and other ARQ modes. The setup consists of 2 PC's along with VAC to route the audio between the path simulator and various digital mode apps. Transformers on all cables eliminated a mild ground loop issue which made for an very clean signal.

There are some limitations to this setup due to VAC latency: I managed to work around it for now. I'm currently gathering some preliminary data which I'll post ASAP.

Tony -K2MO





Tony
 

Rick:

I noticed some audio level variation with Ionos 1.8 which was enough to cause inconsistent test results. Switching to PathSim solved the problem, but it does tend to overload during the first transmission. That issue disappears with subsequent transmissions once the test is underway so it appears to be an "AGC" issue. As far as Ionos is concerned, it could be the old laptop I'm using.

> Could you send me or post a simple sketch of your setup?

I don't have a sketch, but it does eliminate the need for an audio mixer. Basically, the modems output feeds the path simulator via VAC. Windows mixer is then used to mute VAC's output and enable the path simulators output which is then sent to the other PC. I can put something together for you if my feeble attempt of explaining didn't help : ). 

I'd like to see the sketch of your setup when you get a chance.

Thanks,

Tony -K2MO



On 3/14/2018 7:16 AM, Rick Muething wrote:
Great work Tony.

Are you using the HF simulator Ionospheric Simulator by Michael DL6iak?  He recently made an update for me (version 1.8) that allows you to save up to 4 commonly used setups which is helpful in running over 4 standard paths.

Could you send me or post a simple sketch of your setup?  I can supply you one that I have been working with using the HF simulator above.

Thanks,

Rick Muething, KN6KB  rmuething@...



On 3/13/2018 8:29 PM, Tony wrote:
All:

After much experimentation, I am now able to run HF path simulations with Vara, ARDOP and other ARQ modes. The setup consists of 2 PC's along with VAC to route the audio between the path simulator and various digital mode apps. Transformers on all cables eliminated a mild ground loop issue which made for an very clean signal.

There are some limitations to this setup due to VAC latency: I managed to work around it for now. I'm currently gathering some preliminary data which I'll post ASAP.

Tony -K2MO













Tony
 

my predictions are that under tough conditions ARDOP fairs slightly
better.
Under good conditions VARA wins easily. Andy K3UK.
I think that's a fairly accurate prediction Andy. My preliminary HF path
simulations show that VARA and ARDOP do very well under good conditions
with each reaching 1/3 to 1/2 their respective top speeds. With VARA
being something like 3 x faster than ARDOP, the outcome is predictable.

Under poor conditions, ARDOP obviously has the ability to switch to more
robust, narrow-band protocols while VARA takes a different approach: it
adapts to conditions, but it does so using the same wide-band,
multi-carrier mode that generally requires good s/n ratios.

As you can see from the examples below, the differences show up on the
simulator. In the first example, each mode is subjected to a good path
model with a strong +16db s/n ratio.

Path: CCIR Good / +16db s/n

VARA:

*** MBXPYEDWR1ZU - 30897/27553 bytes received
*** Bytes: 27790,  Time: 01:43,  bytes/minute: 16172
Total session time: 02:41,  bytes/minute: 10349

ARDOP:

*** 3MR6VK3I1DVZ - 30897/27550 bytes received
*** Bytes: 27787,  Time: 06:23,  bytes/minute: 4348
Total session time: 07:30,  bytes/minute: 3699


In the next example, the speed-gap closes as the s/n ratios drop from
+16db to +6db while using the same path model.

Path: CCIR Good / +6db s/n

VARA:

*** 233E6HTZC6ZP - 2887/1377 bytes received
*** Bytes: 1409,  Time: 01:42,  bytes/minute: 824
Total session time: 02:39,  bytes/minute: 529

ARDOP:

*** 3843RY160W7H - 2895/1371 bytes received
*** Bytes: 1411,  Time: 00:42,  bytes/minute: 1979
Total session time: 01:54,  bytes/minute: 737

Both modes are fantastic, but as you point out by observation, there are
some differences. I think the lesson here is to chose the right mode for
the right conditions.

Tony -K2MO


Graham
 

Tony , 

To feed this back, do  you  have a  screen  shot  of  VARA , with  just the  audio  routed via the simulator  and   NO path  
restrictions , to  give a  base line and  at the  worst  data  rate, the same shot , showing  waterfall and the  constellation plot  ?

Tnx-Graham
G0NBD


Tony
 



To feed this back, do  you  have a  screen  shot  of  VARA , with  just the  audio  routed via the simulator  and   NO path  restrictions , to  give a  base line and  at the  worst  data  rate, the same shot , showing  waterfall and the constellation plot? Tnx-Graham G0NBD

I can certainly do that for you Graham. A direct path with a high signal-to-noise ratio will yield the fastest data rates and picture-perfect constellations. A single screenshot will show the modem at full-throttle with a specific constellation for that speed. If you want to see the constellations in their various configurations, I can record a short video that shows how they change as the modem shifts from level 1 to top speed.   

The slowest data rate via direct path would require a s/n ratio of +3db. At that s/n level, VARA will stay at 113 bps. The the constellation stays the same since there is no change in data speed.

Tony -K2MO

 



Andrew O'Brien
 

To average Pactor 3 type speeds in VARA , what SNR do we need to be achieving ? 


On Mar 17, 2018, at 9:28 PM, Tony <DXDX@...> wrote:



To feed this back, do  you  have a  screen  shot  of  VARA , with  just the  audio  routed via the simulator  and   NO path  restrictions , to  give a  base line and  at the  worst  data  rate, the same shot , showing  waterfall and the constellation plot? Tnx-Graham G0NBD

I can certainly do that for you Graham. A direct path with a high signal-to-noise ratio will yield the fastest data rates and picture-perfect constellations. A single screenshot will show the modem at full-throttle with a specific constellation for that speed. If you want to see the constellations in their various configurations, I can record a short video that shows how they change as the modem shifts from level 1 to top speed.   

The slowest data rate via direct path would require a s/n ratio of +3db. At that s/n level, VARA will stay at 113 bps. The the constellation stays the same since there is no change in data speed.

Tony -K2MO