All,

Interesting behaviour noticed today whilst monitoring closely the receivers over the solar outage.

Of particular interest to me was a change in behaviour from before with the DD DuoFlex S2 V4A. I have no manual MODCOD filtering in the driver despite it having the same chip as a TBS 6903 due to the way the mods to each have been done (ie the ddbridge driver references the original stv0910 file and not 091x and enough changes have been made you can't just point it to the other!)

BAS kept going right down to 5.9 dB C/N without losing packets. I was most surprised! It missed a few around the threshold for HVS-1 however; but the interaction between HVS-1 and HVS-2 happens around there. I guess I will be testing that tomorrow with HVS-2 disconnected to see how it behaves then; whether the lost packets cannot be avoided or they are a part of the interaction at the LDPC/FEC stage with the other service.

 

I'm 99% this was not the behaviour before and points to increased resilience from the new uplink config.

 

Charlie

On Tue, Feb 28, 2023 at 03:44 AM, Charlie wrote:

I'm 99% this was not the behaviour before and points to increased resilience from the new uplink config.

Charlie,

do I understand that right? You basically say that MODCOD filters for C4 on ST-chips might become obsolete?

Regards,
Ernst

Ernst,

Just did the test manually (disconnected the second feed so we are only looking at C4) and used the tripod dish so I could readily de-point it and force the SNR.

No losses right to 9.3 dB C/N, if you dip below that you see losses on both for 0.5 dB or so, in the same way as the HVS-1/2 interaction seems to work on it. Once you are clear of that point it isn't trying to get the 16APSK signal, HVS-1 fully drops off and BAS runs cleanly with no losses right down to 5.9 dB.

 

Am I right in saying this is roughly similar behaviour to the existing driver with the threshold MODCOD filtering, or does that cut cleanly at 9.3 dB and not have that weird "crossover"?

Charlie

 

 

On Tue, Feb 28, 2023 at 05:29 AM, Charlie wrote:

Am I right in saying this is roughly similar behaviour to the existing driver with the threshold MODCOD filtering, or does that cut cleanly at 9.3 dB and not have that weird "crossover"?

Charlie

Charlie,

under GNU/Linux (for ST chips) there is no threshold value. MODCOD filters are either on or off. The TBS-6909X driver has an additional possibility I never used. It is supposed to cut off automatically any MODCOD all the way down looking at the SNR. Not sure what BDADataEx can do under Windows. The SR1 has two threshold values that you can set (IIRC given as deltas). SNR below Nr one it cuts HVS-1 off. SNR must again mount above Nr. two to switch on HVS-1 (hysteresis). No idea how the Novra-300 handles MODCOD filters.

Regards,
Ernst

Charlie,

what I said about GNU/Linux is of course for TBS cards/boxes using the CrazyCat enhanced stv090x.c and stv091x.c drivers. As you use a patched stv0910.c DD version I wonder whether this has some kind of automatic MODCOD filter built in (like the TBS6909X driver, the option I never used). All our experience so far with TBS cards/boxes and ST chips (and AFAIK including the SR1 that uses the same chip as the TBS-69/5925) is that if you do not set a MODCOD filters at low SNR then HVS-1 will finally also tear down completely BAS.

Regards,
Ernst

Charlie,

What I will try next is the TBS-6909X on C4 without the ts_nosync.
This driver has it as a module parameter. I expect no change for the
SkyStar2 eXpress. AFAIK it only understands 8psk anyway. But IIRC
I had to set ts_nosync for stv090x as well. The stv091x driver has
automatic ts_nosync you probably copied for your stv0910 DD driver.

Ernst

Am 2023-02-28 19:56, schrieb Charlie: