[RX888 mk II / rx888_stream] Parameter tweaking

[hb3xdc]

rx888_stream -f ./rx888_stream/SDDC_FX3.img -s 64000000 -g 90 -m low -d --pga -o

You cannot view this attachment.

Does anyone have experimented already with the parameters in rx888_stream application, and know which would maybe be interesting to tweak a bit?
Which sample rate is recommended to not overwhelm rx888's USB controller?

[Bas ON5HB]

I have removed -d and put -r in place.

As -d adds a bit of noise to the signal, where randomizing adds nothing and has the same effect.

It's a function of rounding a sample either being high or low.

E.g. if a sample is near between value 3025 and 3026, then the default is to round down to 3025.
With randomize it random round up or down....typical the best.

With dithering it looks at other samples to decide to go up or down.

Because we have 16bit, the function is best off or random...not dithering.

I also read that the lowpass-filter is just 30MHz, so with 64MPSP you may remove it.
As such I use 60MSPS

I prefer random.

[hb3xdc]

Did you keept gain on 90 and gain mode on low?

[Bas ON5HB]

High mode works weird....my gain is 60 and low mode...more then enough signal.

Maybe high mode works on VHF+, not seen it working well on HF.

[hb3xdc]

Just changed it on mine, looks like a bit clearer as you said with -r
also switched to 60MSPS, guess I will have to get an FM BC filter as on uppper part some FM Stations come through

[Bas ON5HB]

Are you sure it's not AM stations? As FM is difficult to mix with other signals, apart from overloading.

But overload is just remove some gain until it stops.

[hb3xdc]

http://websdr.hb3xdc.ch:8074/
You can see it at close to 28MHz, to me it looks like a FM BC

[Bas ON5HB]

Yeah it look like a radio station, FM Wideband, probably 3M with a multiplier/mixer to get at 3M band.

If you receive that, and 3M band blocker is the first thing I would do.

Else reduce the gain...10 points....your big antenna may get TOO much signal....try gain of 60....see what happens.

SDR receivers should never be set high in gain, as they have about 140~120db range...you can correct the waterfall and s-meter if not good. But start at LOW gain with weak stations....focus there.

Try gains until you lost very weak stations...then up a little. Then you know your optimal gain.

[hb3xdc]

Good hints Bas.
Went down to -g 40 now for testing, still there on 28MHz, now with gain on 60 its bit weaker than on gain 90

[hb3xdc]

Just added an external LPF Filter, now its gone

[Bas ON5HB]

May work...

[linkz]

I also have images of 34/35MHz signals, transposed to ~24MHz
I have a 30MHz LPF from SV1AFN but sigs are too strong and too close in frequency (almost Line-Of-Sight with repeaters)

But the good thing is that you can hear the Lyon-Marseille "Motorway of the Sun" A6 activity on this HF receiver, HI

Regards
linkz

[magicint1337]

To get high gain mode working you can use -g 20 and put the attenuator to like 10, it wont reduce signals more of an issue with the drivers i guess. You can try i use it.

So

-g 20 -a 10

[Martin G8JNJ]

I have been experimenting with the parameters, but many don't seem to behave as expected.

When using HF, the gain is too high for my liking, even with a value of 10, I see and hear spurious products. I also think this is why the AGC is problematic.

In particular

-m high gain mode is too high, which is to be expected

-g values below 10 seem to result in excessive gain, values above 10 seem to work as expected.

-a value of 10 or higher results in excessive gain, this is not as expected, as larger values should reduce signal levels.

--pga how do you activate that ? I have tried - 1 and true, but all produce an error.


Using the RX888 with HDSDR, I can adjust the RF and IF gain separately, and this seems to produce better results than Phantom.

Regards,

Martin

To get high gain mode working you can use -g 20 and put the attenuator to like 10, it wont reduce signals more of an issue with the drivers i guess. You can try i use it.

So

-g 20 -a 10

[Bas ON5HB]

PGA is a toggle, --pga activates it, without it; it's disabled.

See code:

Code Select Expand

/// PGA enable
    #[arg(long, global = true, default_value_t = false)]
    pga: bool,

Combined with --pga you use; -d (dithering) and/or -r (randomize) rounding to reduce sample-errors.
Just test what works the best for you, but they don't work if --pga isn't used as far as I understand it.

As for the gains:

Code Select Expand

/// VGA gain setting 0-127
    #[arg(short, long, global = true, default_value_t = 1, value_parser = value_parser!(u8).range(0..=127))]
    gain: u8,

    /// VGA gain mode high or low
    #[arg(short = 'm', long, global = true, default_value = "high")]
    gain_mode: GainMode,

    /// Attenuator setting 0-63
    #[arg(short, long, default_value_t = 0, value_parser = value_parser!(u8).range(0..=63))]
    attenuation: u8,

The 0db gain isn't at value 0 as far as I understand it, it seems to go 2 ways....see here:

https://swling.com/blog/2021/02/new-rx-888-mk-ii-sdr-receiver-on-the-way/

1. Add a tunable (variable) attenuator in the HF path, which can tune from 0 to -31.5dB.

2. Change the fixed LNA of RX888 to a VGA, which gives the -10dB to +33dB range. VGA applies to both HF and VHF.

The LNA can attenuate as well as amplify the signal.
If that is the case, I think so, it explains your 'weird' findings with the setting.

As for the attenuator, I noticed it too, I presume it's enabled by default and reduced as you give numbers?

As for Phantom, Phantom doesn't control any of these values, it just receives samples from the RX888STREAM-driver. It's piped from the driver into Phantom without any control by the last one.