Replacing the MiniTiouner

[g0mjw]

Analogue would be mission creep. There are plenty of solutions already for that.

Mike

[g8gtz]

Great to see this design progressing - thanks Martin.

I agree not to put an FM demod on the board but providing an IF tap out after the 1st converter could be useful to feed an SDR or an FM demod.

73
Noel - G8GTZ

[g0mjw]

I agree not to put an FM demod on the board but providing an IF tap out after the 1st converter could be useful to feed an SDR or an FM demod.

Certainly, we will need it for testing if nothing else, but band viewers etc would also benefit from an IF output.

[M0LNG]

An IF output would be a useful feature, but unless some amplification is added in front of the mixer, the IF signal may be quite weak, and splitting it would degrade it further, so some thought needs to go into how to do this.

In the band viewer case both IF branches would be needed simultaneously, which rules out a simple 2-way switch.

If an RF amp is used ahead of the frontend mixer, there would be plenty of IF power and a passive splitter could be used at the IF side, but there could be weird effects if this were combined with IF filtering on either or both branches.

One solution would be a 2-way active splitter after the mixer, such as ADA4303-2. This would have the advantage that both branches could be used simultaneously with different filtering arrangements, allowing a tight filter to be used at the demodulator but with full bandwidth available on the IF port.

Edit: but that splitter is not good for the frequency range, and there's a shortage of better options...

[G8GKQ]

Hi Martin

This looks very promising.

We may need to get some tests done by an SF8008 owner to check the Si2166D performance when demodulating a 333 kS signal with 2 other equal-strength 333 kS signasl 500 kHz spaced either side. Poor performance in this situation is what made the previous BATC-supplied Sharp tuner (using an STV0903 demodulator) unusable on QO-100. The SF8008 also suffers from the issue of 1 MHz minimum tuning resolution which complicates testing this performance parameter.

The MAX2120 also seems to suffer from 1 MHz minimum tuning resolution which makes the RFFC5072 pre-mixer essential to provide fine tuning. I do think it would be worth at least putting pads on the PCB for some gain in front of the RFFC5072. This would enable an optional RF tap on the output of the RFFC5072.

Could any SF8008 owners out there advise on the results of the test above?

Thanks

Dave

[M0LNG]

There's a firmware update which enables 1kHz tuning resolution on the SF8008, it's discussed in the AMSAT-DL thread.

Hopefully being able to input the exact frequency will avoid the demodulator trying to jump to another signal. If it needs further encouragement, it may be possible to help it by adding some tighter filtering on the baseband side.

[g0mjw]

I have an SF8008, but have no idea how to check that.

[g0mjw]

There's a firmware update which enables 1kHz tuning resolution on the SF8008, it's discussed in the AMSAT-DL thread.

Hopefully being able to input the exact frequency will avoid the demodulator trying to jump to another signal. If it needs further encouragement, it may be possible to help it by adding some tighter filtering on the baseband side.

It didn't when I tried it in 2019. MAybe there is a newer version?

[M0LNG]

I suspect there may be an easy software fix for the jumpy locking behaviour seen with the Si2166.

The header files in the reference code give quite a lot of detail about the available settings on the chip. I had a look through them and the DVBS2_AFC_RANGE setting looks like the obvious culprit.

This is a value in kHz which as I understand it sets the frequency range the demodulator will auto-correct by. The default is given as 4000kHz, and other settings used in the reference code are 2000 to 5000 kHz. Obviously these figures are quite reasonable for broadcast signals but inappropriate for RB DATV.

This value is also used in calculating what LPF cutoff to configure, here in the `Si2183_L2_lock_to_carrier` function.

Note that these details reference the Si2183, but this code is what the Si2166 driver in the SF8008 uses, as it is a closely related part.

It should be possible to patch the SF8008 firmware to reduce this value, but it will take some reverse engineering and faffing around.

[g0mjw]

IF we can get the Octagon to work properly, that's the problem solved for most - until it goes out of production.