Birdies on the RX888?

[Philip Gladstone]

I finally got the web interface to ka9q-radio working and I can see birdies on 10kHz multiples:

Do other people see this? The fact that they are so precisely on 10kHz multiples makes me think that they are either coming from the rx888 itself, or are some type of decoding artifact. 

Any thoughts?

Philip

[Clinton Turner]

Hi,

The first test that needs to be run is to view the waterfall with NOTHING connected to the '888 other than the USB cable - no antenna, no external clocking - and then use this as your "baseline" reading.

Everything that appears on the waterfall with other cables connected would be conducted into the '888 via those cables.  It's important to remember that a coaxial cable has THREE conductors:

• The center conductor.
• The inside of the shield
• The outside of the shield.

It's nearly an inevitability that undesired RF currents will flow on the outside of the cable, and - either due to imperfections in the coax (lack of true, 100% absolute shielding, ohmic losses in the cable/connector) or those currents being conducted to the antenna and from there, onto the body of the antenna itself and then into the coax - will find their way back into the receiver.

With the '888 - or pretty much any receiver, really - you will have current flowing through the '888 and onto the coax shield - from whatever else to which it is connected, and in the case of the '888, it is going to be a (likely) very noisy computer+power supply.  If you add an external clocking source, you have yet another possible source of circulating RF currents on the antenna coax shield getting directly into the cable or being conducted out to the antenna and back.  To mitigate this - albeit imperfectly - you have several choices:

• The addition of a common-mode choke.  These are typically 10-15 turns wound on an FT240-31, FT240-43 or FT140-43 (the latter being practical with small diameter coax like RG-316 or RG-174) to introduce a high common-mode impedance, which can reduce the common-mode current by several orders of magnitude at some frequencies.
• The addition of an RF amplifier.  Unless you live in an extremely noisy RF environment, you'll need to add 15-20dB of RF gain in front of the '888 to reach the noise floor of 10 meters under quiet conditions.  By placing this amplifier somewhere "quiet" (e.g. on the coax feed, closer to the antenna than the '888) it can amplify signals prior to the shield being conducted with common-mode noise and override those signals and that which might be ingressing for the imperfections in the coax cable.

I hope that this helps.

73,

Clint, KA7OEI

[Philip Gladstone]

It turned out that these birdies were caused by (I think) an overload in the rx888. When I added a 6db attenuator in the antenna feed, then the birdies were much reduced. I actually moved the ka9q-radio gain from 30 to 15dB (and removed the attenuator) and the problem seems basically resolved.

I think I do need something to block the AM band as well.

Thanks.

--
You received this message because you are subscribed to the Google Groups "ka9q-radio" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ka9q-radio+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ka9q-radio/71c9693e-d739-4c1a-ab8c-de22712e8aa8n%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Rob Robinett]

I have found two potential sources of overloads in the RX888 + KA9Q-radio systems:

1). ADC overloads as reported in the "Overrange" field in 'control'.  They should be .1% or less of the total sample counts on the line above it

     Clint KA7OEI has established that a FE gain setting of +20 dB is optimal for the RX888 noise figure.  Above +20 you are amplifying internal noise as much as the input signals. 

      Never attempt to use the internal attenuator, it will only impair the already poor 20 dB NF of the RX888 unless there is a (highly recommended) LNA ahead of the RX888.

      If the LNA produces overloads, then one should reduce the FE Gain.

2) Decoding gain (i.e. AGC)

     if AGC is on, the KA9Q will ensure that there are no full scale 16 bit samples in the PCM output stream.

    If AGC is off, then you need to monitor the PCM stream for full scale samples and adjust the AGC gain appropriately.

I have found at over 20 RX888 sites in the US it is rarely necessary to reduce the FE gain even when there is a 15 dB LNA ahead of the RX888.  Of course sites vary widely.

But WD runs KS9Q  with AGC off so it can measure diurnal variation in background noise level,  So in some extreme cases where there are nearby transmitters I have needed to adjust the decoding gain from the default +60 dB down to 20-30 dB.   Leave AGC on and you shouldn't need to worry about PCM overloads.