[Skimmertalk] red pitaya noisy

[Nathaniel Frissell]

Hi, Dave. Thank you very much for this work. It is very encouraging to see that it is possible to make the RP have similar sensitivity to the QS1R.

I’m working on getting some more RBN nodes set up here at NJIT (Newark, NJ), and Blairstown, NJ. So, hopefully I will be able to try these things out in person fairly soon! (a month or two, maybe?)

73,
Nathaniel, W2NAF


> On Jul 31, 2016, at 9:03 PM, Pete Smith N4ZR <n4zr at contesting.com> wrote:
> 
> I believe that the schermatic shows that there is an amplifier with a gas protection tube right after the jumpers.
> 
> 73, Pete N4ZR
> Download the new N1MM Logger+ at
> <http://N1MM.hamdocs.com>. Check
> out the Reverse Beacon Network at
> <http://reversebeacon.net>, now
> spotting RTTY activity worldwide.
> For spots, please use your favorite
> "retail" DX cluster.
> 
> On 7/31/2016 5:10 PM, W3OA wrote:
>> 
>> Hi Dave -
>> 
>> This looks like great progress in using the Red Pitaya.  Thank you.
>> 
>> But I'm concerned because of the following FAQ on http://redpitaya.com/faq-page/
>> 
>> Do you think it would be advisable to provide some external protection circuit on the antenna input?
>> 
>> 73 - Dick, W3OA
>> 
>> 
>> 
>> 
>> On 7/31/2016 11:28 AM, David Robbins wrote:
>>> 
>>> Previously for those not included in my discussion with pete:  “Got it running now with the hermes interface dll, but I am plugged in to the router.    Running it on the splitter with a qs1r on the other port, both running cw skimmer server 1.50.  the rp receive seems much less sensitive. On 20m the qs1r has 340 decoders running, the rp only has 45 and is generating a lot less spots.  The ones that both of them spot about the same time the snr is 2-6db lower on the rp.   More later, going to cut some grass or something while the wx is nice.”
>>> 
>>> Then when I dug into it some more:  “While I would like someone with professional instrumentation to repeat this I think you can see a clear difference in the narrow frequency displays of hdsdr below.  The one on the left is the qs1r, the right is the rp.  It looks like the rp is about 5db noisier.  I zoomed in on the signal peaks to compare them closer and the amplitudes are almost identical as close as I can tell, so the signal strength calibration is about the same.
>>> 
>>> cid:image001.png at 01D1E020.4FD7C0D0
>>> 
>>> There are of course many things that could contribute to this.
>>> 
>>> 1. Power supplies are different, I am using the supplied rp supply for it but an after market supply for the qs1r.
>>> 
>>> 2. Front end or adc noise
>>> 
>>> 3. The rp is not in an enclosure like the qs1r maybe some local noise pickup
>>> 
>>> 4. The sample rate conversion could contribute noise
>>> 
>>> p.s. that was a xg2 at 1uv on 40m going into a dxengineering splitter feeding each of the receivers.  The rp is using the hermes interface dll to cwsl_tee to hdsdr and skimmer server.  the qs1r is the same software setup except using the qs1r interface dll.
>>> 
>>> ”
>>> 
>>> Today:
>>> 
>>> Hmmmm, nice rainy morning so played some more with the rp.  I think I found the problem.  on the input lines there are two sets of jumpers… I didn’t quickly find a final schematic, apparently they are keeping those private, but I did find a ‘development’ schematic, excerpt shown here:
>>> 
>>> So jp1 sets the input to one of two attenuators, and jp2 sets the output.  I thought I would try the HV setting, which attenuated my 1uv to below the noise level.  But I noticed when I pulled the jp2 jumper off the output noise went well below the display level I had set on the hdsdr.  So after playing around injecting my signal to pin 5 of jp2 through capacitors and jumpers and generally making a lot more noise, I just put one of their jumpers from pin 2 to pin 5 which bypasses the two input attenuators and voila, now the two cw skimmers are generating almost identical spots.  In fact some of them from the rp are 1db higher snr, and a few are a bit lower, but that is bound to happen just from the skimmer software timing of when it generates the snr value,  one may sample before or after the other and get a noise spike or not.
>>> 
>>> So my guess is that somehow the input attenuator is coupling noise from the board ground into the input signal going to the amp/filter block.  But it would take someone who knows more about the low level design of the amp to figure it out.  I guess I would have expected to see the inputs isolated from the digital board ground and going into a differential amp, but I can’t tell that from the schematic I found.   it could also just be that the attenuation of the input before the amp dropped the signal closer to the amp noise floor and all I’m doing is boosting it a bit, but whatever the mechanism it seems to have helped a lot.
>>> 
>>> David Robbins K1TTT
>>> e-mail: mailto:k1ttt at arrl.net
>>> web: http://wiki.k1ttt.net <http://wiki.k1ttt.net/>
>>> AR-Cluster node: 145.69MHz or telnet://k1ttt.net:7373
>>> 
>> 
> 
> _______________________________________________
> Skimmertalk mailing list
> Skimmertalk at contesting.com
> http://dayton.contesting.com/mailman/listinfo/skimmertalk