P3 Measurements with new K3SYNA Boards

[Jim Brown-10]

I finally got around to repeating a few measurements using with the new
K3SYNA boards. I've added the results near the beginning of
P3_Spectrum_Measurements.pdf

There's also a measurement of an un-keyed, unmodulated carrier.

Executive summary -- the screen of the P3 can display only 80 dB of
dynamic range, and within that displayed range, I cannot see any
difference in signal bandwidth between the new synth board and the old
one. There ARE differences, and I've seen measurements showing
significant reduction in phase noise, but it's another 10-20 dB down
from the peak.

It would be possible to see those differences by lowering the P3
reference level, letting the upper 20 dB of dynamic range run off scale.
I've done that with the new K3SYNA boards, but did not do so with the
older K3SYN, so I have noting to compare to. :)

Note that the only new measurements in the file are those showing the K3
steady carrier and a K3 sending dits at 25W for comparison with older data.

73, Jim K9YC
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Phil Wheeler-2]

Oh oh, Jim: We need a P3A :-)  Good you took this on.

73, Phil W7OX

On 7/27/15 10:05 PM, Jim Brown wrote:

> I finally got around to repeating a few
> measurements using with the new K3SYNA boards.
> I've added the results near the beginning of
> P3_Spectrum_Measurements.pdf
>
> There's also a measurement of an un-keyed,
> unmodulated carrier.
>
> Executive summary -- the screen of the P3 can
> display only 80 dB of dynamic range, and within
> that displayed range, I cannot see any
> difference in signal bandwidth between the new
> synth board and the old one. There ARE
> differences, and I've seen measurements showing
> significant reduction in phase noise, but it's
> another 10-20 dB down from the peak.
>
> It would be possible to see those differences by
> lowering the P3 reference level, letting the
> upper 20 dB of dynamic range run off scale. I've
> done that with the new K3SYNA boards, but did
> not do so with the older K3SYN, so I have noting
> to compare to. :)
>
> Note that the only new measurements in the file
> are those showing the K3 steady carrier and a K3
> sending dits at 25W for comparison with older data.
>
> 73, Jim K9YC

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Jim Brown-10]

On Mon,7/27/2015 10:28 PM, Phil Wheeler wrote:
> Oh oh, Jim: We need a P3A :-) Good you took this on.

Perhaps P3 firmware can be re-written to allow display of greater
dynamic range to the extent that electronics headroom supports it?

73, Jim K9YC
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Alan Bloom-2]

On 07/27/2015 10:46 PM, Jim Brown wrote:
> On Mon,7/27/2015 10:28 PM, Phil Wheeler wrote:
>> Oh oh, Jim: We need a P3A :-) Good you took this on.
>
> Perhaps P3 firmware can be re-written to allow display of greater
> dynamic range to the extent that electronics headroom supports it?

The limit is the 32-bit arithmetic in the DSP firmware.  32 bits gives
you 96 dB less some headroom to allow for the averaging function.  I
tried floating-point arithmetic but it increased the computation time by
more than an order of magnitude.

As you said, you can see the improved phase noise performance of the new
synthesizer by simply cranking up the reference level and allowing the
carrier to go off the top of the screen.  Perhaps someone will do that
when upgrading to the new synthesizer and post before and after plots.
(Didn't somebody already do that?)

Alan N1AL

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Alan Bloom]

On 07/28/2015 07:38 AM, Alan Bloom wrote:

> The limit is the 32-bit arithmetic in the DSP firmware.  32 bits gives
> you 96 dB less some headroom to allow for the averaging function.

I should clarify that these are calculations that are done after the
conversion from voltage to power, so you only get 3 dB/bit.

Alan N1AL
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Jim Brown-10]

On Tue,7/28/2015 7:38 AM, Alan Bloom wrote:

> On 07/27/2015 10:46 PM, Jim Brown wrote:
>> On Mon,7/27/2015 10:28 PM, Phil Wheeler wrote:
>>> Oh oh, Jim: We need a P3A :-) Good you took this on.
>>
>> Perhaps P3 firmware can be re-written to allow display of greater
>> dynamic range to the extent that electronics headroom supports it?
>
> The limit is the 32-bit arithmetic in the DSP firmware.  32 bits gives
> you 96 dB less some headroom to allow for the averaging function.  I
> tried floating-point arithmetic but it increased the computation time
> by more than an order of magnitude.

Understood, and clearly the right design compromise.  My data is the
accumulation of peaks, without averaging.  Might it be possible for a
special "measurement version of the firmware that would squeeze another
10 dB?

> As you said, you can see the improved phase noise performance of the
> new synthesizer by simply cranking up the reference level and allowing
> the carrier to go off the top of the screen.  Perhaps someone will do
> that when upgrading to the new synthesizer and post before and after
> plots. (Didn't somebody already do that?)

I did see some work like that posted done with an analyzer with less
frequency resolution, and the results clearly showed the improvement.

Thanks and 73, Jim
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]