20MW

[Pamela J. Filicky]

GM Glenn - I am still waiting for my "Cobia Cigar" for copying your 20 Mill watt peanut whistle on 2 meter WSPR, about the only time I didn't copy was when John/PZB was

transmitting at the same time. Really surprised me.  CUL Steve/N6KOG 

[Glenn Elmore]

Steve,
I saw your spots yesterday - almost every transmission!  It looks like I might have dropped the power down to 5 mW and still have had spots from you. During some tropo enhancement I bet things get even better.

Well, you can settle for being at the top of the QRP Challenge listing in the next publication of the 2m QRP Contest Journal or you can double down and try to hear the 200 mW signal on 10m from the same transmitter.  If you can identify the frequency you can get an honorary endorsement on the certificate (suitable for framing) as soon as it is issued. :-)

As it turns out, I've been corresponding with G0UPL, Hans Summer, who provides the kits and writes code for the U2 and U3 beacons. It is beginning to look like the U2's (Analog Devices AD9850 DDS synthesizer) all have slightly incorrect WSPR tone spacing and that the U3 (Silicon Labs SI5351 fractional-N synthesizer) probably has the correct tone spacing but still does not have precise frequency settability. Thus the endorsement as described above.

It also appears after a few days of data taking like the WSPR decoder in K1JT code is not as good at copying the slightly off tones of the U2.  K9AN has written a new decoder which gets almost 2.5 times the number of decodes of N6GN (on 10m) compared to standard WSPR code.  This could indicate that a significantly better decoder is possible for all of us - HF and VHF.

Glenn

PS
(You should anticipate significant delay in receiving your certificate. Also during this time, it is possible that someone else from further off may spot it and preempt you! :-) )

[Joe Pankow]

FYI I am using an SI5351C (with external clock input) to generate my FT-897's 22.625MHz refclk. Or at least I was until I bumped it last night. I'll have to reset it when I get home from work.

Right now I am connecting to the 5351C I2C interface via a Raspberry PI to learn how to program it without the Silabs ClockBuilder application. Once I work out the programming sequence, I'll implement an I2C master controller in my FPGA so that I'll have a self-contained solution consisting of GPS / FPGA / OCXO /  SI5351. 

Joe W6JHP formerly KJ6QBA

--
You received this message because you are subscribed to the Google Groups "2 Meter WSPR" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 2-meter-wspr...@googlegroups.com.
To post to this group, send email to 2-mete...@googlegroups.com.
Visit this group at http://groups.google.com/group/2-meter-wspr.
For more options, visit https://groups.google.com/d/optout.

[Glenn Elmore]

Joe,
That sounds great! Yours is the second response regarding Si5351 solutions for WSPR. RA3APW, also on list, has breadboarded a solution too. He and I both have used AdaFruit breakout boards.
I compared the Si5351 phase noise with the AD9850 DDS and found both to be fine (IMO) for amateur WSPR use. 

I've been thinking about using one of the 5351 outputs along with a GPS time pulse (U-Blox models let you go from 1 Hz to 10 MHz, more I think on Glonass) to lock up the 5351's 27 MHz oscillator with a heater soldered to the case. It won't be a lot of PLL bandwidth but might be enough with a little styrafoam added to get ppb sorts of stability. Then the other (7?) 5351 outputs are available to lock beacons, transceivers and transverters so that the whole shack is GPS referenced. Obviously Rubidium could be substituted as the shack reference too.

In my copious spare time :-)  I'll try my heater idea on an AdaFruit 5351 board crystal to see what the thermal time constant is.

If this idea works, it would be a very inexpensive solution not just for VHF/UHF WSPR but for general use as well.

Glenn

(*Now* I know who W6JHP is!!  )

[Line Printer]

On Tue, 24 Feb 2015, Glenn Elmore wrote:

> It also appears after a few days of data taking like the WSPR decoder in K1JT
> code is not as good at copying the slightly off tones of the U2. K9AN has
> written a new decoder which gets almost 2.5 times the number of decodes of
> N6GN (on 10m) compared to standard WSPR code. This could indicate that a
> significantly better decoder is possible for all of us - HF and VHF.

Hello,

Not for lack of listening but I haven't found any N6GN signals since they
were 43 dBm. I also tried 10 meter WSPR action but the exception of
WB6HYD, the range of all other stations were at least 1424 miles away.

In an effort to balance the QRP fun, I've reduced my signals to 30 dBm -
which is QRP for an IC-7000 anyway. ^-^

The K9AN k9an-wsprd decoder has been working well for me. Especially after
the floatydouble and %10.7f changes. The frequency output from WSPR spots
now looks reasonable and there is an additonal digit showing tenths of a
Hertz. I have it stored in the ALL_WSPR.TXT file ouput from the
wsprlinux-2.1.exe application. Although I am not sure of the precision of
the measurements to a tenth of a Hertz but I have added info from that
ALL_WSPR.TXT file attached for your examination. It is not a complete log
from 20150222 0548 UTC as there were times when HF was being worked. Some
say "nothing is impossible in Militas" but decide for yourself....

73,
Kevin Martinez
KI6STW

PS
All of my self spotting is not included becuase they always appear similar
to the folowing:

150224 1800 10 18 -0.9 144.4904700 KI6STW CM97 30 0 1 0

>
> Glenn
>
> PS
> (You should anticipate significant delay in receiving your certificate. Also
> during this time, it is possible that someone else from further off may spot
> it and preempt you! :-) )
>
>> GM Glenn - I am still waiting for my "Cobia Cigar" for copying your 20 Mill
>> watt peanut whistle on 2 meter WSPR, about the only time I didn't copy was
>> when John/PZB was
>> transmitting at the same time. Really surprised me. CUL Steve/N6KOG
>>
>
>

--
------------------------------------------------------------------------
Line Printer System | HODIE NATUS EST RADICI FRATER
lps at rahul 'dot' net | 645/6180 BOS/BCE
------------------------------------------------------------------------

[Glenn Elmore]

Kevin,

N6GN transmit power has once again dropped to the "not noticeable"
level. There's a difficult-to-find intermittent in the transceiver that
I am still chasing.

The addition of tenth Hz resolution to the K9AN code is a result of our
ongoing studies on 10m, first trying to determine frequency spreading
and shifting due to the ionosphere, then trying to understand U2 and U3
beacon vagaries and finally involving characterizing the different WSPR
decoders with less than perfect signals.

What I know from recent mail with Hans is that the tone spacing in newer
U3s code is correct but little else may be.

Thanks for sending the .txt file, I'll have a look at it.

Now that the QRP craze is beginning, maybe I'll see if I can make a
light weight version of the 2m WSPR beacon to lift aloft, perhaps a few
hundred meters, to see if we can increase the DX spots!

Glenn

[n3...@aol.com]

Maybe make a high altitude balloon that we can track. :)

Kevin,

N6GN transmit power has once again dropped to the "not noticeable" 
level. There's a difficult-to-find intermittent in the transceiver that 
I am still chasing.

The addition of tenth Hz resolution to the K9AN code is a result of our 
ongoing studies on 10m, first trying to determine frequency spreading 
and shifting due to the ionosphere, then trying to understand U2 and U3 
beacon vagaries and finally involving characterizing the different WSPR 
decoders with less than perfect signals.

What I know from recent mail with Hans is that the tone spacing in newer 
U3s code is correct but little else may be.

Thanks for sending the .txt file, I'll have a look at it.

Now that the QRP craze is beginning, maybe I'll see if I can make a 
light weight version of the 2m WSPR beacon to lift aloft, perhaps a few 
hundred meters, to see if we can increase the DX spots!

Glenn


On 02/24/2015 10:20 AM, Line Printer wrote:
> On Tue, 24 Feb 2015, Glenn Elmore wrote:
>
>> It also appears after a few days of data taking like the WSPR decoder 
>> in K1JT code is not as good at copying the slightly off tones of the 
>> U2.  K9AN has written a new decoder which gets almost 2.5 times the 
>> number of decodes of N6GN (on 10m) compared to standard WSPR code.  
>> This could indicate that a significantly better decoder is possible 
>> for all of us - HF and VHF.
>
>
> Hello,
>
> Not for lack of listening but I haven't found any N6GN signals since 
> they were 43 dBm.  I also tried 10 meter WSPR action but the exception 
> of WB6HYD, the range of all other stations were at least 1424 miles away.
>
> In an effort to balance the QRP fun, I've reduced my signals to 30 dBm 
> - which is QRP for an IC-7000 anyway. ^-^
>
> The K9AN k9an-wsprd decoder has been working well for me. Especially 
> after the floatydouble and %10.7f changes.  The frequency output from 
> WSPR spots now looks reasonable and there is an additonal digit 
> showing tenths of a Hertz.  I have it stored in the ALL_WSPR.TXT file 
> ouput from the wsprlinux-2.1.exe application.  Although I am not sure 
> of the precision of the measurements to a tenth of a Hertz but I have 
> added info from that ALL_WSPR.TXT file attached for your examination.  
> It is not a complete log from 20150222 0548 UTC as there were times 
> when HF was being worked. Some say "nothing is impossible in Militas" 
> but decide for yourself....
>
> 73,
> Kevin Martinez
> KI6STW
>
> PS
> All of my self spotting is not included becuase they always appear 
> similar to the folowing:
>
> 150224 1800  10  18 -0.9 144.4904700  KI6STW CM97 30 0     1    0
>
>
>
>
>
>
>
>
>>
>> Glenn
>>
>> PS
>> (You should anticipate significant delay in receiving your 
>> certificate. Also during this time, it is possible that someone else 
>> from further off may spot it and preempt you! :-) )
>>
>>> GM Glenn - I am still waiting for my "Cobia Cigar" for copying your 
>>> 20 Mill watt peanut whistle on 2 meter WSPR, about the only time I 
>>> didn't copy was when John/PZB was
>>> transmitting at the same time. Really surprised me.  CUL Steve/N6KOG
>>>
>>
>>
>

-- 
You received this message because you are subscribed to the Google Groups "2 
Meter WSPR" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to 2-meter-wspr...@googlegroups.com.
To post to this group, send email to 2-mete...@googlegroups.com.
Visit this group at http://groups.google.com/group/2-meter-wspr.
For more options, visit https://groups.google.com/d/optout.

[Glenn Elmore]

This is already under discussion at the weekly Sonoma County 2m WSPR meetings (not really, but sort of).
It looks like we might be able to architect a GPS referenced 2m WSPR station nearly as light weight as the 20m/30m party balloon WSPR launched by VK3YT.  I think that payload was only 15 grams (1/2 ounce). We would probably have some extra weight involved in stabilizing the reference for 2m (or higher) so perhaps something slightly larger than a party balloon would be necessary, but the U-Blox receiver I got is under 4g once the patch antenna is removed. We could replace that with a very lightweight metalized paper/plastic eggbeater or clover design, I think.  Other items are the Si5351 board, maybe a FET to make 100 mW or so, one chip controller like an Arduino Nano, PIC or similar and a flexible solar panel/battery.  If it were expected to stay up overnight or for several days, we'd probably want to write some code to turn it off at night and do a little power supervision. Most of this is already done or in libraries somewhere.

I previously wrote to one of the amateur near-space balloon groups (in Freemont, if I remember right) who no doubt know a lot about the art of properly sizing and helium-filling really high altitude balloons.  I offered a payload but never heard back.  Maybe someone on the list knows of a similar group that might be receptive.  It would really be fun to see what kinds of spots are possible from a 2m WSPR station at ~80,000 feet.  LOS to the radio horizon (4/3 earth radius) should be around 900 miles.  It might spottable well beyond that.

A project like this might also get some more stations on 2m WSPR.  Who wants to take it on?

Glenn

[Line Printer]

On Tue, 24 Feb 2015, Glenn Elmore wrote:

> I previously wrote to one of the amateur near-space balloon groups (in
> Freemont, if I remember right) who no doubt know a lot about the art of
> properly sizing and helium-filling really high altitude balloons. I offered
> a payload but never heard back. Maybe someone on the list knows of a similar
> group that might be receptive. It would really be fun to see what kinds of
> spots are possible from a 2m WSPR station at ~80,000 feet. LOS to the radio
> horizon (4/3 earth radius) should be around 900 miles. It might spottable
> well beyond that.
>
> A project like this might also get some more stations on 2m WSPR. Who wants
> to take it on?

I've previously tracked balloons carrying APRS hardware that were launched
in south San Jose by Ron Meadows K6RPT. Some of them attained fairly high
(over 100,000 feet) and traveled quite far. An image from my radio on
December 2 2012:

https://picasaweb.google.com/112916124640757906440/FTM350#5817643824706513954

The last packet I got from the CNSP-18 balloon was from 436 miles away:

https://picasaweb.google.com/112916124640757906440/FTM350#5817643510479352370
https://picasaweb.google.com/112916124640757906440/FTM350#5817643528266187538

I'm not sure if it is still flying but searching for K6RPT-11 should show
the last APRS data from the flight. It was over Greece as of Feb 28......

Try:
http://aprs.fi/#!mt=roadmap&z=1&call=a%2FK6RPT-11&timerange=518400&tail=518400


The high altitude gives a large footprint of area for potential receivers
that can be reached. The downside to existing at high altitudes is the
ultraviolet exposure from the sun. What little atmosphere found above
100,000 feet does little to reduce ultraviolet light and harsh radiation
limits the lifetime of the actual balloon material.

Ron has been using "super pressure" balloons lately. The thinking behind
these is to have a constant volume balloon inflated with light gases.
This limits the altitude reachable by the balloon and reduces excursions
in altitude that are produced by the heating/cooling of the gas in the
day/night temperature changes. This combination is selected for long
duration flights and are usually found between flying 30,000 and 40,000
feet in altitude. If the K6RPT-11 balloon is still flying, the APRS info
should show traveling velocity at reasonable speed despite being less than
100,000 feet.

If there is one thing I've learned about these balloon adventures is that
there is much more to learn than I had guessed. Much of what I have
learned has been from a Yahoo group called "Great Plains Super Launch":

https://groups.yahoo.com/neo/groups/GPSL/conversations/messages

It could be very interesting but I know enough to know that more
knowledgable people than myself would be needed for functionality.

73,
Kevin Martinez
KI6STW

[Pamela J. Filicky]

This sounds like a interesting project, I would be interested in helping out & gladly provide some funding.

                                                                 Regards Steve/N6KOG

[Line Printer]

Since February 22 2015, the wsprlinux.exe app (now at version 2.2) working
with the k9an-wsprd tool has been recording the frequency spots to the
tenth of a Hertz. Although it exists in the local ALL_WSPR.TXT file, the
data found at wsprnet.org does not show any fractions of the Hertz digit.

For fun, I've gathered my data from the 432.3 MHz WSPR spots on February
25 2015 into two files (attached). One is just the chronological sorted
spots as found in ALL_WSPR.TXT with the deletion of self-spotting. The
second file is the same data from ALL_WSPR.TXT but sorted by the spotted
station.

Perhaps useful ideas can be found after looking through this data. I can
also post or send the same sorts of data from the 144 MHz spotting if
there is any interest.

I wonder how many stations are spotted anyway?
more than striped stations...... ^-^

73,
Kevin Martinez

PS
Since self-spotting is gratuitious anyway, I'll just include one because
they all read 432.3014700:

50225 0628 10 18 -0.9 432.3014700 KI6STW CM97 37 0 1 0


--

[Glenn Elmore]

Just for comparison, here is a plot of several days of spots by K9AN of N6GN on 10m WSPR

This is a U3s WSPR beacon at 200 mW, using a GPS-coherent version of the shack standard.  Frequency variations are either due to Doppler shift on the 10m path, you can see the diurnal changes as the layer moves down and then up, or they are due to characteristics of the WSPR decoding software.  Actual transmit frequency is probably close to 28.1251007, slightly below midway up the Y axis.  This offset from even Hz is a result of imprecision in the U3s software which was designed to have the correct tone spacings but not correct center frequency.

The variations on your 432.3 spots may be in part due to some of these same measurement characteristics or they may also be due to GPS satellite number/quality.  I estimate that the GPS10s involved are probably on the order of .1 ppb, .04 Hz at 432, but if there are room temperature changes or one of the OCXOs isn't working particularly well, it could be more than this.

Glenn

[Glenn Elmore]

> It could be very interesting but I know enough to know that more
> knowledgable people than myself would be needed for functionality.

It looks like it may still be up, somewhere NE of Greece perhaps. These
might be just the people to enable a WSPR beacon. The advantage I see
over APRS is the ~40 dB extended S/N range. Not sure if we could easily
do it within their weight budget. If you want to take on making contact
and coordinating with the group and opening discussions, I might be able
to help with hardware. Not necessarily a time frame of this year but
perhaps.

Glenn

[Karen, RA3APW]

Hi All,

 

as Glenn, N6GN mentioned before -  some modification on U3S with Si5351A synth have been made to get more frequency stability (especially on 2m WSPR)  using GPSDO way.

u-blox LEA-M8F provides a reference signal 30.72 MHz (from internal TCXO) which used for Si5351 (instead of 27 MHz xtal). 

Allan deviation is about 2.5E(-10) for tau=100 sec. Reference (30.72MHz)  phase noise is -143 dBc/Hz @ 10 kHz.   

Description of U3S/GPSDO solution  with measurement results (in Russian, sorry) can be found on  http://www.ra3apw.ru/wspr-u3s-gpsdo/

 

Hope Google should help with translation here - https://translate.google.com/translate?sl=ru&tl=en&js=y&prev=_t&hl=ru&ie=UTF-8&u=http%3A%2F%2Fwww.ra3apw.ru%2Fwspr-u3s-gpsdo%2F&edit-text

73 de Karen, RA3APW

вторник, 24 февраля 2015 г., 20:33:43 UTC+3 пользователь n6gn написал:

[Glenn Elmore]

Karen,

Thanks for sharing the U3s/Si5351 progress.  As you know, I've been using substitute GPS-derived 27 MHz on a U3s on both 10m and 2m.  I do want to mention that the U3s v3.07 still has a problem though.  While the tone spacing in WSPR is correct for the U3s, unlike the U2 v2.07 which is wrong and results in fewer spots with K1JT decoders, the center frequencies of both the U2 and the U3s are not precise.  That is, the U3s, with Hans' latest code does not produce correct center frequencies as you stated on your pages. In fact, they can be in error by quite a lot, certainly more than 1 Hz at 144 MHz.  The algorithm is not monotonic either. Sometimes raising the set frequency by 1 Hz   results in the actual frequency going down!  I've had email with Hans about this and he is aware of it. The issue is no longer one of path precision in the code, as was the problem in the U2/U3 AD9850 versions but rather in his attempt to maintain small, even integers in for the fractional-N divide numbers in the Si5351 - per Silicon Labs suggestion. This is suggested as a means to improve spurious and phase noise performance.

I would be interested to see if you have measured the close-in spectrum of the U3s at 144 MHz with the 32 MHz clock.  If you are able, a total spectrum analyzer bandwidth of, perhaps, 10 kHz would be interesting. 

Best
Glenn n6gn

[Steven Hess]

I picked up a Si5351A breakout board http://docs.etherkit.com/Si5351A_Breakout_Board on a whim just to help fund this https://www.indiegogo.com/projects/si5351a-breakout-board

I wonder how useful it would be if applied to this?  I've never even worked with surface mount stuff. I got the kit with the TCXO and SMA connectors. 

____________
Apply appropriate technology. Use what works without prejudice.  
Steven L Hess ARS KC6KGE DM05gd22
Owner Flex-1500 and Flex-3000, FT-857D, FT-817ND
openSUSE Linux 13.1 KDE
Known as FlameBait and The Sock Puppet of Doom.

[Glenn Elmore]

Well, it's more expensive than http://www.adafruit.com/products/2045 which does about the same thing but I understand the altruistic motive.  In this case you don't have to assemble anything SMD,  anything at all except maybe connectors, if you want them.
The question around the Si5351, as I see it, revolves around spectral purity. The M/N technique, like the DDS method, has its own set of issues when it comes to generating very fine steps. They are different but each has discernible phase noise and/or coherent/structural spurious.
Just for reference, here is a Si5351A in the U3s generating 10m WSPR, 7 kHz  at 2m and 300 MHz spans at 10m  so that you can see close-in as well as broadband effects.  The close-in performance is no-doubt affected by choice of divide numbers.  The very wideband stuff, like harmonics may be filtered out with appropriate transmitter filters. The wideband but closer in stuff may be more of a problem.  
The numbers aren't necessarily awful, compared to third order IMD for typical SSB transceivers (though we are operating in CW portions) but if these were sourcing a high power transmitter, the unwanted stuff might still be worse than desired, even for amateur use. I wouldn't like to see these at 1 kW into significant antenna gain. That could generate some ill will if not a ticket.

Glenn n6gn

[Karen Tadevosyan]

Glenn,

please find the spectrum from Si5351A (pin 10) with a standard 27 MHz clock (xtal) at 144 MHz for different spans

http://www.evernote.com/l/APfQGGuJTHFBPZXGWzPbcf6qYZJJRyfQp0k/

Please note that 10 dB attenuator has been installed between Si5351A and spectrum analyzer.

 

I will measure the close-in spectrum of U3S at 144 MHz with 32 MHz clock shortly.

 

73 de Karen, ra3apw

[Bo, OZ2M]

Hi

I have a Si5351A here, that I plan to include on the PI4ino generic beacon controller www.rudius.net/oz2m/pi4ino , and my measurements confirm more or less what Karen, RA3APW, shows. I have spurious at -45 dBc vs Karen's -50 dBc but that could be related to PCB layout and output matching. I use an Adafruit breakout board which from my perspective cannot be called a RF design, e.g. no supply partitioning, no load optimization ...

So are spurious, not just harmonics, an issue? We hams have relaxed requirements I believe at least in my neck of the woods. However, in an ETSI context the most relaxed specification is -36 dBm below 1 GHz and -30 dBm above 1 GHz. So if the Si5351A carrier is 20 mW/13 dBm the spurious would be -37 dBm to -32 dBm on a 2 m signal. [Amplifying the carrier output to 20 MW/93 dBm would mean 53 dBm/200 to 58 dBm/630 W. :-) Let's hope this will never happen. On the other hand who can afford the electricity bill?] So 20 mW is probably the sound limit one can take an Si5351A. But it is good fun.

73
Bo, OZ2M

[Glenn Elmore]

Karen,Bo,
Thanks for the additional measurements and comments. Yes, this agrees pretty much with what we see.  Silicon Labs suggests special selection of divide numbers, particularly for the first PLL, I think.  This can perhaps improve performance but perhaps still doesn't leave it good enough for more than a QRP transmitter.  I don't think board lay out, per se, may help enoough.

OTOH, supply partitioning might help somewhat.  I think actual harmonics can be dealt with the same way we do for any other radio, conventional superheterodyne or SDR, with suitable low pass or other filtering. 
These spurs we see in the measurements are, to a degree, part of the fractional-N design itself.  It may be part of the reason that Si Labs added the spread-spectrum feature- to spread the energy of these structural spurs out so they wouldn't be so objectionable.  

There is also the question of non-coherent spurious. This would be the "dithered noise" that we see in the very close in measurements. It looks like this is perhaps withing 1 to a few kHz of carrier and is the noise pedestal seen in one of the measurements I sent previously.  Would this be a problem in amateur service?  Probably not so much. Third-order IMD is no doubt worse for a SSB station.

So we have

• harmonics
• distant coherent spurs
• close-in spurs
• pedestal noise

Some of these can be improved by transmitter design, some by divide-number selection, possibly some by power supply decoupling. I suspect we are still left with a mediocre signal as far as spectral purity is concerned.  

But put in context and considering the applications at QRP, it may be fine.  I suspect if we look closely at many amateur radios we would also find some similar "surprises". 
 -40 to -60 dBC really isn't state of the art. But theSi5351's certainly are an easy part to use and as long as care is taken, they *may* have particular uses that are fine. QRP may be one of these.

Glenn n6gn

[Karen Tadevosyan]

Hi Glenn, Bo and All,

>I would be interested to see if you have measured the close-in spectrum of the U3s at 144 MHz with the 32 MHz clock.  If you are able, a total spectrum analyzer bandwidth of, perhaps, 10 kHz would be interesting.  

Please find the result of a mesurement - part 4  the fig. 19-21 on the link http://www.ra3apw.ru/proekty/si5351-spectrum/

Phase noise is about -110 dBc/Hz @ 10 kHz at 144.490 MHz - not bad and more than enough for this application.

But I am not sure that we can be happy with spures (-41dBc/997kHz) closed to carrier ...

73 de Karen, ra3apw 

[Karen Tadevosyan]

Hi Bo and All,

>So are spurious, not just harmonics, an issue? 

Yes, correct, the main issue is spurious near carrier which can not be sufficiently filtered.

>However, in an ETSI context the most relaxed specification is -36 dBm below 1 GHz and -30 dBm above 1 GHz.

This is ETSI technical requirement for professional radio communication systems. 

If you look at amateur radio requirements (for example from ICOM) for VHF/UHF portables and mobiles - "Spurious emissions : Less than –60dB"...

I believe that the level of spurios from Si5351A at 144 MHz can be improved by 15-20 dB - we are working on it.

>I have a Si5351A here, that I plan to include on the PI4ino generic beacon controller www.rudius.net/oz2m/pi4ino

Your "Next Generation Beacon" design (http://www.rudius.net/oz2m/ngnb/index.htm) looks really nice and professional...  

73 de Karen, ra3apw

[Bo, OZ2M]

Hi Karen and the list

Interesting with findings on the SI5351A and possible improvements.

Indeed ETSI is for PMR. But to get a number it is better than pulling a white rabbit out of a hat :-) I don't know where e.g. ICOM got the -60 dBc from. But perhaps some kind of manufacturers' standard or what is technically feasible without improving anything. For OZ7IGY I can say that -60 dBc will not be accepted. There is a contest station located 10 km away and one DXer 11 km away. Two years ago we have an issue with the switching power supply, on the 2 m beacon at some -70 dBc, which created havoc. Fortunately we solved the problem with a better PSU. The side band noise of the 1 GHz clock is excellent too and the DDS even better. So now we are are on the safe side.

The Next Generation Beacons project is semi professional yet this was not the target as all. But we wanted to design and build what was economical feasible and "RF environmentally friendly". The latest development in digital RF chips has been good for the amateur radio, but there are also a lot of designs that are on the air with a really lousy spectral performance. Unfortunately many focuses only on the harmonics and not the close in spectrum performance which will actually hurt other hams. Therefor your work on the Si5351A is highly valuable. We have learned a lot from the Next Generation Beacons project and so far I believe it has been very successful not just on the H/W side but also on the mixed mode PI4 + CW + carrier sequence. Units for seven beacons,  including USA, have been supplied so far, PCBs to a handful of beacon around the world and a national research institution is using the DDS units for a hydrogen maser time standard with an accuracy of 5 x 10^-16. Some of the key components are not cheap though: 1 GHz VCO is 112 USD, the AD9912 DDS is >57 USD and the 10 MHz OCXO is ~100 USD. But quality does not come cheap.

I am looking forward to your Si5351A improvements. I know other PI4ino users are too :-)

73
Bo, OZ2M

[Glenn Elmore]

I just found this recent presentation from NT7S that includes some
spectrum plots for the Si5351. It seems to pretty well agree with our
experience but adds a bit more information about crosstalk and terminations.
http://nt7s.com/

I think I'm going to have to pass on the Si5351 as an LO for anything
other then lo/mid performance hardware, unless I'm prepared to be
particularly careful about divider selection and also able to
post-filter the output.

Glenn n6gn