Transmitter Spectrum Analyzer results

[Joe]

I took some photos of the transmitter output using my spectrum analyzer no filters of any kind 

in these results. I use a surplus 30MHZ lowpass on my own board it does a good job cleaning up signals

above that frequency keep in mind that this filter wasn't  used for the pictures.

Also note that on all bands below 15 meters most non harmonic spurs where below -70db.

But on 15 meters I fond a spur at about 16.6MHZ about -60db down  12meters was OK other than harmonics 

and on 10meters their where 2 spurs below 28MHZ the worst at 22.7MHZ -55db down.

Joe   wa9cgz

[Joe]

The photos show remarkably good results for this project and they are for my setup only using

the SDK and PowerSDR only others may have different results.

After studying the pictures I believe a good 30MHZ low pass filter is needed before any transmit amplifier if for no other reason to 

clean up strong signals above 30MHZ. At my QTH I use a relay driver from the PTT line to convert the RF in and out to a single coax.

the 30 Mhz low pass filter is inserted into this line and is used on both RX and TX. Both 15 and 10 meters have some spurs below

their operating frequency 10 being the worst at -55db. For the purest simple HI pass filters could be added to reduce them,

or maybe a firmware change could help since as I don't understand how they are generated. The Hermes-lite does a remarkable

job for such a single device and other than the stuff above 30MHZis remarkable. I added a photo showing 10 Meters after the addition 

of the low pass filter showing how much better things look.

Joe  wa9cgz

On Monday, February 23, 2015 at 4:07:18 PM UTC-6, Joe wrote:

I took some photos of the transmitter output using my spectrum analyzer no filters of any kind 

in these results. I use a surplus 30MHZ lowpass on my own board it does a good job cleaning up signals

above that frequency keep in mind that this filter wasn't  used for the pictures.

Also note that on all bands below 15 meters most non harmonic spurs where below -70db.

But on 15 meters I fond a spur at about 16.6MHZ about -60db down  12 meters was OK other than harmonics 

[John Williams]

Joe,

In both PAs I have a low pass filter for 60MHz. We had it there when we were considering 6M and decided to leave it. I will rebuild it for 30 MHz. Perhaps can I plug in your values in my design? See attached schematic around L9/L11.

BTW, I want to allow operation to 32MHz. I have a Ten-Tec 10M to 2M transverter and I need 32 MHz to get the full 2M band converted...

Also, The superband PA is getting closer. I just ordered Rev 2 boards this weekend.

John

--
You received this message because you are subscribed to the Google Groups "Hermes-Lite" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[John Williams]

Joe,

You seem to not be bothered by any of the current issues. I am beginning to wonder if I should have replaced my damaged CV with a SDK.

Today running 20150222 testing the amp, the power output was intermittent, while tune was on solid. Backed up to 20150210 and tune was solid and power output was solid as well.

I have a new CV coming on Thursday.

I have the new values for a 32 MHz LPF, so will take your suggestion and change the design.

John

On 2/24/2015 9:22 AM, Joe wrote:

--

[Steve Haynal]

Hi Joe,

Many thanks for the spectrum plots! I'd like to eventually put these on the wiki. There may be some tweaks to try and would appreciate your analysis with those when they happen.

73,

Steve

KF7O

[Jacinto Rebelo]

Hi John, 

can you share you filter design ? I also have a transverter and would like to build one ...

73's

Jacinto

cu2ed

[Joe]

John, 

I don't believe with the current layout/firmware the Hermes-lite will transmit above about 29MHZ even with a GOOD 30MHZ lowpass 

filter. It appears to me doing spectrum analyzer checks that what may be aliases slide down from frequencies above 30.

I find that even with a good 30MHZ filter about the lowest you can transmit is about 29MHZ without their being strong out of band signals.

With the current setup I would forget about anything above 29MHZ. Would it be possible to change the xtal in you transverter to use a lower IF frequency? I bread boarded up 2 hi-pass filters for 15 and 10 transmit, they roll of the unwanted lower frequency to lower than -65db  

but as it stands 15 maybe OK as the lower spur is already -60. I wouldn't count on the transmit amplifier harmonic filter alone

to remove signals above 30MHZ  particularly on the higher HF bands. 

 All the above test done with a SDK and PowerSDR which has been stable with no dropouts. 

  

Joe   wa9cgz

[John Williams]

Joe,

So, looks like my design for Superband, that uses bandpass filters, works well for this issue. Take a look at this elsie printout to see the calculated band pass and see if it resolves your issue. For the Megaband, we are using the overall 30MHz low pass then LPFs per band group on transmit. May have to change them to BPFs for 17/15 and 12/10.

Can you share your 30MHz filter with us so that Jacinto can build one. I will try it also on the next revision of Superband. My 30 MHz design is attached. May not be aggressive enough. Anyway, for Jacinto, the inductors are 9 turns of #26 wire on a T37-6 Toroid...

What is your objective on harmonic suppression. FCC says -43db for HF and -60db for 6M, which is not an issue for us. What should our overall objective be? Also, the single FET amplifier will make this issue worse, I fear. Once I get a good, functioning Superband for 17-10, I will send it to you to have you examine it more carefully. My boards are at the fab now, so will probably be 2-4 weeks from now...

I am going to hold off on the Megaband fab until we get these design issues flattened. The 2 designs share the amplifier layout so no need moving too fast if we are not firm on our design...esp if I have to redesign some of the TX filters.

John

[John Williams]

Joe,

Thinking about this further, the use of a transverter should be subordinate to us obtaining good 10M band capability. We should design for 28 to 29.7 and insure that our basic design gives full 10M band support. On the transverter, I can use a superband board, purpose built for 28-32Mhz. As we progress, I am sure I will have extra Hermes boards and CVs to have a special purpose radio for 2M All-mode.

So, I am more concerned about your 29Mhz statement. Any way to get legal performance up 600 Khz further? That should be our focus.

Also, As I watch the transmit panadapter showing me I/Q data, I see small spurs appear across the band near the transmit freq that must be generated by the software. Are these perhaps being amplified by Hermes and resulting in the spurs you see?

John

[John Williams]

Joe,

On a related topic - what spectrum analyzer are you using?

Just a thought - is Hermes Lite far enough along that we could trust using it along with the bandscope capability in g0orx.openhpsdr as a spectrum analyzer. My injured CV is hampered by no PTT signal, but it could still be used RX for a bandscope. I have a Nexus 7 tablet also that I could use for the interface. We could fork the code base and add dialogs to allow for setting sweep parameters, etc.

Sid, what do you think?

Just thinking out loud.

John - AC9HY

On 2/25/2015 9:21 AM, Joe wrote:

[Sid Boyce]

Sounds a good move, bandscope should work.

There is no openHPSDR VNA software for Linux or Android. My only experience is with HiQSDR which works fine, openHPSDR Windows VNA was based on the HiQSDR VNA Linux implementation.
73 ... Sid.

On 25/02/15 16:20, John Williams wrote:

Joe,

On a related topic - what spectrum analyzer are you using?

Just a thought - is Hermes Lite far enough along that we could trust using it along with the bandscope capability in g0orx.openhpsdr as a spectrum analyzer. My injured CV is hampered by no PTT signal, but it could still be used RX for a bandscope. I have a Nexus 7 tablet also that I could use for the interface. We could fork the code base and add dialogs to allow for setting sweep parameters, etc.

Sid, what do you think?

Just thinking out loud.

John - AC9HY

On 2/25/2015 9:21 AM, Joe wrote:

John, 

I don't believe with the current layout/firmware the Hermes-lite will transmit above about 29MHZ even with a GOOD 30MHZ lowpass 

filter. It appears to me doing spectrum analyzer checks that what may be aliases slide down from frequencies above 30.

I find that even with a good 30MHZ filter about the lowest you can transmit is about 29MHZ without their being strong out of band signals.

With the current setup I would forget about anything above 29MHZ. Would it be possible to change the xtal in you transverter to use a lower IF frequency? I bread boarded up 2 hi-pass filters for 15 and 10 transmit, they roll of the unwanted lower frequency to lower than -65db  

but as it stands 15 maybe OK as the lower spur is already -60. I wouldn't count on the transmit amplifier harmonic filter alone

to remove signals above 30MHZ  particularly on the higher HF bands. 

 All the above test done with a SDK and PowerSDR which has been stable with no dropouts. 

  

Joe   wa9cgz

On Tuesday, February 24, 2015 at 7:15:57 PM UTC-6, John Williams wrote:

Joe,

You seem to not be bothered by any of the current issues. I am beginning to wonder if I should have replaced my damaged CV with a SDK.

Today running 20150222 testing the amp, the power output was intermittent, while tune was on solid. Backed up to 20150210 and tune was solid and power output was solid as well.

I have a new CV coming on Thursday.

I have the new values for a 32 MHz LPF, so will take your suggestion and change the design.

John

On 2/24/2015 9:22 AM, Joe wrote:

The photos show remarkably good results for this project and they are for my setup only using

the SDK and PowerSDR only others may have different results.

After studying the pictures I believe a good 30MHZ low pass filter is needed before any transmit amplifier if for no other reason to 

clean up strong signals above 30MHZ. At my QTH I use a relay driver from the PTT line to convert the RF in and out to a single coax.

the 30 Mhz low pass filter is inserted into this line and is used on both RX and TX. Both 15 and 10 meters have some spurs below

their operating frequency 10 being the worst at -55db. For the purest simple HI pass filters could be added to reduce them,

or maybe a firmware change could help since as I don't understand how they are generated. The Hermes-lite does a remarkable

job for such a single device and other than the stuff above 30MHZis remarkable. I added a photo showing 10 Meters after the addition 

of the low pass filter showing how much better things look.

Joe  wa9cgz

On Monday, February 23, 2015 at 4:07:18 PM UTC-6, Joe wrote:

I took some photos of the transmitter output using my spectrum analyzer no filters of any kind 

in these results. I use a surplus 30MHZ lowpass on my own board it does a good job cleaning up signals

above that frequency keep in mind that this filter wasn't  used for the pictures.

Also note that on all bands below 15 meters most non harmonic spurs where below -70db.

But on 15 meters I fond a spur at about 16.6MHZ about -60db down  12 meters was OK other than harmonics 

and on 10meters their where 2 spurs below 28MHZ the worst at 22.7MHZ -55db down.

Joe   wa9cgz

--
You received this message because you are subscribed to the Google Groups "Hermes-Lite" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Hermes-Lite" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Hermes-Lite" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

-- 
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support
Senior Staff Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks

[Steve Haynal]

Just a few comments on this thread. First, the ADC on the Hermes-Lite samples at fs=61.44 MHz. According to the Shannon-Nyquist sampling theorem, the maximum RX frequency in the first Nyquist zone is fs/2=30.72. Although it is possible to receive signals in the second Nyquist zone (30.72-61.44 MHz) it does require a new frontend with the proper filter which is a bit of work. John - I think you are better off lowering the frequency of your transverter.

Second, the DAC on the Hermes-Lite runs at a separate higher frequency up to 200 MHz. The current configuration is pretty basic, but in the future we can try to tweak the frequency of the DAC and the AD9866 interpolation rates. This may allow for cleaner TX up to 29.7 MHz.

73,

Steve

KF7O

  

[John Greusel]

John, Another solution for your transverter is to scale the frequency of the transverter crystal down a few mhz. That's what I did with my 6M Tentec transverter.

John
KC9OJV

[John Williams]

That is what I will do, move it down to the middle of the 12/10 passband... Did you find a new crystal or devise a way to use one of these inexpensive VFOs that they talk about a lot on the Softrock list?

John

[Joe]

I did a quick Two tone transmitter intermod test of the Hermes-lite board using the Setup Test menu on PowerSDR. I have both a HP spectrum analyzer and a well shielded home brew softrock superhet  which I believe works better for narrow band display because of band width and

stability issues on the old HP. 

The results are shown below on a screen shot using Rocky to display the results which in my option look very good. You can see the fundamental 

with the intermod down greater than -50db this is for the AD9866 only no amplifiers on 40 meters. BTW I also checked on the low end of 10 and the results where the same.

The more I use and experiment with this project it's amazing how well it works.  Thank you Steve for your idea and perseverance to see it 

through.  

Joe   wa9cgz

[Steve Haynal]

Hi Joe,

Thanks for the encouraging words. I've actually been very busy with work and family the past few weeks, so haven't had too much time to spend on the Hermes-Lite. Now that the firmware is hopefully more stable for C8 users, the next items on my to do list are:

• Experiment with the TX interpolation settings to see if we can reduce the harmonic you saw on 10M and also increase the TX range up to the top of 10M. I'd appreciate your help with some spectrum plots when I have something to test here.
• Investigate why we are seeing more TX wspr drift now. I suspect the second current mirror that I activated. This requires modifying the RTL to write the gain settings differently and we may need to test that the gain adjustment still works properly.
• Investigate the artifacts at 3.84 and on 384kHz bandwidth with no antenna connected. This is probably related to decimation choices made when enabling 384kHz for the new 61.44MHz crystal.
• Overall simplification to make room for additional receivers.

73,

Steve

KF7O