Pure Signal with HL2

[Don Solberg]

I am still working on getting Pure Signal to work with the HL2.  I was not able to get enough leakage feedback to get any indication on Power SDRmxPS that it was getting any feedback.

 

I have since purchased an external bi-directional coupler with a PureSignal feedback port from Wavenode.  I worked with Alan from Wavenode to increase the output, as I wanted to test with the HL2 barefoot before I tried it with my amplifier.  I will add external attenuation when I use the amplifier.

 

I measured the output from the Wavenode sensor while running "full power", 5.9 watt tests on 75M.  I measured the feedback with my spectrum analyzer: -17.1dbm with a single tone, and -24dbm with a dual tone.  From what I have read, -17dbm should be in the ballpark.

 

My plan is to connect a coax to  RF3 on the PCB and bring this out to an SMA panel connector.  I hope to test that tomorrow.  Will this work or do I have to make any other changes?

73,

Don

K9AQ

 

[Steve Haynal]

Hi Don,

I think it should just work. I would test this new input with general receive (it is an alternate receive input) just to make sure the path is good. Also, when the gateware went up to 4 receivers, we started using the 4th receiver instead of the 2nd for PureSignal. For general information, people will have to select a larger radio model (anan100d) in PowerSDR for PureSignal to use the 4th receiver.

73,

Steve

kf7o

[Don Solberg]

Steve,

This is were I get confused.  I can't find any reference in the PowerSDR forum for using a 4th receiver for Pure Signal.  On the You Tube video on setting up PowerSDR for an Anan 100, they tell you to select Ext 1 on transmit.  This is done on the Ant/Filters tab.  

I have attached a screen shot of that tab.  As you can see there are two options: Ex2 "enable RX2 in on Alex or ex1 on Anan during transmit", or Ext2 on Transmit "enable RX1 in on Alex or Ext 2 on Anan during transmit".  Changing from a radio type of Hermes to an Anan 100D does not change these options.

What was the reason for changing from RX2 to RX4?

There is supposed to be an indication on the main screen in PowerSDR to tell you if Pure Signal is turned on.  I see that and the color is green, which is correct.  There are also supposed to be indicators for Feedback and Correcting.  Neither of these display on my screen. They should show to the right of the Pure Signal text.

Don

K9AQ

[Alan Hopper]

Hi Don,

receiver 4 is used to feedback the signal that is sent to the dac so it can be compared to that measured by the adc. Receiver 4 is used by all but the very small anan radios, There is no mention or control over this in powersdr as it is an internal thing that should just work.

73 Alan M0NNB

[Don Solberg]

Thanks Alan for clearing this up.  Do you know why I am not seeing the Feedback level and Correcting on PowerSDR.  I do see "Pure Signal" displayed in green when it is enabled.

73,

Don K9AQ

On Thursday, November 7, 2019 at 5:23:25 PM UTC-6, Don Solberg wrote:

[Don Solberg]

I found this video on You Tube I found a You Tube video that shows Pure Signal 2 in operation on Power SDR mXPS v3.4.9 https://www.youtube.com/watch?v=QcBV3oryEwg

This clearly shows that the feedback and correcting indicators don't display until Pure Signal receives feedback.  I have been testing my setup without the external coupler feedback connected, hoping to have enough internal feedback for Pure Signal to marginally work.  I am away from my bench for a couple of days.  When I return I will wire the coax to RX3 and we will see what happens.

73,

Don

K9AQ

On Thursday, November 7, 2019 at 5:23:25 PM UTC-6, Don Solberg wrote:

[Steve Haynal]

Hi Don,

I checked the Hermes and Angelia RTL today. The signal sent to the DAC as Alan mentioned is on RX4 for the Hermes and RX5 for the Angelina. You will want to select Hermes in PowerSDR. If there is a problem here, I can send you some gateware using different receivers.

For the feedback from the output of the PA, any other RX is fine as there is just a single ADC in the HL2.

73,

[Don Solberg]

Steve,

I connected an SMA panel connector with coax to RX3 on the PCB.  I confirmed with my signal generator that the RX3 input was working.  I then hooked up my Wavenode bi-directional coupler with the Pure Signal sample output to RX3.  I get a signal on the sprectrum display but Pure Signal is not working.  The feedback and correcting displays are not being displayed.  I am also not seeing any activity in the Amp View. Attached is a screen snapshot showing the sampled RF being displayed.  I also attached another screen snapshot that I found on the Internet that shows the feedback and correcting displays that should be there when Pure Signal is working.

I would appreciate any help that you can provide.

73,

Don

K9AQ

On Thursday, November 7, 2019 at 5:23:25 PM UTC-6, Don Solberg wrote:

[Steve Haynal]

Hi Don,

I looked at PureSignal tonight and found an issue with the gateware. The DAC signal feedback was being sent on RX 5 and 6 if present. I moved that to RX 2 and 4. For the Hermes, it expects the DAC feedback on RX4 and the real RX on RX3. There are a few timing paths I need to clean before I release a new testing, but I will send you the .rbf and .jic privately for your early testing. 

A key adjustment for me was to manually set the SetPk value, typically to something very close to the GetPk that is being measured. Also, I did not use Auto-Attenuate but adjusted the S-ATT until I saw green feedback and correcting. A PureSignal expert could provide more details on adjustment.

Below is the low power PA uncorrected. Here I have a cable with 30dB attenuation from the low power output to the general RF connector. The top spectrum is the real RX. The bottom spectrum is the signal being sent to the DAC.

Here is the same low power but with predistortion. Note that the low power output is pretty clean to begin with and the amp view doesn't show much differentiation.

Below is with the 5W PA on and connected to a dummy load, no predistortion. No wired feedback is used, only stray pickup.

Now the same 5W setup but with predistortion on. Note that the AmpView shows more variation.

73,

Steve

kf7o

[Don Solberg]

Steve,

That is GREAT news!  Thanks for your help with this.  The HL2 with Pure Signal should be one of the best xcvr's for those of us with LDMOS amps that are looking to improve the IMD. Since I am in the midst of moving, I won't have my LDMOS amp ready for Pure Signal testing until late this winter.  I can however test it with a 1KW VMOS amp.  I will do that testing once I have the Pure Signal working on the barefoot HL2.  It is very exciting to see that it will work.

No rush on cleaning up the gateware.  I will be gone this weekend and pretty busy with the move next week.

Thanks again,

73

Don

K9AQ

On Thursday, November 7, 2019 at 5:23:25 PM UTC-6, Don Solberg wrote:

[Don Solberg]

Thanks to help from Steve,  I have Pure Signal working on my HL2. Steve furnished a modified version of the gateware which I uploaded with SparkSDR.

I followed Steve's instructions on how he got it to work, and I was know able to now see the "feedback" and "correcting" displays.  Using just internal feedback, I was not able to get a "green" indicator for feedback, which indicates the correct feedback level.  My display is "yellow", which is marginally low, but Purse Signal is still working.

You can see the improvement on the PowerSDR screen but it is much easier to evaluate on a spectrum analyzer. There was almost an 11db improvement on the 3rd order IMD, and the 5th order almost disappears in the noise floor. I assume that with a higher sampling level the correction would be even better.

This is what the IMD looks like without PureSignal on.

This is what it the IMD looks like with Pure Signal on.

My next experiment will sampling with the external coupler in to ALT RX antenna connection.  My concern is the sampling level may be too low with 5W to override the internal coupling.  I want to test this with my 1KW homebrew solid state amp, but this will have to wait until we are done moving, which will be the first week of December.

Thanks Steve for all of your help.

PureSignal does indeed work on an HL2 with PowerSDRmXPS!

73,

Don

K9AQ

[Don Solberg]

I connected a Wavenode bi-directional coupler, with a PureSignal sampling port, to the output of the HL2.  I then fed the sampling port output through a step attenuator, and then into RF3, the optional RX only antenna connection.  I installed an SMA connector on the HL2 panel and ran coax to the RF3 pin on the PCB.

This gave me a larger feedback signal than what I was getting with the internal feedback alone.  I found that running the HL2 barefoot that I didn't need any additional attenuation.  I am now getting a "green" display for the "feedback".  I changed the configuration to enable auto attenuation, and it is setting the S-ATT  to 4 when I am transmitting.

I can't wait to hook this up to my amplifier.

73,

Don

K9AQ

[Steve Haynal]

Hi Don,

Looks good! On the new back panel I've included a hole to install an additional SMA for PureSignal use. Which SMA connector did you use?

73,

Steve

kf7o

[Don Solberg]

Steve,

I bought a package of 5 female SMA connectors with bulkhead nut and sold cup terminals on Amazon https://www.amazon.com/gp/product/B078H4F8R6/ref=ppx_yo_dt_b_asin_title_o02_s00?ie=UTF8&psc=1

73,

Don

K9AQ

[Don Solberg]

Pure Signal with the HL2 and an external amplifier is working!

The output of the HL2 goes through a Wavenode bi-directional coupler, with a Pure Signal sampler port, to my homebrew SD2943 amplifier.  Normally I drive this amp with an Elecraft K3-10, which outputs 12 watts.  I routinely get about 850 watts on 75M SSB, with 5 watts from the HL2, I can only get about 350 watts.

Alan from Wavenode increased the feedback from his sensor so that I could get more feedback when I was testing the HL2 barefoot.  It is putting out about 15% more power out of the sampling port.  I used a step attenuator to reduce the sampling feedback by 12db. This got the feedback sample to a level that the auto attenuator could handle and give me a Green feedback 

indicator on my voice peaks.

Steve found that you need to manually set the SetPk setting to be close to the GetPk reading.  I have found that I have to do this every time I restart PowerSDR.  I don't know why you have to do that, none of the Pure Signal set up instructions talk about that setting. 

I wasn't able to look at the output on my spectrum analyzer to accurately measure the IMD improvement, but I "tested" it on a 75m SSB net that I routinely check into.  I got all positive comments on the quality of my audio with the HL2 and the amp.  Even without PS on, the audio quality of the HL2 with PowerSDRmxPS is really good.  Hams who know my voice felt that the HL2 had better sounding audio than the K3.  This web site gives detailed instructions on setting up the TX audio.  http://www.w1aex.com/anantxgain/anantxgain.html You have to do this before you start setting up Pure Signal.

I still have a lot more testing that I want to do, but I am in the middle of a house move so I don't have much time to work on this.  I will probably put it on hold until later this winter.  My goal was to use Pure Signal to improve the IMD from my dual LDMOS homebrew amplifier.  I have the PA ballet working, and the LPF is tested and tuned, so the next step is to put everything in the cabinet.  This amp only needs a couple of watts to drive it to 1,500 watts so the HL2 is ideal for use with this amplifier. The IMD from this amp is not as clean as my SD2943 amp so Pure Signal should help a lot.

Here is a screen snap shot of the Amp View in Pure Signal 2.  It clearly show the adjustments that Pure Signal made to make the amplifier more linear.

One of the hams on the net took a screen snapshot of my signal on his Flex 6500's panadapter.  You can clearly see how clean the signal is.

Thanks again to the help that I got from Steve and from Alan at Wavenode. 

73,

Don

K9AQ

[Chris Moore]

Hi Don - 

I've been following your progress closely - I have a coupler on the way and look forward to trying it out once the gateware is released. 

 A question - shouldn't the coupler be sampling the output from your SD2943 (and not the HL2), otherwise you're linearizing the HL2 but not the amp? 

[Bob A. Booey]

Hi Don,

1.  X2 on the coupler location.

2. Looked over the website for Wavenode (wavenodedevelop.com) but find no mention of couplers, is your coupler custom made/not “off the shelf”?  Could you provide info on it?

3.  Do you have a RF2K+ amp as well?  I recognized your call and found you were a RF2K forum member as well.  I have an RF2K+ and plan on using my HL2 Plus with it eventually.

Thanks and watching your work with interest.

73,

Robert, WA2T 

Sent from my iPhone

On Nov 26, 2019, at 10:00 AM, Chris Moore <rac...@gmail.com> wrote:



--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/1cae6dfd-087e-405e-8fa3-ebb748f5f35a%40googlegroups.com.

[Don Solberg]

I misspoke.  The Wavenode sensor is at the output of the amplifier. I am using an HF-1 with the RFView option https://wavenodedevelop.com/sensors/

I am a member of the RFKits forum but I do not own one.  I have a dual BLF188XR amplfier that I homebrewed.  It is similar to the RFKit but the LDMOS transistor are soldered to the copper heat spreader and I use a 16:1 output transformer that is wound with low ohm coax.  The HL2 should be an ideal exciter for your RFKit amp.

73,

Don

K9AQ

On Tuesday, November 26, 2019 at 10:21:13 AM UTC-6, Bob A. Booey wrote:

Hi Don,

1.  X2 on the coupler location.

2. Looked over the website for Wavenode (wavenodedevelop.com) but find no mention of couplers, is your coupler custom made/not “off the shelf”?  Could you provide info on it?

3.  Do you have a RF2K+ amp as well?  I recognized your call and found you were a RF2K forum member as well.  I have an RF2K+ and plan on using my HL2 Plus with it eventually.

Thanks and watching your work with interest.

73,

Robert, WA2T 

Sent from my iPhone

On Nov 26, 2019, at 10:00 AM, Chris Moore <rac...@gmail.com> wrote:



Hi Don - 

I've been following your progress closely - I have a coupler on the way and look forward to trying it out once the gateware is released. 

 A question - shouldn't the coupler be sampling the output from your SD2943 (and not the HL2), otherwise you're linearizing the HL2 but not the amp? 

On Tuesday, November 26, 2019 at 8:35:43 AM UTC-6, Don Solberg wrote:

Pure Signal with the HL2 and an external amplifier is working!

The output of the HL2 goes through a Wavenode bi-directional coupler, with a Pure Signal sampler port, to my homebrew SD2943 amplifier.  Normally I drive this amp with an Elecraft K3-10, which outputs 12 watts.  I routinely get about 850 watts on 75M SSB, with 5 watts from the HL2, I can only get about 350 watts.

--
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 herme...@googlegroups.com.

[Duncan Clark]

In message <1cae6dfd-087e-405e...@googlegroups.com>,
Chris Moore writes

>I've been following your progress closely - I have a coupler on the way
>and look forward to trying it out once the gateware is released. 

Just tested for real with QRO this afternoon a homebrew coupler using
the design by VK1HW that he uses with his ANAN and SPE 1.3K Amplifier.

https://sites.google.com/view/vk1hw/homebrew/taps/rf-sampler

I get -50dB good past 50MHz and it didn't blink at the 400W I fed mine
this afternoon for quite while whilst looking at my harmonic levels
using an old spectrum analyser.

Duncan

--
Duncan Clark
G4ELJ

[ROBERT ENTWISTLE]

Hi Duncan,

What were you using for an exciter, an Anan transceiver? Guessing something other than an HL2 since you mention past 50 MHz.

Thanks for the link to the coupler designs.

73,

Robert, WA2T

> --
> 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.
> To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/1PVaHqBf%2BW3dFwQQ%40ntlworld.com.

[Duncan Clark]

In message <1972616170.2295...@connect.xfinity.com>, ROBERT
ENTWISTLE writes

>What were you using for an exciter, an Anan transceiver? Guessing
>something other than an HL2 since you mention past 50 MHz.

I was using my main rig, a modified Pic-A-Star txcvr.

Designed by G3XJP and published in the UK's RSGB Radcom magazine over
several issues maybe 15 or so years ago. Effectively think a homebrew
Elecraft K3 before the K3 appeared.

High IMD narrow bandpass filters, v.high IP H-Mode mixer front end,
AD9951 DDS mixed down to narrow Xtal filter at 10.695MHz, low amount of
IF amplification then another H-Mode mixer down to 15kHz and into an
AD1885 CODEC plus ADSP2181 processor (old by today's standards) where
everything is done in DSP. All controlled by a 16F870 PIC. All normal
(and more) bells and whistles as available on any current commercial
txcvr.

Designed originally for 160m-10m, mine does 6m as well and has a more
PA3AKE-like H-mode mixer using FSA3157's plus various other bits like a
Softrock based panadaptor.

It's all homebrew including the pcbs. Mix of leaded and SMD components

10W PEP out drives my homebrew linear to 400W on all bands using a
Russian GU43B tetrode.

>
>Thanks for the link to the coupler designs.

Ignore his comment on the alternative European ferrite to use. No way
will that fit on any RG8 or similar sized coax.

I used FT82A-43 bought from Qubits in the UK for anyone interested over
here. http://radio-store.co.uk/

[Steve Haynal]

Hi Group,

I moving this thread below back to the list. I had sent Duncan the PureSignal-capable gateware with 4 receivers and he made some experiments with his HL2. He found that a 470Ohm resistor was useful to not degrade the main RX signal.

73,

Steve

kf7o

Steve Haynal writes
>Attached is the .rbf. I hope to release this weekend.
>
>I think a resistor in the 500-2000 Ohm range may be a good start. That
>should keep the overall RX path impedance close to 50 Ohms. I'm not
>sure if/what difference this will make, but am interested in any
>experimental or analytical results you may come up with.

OK gateware loaded and tested.

Definitely needs a resistor in the signal path from the coupler. No
resistor drops sensitivity by 3dB. 470 Ohm has no effect on sensitivity
so am using that and feeding the feedback signal into RF5.

The attached screen grabs show the effect at 400W PEP on 15m and I've
tested from 17m-10m. Minimum 20dB improvement in all IM3 products.
Auto-attenuate sits at around 0dB.

Normally I would use a US based WebSDR to observe my signal but there
don't seem to be any around today with low noise and decent antenna's.

However from from QSO's it all looks and sounds good.


Enjoy thanksgiving.

73's

Duncan

[Steve Haynal]

Hi Group,

I created another two quick PCBs with PureSignal feedback in mind. They have circuitry for either a Pi or T attenuator. 

https://chemandy.com/calculators/matching-pi-attenuator-calculator.htm

https://chemandy.com/calculators/matching-t-attenuator-calculator.htm

This is to provide a better match to the PA tap and still provide a high impedance sideload to the HL2 main RX path.

This board is meant to mount on top of the existing N2ADR/HL2 small connector board. There are half holes on the edges to facilitate soldering it to the connector board which don't show up in the 3D view.

https://github.com/softerhardware/Hermes-Lite2/raw/master/hardware/companions/psfeedback/psfeedback.pdf

This board is an entire replacement for the N2ADR/HL2 jumper. It also is an example of adding an I2C chip. In this case there is a MCP23008 port expander. 

https://github.com/softerhardware/Hermes-Lite2/raw/master/hardware/companions/io_jumper/io_jumper.pdf

73,

Steve

kf7o

[Duncan Clark]

In message <2fb9b62b-3de0-4139...@googlegroups.com>, Don
Solberg writes

>Steve found that you need to manually set the SetPk setting to be close
>to the GetPk reading.  I have found that I have to do this every time I
>restart PowerSDR.  I don't know why you have to do that, none of the
>Pure Signal set up instructions talk about that setting. 

I find exactly the same. However it is writing the correction values to
the xml database.

Maybe there is a buglet in the 3.4.9 version that simply doesn't re-read
the values. No idea where it is getting the default 0.4072 value from.
If we could alter that to 0.233 (the figure I come out with or
thereabouts I would be happy.

I'm not a programmer so although the PowerSDT source code is there I've
no idea where to begin.

I thought about e-mailing W5WC but he has no e-mail address showing on
qrz.com

[Don Solberg]

I have not seen any reference to this on the PowerSDR forum.  Is it possible the the HL2 gateware is resetting this value?

Don

K9AQ

On Thursday, December 5, 2019 at 1:17:37 PM UTC-6, Duncan Clark wrote:

In message <2fb9b62b-3de0-4139-b978-647038...@googlegroups.com>, Don

[Duncan Clark]

In message <03786be1-cc37-46a9...@googlegroups.com>,
Steve Haynal writes

>Normally I would use a US based WebSDR to observe my signal but there
>don't seem to be any around today with low noise and decent antenna's.

Finally found a WebSDR in the US that was rxing me pretty well this
afternoon on 17m.

My signal on 18125 looked nice and clean on SSB, especially compared to
the signal on 18150, and from 2 tone testing I can see the improvement
when running Pure Signal. All good.

[Steve Haynal]

Hi Don,

The HL2 doesn't have direct control over any of those values. From the gateware perspective, there is one control bit that is set by software to enable PureSignal. When set, the HL2 then sends back the TX output signal back to the host on receiver. The software monitors this internal TX loopback as well as an external TX loopback on two receivers. There is nothing more that the gateware does. Maybe the HL2 setups are just producing values in a range not expected by the software.

73,

Steve

kf7o

[Don Solberg]

We completed our move so I am back to working on radio projects.  

I used a T connector on the output of the step attenuator from the Wavenode RFView sampling port.  This allows me to monitor the transmitted signal with my spectrum analyzer.  I am seeing a lot of artifacts on PowerSDR outside of my signal.  I was pretty sure these were not on my signal but rather crosstalk in the HL2.  The spectrum analyzer output is clear and shows about a 10db improvement with PureSignal on.

With 5 watts output from the HL2, I can only get 325 watts out of my 1KW homebrew solid state amplifier.  This is not the "sweet spot" for this amp, so I suspect the IMD improvement could be better at 800 watts or more.  When I finish my LDMOS amplifier, I should be able to get full output with the HL2.

On-air reports of the audio have been excellent.  The audio quality is noticeably better with PureSignal on, and the signal strength is about 3db greater at the same power level.

I think there is still an issue with internal cross talk in the HL2 when using the optional RX only port for the Pure Signal feedback.  Attached are a screen snapshot showing the PowerSDR panadapter, as well as on from my spectrum analyzers.  The spectrum analyzer has a two tone display at 3.940 with out PureSignal and one at 3.985 with PureSignal on.

I would welcome any comments on the artifacts on the PowerSDR display.

73,

Don

K9AQ

[Duncan Clark]

In message <2b7db5a9-a881-4ad4...@googlegroups.com>, Don
Solberg writes

> used a T connector on the output of the step attenuator from the
>Wavenode RFView sampling port.  This allows me to monitor the
>transmitted signal with my spectrum analyzer.  I am seeing a lot of
>artifacts on PowerSDR outside of my signal. 

PureSignal will only correct about 20kHz of total signal width so
anything outside of that won't change.

>I was pretty sure these were not on my signal but rather crosstalk in
>the HL2.  The spectrum analyzer output is clear and shows about a 10db
>improvement with PureSignal on.
>

>I think there is still an issue with internal cross talk in the HL2
>when using the optional RX only port for the Pure Signal feedback. 
>Attached are a screen snapshot showing the PowerSDR panadapter, as well
>as on from my spectrum analyzers. 

I've the same RSP with the same Spectrum analyser software but I haven't
tried using it with the HL2 and linear combo.

I too have been wondering about pickup and whether the HL2 really is
sensing off my 400W linear output or pickup from the 5W HL2 PA itself
with it being in the same box.

Testing tonight, there is enough pickup from the 5W HL2 PA even at 0.5W
out (2 tone) to give a Green correcting light from PureSignal. Auto
attenuation setting is zero at 0.5W, as is 5W output.

Therefore the key is the ensure that the output from the sampler/coupler
after your linear is feeding enough signal back into your HL2 so that
you override the 5W PA pickup. My VK1HW sampler with ~400W at the linear
output gives enough output signal to cause 9dB of auto-attenuation in
PowerSDR.

From the Ampview graph in the PureSignal window I can see that the
curves look very different from 400W to just the 5W output. If I add in
an additional 10dB attenuation to the sampler the curves revert to more
like the 5W PA on it's own. So IMHO make sure you have enough output
from your coupler to force HL2 to use that as the dominant signal for
PureSignal to work on. I might change my coupler from -50dB to -45dB or
so just to give me an approx. +10dBm output at 400W i.e. a little more
headroom.

I will hook my RSP up as you have done with the Spectrum Analyser
software and measure it all properly at the weekend.

[Don Solberg]

Duncan,

I am very sure that the output from my sampler is high enough to overcome the internal feedback.  I really don't care what the PowerSDR panadapter looks like as long as the output is clean.  The signal looks very good on my spectrum analyzer.  The screen shot that I shared was used a peak setting on the spectrum analyzer, if I set it for averaging the IMD improvement is even better.

I consider this to be a successful test of Pure Signal with the HL2 and an external amplifier.  I really can't wait to get my LDMOS amplifier completed so that I can test it with Pure Signal.  The IMD of the dual BLF188XR amp is not as good as the (4) SD2943 amp that I used for these tests.

Since I moved up to my cabin I no longer need to run an Internet remote control station so I have been using the HL2 and my SD2943 amp as my primary station.  I really love the HL2 and Power SDR for a SSB station.  The receive and transmit audio quality is much better than my Elecraft K3.

73,

Don

K9AQ

On Tuesday, December 10, 2019 at 11:06:03 PM UTC, Duncan Clark wrote:

In message <2b7db5a9-a881-4ad4-95e4-2f70e0...@googlegroups.com>, Don

[Steve Haynal]

Hi Don,

There are several places artifacts could originate from: RF crosstalk in HL2 or ambient, RF crosstalk inside the AD9866 IC, gateware digital signal processing inside the FPGA, software digital signal processing. Do you see the same on 80M and 15M? If it is frequency dependent, I'd suspect the first two. If no change between bands, then something in the DSP.

73,

Steve

kf7o

[Duncan Clark]

In message <2b7db5a9-a881-4ad4...@googlegroups.com>, Don
Solberg writes

>The spectrum analyzer output is clear and shows about a 10db
>improvement with PureSignal on.

I am seeing exactly the same 10dB improvement at 400W on the my SDRplay
RSP-1a with the latest version of the spectrum analyser software. Tested
from 17m through to 10m. PowerSDR shows a lot larger improvement in IMD.

Artifacts, those big spikes every few kHz, are showing on PowerSDR and
do not show on the analyser.

One question for you. Mic gain?

I noticed you have yours set at 16dB. With more than one powered
electret microphone tested, the mic gain has no effect for me. Doesn't
matter what I set it to, it does nothing. Mic boost does nothing either
but I believe that is correct as it would apparently normally turn on a
20dB analog preamp that is contained in a TLV320 CODEC chip that we
don't have on the HL2.

What I can do on the microphone gain is switch to Digi mode which opens
a box that allows me to adjust the VAC gain upwards.

I have been using the following to set it all up.

http://www.w1aex.com/anantxgain/anantxgain.html

Like you found on-air reports of the audio have been great but having a
mic gain control that does nothing is a little disconcerting.

[Don Solberg]

Duncan,

I set up my audio using the same link.  If I set the RSP Spectrum Analyzer to "avg" instead of "peak" the IMD improvement seems to be higher.  I really appreciate the SDR Play Spectrum Analyzer software.  It is amazing how well this low cost SA works.  I had an older HP SA and the SDR play offers a lot more features.  I don't know what I would do without it.

I will double check my Mic gain tomorrow, but I thought it had been working.  I am using a Heil Proset Media headset, which is designed for use with sound card input to an SDR. 

The artifacts that we are seeing in the Power SDR panadapter don't seem to "really" exist, and they don't hurt the signal quality, so I don't really care if they are displayed.

73,

Don

K9AQ

On Friday, December 13, 2019 at 12:30:53 PM UTC-6, Duncan Clark wrote:

In message <2b7db5a9-a881-4ad4-95e4-2f70e0...@googlegroups.com>, Don

[Don Solberg]

Duncan,

I had an "old age" minute yesterday and forgot how you have to setup the mike gain for the HL2 with PowerSDR.  The ANAN radios have a mike jack that feeds directly into the radio, this is what the front panel mic gain controls.  When you use an external sound card, you have to treat the mike like line input.

Go to the Transmit setup tab in Power SDR.  There you will see how you control the mike gain.  It is not as handy as having it on the main screen but once you get the entire transmit audio chain setup you shouldn't have to change it.  If you use different microphones, you can set up a profile for each one,

73,

Don

K9AQ

On Fri, Dec 13, 2019 at 7:39 PM Don Solberg <dsolbe...@gmail.com> wrote:

Duncan,

I set up my audio using the same link.  If I set the RSP Spectrum Analyzer to "avg" instead of "peak" the IMD improvement seems to be higher.  I really appreciate the SDR Play Spectrum Analyzer software.  It is amazing how well this low cost SA works.  I had an older HP SA and the SDR play offers a lot more features.  I don't know what I would do without it.

I will double check my Mic gain tomorrow, but I thought it had been working.  I am using a Heil Proset Media headset, which is designed for use with sound card input to an SDR. 

The artifacts that we are seeing in the Power SDR panadapter don't seem to "really" exist, and they don't hurt the signal quality, so I don't really care if they are displayed.

73,

Don

K9AQ

On Friday, December 13, 2019 at 12:30:53 PM UTC-6, Duncan Clark wrote:

In message <2b7db5a9-a881-4ad4...@googlegroups.com>, Don

--
You received this message because you are subscribed to a topic in the Google Groups "Hermes-Lite" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/hermes-lite/03cl7KLlbqM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to hermes-lite...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/eae20185-3b11-4471-8aaf-ef3c27a7f806%40googlegroups.com.

[Don Solberg]

 I reread one of the web sites for setting up Pure Signal and they recommended setting the TX panadapter for a minimum signal display of -60db.  I played around with it, and found that -70 worked good for me.  This really cleaned up the TX pandapater.

73,

Don

K9AQ

On Friday, December 13, 2019 at 12:30:53 PM UTC-6, Duncan Clark wrote:

In message <2b7db5a9-a881-4ad4-95e4-2f70e0...@googlegroups.com>, Don

[Duncan Clark]

In message
<CAGjZ1-qfGXiPR=ovzn5HUMe3Fxm-fWuk...@mail.gmail.com>,
Don Solberg writes

>Go to the Transmit setup tab in Power SDR.  There you will see how you
>control the mike gain. 

Got it.

Ta.

[Andreas]

Hi,

in between I also made some tests with MRF300 and MRFE6VP61K25H transistors and found out, that HL2 output is

clean up to 2.6W output (SDR Console power level @ 0.7). I have posted measurements on my github.

https://github.com/dl9lj/pa

Did you make the same experience?

73, Andreas (DL9LJ)

--
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.

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/03786be1-cc37-46a9-b7ac-d868ecc0ded4%40googlegroups.com.

[Steve Haynal]

Hi Andreas,

Very interesting work. May I post a link to this on the wiki?

I have used PureSignal with 5W out with the standard HL2 PA. You mention 2.6W below, but doesn't your build produce higher power?

73,

Steve

kf7o

To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite+unsubscribe@googlegroups.com.

[in3otd]

Hello Andreas,

if you mean that the TX IMD with a two-tone signal is good up to 2.6 W (average) all should be fine, the PA is expected to provide 5 W PEP (and PEP is two times the [average] power for a two-tone signal) - or also 5 W CW.

You can find some measurements of an H-Lv2b8 PA output power and IMD here.

73 de Claudio, DK1CG / IN3OTD

To unsubscribe from this group and stop receiving emails from it, send an email to herme...@googlegroups.com.

[Andreas DL9LJ]

Hi Claudio, thank you for the link. I use H-Lv2b8 with old transformer - at 34dBm pout I measured -40dBc -

at higher power levels it gets worse. With the new transformer the signal will get better at higher power levels - right?

To measure the IMD of my PA I need to get a clean input. First I thought, that the very clean  measurements from Simon

are behind the 5W PA. https://www.sdr-radio.com/Radios/Hermes-Lite2

So now I made my measurements with 2.6W (34dBm) and found out, that the MRFE PA is also clean at this

level -38dBc @130W/51dBm. The gain of the transistor is 27dB but I have a 10dB attenuator at the input.

Next step is to wind up a new transformer and reduce the input attenuation to find out until which output power the

MRFE PA stays clean (< -35dBc). After that I will continue with PureSignal.

73, Andreas (DL9LJ)

   

To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/ef08cc51-ab40-469c-b92b-5ac201faf6e5%40googlegroups.com.

[in3otd]

Hello Andreas,

rewinding the transformer may help to get 5 W also at 30 MHz but the IMD performance will not necessarily improve at the lower end of the HF - in fact they were actually better with the old transformer at 1.8 MHz here, for some reason. You can find some measurements with the H-Lv2b8 old transformer at the end of the same webpage.

The easiest way to get a very clean signal for testing your high-power PA could be to use the H-L low power output, which is extremely clean (IMD3 is lower than 60 dBc at full output, IIRC), to drive one of the high-power PAs which will in turn fully drive the PA under test.

73 de Claudio, DK1CG / IN3OTD

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/ef08cc51-ab40-469c-b92b-5ac201faf6e5%40googlegroups.com.

[Chris Moore]

Hi Steve (or Don) 

Can you confirm if gateware 20191110_68p4 includes the PureSignal fixes you mention below?  I have a coupler arriving today but am running this gateware and can't seem to get PS working via the internal feedback.

Cheers

On Friday, November 15, 2019 at 12:16:54 AM UTC-6, Steve Haynal wrote:

Hi Don,

I looked at PureSignal tonight and found an issue with the gateware. The DAC signal feedback was being sent on RX 5 and 6 if present. I moved that to RX 2 and 4. For the Hermes, it expects the DAC feedback on RX4 and the real RX on RX3. There are a few timing paths I need to clean before I release a new testing, but I will send you the .rbf and .jic privately for your early testing. 

A key adjustment for me was to manually set the SetPk value, typically to something very close to the GetPk that is being measured. Also, I did not use Auto-Attenuate but adjusted the S-ATT until I saw green feedback and correcting. A PureSignal expert could provide more details on adjustment.

Below is the low power PA uncorrected. Here I have a cable with 30dB attenuation from the low power output to the general RF connector. The top spectrum is the real RX. The bottom spectrum is the signal being sent to the DAC.

Here is the same low power but with predistortion. Note that the low power output is pretty clean to begin with and the amp view doesn't show much differentiation.

Below is with the 5W PA on and connected to a dummy load, no predistortion. No wired feedback is used, only stray pickup.

Now the same 5W setup but with predistortion on. Note that the AmpView shows more variation.

73,

Steve

kf7o

On Thursday, November 14, 2019 at 12:12:30 PM UTC-8, Don Solberg wrote:

Steve,

I connected an SMA panel connector with coax to RX3 on the PCB.  I confirmed with my signal generator that the RX3 input was working.  I then hooked up my Wavenode bi-directional coupler with the Pure Signal sample output to RX3.  I get a signal on the sprectrum display but Pure Signal is not working.  The feedback and correcting displays are not being displayed.  I am also not seeing any activity in the Amp View. Attached is a screen snapshot showing the sampled RF being displayed.  I also attached another screen snapshot that I found on the Internet that shows the feedback and correcting displays that should be there when Pure Signal is working.

I would appreciate any help that you can provide.

73,

Don

K9AQ

On Thursday, November 7, 2019 at 5:23:25 PM UTC-6, Don Solberg wrote:

I am still working on getting Pure Signal to work with the HL2.  I was not able to get enough leakage feedback to get any indication on Power SDRmxPS that it was getting any feedback.

 

I have since purchased an external bi-directional coupler with a PureSignal feedback port from Wavenode.  I worked with Alan from Wavenode to increase the output, as I wanted to test with the HL2 barefoot before I tried it with my amplifier.  I will add external attenuation when I use the amplifier.

 

I measured the output from the Wavenode sensor while running "full power", 5.9 watt tests on 75M.  I measured the feedback with my spectrum analyzer: -17.1dbm with a single tone, and -24dbm with a dual tone.  From what I have read, -17dbm should be in the ballpark.

 

My plan is to connect a coax to  RF3 on the PCB and bring this out to an SMA panel connector.  I hope to test that tomorrow.  Will this work or do I have to make any other changes?

73,

Don

K9AQ

 

[Steve Haynal]

Hi Chris,

20191110_68p4 has a PureSignal bug that was introduced when increasing the receiver count to 4. I sent a patched version to Don and a few others, but that had some logic timing violations which I wanted to clean before making a general release. The next release which fixes both the PureSignal bug and the logic timing violations should be released in the next few days. I am currently testing things like PureSignal and remote gateware upgrades. Sorry for the confusion.

73,

Steve

kf7o

[Don Solberg]

Chris,

The default setting for SetPk in the PureSignal advanced setting doesn't work with the HL2.  You need to change that value to something close to the value the PureSignal shows for GetPk.  For me, I set it to 0.23.  You will have to reset this every time you close down PowerSDR and then restart it.  I don't know if there is anything in the gateware that could change this value.  My understanding is that this in PowerSDR.  Maybe Steve can add some more information.

73,

Don

K9AQ

--
You received this message because you are subscribed to a topic in the Google Groups "Hermes-Lite" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/hermes-lite/03cl7KLlbqM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to hermes-lite...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/cafd934c-bf5f-49de-b808-6e8a52a800b3%40googlegroups.com.

[Steve Haynal]

Hi Group,

There really isn't anything in the gateware for PureSignal. When PureSignal is enabled, one of the receivers processes the data sent to the DAC instead of the RX ADC as feedback to the software. There could possibly be some differences here related to the 12-bit versus 16-bit and the final amplitudes seen by the radio.

I have started a wiki page for PureSignal documentation. Since I am not a PureSignal user (I used Linux), I'd appreciate help with the documentation:

https://github.com/softerhardware/Hermes-Lite2/wiki/PureSignal

73,

Steve

kf7o

To unsubscribe from this group and all its topics, send an email to hermes-lite+unsubscribe@googlegroups.com.

[Chris Moore]

Thanks Don/Steve -

For another data point, I now have PS working with the 69p0 gateware, an Xtronic -44db coupler and Elecraft 100W amp.  Using the RSP spectrum analyzer (thanks for the tip Don), looks like I'm seeing similar results of ~10-12db down for the 3rd order IMD product with PS on.  At ~80-100W output, S-ATT auto-attenuates to 10, and I'm using a 680 ohm resistor between the extra SMA connector I added and RF3.  The cleaner signal is really apparent on a remote SDR:

Cheers - Chris

AG5RR / VK3ICM

[Don Solberg]

Chris,

Welcome to the PureSignal club! Once I got the new gateware from Steve PureSignal has worked great.  I use it all the time now.  Between Pure Signal and excellent TX audio features in PowerSDR, I get consistently great audio reports.

With my experience with PureSignal, I don't think I would buy a new radio that didn't offer this.  I know that really narrows down the list of available radios!  I plan to stay with the HL2.  My Elecraft K3 has been left sitting unused on the desk.  Hopefully in the future more vendors will see the benefit of adaptive predistortion in reducting IMD.

73,

Don

K9AQ

[Chris Moore]

I'm curious if anyone has successfully used PS with pihpsdr?  I've tried v2.0.0-rc1 and 2 and it doesn't look like it's seeing any feedback.

Cheers - Chris

[Steve Haynal]

Hi Chirs,

Are you using the latest testing gateware? There was a PureSignal bug. Also, with PowerSDR we have to tweak a few settings (see older posts) before proper feedback. Maybe something similar to this is required for pihpsdr.

73,

Steve

kf7o

[Chris Moore]

Hi Steve -

Yes, using 69p0 and manually setting the value for SetPk in Pihpsdr (I have it working  in PowerSDR/Windows..). 

Sent from my iPhone

On Feb 4, 2020, at 11:28 PM, Steve Haynal <softerh...@gmail.com> wrote:



--

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.
To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/9bb9bc0c-9f9b-4633-bcd8-d31d8c08d4ab%40googlegroups.com.