N2ADR observations

289 views
Skip to first unread message

John Williams

unread,
May 29, 2019, 6:41:21 AM5/29/19
to Hermes-Lite
Having the ability to read power data using my board designs for HL1, I
have been using PowerSDR and PiHPSDR directly with the power readings
tracking within 10% of being correct. SparkSDR has always been about 10X
low using my boards. As I have now transitioned to HL2 I note that the
power readings are correct on Spark and Quisk (never used Quisk with
HL1). PowerSDR shows an interesting behavior. The power reading is 10X
too large i.e. shows around 50W when being driven with around 3.5 actual
watts. Also, the power is reversed. i.e. the forward is low but the
reverse is high. If you view the actual ADC readings on the PA-Watt tab
you can see this behavior. In using SparkSDR with my HL1 I did not see
any reversal. I am wondering if the way Apollo reads the power is
reversed. I was using the Alex setting to enable power reading for HL1
and it worked FB.

Interesting. I plan to use the external power connector on the N2ADR
board so I can continue to use my power sensing board after my linear
amp so this is somewhat moot for my needs.

John

Steve Haynal

unread,
May 29, 2019, 3:07:52 PM5/29/19
to Hermes-Lite
Hi John,

I think Quisk and maybe SparkSDR allow for two ways of winding and assembly of the power sense. The larger value is forward. Maybe PowerSDR expects only one way.

Also, there was no attempt to calibrate the reading for what PowerSDR expects. So far, this is a Quisk and SparkSDR supported feature.

Finally, with both Quisk and SparkSDR, you can create a power profile by taking ~10 readings at different power levels and calibrating against a known better measurement method. This will improve the accuracy in Quisk and SparkSDR.

73,

Steve
kf7o

Alan Hopper

unread,
May 30, 2019, 4:53:04 AM5/30/19
to Hermes-Lite
Hi John, All,
Spark does indeed assume the larger off the two signals is fwd power so copes with different windings.  The default calibration table In Spark is the one provided by Jim.
As Steve says this table can be edited, it needs at least 2 values.

There is a bug in the current version of SparkSDR that can stop the power sensing working, a work around is described here https://groups.google.com/forum/#!topic/sparksdr/WgdFhBd0_Qg
73 Alan M0NNB

rentwist

unread,
Jun 3, 2019, 9:53:37 AM6/3/19
to Hermes-Lite
Hi John,

Do I understand that you are successfully controlling a HL2 using PiHPSDR?  I have PiHPSDR running on a Rock64 SBC but have had some trouble getting some HL2 things to work.  I selected the Apollo board option (which seems to be exclusive to the Alex board option) and so far I am only seeing about 1.25W out which leads me to believe the N2ADR board is not being controlled (and I hear no clicks when I change bands).  I am feeling my way around in the dark with the PiHPSDR for HL2 use although I have used it successfully with the Radioberry - my RB at the moment has no filter board or PA so maybe that's why I have had success!  Probably some less than obvious settings I am missing?

Any advice or settings guidance for PiHPSDR with HL2 would be much appreciated!

73,

Robert, WA2T

Walter Holmes

unread,
Jun 3, 2019, 2:08:07 PM6/3/19
to Hermes-Lite
Robert,

 In PowerSDR, under Hermes control, did you set the filter options for the different bands, and press the lower left button for Ext Control?

That should then start to set the filters as you change bands etc...

Hope that helps,
Walter/K5WH

ROBERT ENTWISTLE

unread,
Jun 3, 2019, 2:16:08 PM6/3/19
to Walter Holmes, Hermes-Lite
Hi Walter,

I did indeed and PSDR is working fine (for the most part), thanks. I was asking about PiHPSDR as a controller for the HL2. I have been working on a more-powerful-than-RPi-3B+ PiHPSDR controller and would like to try it with the HL2.

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/7568b847-6b1f-4b8b-8fe2-1be3ead1627b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

 

Walter Holmes

unread,
Jun 3, 2019, 3:24:40 PM6/3/19
to Hermes-Lite
ahh, sorry for my confusion.

I was just trying the HPSDR app for Android a few minutes ago, and it seems it wants to see the Atlas environment, and not the Hermes.

I look forward to hearing more about your PiHPSDR enhanced version as you progress. :)

Walter/K5WH
To unsubscribe from this group and stop receiving emails from it, send an email to herme...@googlegroups.com.

John Williams

unread,
Jun 4, 2019, 6:16:28 AM6/4/19
to herme...@googlegroups.com

I have not tried PiHPSDR with my HL2 yet. Still in the queue.

--
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/9063e07b-08f2-4615-9cb9-39ac7b36cb44%40googlegroups.com.

ROBERT ENTWISTLE

unread,
Jun 4, 2019, 9:05:12 AM6/4/19
to John Williams, herme...@googlegroups.com
Thanks John. I have tried the HL2 with the PiHPSDR running on a Rock64 but having trouble getting the filter board to be controlled by it. No LPF board relay clicks regardless of selection among NONE, ALEX, APOLLO or CHARLY25 (maybe another selection or two I'm forgetting). Selecting APOLLO does make a difference in that I then get a tiny bit of power out, around 1.25W (with Drive level at 100%). This makes sense as APOLLO refers to a power amplifier but I am left wondering why can't I get full power out. What does not make sense is why the APOLLO and ALEX options seem to be mutually exclusive meaning only one can be selected, not both. ALEX refers to a filter board so why can't both APOLLO and ALEX be selected? These and a myriad of other questions that I struggle to answer in the world of SDR's which ride on the bleeding edge...

So PiHPSDR works "a little" thus far for me. Power SDR seems to much more functional for HL2. I plan on trying Quisk (have that ready and working on a W10 laptop) and Spark SDR which I have yet to dig into. I get the impression that Quisk may be the hot ticket for HL2 in that it has native support for it.

Seems the lion's share of time in my SDR work is spent searching for the "killer app" for whatever platform I am fiddling with at the time.

73!

Robert, WA2T

John Williams

unread,
Jun 4, 2019, 7:38:24 PM6/4/19
to Bob A. Booey, Hermes-Lite

Opening it up to the group.

I cannot get PiHPSDR to get any significant power out either. Around 1W. Worked fine on HL1. Tried everything I know to try. Tried LinHPSDR as well with same results. Set OC correctly and set PS power to 38.8 across the board. No Joy.

John

On 6/4/2019 5:00 PM, Bob A. Booey wrote:
Hi John,

Tried the OC stuff.  Relays click now but power out is still the same, so maybe the wrong relays are activating.  Pegasi because I am using a Rock64 (faster proc and Gig Ethernet).  I can try it with a regular RPi and see how that goes (although my goal is the R64), might shed some little light on what I’m seeing.

Thanks fur the suggestion and anything else you come up with/observe.

73,

Robert, WA2T 

Sent from my iPhone

On Jun 4, 2019, at 3:50 PM, John Williams <jswi...@gmail.com> wrote:

I will look into this and get back to you. If you’re game, setup the OC like you did for PowerSDR and enable Apollo. 

On Tue, Jun 4, 2019 at 2:06 PM ROBERT ENTWISTLE <rent...@comcast.net> wrote:
Hi John,

I am measuring power with my LP-100A. Interesting on the OC page, I (apparently incorrectly) thought these were GPIO pins of the RPi and so not applicable to the N2ADR board (which uses I2C). But then again I seem to recall that a couple of the GPIO pins can talk I2C (think that is how the Radioberry is controlled by a RPi). Going by memory, there was nothing on the OC page that would relate directly to the N2ADR board, any idea where I should be nosing around for that sort of info?

Oh, and let's say I did get the OC page programmed correctly, wouldn't I need the ALEX (for filters) and APOLLO (for PA) options of the PiHPSDR SW to be enabled simultaneously?

I'm kind of feeling my way around in the dark so please forgive anything I've written that is way off base!

Thanks and 73,

Robert, WA2T


On June 4, 2019 at 2:39 PM John Williams <jswi...@gmail.com> wrote:

How are you measuring your power? Did you set the proper filter bits in the open collector page?

rentwist

unread,
Jun 4, 2019, 8:32:42 PM6/4/19
to Hermes-Lite
Hi John and all,

Here's what I believe is a meaningful data point (on PiHPSDR).  Using the CW key line does yield "normal" power out, (over 5W +/- depending on band), Tune ~1W.  This is of course different than what is seen in PowerSDR.  Going by memory, I saw upwards of 7W on 160, about 6.82W, so even more than I remember seeing on PSDR (unless this additional is a result of incorrect LPF filter selection).

Also, I notice what appears to be a RX instability on 160m, wobbly noise floor peaks at 1.8 and 1.85 MHz, they are present on PSDR as well.

73,

Robert, WA2T
To unsubscribe from this group and stop receiving emails from it, send an email to herme...@googlegroups.com.
--
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.

f6ehp

unread,
Jun 5, 2019, 12:47:15 AM6/5/19
to Hermes-Lite
Hi
Same issue here : LinHPSDR gives full power on HL1 on a tune or pressing on "Mox", but no ssb signal, and no power out at all on the HL2b6.
Pascal

Steve Haynal

unread,
Jun 5, 2019, 12:53:30 AM6/5/19
to Hermes-Lite
Hi All,

Okay. I will have to try LinHPSDR and PiHPSDR. I haven't gotten around to it as I mainly use Quisk and SparkSDR. You will find the best TX support in those two programs.

73,

Steve
kf7o

Jonas Sanamon

unread,
Jun 5, 2019, 1:59:33 AM6/5/19
to Hermes-Lite
Hi Pascal,
Just to be sure: You have enabled the apollo in LinHPSDR (to enable the PA) right? 
Also, there seem to be a bug in LinHPSDR in that the apollo setting is reset every time You open the settings, so be sure to set apollo after every time you change something in the configuration dialog and after starting LinHPSDR

@Steve: Do you know if John Melton has access to a HL2 now?  If not I could be willing to lend him one of mine...

Best Regards,
Jonas - SM4VEY

--
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/172fc3e5-f4d5-47af-b931-d3d76586d4fb%40googlegroups.com.

Jonas Sanamon

unread,
Jun 5, 2019, 2:10:02 AM6/5/19
to Hermes-Lite
Hi Steve,

Thats a good idea :-)  In my opinion LinHPSDR is very promising. 
+ Visually pleasing scope and waterfall, beats everything else i've seen incl. PSDR
+ Crisp receiver with lots of features like for instance NR2 (like PSDR)
- low power on transmit and some generic transmit issue (underrun?)

Best Regards,
Jonas - SM4VEY

--
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/7e4d6c65-9894-41da-a44f-9856720a3d59%40googlegroups.com.

Jonas Sanamon

unread,
Jun 5, 2019, 3:04:31 AM6/5/19
to Hermes-Lite
Oh, and forgot:  As far as I know, LinHPSDR is the only SW option for HL if you're on a Mac. No transmit yet though, but I recently heard John was working on that.
(There used to be a HPSDR SW for Mac called Heterodyne, but I don't think that is available anymore, and seems like it hadn't been updated for years before it disappeared)

Regards,
Jonas - SM4VEY

Alan Hopper

unread,
Jun 5, 2019, 3:49:59 AM6/5/19
to Hermes-Lite
sorry I posted this direct rather than to  the group earlier
Hi Jonas,
John lives close to me and I have offered him mine, In fact he came over a few weeks ago to sort out a problem with LinHPSDR starting the HL2 which turned out to be caused by the watchdog.  Should the group ever end up with surplus HL2 I'm sure he could find a use for it:)
73 Alan M0NNB

On Wednesday, June 5, 2019 at 6:59:33 AM UTC+1, SM4VEY wrote:
Hi Pascal,
Just to be sure: You have enabled the apollo in LinHPSDR (to enable the PA) right? 
Also, there seem to be a bug in LinHPSDR in that the apollo setting is reset every time You open the settings, so be sure to set apollo after every time you change something in the configuration dialog and after starting LinHPSDR

@Steve: Do you know if John Melton has access to a HL2 now?  If not I could be willing to lend him one of mine...

Best Regards,
Jonas - SM4VEY

Den ons 5 juni 2019 kl 06:47 skrev f6ehp <pascal....@gmail.com>:
Hi
Same issue here : LinHPSDR gives full power on HL1 on a tune or pressing on "Mox", but no ssb signal, and no power out at all on the HL2b6.
Pascal

--
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.

Jonas Sanamon

unread,
Jun 5, 2019, 5:12:29 AM6/5/19
to Hermes-Lite
Hi Alan,

Well, I did the same :-)

Thanks for answering.   How about this idea then:
We set up a collection/fundraiser for getting John (G0ORX) a HL2,  hopefully Steve will have some extra of the spares to sell in a month or two. In the mean time I can lend John one of mine HL2s.   I'll throw in the first $10.

Thoughts?
Steve,  possible?

Best Regards,
Jonas - SM4VEY
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/f29b5e6e-3789-4832-bc6f-f601586d0566%40googlegroups.com.
Message has been deleted

rentwist

unread,
Jun 5, 2019, 9:21:50 AM6/5/19
to Hermes-Lite
Here's a question.  Is there an "interface control document" that all of these SDR's that are controlled over Ethernet are programmed to follow?  And/or vice versa for the various software available to control SDR's?  I'm guessing there is some sort of de facto standard but perhaps not official.  Knowing little to nothing about this aspect of SDR's my feeling is that a lot of the questions that are coming up about what works with what software lead back to the control interface implementation.

This puts me in mind of 3GPP which is a standards organization in the wireless world.

73,

Robert, WA2T

f6ehp

unread,
Jun 5, 2019, 3:02:24 PM6/5/19
to Hermes-Lite
Hi Jonas,
OK with apollo enabled, I also get a small 1W output on a tune or when I press the mox, but no modulation output.
Pascal

Steve Haynal

unread,
Jun 9, 2019, 2:00:31 AM6/9/19
to Hermes-Lite
Hi Robert,

The wobbly noise floor peaks you see at 1.8 and 1.85 MHz are probably harmonics from the onboard switching power supply. If you search this list, you will find early discussion about this. The short story is that we noticed them once the N2ADR filter board was done. The proximity of traces or relays on the N2ADR filter board picked up these harmonics which we weren't seeing before. We decided that given the high noise floors typical of 160M, that these would still be below the noise floor so wouldn't matter. They are clearly visible without an antenna, but should be gone with an antenna. Were you using a proper 160M antenna when you observed them?

Also, there is an option on the HL2 to install R13 and R16, and remove J2 and J11, so that the FPGA can provide the switching clock to the regulators. This will allow us to move these noise humps. The support for this has not yet been tried in the gateware, although it is still on my list. In the simplest use, I envision a checkbox in software that allows you to choose one of two switching frequencies, both of which should place noise humps outside of the 160M band.

73,

Steve
kf7o

Bob A, Booey

unread,
Jun 9, 2019, 8:54:18 AM6/9/19
to Steve Haynal, Hermes-Lite

Hi Steve,

I believe I was using a dummy load since my antenna needs tuning depending on band of interest, have not gotten as far as putting the HL2 on the air yet.

Thank you for the explanation!

73,

Robert, WA2T

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/17f4e951-d479-4ae3-81b3-924ca25da8d4%40googlegroups.com.

Matt

unread,
Jun 10, 2019, 3:56:02 PM6/10/19
to Hermes-Lite
I have recently been trying to improve my MF/LF receive antenna and looking to do some noise cancelling to remove local QRM on 160m. I've been having success with a homebrew active antenna high above the ground. I've been experimenting with some other antennas and low noise pre-amps on 160m. I believe this is how some of the (much) more serious DXers operate too with a low-noise antenna.

These wobbly peak do cause me problems around the CW section on 160m. To be honest, I didn't realise they were coming from the HL2 and it had been on my todo list to track these down! I would welcome being able to move these humps.

73 Matthew M5EVT.

Steve Haynal

unread,
Jun 11, 2019, 1:28:54 AM6/11/19
to Hermes-Lite
Hi Matthew and Group,

This evening I finally tried the experiment of generating the switching clock from the FPGA, and it worked beautifully! Right now I have 125MHz/120 = 1.04166MHz. The second harmonic falls outside of the 160M band. The spur is much narrower and defined "dither." I may try to add some spread spectrum. The 7th harmonic is calculated in the upper part of 40M although I have not observed it. If I switch to 125MHz/119 = 1.0504MHz, then nothing is in amateur radio bands until the 20th harmonic in 15M, which probably doesn't exist. I'll experiment a bit more with frequencies and spread spectrum but will probably settle on two settings that avoid amateur radio bands. This way there will always be an option to move a spur out of the way for SWL or other uses. I will try to release new gateware and mod instructions this weekend.

There is a circular dependency here where the power supply powers the FPGA but the FPGA generates the switching clock that makes the power supply function. This works as the power supply always starts with the default internally generated switching clock and thus breaks the circular dependency. I need to make sure that this always happens reliably, for example, when the FPGA is not programmed at all.

73,

Steve
kf7o

in3otd

unread,
Jun 11, 2019, 1:01:38 PM6/11/19
to Hermes-Lite
Hello,
I see the DC/DC converters noise peaks at around 2 MHz and I noticed that their amplitude depends strongly on the selected filters; without any filter selected they are at the radio internal noise floor with a LNA gain of 10 dB and they are stronger when the lowest frequency filters or the HPF are selected (at about 20 dB above the radio noise floor). Maybe the bigger inductors used there make also good antennas and a shielded version of them - if it exists - may help. Touching the RX input trace with a short wire does not change their amplitude much so I'd guess they are mostly due to the SMPSs inductors magnetic field. Placing a big piece of "EMI absorber" (lossy ferrite-loaded plastic sheet) over the converters inductors decreases the noise amplitude a few dBs.
Certainly the solution Steve implemented by controlling the converters switching frequencies is much more effective and practical.

73 de Claudio, IN3OTD / DK1CG

Steve Haynal

unread,
Jun 12, 2019, 2:06:16 AM6/12/19
to Hermes-Lite
Hi Claudio,

That is a good idea. People can set up software to not use the 160M filter on 160M RX and reduce these spurs.

I discovered I don't have the resolution to do any decent spread spectrum, so will just add the option for a pure frequency outside of any amateur radio bands. The default will be as is. The user can select via software to enable the FPGA-generated signal, maybe by checking dither since this is a type of dither...

73,

Steve
kf7o

in3otd

unread,
Jun 12, 2019, 4:33:32 AM6/12/19
to Hermes-Lite
Hello Steve,
out of curiosity, how were you planning to implement the spread-spectrum? I took a quick look at the Cyclone IV PLLs and while they can not generate a spread spectrum clock it seems you can dynamically change their output delay a bit but I did not try to do any calculation to see if that would give enough spreading.

73 de Claudio, DK1CG / IN3OTD

James Ahlstrom

unread,
Jun 12, 2019, 6:31:01 AM6/12/19
to Hermes-Lite
Hello Steve,

Maybe the default should be to move the spur. I don't see a reason not to.

Jim
N2ADR

Steve Haynal

unread,
Jun 13, 2019, 12:37:26 AM6/13/19
to Hermes-Lite
Hi Claudio and Jim,

Yes, I can't generate spread spectrum with a Cyclone IV PLL. I tried a semi random sequence of similar integer divisors for the 125MHz clock, but that just resulted in an interesting pattern of spurs. :) One problem is that the difference between dividing by 123 and 124 is a relatively significant change in frequency. So I am sticking with a single frequency. Even then, given the square wave and the jitter in the FPGA, I am seeing new occasional small spurs here and there when observing the radio noise floor with no antenna. Also, the internal switching of the regulators does not appear to go completely off. On some units, I can still barely see some of the original noise humps. They are at least 25dB lower now though. It will be interesting to see what people find when they start playing with this. I definitely want to have the option to turn it off. I can make it default on in the next gateware as it is very easy to reprogram the gateware now. Maybe the "random" checkbox can be used to turn it off.

Your original observation to not use the 160M filter on 160M RX is still very good to reduce and pretty much eliminate these spurs.

73,

Steve
kf7o
Reply all
Reply to author
Forward
0 new messages