Spur on 160m new Hermes Lite 2

272 views
Skip to first unread message

Wayne Roth

unread,
Apr 13, 2024, 1:27:24 AM4/13/24
to Hermes-Lite
Hello I received a Hermes lite 2 "kit" with the optional I/O board which has a significant spur on around 1.89 MHz that does not go away with the antenna input grounded.  I read that this was common around 4 years ago and earlier and there was an option to switch a firmware setting related to the FPGA clock that moved it.  The "firmware" is 7.4.2 as it came on the board and I'm running the hermes-specific version of Thetis v 2.10.3.5 Beta 2.  Apparently this is possibly coming from a regulator on the board according to what I've read.  Is this common again with the latest boards built by Makerfabs shipped in March after their brake?  Is there a fix to move the spur on the latest board?  Thanks, Wayne WA2N.
spur.jpg

Mike Lewis

unread,
Apr 13, 2024, 5:20:13 AM4/13/24
to Wayne Roth, Hermes-Lite

--
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/42df9e2e-ef61-495d-aa85-d8a09b3c04e9n%40googlegroups.com.

Duncan Clark

unread,
Apr 13, 2024, 5:31:08 AM4/13/24
to Hermes-Lite
In particular:

as in:

"By default, the gateware has the FPGA-generated switching frequency on. It can be disabled by setting the "LT2208 Random" bit which is unused in the HL2. You can find the checkbox control for "LT2208 Random" in PowerSDR and other openhpsdr software configuration screens. On Quisk, there is Config->->Hardware->Disable Power Supply Sync to turn this feature off or on."

I'm not sure if that checkbox control is still in Thetis?

Duncan

Wayne Roth

unread,
Apr 13, 2024, 9:10:01 AM4/13/24
to Hermes-Lite
Thanks, it is still there in the latest Thetis by MI0BOT and while it was unchecked originally with the spur present, I toggled the state on and off and the spur was reduced from -103dbm to -125dbm and moved up in frequency.  Not out of the 160m band unfortunately, but better than it was.  Not sure whether this will persist across power supply turned on and off since it was off originally with the loud spur present.
spur mitigation.jpg

John Williams

unread,
Apr 13, 2024, 9:29:45 AM4/13/24
to Wayne Roth, Hermes-Lite
Shut off 160M receive filter and give that a try. Also, filter setting is a mix of VFOA and VFOB. RX filters will get set to allow highest frequency between the two. I try to keep VFOB set to 1.900 so that the RX filters get set correctly for each band on VFOA. You can see what filters are set by monitoring the hardware pin state. 



From: herme...@googlegroups.com <herme...@googlegroups.com> on behalf of Wayne Roth <wa2...@gmail.com>
Sent: Saturday, April 13, 2024 8:10 AM
To: Hermes-Lite <herme...@googlegroups.com>
Subject: Re: Spur on 160m new Hermes Lite 2
 

John Williams

unread,
Apr 13, 2024, 9:42:28 AM4/13/24
to Wayne Roth, Hermes-Lite
Read the discussion of this problem in the Options section of the Wiki in the Archive category.
From: John Williams <jswi...@gmail.com>
Sent: Saturday, April 13, 2024 8:29 AM
To: Wayne Roth <wa2...@gmail.com>; Hermes-Lite <herme...@googlegroups.com>

Wayne Roth

unread,
Apr 13, 2024, 10:03:30 AM4/13/24
to Hermes-Lite
I checked the Hermes board and there are 0 ohm resistors across R16 and R13 and the J11 and J2 are populated with resistors marked O1C (10k-ohms?) as suggested in the Wiki.  Turning off the 160m filter does reduce the spur from -103 to -111 which helps but does not eliminate the spur.  The FPGA apparently defaults to the state where Disable PS Sync is set which "turns on" this spur at 1.89 MHz. Despite the checkbox being off initially in the Thetis software, it's off state is apparently not written to the board at startup.  Thus turning it on, then off, does reduce the amplitude of the spur to -125db and moves it above 1.92 MHz or so where it floats around.  I added a bug report to the latest MI0BOT Beta 2 bug list on Github suggesting it should be written at startup whether on or off to the board.  Thanks for all the help folks.

John Williams

unread,
Apr 13, 2024, 10:22:17 AM4/13/24
to Wayne Roth, Hermes-Lite
In practical terms and in my opinion, at -125db on most 160M nights, you will not see the spur. The noise level will be much higher. On my inverted L I do not see the spur. I have a build8 board without the change. Shutting off the RX filter is sufficient. I also have a beverage antenna and can occasionally see the spur if it is quiet. I did have to build a high pass filter to block BC stations with the beverage. My spur shows up around 1.950 and wobbles about. I just live with it. 
Sent: Saturday, April 13, 2024 9:03 AM

Wayne Roth

unread,
Apr 13, 2024, 10:27:54 AM4/13/24
to Hermes-Lite
Agreed as long as the spur is reduced to -125dbm it's not a problem.  On my EFHW that I use for 160 that's about where the noise floor is so it's not noticeable.  The main issue seems to be that the Thetis software doesn't properly set the state of that switch when the program is initialized to reduce it from -103 dbm (S9+ here) which is definitely noticeable!  For now I can just cycle that option switch on then off if I'm on 160m to get rid of it.  Good until the next power cycle.

Steve Haynal

unread,
Apr 14, 2024, 7:42:30 PM4/14/24
to Hermes-Lite
Hi Wayne,

As discussed already in this thread, the 160M spur is expected and is from the cost-saving switching power supply on the HL2. The HL2 philosophy is that it is easier and more cost effective to move the spurs. It looks like your spurs are about 20dB above the noise floor if antenna is grounded or has a dummy load attached, so I don't think there is any problem with your build. This is what is expected. Quisk with bandscope view provides a nice view of any spurs. Here is a HL2 with dummy load on the antenna and LNA gain set to +20dB. This is without moving the spurs. Note that there are two for two of the regulators in the HL2. The regulator internal oscillator will not be the same from regulator to regulator. You can see that the worse spurs are about 20dB above the noise floor. There are the spurs at ~950kHz, which is the native frequency of the regulators, and then the first and second harmonics. 
nosync.png

If you turn on the regulator clock provided by the FPGA (which is or should be the default in all software), then you see that all regulator frequencies align and none of the spur harmonics are in an amateur radio band.
sync.png

When I connect a real antenna and disable the AM block filter (as it can affect 160M), I have to lower my LNA to +12dB to avoid ADC clipping. The noise floor is now well above any of the spurs. As a side note, when skimming 10 bands, I often keep the very helpful AM blocking filter on even if it attenuates some 160M signals. I don't think this matters as at the +18db LNA gain I usually operate there is still enough gain for 160M.

antenna.png

73,

Steve
kf7o

Wayne Roth

unread,
Apr 15, 2024, 12:32:56 AM4/15/24
to Hermes-Lite
Steve I understand and agree with your assessment on the HL2 160m spur issue.  Unfortunately Thetis 2.10.3.5 Beta2 does not turn on the HSYNC clock, you have to manually go into options and toggle the Disable PS Sync checkbox on (to disable) and then off each time the program is started.  I don't know whether it's writing a disable off state at initialization and ignoring the options setting, or doing nothing at all and the FPGA defaults to let the regulators free-run.

I've not tried quisk yet, just Thetis.  I'm unable to find the "AM Block" filter setting.  Is there a setting for this in Thetis and where?

Wayne

Mike Lewis

unread,
Apr 15, 2024, 12:37:48 AM4/15/24
to Wayne Roth, Hermes-Lite

The “AM Blocking filter is simply the Hi Pass filter enabled by Bit 7 in the Setup->General->OC Control J16 RX pins.  It is normally off for 160M and TX.   Tick the N2ADR Filter button to set the default values correctly for most users.

 

From: herme...@googlegroups.com <herme...@googlegroups.com> On Behalf Of Wayne Roth
Sent: Sunday, April 14, 2024 9:33 PM
To: Hermes-Lite <herme...@googlegroups.com>
Subject: Re: Spur on 160m new Hermes Lite 2

 

Steve I understand and agree with your assessment on the HL2 160m spur issue.  Unfortunately Thetis 2.10.3.5 Beta2 does not turn on the HSYNC clock, you have to manually go into options and toggle the Disable PS Sync checkbox on (to disable) and then off each time the program is started.  I don't know whether it's writing a disable off state at initialization and ignoring the options setting, or doing nothing at all and the FPGA defaults to let the regulators free-run.



I've not tried quisk yet, just Thetis.  I'm unable to find the "AM Block" filter setting.  Is there a setting for this in Thetis and where?

Wayne

On Sunday, April 14, 2024 at 7:42:30 PM UTC-4 softerh...@gmail.com wrote:

Hi Wayne,

 

As discussed already in this thread, the 160M spur is expected and is from the cost-saving switching power supply on the HL2. The HL2 philosophy is that it is easier and more cost effective to move the spurs. It looks like your spurs are about 20dB above the noise floor if antenna is grounded or has a dummy load attached, so I don't think there is any problem with your build. This is what is expected. Quisk with bandscope view provides a nice view of any spurs. Here is a HL2 with dummy load on the antenna and LNA gain set to +20dB. This is without moving the spurs. Note that there are two for two of the regulators in the HL2. The regulator internal oscillator will not be the same from regulator to regulator. You can see that the worse spurs are about 20dB above the noise floor. There are the spurs at ~950kHz, which is the native frequency of the regulators, and then the first and second harmonics. 

If you turn on the regulator clock provided by the FPGA (which is or should be the default in all software), then you see that all regulator frequencies align and none of the spur harmonics are in an amateur radio band.

 

When I connect a real antenna and disable the AM block filter (as it can affect 160M), I have to lower my LNA to +12dB to avoid ADC clipping. The noise floor is now well above any of the spurs. As a side note, when skimming 10 bands, I often keep the very helpful AM blocking filter on even if it attenuates some 160M signals. I don't think this matters as at the +18db LNA gain I usually operate there is still enough gain for 160M.

 

73,

 

Steve

kf7o

 

Reply all
Reply to author
Forward
0 new messages