Drive (50) & Internal Pure Signal with HL2

[Kurt V.]

Hi all, 

Using the Thetis v2.10.2.1 here. 

When I go around on the bands, I can tune the drive slider all the way to the right, giving the nice almost 5 W out. It will display above the Drive slider Driver: 100. But for some weird reason, when I am on 20m, it displays above the drive slider the following: Drive :(50). And when I move the slider beyond 50%, I get a red line and it will not go beyond 50. Did I change some settings to limit it to only 50% drive? How can I resolve. I have been searching for hours and I cannot find any information about this. 

Another question is about the Pure Signal. Some other friends on the radio pointed out the Thetis software does have an Internal Pure Signal feature. How do you enable it? I can't find any info about it, nor on youtube, nor in the manual. 

Kind regards

de ON7OFF

[Ed Stroh]

Hi Kurt! I can help with this.

The red bar is set by right clicking on the drive slider. You can configure a software limit to the drive level by doing this. I suspect you must've right clicked on the drive bar while on 20m by mistake, accidentally setting the limit to 50%. Just right click and slide all the way to 100 and you should get full power.

Pure Signal is a bit more involved. Do you have RF3 soldered to the HL2's jumper bridge already or are you planning to use the internal TR switch for sampling?

73,

Ed KS7ROH

[Kurt V.]

Hi, got it Ed, 

Thanks for the help. I will buy a sampler later, but some gentlement told me that even without a sampler, you can already enable the pure signal feature. He told me to go to Setup/General and then ANT or something. But in my Thetis, I do not see any ANT tab under General. So I don't know if it is something that needs to be enabled in the software or not. 

I have tried on the Thetis general version and the one for HL2. None of them has that ANT tab. 

Maybe he explained it incorrectly. 

The HL2 community is great and I will contribute with a few more explanatory video's in the future. And what a beast, this HL2. The best purchase I have made in radio gear in my entire life!. 

73's 

[Ed Stroh]

Hey Kurt,

Glad to hear you're enjoying the HL2! As am I! Definitely among my favorite ham radio purchases. It's a fantastic radio.

As for getting PS working, I think what the gentleman you spoke to previously may have been referring to is the General>Ant/Filters>Antenna tab. See attached image. I think you might need to enable the RX 1 OUT on Tx checkbox to get the TR relay working for PS but I'm not sure about that. I haven't personally attempted using the TR relay, only RF3 with a sampling port. It's also possible you may need to go into General>ADC and reassign DDC1 and DDC2 to ADC0 to get this working. Hopefully someone can correct me on this if I'm leading you astray. If nothing else, these are the places I would start messing with to get it working. 

Also, make sure the DUP button (on the left of the main window below MOX) is highlighted. Fairly confident this is necessary for PS to work but more importantly, it shows you what the signal actually looks like when transmitting rather than just the spectrum of the audio input going to the radio. 

Hope this helps! Be sure to report back with your progress!

73,

Ed KS7ROH

P.S. - Thanks for your videos on tuning audio! Very helpful for getting a good, clear signal from my station!

[Reid Campbell]

Hi Ed, Kurt,

The only software requirement for PS to work is,

192K sample rate
DUP selected
PS-A selected
Check that SetPK is 0.233 on the PureSignal form (Linearity Menu Item).

There should be no need to touch any of the other settings.

Cheers 

Reid
Gi8TME/Mi0BOT

--
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/4e3cb965-1b19-4aa4-bfa4-c8cf36167d31n%40googlegroups.com.

[Paul-M1TZR]

Thanks for this Reid, I am also new to HL2 with Thetis but do have some experience using Anan hardware.

I was sure I had PS working but was getting distortion when ever I turned it on.

Dropping the sample rate to 192K from 384K resolved it as per your instruction. So I must have previously enabled it with the reduced sample rate.

Which leads me to ask, in the official release (for want of a better description) protocol1 only allows you to go to a max sample rate of 192K.

How come yours goes unto 384K and what other limitations does running this sample rate have along with the PS issue ? if any.

Thanks for your ongoing work.

Paul

[Reid Campbell]

Hi Paul,

Running 384K maxes out 100M Ethernet but the HL2 has 1G implementation.  The PS limitation is the only down size I can think of running 384K.

Cheers 

Reid
Gi8TME/Mi0BOT

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/ceb42d83-dbef-4883-9551-10cfe7716df1n%40googlegroups.com.

[Paul-M1TZR]

Thanks for the explanation Reid, makes sense.

So the next question is there such a thing as a p2 Gateware for these little beasties that can utilise the full gig ?

Thanks again

Paul

[Reid Campbell]

No, but there is no reason that somebody couldn't produce one, if they were minded to. It might take up more resources, so limiting the number of receivers.

Thetis was suppose to be P2 only but some of the older Anan series transceiver's FPGA couldn't meet the time restrictions, I think specify for the 1G Ethernet. Thetis was updated to support the P1 protocol and that is why the HL2 can run with Thetis or we would still be with PowerSDR.

Cheers 

Reid
Gi8TME/Mi0BOT  

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/baad4856-f07a-4233-a81c-ff1d6071dfbfn%40googlegroups.com.

[Jim KD4YLQ]

Reid,

           thanks for the explanation of how to set up PS in the HL2! still need a sampler appropriate for the power output level intended, right?

            73,  Jim  KD4YLQ

[Paul-M1TZR]

Figured that may be the case but worth asking to help me get a better understanding of what's going on.

All the best

Paul

[Reid Campbell]

Hi Jim,

Yes, you will need a sampler for an external amp but I'm not sure of what dBm you need to supply. There is also the other problem of getting your sample back into the HL2. The new I/O board can handle that for you or you can add a connection to the board inter-connector jumper. If you search the messages, you should be able to find the exact details.

I updated the Thetis code a few versions ago to use the full spread of the LNA for PS which has 61dB of range. That should cover most sample inputs.

PS will work with no external amp using cross talk from the Rx/Tx relay. I would recommend checking functionality stand alone to get familiar with its operation.

Cheers 

Reid
Gi8TME/Mi0BOT

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/f5a31156-91ec-4d89-b8d7-de8218752909n%40googlegroups.com.

[Ed Stroh]

Jim,

If you're looking for a sampler for PS, I can recommend this solution from CleanRF: https://www.cleanrf.com/products.html#RF-S2K

It's rated for above legal limit and has variable attenuation so you can set it to whatever power level you might use. 

73,

Ed KS7ROH

[Kurt V.]

Now that we are on the subject of Pure Signal, another factor that will make your signal more clean, as far I understand is setting the SSB filter size as high as possible (depending on the available processing power of the computer using thetis). 

This can be done by going to setup/DSP/Options. 
Look at the filter size column. SSB/AM TX filter. If you have it at 1024, you might creat a lot of splatter. As far as I understand, having this value increased, the TX signal will go through more filters and be cleaner. 
This is information I got from Claude,  VA2CST. He has been very helpful regarding teaching me these things. 

Now that Thetis is getting better by the time versions increment, Pure Signal is getting better and better too. Thank you Mi0BOT, Reid for the wonderful work you do. 

I have one more question Reid not related to PS. Did you check out the RM noise filtering feature yet?

Regards. 

Kurt (de ON7OFF)

["Christoph v. Wüllen"]

The filter sizes (in taps) mostly affect the fall-off at the edges of the passband. I do not think
that what you say makes any sense. Note 1024 is on the small side, I think the "standard" size
should be 2048.

> To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/29f6d7fd-7959-4ccf-b9ed-dc69defa5e23n%40googlegroups.com.

[Reid Campbell]

Hi Kurt,

On 18/10/2023 04:45, Kurt V. wrote:

Now that we are on the subject of Pure Signal, another factor that will make your signal more clean, as far I understand is setting the SSB filter size as high as possible (depending on the available processing power of the computer using thetis). 

This can be done by going to setup/DSP/Options. 
Look at the filter size column. SSB/AM TX filter. If you have it at 1024, you might creat a lot of splatter. As far as I understand, having this value increased, the TX signal will go through more filters and be cleaner. 
This is information I got from Claude,  VA2CST. He has been very helpful regarding teaching me these things. 

Now that Thetis is getting better by the time versions increment, Pure Signal is getting better and better too. Thank you Mi0BOT, Reid for the wonderful work you do.

I would be interested if you have seen any improvement or deterioration with the changes. One of the nice features of Thetis is being able to see the effects of PS on the transmitted signal. 

I have one more question Reid not related to PS. Did you check out the RM noise filtering feature yet?

I think that is the AI noise reduction. To be honest, I don't see much of an improvement over NR2 which is included within Thetis. I think you have to connect to a remote server for the AI version, which seem like a backward step. I think when it can work stand alone it will appear in various SDR software, depending on the licencing conditions.

Cheers

Reid
Gi8TME/Mi0BOT

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/29f6d7fd-7959-4ccf-b9ed-dc69defa5e23n%40googlegroups.com.

[Kurt V.]

Hello Reid,

I will try to get Claude on board of this discussion. He will be the best person to give you comments on the difference in splatter using a different value on the SSB filter size. 

He will give you more details and comparison. 

According his last email he sent me, he do see less bleed over when increasing the SSB filter size value. 

Regarding the RM noise filter, I have suggested in another post to actually have some contact with Randy Williams who is the programmer of RM noise. Maybe he is willing to share the code for the AI filtering feature and we can all let it run locally on our own computers running the Thetis software. I personally like the RM noise AI feature because it does give a very clean, non metallic filtering. NR2 sometimes give a metallic feedback (which I also happen to have limited by changing the SSB RX filter type from Low Latency to Linear Phase). 

I hope one day to talk to you on the radio. You are a genius when it comes to programming, so my gratitute and congratulations on the work you do on the Thetis software. 

Kind regards. 

Kurt

ON7OFF

[Clifford Heath]

It would be excellent to get Claude's perspective.

One hazard is worse with longer filters: it is more likely to cause arithmetic overflow. The more terms are added, the more bits of numeric width are produced. The RF chip only generates 12 bits, but the multipliers in the Cyclone IV are only 18x18, so careful amplitude management is needed to avoid overflow. Using longer filters might require more attenuation at the input, which could make the results worse rather than better.

Clifford Heath 

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/08859fe6-5be5-4e8e-9094-220cf74bfe56n%40googlegroups.com.

[Reid Campbell]

Hi Kurt,

On 19/10/2023 09:43, Kurt V. wrote:

Hello Reid,

I will try to get Claude on board of this discussion. He will be the best person to give you comments on the difference in splatter using a different value on the SSB filter size. 

He will give you more details and comparison. 

According his last email he sent me, he do see less bleed over when increasing the SSB filter size value.

It will interesting to see his results and how he is measuring it.

Regarding the RM noise filter, I have suggested in another post to actually have some contact with Randy Williams who is the programmer of RM noise. Maybe he is willing to share the code for the AI filtering feature and we can all let it run locally on our own computers running the Thetis software. I personally like the RM noise AI feature because it does give a very clean, non metallic filtering. NR2 sometimes give a metallic feedback (which I also happen to have limited by changing the SSB RX filter type from Low Latency to Linear Phase).

I don't know what licencing schedule is used for the noise filter but Thetis is open source and only other open source code would normally be integrated with it. All this type of stuff is looked after by Warren, so I'm sure he is aware off what is out there. It might get added to his DSP library, that Thetis uses, if he thinks it is contributions better performance. 

Cheers 

Reid
Gi8TME/Mi0BOT

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/08859fe6-5be5-4e8e-9094-220cf74bfe56n%40googlegroups.com.

[Reid Campbell]

Hi Clifford,

All these filter are within the PC software, so don't affect the filters on the hardware. The major effect is increased latency of the transmitted signal but that would not normally be an issue with typical QSO's.

Cheers 

Reid
Gi8TME/Mi0BOT

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/CAPnXPY3B4C4%2BGwY6jq2g_8pqnf8T%2Bm91on9dxCs86z_jVU1h5g%40mail.gmail.com.

["Christoph v. Wüllen"]

Dear Reid,

I have now re-compiled the code and tested the audio input with "Audacity"
and made progress. First my jumper settings are (viewed from the top)

a b c
x y z

---------
! MIC !
! JACK !
---------

such that two jumpers are set that connect x-y and b-c

The connection b-c is relevant for the microphone, it connects
the tip of the 3.5mm jack to the mic input. The connection x-y
is for PTT.

I have now wired the pink (microphone) plug of the headphone as follows:

Gnd(Headphone) ==> Gnd(Jack)
Ring(Headphone) ==> Tip(Jack)

I guess for the last connection it does not matter if you connect
with the ring or the tip of the headphone since they are the same.

Note it also works if you just plug in the microphone plug
into the microphone jack, but then you loose PTT.

Two changes had to be made to CWKeyerShield.cpp:

--- a/libraries/teensy/CWKeyerShield/CWKeyerShield.cpp
+++ b/libraries/teensy/CWKeyerShield/CWKeyerShield.cpp
@@ -68,11 +68,19 @@ void CWKeyerShield::setup(void)
wm8960->enable();
wm8960->volume(masterlevel_actual);
wm8960->inputSelect(0); // select and activate microphone input
- wm8960->inputLevel(0.1F, 0.1F); // volume control for mic input (both mic and MEMS)
+ wm8960->enableMicBias(1);
+ wm8960->inputLevel(0.5F, 0.5F); // volume control for mic input (both mic and MEMS)
}

The important thing is the line with enableMicBias, otherwise the headset cannot work properly.
Then, I adjusted the input levels (they were a little weak).


I then unplugged the microphone jack, whistled into the headset mic and re-plugged it after few
second, the recording is as follows:

[Reid Campbell]

Hi Christoph,

On 20/10/2023 13:52, "Christoph v. Wüllen" wrote:
> Dear Reid,
>
> I have now re-compiled the code and tested the audio input with "Audacity"
> and made progress. First my jumper settings are (viewed from the top)
>
> a b c
> x y z
>
> ---------
> ! MIC !
> ! JACK !
> ---------
>
> such that two jumpers are set that connect x-y and b-c
>
> The connection b-c is relevant for the microphone, it connects
> the tip of the 3.5mm jack to the mic input. The connection x-y
> is for PTT.
>
> I have now wired the pink (microphone) plug of the headphone as follows:
>
> Gnd(Headphone) ==> Gnd(Jack)
> Ring(Headphone) ==> Tip(Jack)

Yes, this is the way I have mine setup.

>
> I guess for the last connection it does not matter if you connect
> with the ring or the tip of the headphone since they are the same.
>
> Note it also works if you just plug in the microphone plug
> into the microphone jack, but then you loose PTT.
>
> Two changes had to be made to CWKeyerShield.cpp:
>
> --- a/libraries/teensy/CWKeyerShield/CWKeyerShield.cpp
> +++ b/libraries/teensy/CWKeyerShield/CWKeyerShield.cpp
> @@ -68,11 +68,19 @@ void CWKeyerShield::setup(void)
> wm8960->enable();
> wm8960->volume(masterlevel_actual);
> wm8960->inputSelect(0); // select and activate microphone input
> - wm8960->inputLevel(0.1F, 0.1F); // volume control for mic input (both mic and MEMS)
> + wm8960->enableMicBias(1);
> + wm8960->inputLevel(0.5F, 0.5F); // volume control for mic input (both mic and MEMS)
> }
>
> The important thing is the line with enableMicBias, otherwise the headset cannot work properly.
> Then, I adjusted the input levels (they were a little weak).

Would your binary for this work directly for Steve's board. I have no
way of directly building from source.

>
>
> I then unplugged the microphone jack, whistled into the headset mic and re-plugged it after few
> second, the recording is as follows:
>
>
>
>
>

> In the first two second, you only see signal on the right channel from the built-in
> MEMS microphone. After two seconds, when I plugged in the head-set, I get
> a signal on the left channel (mic jack) as well.
>
> The downside of all this is that there is no software API to make these settings (e.g.
> bias on/off or volume settings), so one has to recompile for a given need.
> What *is* possible to define "typical situations" and generate a hex file for these.
>
> I send a copy to the list, since this may be important.

If there are going to over 50 boards produced, it's important that we
have working software to download on to them with as much functionality
as possible.

Cheers

Reid
Gi8TME/Mi0BOT

>
> Yours
>
> Christoph DL1YCF.
>
>