Problem with Pure Signal using DL1YCF piHPSDR and Hermes-Lite

[Doug Lung]

I've been investigating setting up remote operation using a Raspberry Pi 5 and the DL1YCF fork of piHPSDR. 

Everything seems to work fine except Pure Signal. Googling this I found a posting that piHPSDR Pure Signal would not work with the HL2. I have no problem using Pure Signal with Thetis, so know my feedback levels are within range. 

I'm wondering if Pure Signal has been fixed in piHPSDR and I'm just missing a setting configuration or compile option? I'm using the latest version, downloaded today, from DL1YCF's Github. 

If no one has gotten Pure Signal working on Linux with piHPSDR, I'll go back to Quisk, which has great remote control options and works fine for digital modes, and stick with Thetis on Windows for local SSB operation. 

Here is what I found.

With piHPSDR Pure Signal set to auto-attentuation, I get an error that ADC0 is overloaded. 

Turning off auto-attenuation, I can adjust the TX attenuation to around 8 dB and the overload goes away, but Pure Signal never corrects. Checking auto-attenutation and "restart" it starts to correct but as soon as the TX attention drops to zero it stops and displays "ADC 0 overload". 

The feedbk, dg. cnt, , sin.. chk and cor. cnt  are "0" throughout the process. GetPk is 0.238 and SetPk is 0.240 as recommended for the HL2.  Enable PS and PS MAP are checked. Feedback Lvl says "Correcting" and on the main screen "Correcting" is in red. According to the manual, the PS FeedBk ANT setting shouldn't matter on the HL2 but I tried with "Internal", "Ext1" and "Bypass" with no difference. 

73, Doug, AH6DL

[Doug Lung]

Update! I installed the source code on my Fedora 40 laptop and Pure Signal works great on it, so this problem is only with the Raspberry Pi 5 and the latest Pi OS. 

I'm interested in hearing if anyone else has the DL1YCF version of piHPSDR running on the Raspberry Pi with Pure Signal and if so, what OS are you using? I may try downloading Ubuntu for the Pi and see if that works better. 

73, Doug, AH6DL

[Andrew Kovac]

On my Linux Mint PC's piHPSDR Pure Signal does work but not as good as Thetis on Win11.

I and others observe regular outings and not just at the start of each over, feedback level is good.

This is with both HL2 and ANAN 7000.

A friend has the same issue with Ubuntu.

[Ron Lewkowicz]

With the lower available cpu power perhaps reduce the radio sample rate to 48k and enable the "oneshot"  option in the PS menu.  This option uses the saved calibration obtained during the two tone experiment rather than having PS constantly update the calibration while TX.  You'll need to manually re-calibrate any time you change band or power if this is selected.  Generally Auto Attenuate should always be selected.  For PS to work properly I believe TX audio input must also be set correctly by adjusting mic gain.

I have no issues with PS on Linux Mint.  

[Doug Lung]

Thanks for the comments. As noted it worked great on my Thinkpad X1 with Fedora 40, so perhaps it is a CPU issue. It is odd that I can't get it to work with just the two-tone test. I figured if it wouldn't work with that then there wasn't any point trying with a microphone. I also tried the one-shot mode and that didn't work either, with manual or automatic tx attenuation. If i set the tx-attenuation manually to 8 dB (1 or 2 dB below overload) it doesn't even try to correct. If I turn on automatic tx attenuation then it starts to correct but decreases tx-attenuation to zero and stops working with a partially corrected two-tone test (only one spur attenuated) when "ADC0 overload" appears on the display. The problem appears to be the automatic tx attenuation part is not working as it should. 

The only difference in the config between the laptop and the Pi is I added a make.config.pihpsdr file with "GPIO=OFF" in it. Since I'm not using GPIO on the Pi (thought I might at some point for CW) I'll try recompiling with that option on the Pi and see if it makes a difference. If that doesn't work I'll add a 10 dB attenuator to the sample line which might help, hopefully without messing up Thetis and my laptop piHPSDR. 

  -

["Christoph v. Wüllen"]

> Am 09.12.2024 um 01:34 schrieb Doug Lung <dlu...@gmail.com>:
>
> I've been investigating setting up remote operation using a Raspberry Pi 5 and the DL1YCF fork of piHPSDR.
>
> Everything seems to work fine except Pure Signal. Googling this I found a posting that piHPSDR Pure Signal would not work with the HL2. I have no problem using Pure Signal with Thetis, so know my feedback levels are within range.

PureSignal always has worked with piHPSDR and the HL2, see

https://dl1ycf.darc.de/hl2.htm

It works both with MacOS and Linux on the RaspPi. There is no
compile option for PS, it is always there.

It is recommended to have a proper feedback circuit, but
it even works with the cross-talk from the T/R relay.

If at 48k sample rate it is difficult to ge a correction,
perhaps try 96k, 192k might even work better.

>
> I'm wondering if Pure Signal has been fixed in piHPSDR and I'm just missing a setting configuration or compile option? I'm using the latest version, downloaded today, from DL1YCF's Github.
>
> If no one has gotten Pure Signal working on Linux with piHPSDR, I'll go back to Quisk, which has great remote control options and works fine for digital modes, and stick with Thetis on Windows for local SSB operation.
>
> Here is what I found.
>
> With piHPSDR Pure Signal set to auto-attentuation, I get an error that ADC0 is overloaded.
>
> Turning off auto-attenuation, I can adjust the TX attenuation to around 8 dB and the overload goes away, but Pure Signal never corrects. Checking auto-attenutation and "restart" it starts to correct but as soon as the TX attention drops to zero it stops and displays "ADC 0 overload".
>
> The feedbk, dg. cnt, , sin.. chk and cor. cnt are "0" throughout the process. GetPk is 0.238 and SetPk is 0.240 as recommended for the HL2. Enable PS and PS MAP are checked. Feedback Lvl says "Correcting" and on the main screen "Correcting" is in red. According to the manual, the PS FeedBk ANT setting shouldn't matter on the HL2 but I tried with "Internal", "Ext1" and "Bypass" with no difference.
>
> 73, Doug, AH6DL
>

> --
> 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 visit https://groups.google.com/d/msgid/hermes-lite/90427af8-a7af-4b41-8d41-b7edd2221fb2n%40googlegroups.com.

["Christoph v. Wüllen"]

If the mic is not properly levelled, adjusting mic gain is important.
The TX peak ALC reading should scratch 0 dB (which correspondes to
"full amplitude" mic input). Not mic gain has no effect during a
two-tone experiment since the "audio" is generated internally there.

> --
> 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 visit https://groups.google.com/d/msgid/hermes-lite/e0839da3-e146-4201-84ea-575febd08a88n%40googlegroups.com.

[Doug Lung]

I'll keep trying Christoph. If you have it running on a Raspberry Pi with the latest updated Raspberry Pi it should work for me, as it does on my other devices. So far I've only tested with the two-tone test but will hook up a mic, check the levels and ALC, and see if that allows it to run without the ADC0 overload error. 

[Doug Lung]

Since some have noted that a too high sample rate might be causing the problem, all my testing has been at a 48k rate, nothing higher. 

On Sun, Dec 8, 2024 at 10:21 PM "Christoph v. Wüllen" <DL1...@darc.de> wrote:

[Doug Lung]

With Christoph's help I was able to get Pure Signal working on my Raspberry Pi 5. The solution turned out to be quite simple -- increasing the sample rate from 48K to 96K allowed it to work fine and provide a significant improvement in the output from my 50 watt amplifier. I did not have to add any additional attentuation, and the automatic tx-attenuation ended up at 9 dB. If anyone else finds this thread in the future, I hope increasing the sample rate works for you too! 

Thanks to all for their comments and especially to Christoph for his excellent port of piHPSDR! 

73, Doug AH6DL