Decode Problems SparkSDR und WSJT-X under Ubuntu

[Juergen, DL8LE]

Hello,

I want to use a Hermes Lite II as receiver together with my good old YAESU FT1000MP MKV as transmitter because the receiver performance of the YAESU unit is very poor in FTx and on the other hand for some stations it is necessary to have a higher output than just the 4 -5 W out of the HL II. In a first step I have installed WSJT-X 2.5.4 and SparkSDR 2.0.17 on my Ubuntu pc (new version, all updates installed).

SparkSDR runs very well with the HLII and WSJT-X with the FT1000 if operated separately . When using the audio input from the SparkSDR audio for WSJT-X, however, the waterfall will display any signals in WSJT-X only, if I select DigiU instead of FT8. Unfortunately the WSJT-X instance doesn't decode anything despite of the clear and normal looking waterfall, but at the same time the waterfall is running very fast in SparkSDR. There is no cat control possible for the HLII (hamlib error) even though I am using the address of the pc and the port shown by SparkSDR (e.g. 51112 or 51113 or 51111). WSJT-X shows the receiving input as green as it should be. I am using Pulseaudio and I have tried to set up a virtual sink monitor but no change.

The HLII works as it should be when I use my Windows 11 notebook (i.e. WSJT-X is decoding etc. with SparkSDR in DigiU mode). WSJT-X just for my FT1000 only and no opening of SparkSDR and the HLII is also ok.

I have gone back to an earlier date using timeshift and re-installed SparkSDR and WSJT-X. I have the same hamlib error for the HLII, the FT1000 is ok and works stand alone with WSJT-X ok but with poor receiving performance, the HLII is working very well with SparkSDR only in FT8 (and in CW, I tested it as well). So the question is what I could do to have my HLII working together with WSJT-X and SparkSDR in a first step, and in a second step what needs to be done so that I have an instance in WSJT-X not in conflict with SparkSDR which is controlling my FT1000 with the audio input of the HLII (controlled by SparkSDR).

Any help would be very much appreciated.

73

Juergen, DL8LE

[Juergen, DL8LE]

Here is some additional information after I have restored again a previous version with Timeshift and installed SparkSDR 2.0.17 and WSJT-X 2.5.4 :again: SparkSDR in FT8 is showing many decoded signals, but when I use DigiU and open WSJT-X there are only very few and faint traces in the waterfall of WSJT-X and nearly no decoded signal. Going back from DigiU to FT8 the same large number of traces in the waterfall of SparkSDR and many decodes.

When I open a separate instance of WSJT-X named WSJT-X FT1000 the waterfall is showing many traces but much less than I see in SparkSDR's waterfall with the HLII (many more, however, than in the DigiU-Mode of SparkSDR and the HLII). There was no way, however, to use the audio information showing in the SparkSDR waterfall in this FT1000 WSJT-X instance.

73

Juergen, DL8LE

[ahop...@googlemail.com]

Hi Juergen,

I'm not quite sure I understand what is happening but here are a few comments that might help:-

The ft8 and digiu mode displays should show the same thing if you zoom in on the digiu mode as its defaults are different.

The ft8 mode is by default muted and will only output audio to the selected audio device if un muted.
I have had better results using asio loops rather than pulse on linux, there is a guide some where for pihpsdr on doing this.

Does the hamlib connection ever work for you? I'm not aware of any problems with it other than wsjtx not retrying after timeouts.

I'm not sure if any of this helps but I'm sure we will get there in the end.

73 Alan M0NNB

[Juergen, DL8LE]

Hello Alan,

I have tried the setup with asio loops as well but the result is the same as before: the number of traces shown in the SparkSDR waterfall is about 30 (and the equivalent number of decoded signals) in FT8 and in SparkSDR DigiU there are max 10 or 12 faint traces in the WSJT-X waterfall, the difference in the SNR between SparkSDR FT8 and Digi-U in WSJT-X is more than 10 - 15, i.e. much weaker in WSJT-X. The green slider in WSJT-X for the receiving level shows about 49/50 dB, which is good.

I have also tried to compare the transmission behavior: when I transmit via WSJT-X in Digi-U mode of SparkSDR the current displayed is 160 mA, in FT8 it is 1,4 A, the signal level in my FT1000 I use without antenna for controlling the signal is 9+30 for WSJT-X transmission in Digi-U and full scale (>9+60) for transmission via FT8 mode in SparkSDR

I have now set up SparkSDR 2.0.17 on my AsusZenBook together with WSJT-X. I have installed VAC and used a virtual cable for the input instead of the built-in soundcards (Realtech for the output and Intel intelligent ... for the input). After switching on the first time the result in Digi-U was exactly as the result in FT8. Therefore I was quite happy. Unfortunately I was unable to repeat this after I have tried to install a second virtual cable for the output. So I have changed the settings again, deactivated all virtual line 1 and 2 devices in the audio setup but not any audio now. Therefore I cannot provide additional information before I have deinstalled the virtual audio cables and start again from scratch. The waterfall in WSJT-X, however, now looks the same when there is no audio when I am using a remote Flex 6600 at our university clubstation (I am allowed to do this from time to time).

I hope that this information can help you a little bit better to understand my problems. If there is no other way I will remove the present Linux-System from my PC and install it again from scratch, but this should be the final option in case nothing would work (my present Linux knowledge is rather limited). Regarding my notebook I hope that after the deinstallation of VAC I am able to start again (might be that I don't need it, but I am not sure). I will post the information here.

73

Juergen, DL8LE

[Juergen, DL8LE]

Unfortunately no luck at all, even after de-installing VAC in Digi-U (now in Windows 11). After doing so there is no relevant signal in the waterfall, when in Digi-U also no audio-signal which is transmitted. When I try to send a message from WSJT-X, e.g. CQ, however, and SparkSDR is in FT8 then there is full output. I have done many tests with different sound cards setting, but no success. At the end I have used the echo-function of Skype and there is absolutely no problem with audio in and audio out. The soundcards specified in the virtual rx in SparkSDR (FT8 and Digi-U) are defined as standard device (the drop down menu in Sound - additional settings for both devices shows grey settings with Standard Device, there is no possibility to select a Standard Communication Device. Even after deactivting and activating again there is no other selection than just standard.

There is full volume iat the speaker symbol in SparkSDR's menue, also no red x, the sound devices selected in SparkSDR and in WSJT-X are identical. CAT is working ok in both ways when in Digi-U.

I have tested WSJT-X with a Flex and their DAX audio, no problem. Due to missing hardware (cable and connectors) I was unable to connect my old FT1000 MP to the audio of the notebook.

I was unable to find the folder which stores something like an ini-file for SparkSDR. So I have de-installed SparkSDR and WSJT-X, but no luck either, I have also tried with 2.0.15 again, results are the same. So I am lost somehow.

73

Juergen, DL8LE

[Juergen, DL8LE]

Hello,

following is a summary of my attempts get Digi-U in SparkSDR working together with WSJT-X on my Asus ZenBook (Windows 11). Even though it is somehow special for this notebook setup it might be helpful for others with similar problems.

1. The available sound devices in my notebook are Microphone Array (Intel Smart Sound Technology) and Speakter (Realtek Audio). I have to add other devices such as a bluetooth headset or similar via the corresponding settings. SparkSDR and WSJT-X are listed as apps which want to use the built in sound devices.

2. There is no hardware switch for switching off the speaker or microphone, it is possible via the x in the settings.

3. Digi-U can be used with speaker in the audio output and microphone in the audio input of SparkSDR and WSJT-X. This means, however, that you always listen to the noise on the channel because the input into WSJT-X is via the real speaker of the notebook into the real microphone of the notebook. This will result in the loss of some decoded signals, by the way.If you switch off the volume in SparkSDR or the speaker in the notebook there is no input into WSJT-X other than the noise in the room.

4. Trying to resolve this problem by using a virtual audio cable (tried VAC and VB) and connecting the output of SparkSDR with the input of WSJ-X via cable 1 (audio input of SparkSDR and audio output of WSJT-X not connected but both set to none for the first trials) did not provide any signal in WSJT-X (used 16/48000 on both input and output and checked this in Windows sound settings, used several others as well). The control panel of VAC as well as that of Voicemeeter displayed a sufficient input level.

5. There were blue screens in between when using the virtual cables  so I had to restart the notebook together with the apps several times, the exact conditions are not clear.

6. Remark: When I used SparkSDR and Digi-U on my Raspi I never had any problems in this Digi-U mode.

I hope the information is clear enough now (I have also found the settings in the m0nnb folder in between, however no solution, but I have stored a couple of different profiles there now). I have also changed the title of this thread in my response from Decode to Audio Problems. It would be nice if in one of the future versions of SparkSDR which I like very much a solution for my specific problem could be provided. And: I don't want to give up my project yet to have the HLII as receiver and my old FT1000 as transmitter as standard setup on my Ubuntu Radio pc ....

73

Juergen, DL8LE

[ahop...@googlemail.com]

Hi Juergen,

a screen shot of the digiu mode using vac might just give a clue.  This is a common setup on windows and I've not had other reports of it not working so something odd is going on.  

Back to the hamlib connection issue, one thing you could try is turning on extended logging in settings, restart spark and then try to connect and send me the log file. 

73 Alan M0NNB

[Mike Brown]

I'm not sure my experience is necessarily relevant but I often have problems getting DigiU TX audio out of SparkSDR via VAC into WSJT-X. I think I saw the same issue with VB and both W10 & W11. After fiddling around I can always get it to work though.

It always works fine if I select speaker as SparkSDR's output. Sometimes just changing the output back to VAC sorts the problem, other times changing the WSJT-X input from VAC & back again fixes it. Other times I have had to restart WSJT-X and, I think, SparkSDR.

I don't think the error log shows anything but I'll double-check next time it happens.

73

Mike

G4RAA

--
You received this message because you are subscribed to the Google Groups "SparkSDR" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sparksdr+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sparksdr/e9aeb939-827c-4186-a835-ca51e8ac1034n%40googlegroups.com.

[ahop...@googlemail.com]

Hi Mike,

that is interesting, I shall add some extra logging as it always just works for me.  I have an alternative audio backend mostly done, maybe I should finish it if this is not an isolated issue.

73 Alan M0NNB

[Juergen Sturhahn]

Hello,

I have attached three screenshots to this mail, I hope they will be displayed or delivered to your mailbox. As Mike mentioned before: after some fiddling around I was able to get Digi-U working with VAC (but still with a loss of few signals which were decoded by SparkSDR but not in WSJT-X).

The signals displayed in the WSJT-X waterfall as well as the number of signals decoded is very much depending if AGC is switched on or off. With AGC ON the decoding there are many more signals displayed on the WSJT-X waterfall but also some noise between 0 and about 450 Hz with the filter settings in SparkSDR as the default values 150 and 3000. When I change those settings to -10hz and 3300 Hz there is not such a noise between 0 and 3000 Hz in the WSJT-X waterfall.

I have not been able to find a solution for my audio problems in the Ubuntu pc except that SparkSDR is working in FT8 (but not in Digi-U). Unfortunately I don't have any knowledge of Linux (have to look into tutorials or other information in the web every time when I want to try something new, here I did not find the information). If there is somebody who could tell me how to reset all the changes I have done in between in the audio and permission settings so that I could start from scratch again would be great. Otherwise I will format the drive and re-install Ubuntu (something I would prefere to avoid if ever possible).

73

Juergen, DL8LE

You received this message because you are subscribed to a topic in the Google Groups "SparkSDR" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/sparksdr/VolCCwAD0sw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to sparksdr+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sparksdr/99e1c252-4851-4982-b2df-9aaee75174b0n%40googlegroups.com.

[Mike Brown]

Hi Juergen (and Alan)

It sounds like you have yours working like mine on Windows. I wonder if the missing decodes in WSJT-X might be because of the extra VAC + WSJT-X processor workload?

Your screenshot shows you using VAC Line 1 for both RX and TX. I never got that to work, so use Line 1 VAC for SparkSDR Audio Output and Line 2 VAC for Audio Input.

I have attached a screenshot of my VAC control panel for reference. There are differences but nothing apparently critical aside from the number of active channels.

73

[ahop...@googlemail.com]

Hi Juergen and Mike,

I'm not surprised that there is sometimes a difference in decodes between digiu/wsjtx and the modes in spark. The vac, resampling and agc can all have an effect in digiu.  In the integrated modes these are all eliminated, a special agc is used that operates on the recorded signal so it knows what the future holds.  It might be worth trying with the  audio resampler turned off.  

I'm not sure what could be causing noise with different filter settings, I'll see if I can recreate that.

73 Alan M0NNB

[Mike Brown]

Just to note I recognise the noise Juergen is talking about. Adjusting the bandwidth sorted out, so I didn't worry. WSJT-X does odd things if the bandwidth is wider than the incoming audio anyway.

73

Mike

G4RAA

To view this discussion on the web visit https://groups.google.com/d/msgid/sparksdr/3a95cae3-ab70-474d-aa1c-03868467294en%40googlegroups.com.

[Juergen Sturhahn]

Hi Mike and Alan,

I have tried the VAC settings on your screenshot as well, but my Asus ZenBook does not like them, don't know why, might be because of specific modifications for some driver software they have implemented. But I am very glad that it works now with my VAC settings and I can use WSJT-X in Digi-U (WSJT-X because it allows me to use DXLab, which is very helpful).

I will wait now for the result of Alan's investigation regarding the filter bandwidth.

Just for info: I have been running one HLII at my home (good HF site and antenna) and compared it with the Flex6400 in our clubstation in the university (15km away, medium HF site but good antenna). It is interesting that the number of decoded signals is very comparable for both rigs.

Unfortunately my second HLII has a noise floor of up to 30 dB higher than the first one, i.e. decoding capability is very poor. Any idea how to fix the hardware problem other than returning it to makerfabs (there is still warranty, but they will only reimburse the money, no repair) would be very nice.

73

Juergen, DL8LE

To view this discussion on the web visit https://groups.google.com/d/msgid/sparksdr/CAP2QPQz%3DkEerDJhaRdojK37VuNLu3igb57hKki_cu4caUpTmQQ%40mail.gmail.com.

[Mike Brown]

Hi Juergen

The HL2 RX problem may be a faulty Omron relay. The High Pass filter relay in my N2ADR filter board went bad and was attenuating signals be around 30dB. AGC made it difficult to detect initially. You can check it by switching RX filtering off. Some people have had a similar problem with the TRX relay on the main board.

If you do have a faulty relay I suggest you replace with a Kemet (NEC) relay if you can. The Kemet has gold plated contacts so a better bet long term.

See the HL2 user group for more info and a source of free replacement OMRON relays.

73

Mike

To view this discussion on the web visit https://groups.google.com/d/msgid/sparksdr/dd206c7c-138d-a9f2-6fb2-8be168c56cda%40yahoo.com.

[Juergen Sturhahn]

Hi Mike,

I have checked the filter settings. There is no change depending on the filter settings (and also no change depending on the settings of the LNA). The difference between the two units remains unchanged (today with indoor antenna because of thunderstorms outside below -140dB for the good one and below -105/110 dB for the device with much lower sensitivity.

I will raise my question in the HL II User Group (registered today following your message).

73

Juergen, DL8LE

To view this discussion on the web visit https://groups.google.com/d/msgid/sparksdr/CAP2QPQwG%2BWg6ruv%2BY%3D_mg0TC9ZM19ShSCCg6XZ%3Dq_oe9%2BN7x3Q%40mail.gmail.com.