Quisk

[F5NPV Amateur Radio]

Hello,

I am using the RF-HFIQ product.

I have downloaded the Raspbian Image from the Hobbypcb Wiki .

I am using the StarTech USB audio interface and manage to configure Quisk accordingly

Currently all working well and I am using a homemade mosfet amplifier controlled by the J3 ext Ptt which providing me up to 70watts output power.

Great product and really work smoothly.

Regarding Quisk in fact this software is quite new for me and currently I need some help or clarification.

For example on 7Mhz ,The graph screen is separated in 2 with in the middle a mark centered on 7080Mhz. Both sides are in fact identical and for example if I am monitoring a signal on 7074Mhz it is active on both side of the central mark. I try my best to find some information on Quisk google group without any success. I guess it is related to dual receive or split mode but not sure at all. I can play around with the right/left arrow but to be honest it is not clear for me.

In addition when i start quisk , i have popup Python window related to STDOUT witch indicate a tune frequency. To be honest i have no idea regarding  the purpose of this additional windows.

All working well , but i am just a bit confuse by Quisk and how to use it.

I will put tonight some pictures in order to elaborate.

73s from F5NPV / Didier

[Björn]

I'm not using the provided rPi image myself but your problem probably have to do with the audio signal chain. If your left and right sides of the spectrum are exact images of each other you are probably missing left or right channel (i.e. I or Q signal) or you might be using a mono signal as I/Q.

If I lift one of the jumpers to disconnect I or Q I get an exact mirror of left and write hand side of the spectrum, but with lower amplitude as the energy used per half is in fact halfed.

Check if there is something in your setup that make you miss the IQ-nature of the received signal. Perhaps the sound card is set to mono for some reason? If you have an oscilloscope and a signal generator you should be able to see the phase shift between I and Q in the audio spectrum. Also check that your 3.5mm plugs are properly inserted and that you in fact are using proper stereo cables and not the mono variant.

You could of course have a faulty op-amp on the output of the RS-HFIQ but that you should be able to detect by simply inserting a speaker or headphone in the receive out and assert a signal a few khz offset from your listening frequency. You should hear a tone in both ears/channels.

73 de

Björn, SM0SBL 

[WA2EUJ]

If you are using the Startech 96kHz 24Bit USB sound device and it has 'REV B' on the back, you'll have to go to the StarTech website and upgrade the FW or it only creates a signal channel input.

Here is the link:

ttps://www.startech.com/Cards-Adapters/Sound/USB-Audio/USB-to-Audio-Converter-External-SPDIF-Sound-Card~ICUSBAUDIO2D#dnlds

73,

Jim WA2EUJ

[F5NPV Amateur Radio]

Hello Guys,

Many thanks for your responses which make sense to me . I will check tonight after office hours .

I have the startech USB for a long time now and there is a good chance it is outdated.

Btw , the product is working really well , good sensitivity , the output power is more than enought since i am using a tiny 50 to 70 watts Amp. Quisk is a bit tricky , but afer a while you can get along.

Regarding the USB device , i have noticed on the wiki you are recommending the Behringher UFO202 but your link is pointing on the behringer U-Control UCA202. Are they identical ? and which one your recommending .

73s Didier (F5NPV)

[F5NPV Amateur Radio]

Hello guys,

I manage last night to perform some investigation regarding the Quisk issue.

-Firmware update --> In fact my startech interface have already the correct firmware but I upload it in order to secure my investigation;

-Cables --> OK (but I order to make me confident I use 2 new cables from my stock)

-Jumpers : I play with the main board jumpers especially the IQ output and whatever I remove the I or Q Jumpers the problem still there and I still have the double spectrum

-Audio IQ output check from the RS-HFIQ : from the IQ output I plug a speaker and the oscilloscope and so far all good

-Audio IQ input in the Startech USB --> KO , the signal is only in the left channel

I have no time to investigate regarding the pulseaudio configuration and will do it this coming week-end.

In order to move further the investigation, I use my Windows 10 laptop with HDSDR and so far all good , the receive is perfect and the IQ signal from the input of the Startech is OK

Summarize:

-Raspbian --> KO so it means there is some issue regarding the STARTECH configuration or maybe de IQ signals are totally out of phase and need to be adjusted.

-Windows --> All OK without any issue and the receive/transmit signal are crystal clear.

BTW many thanks for all your inputs and the investigation was interesting and assist me a lot for all it works.

Yesterday I have ordered the UCA202 and will keep the Startech for the my windows laptop since I am planning to use is for portable.

73s Didier F5NPV

[Rob French]

A couple of notes.

(1) I had issues with the StarTech card with every Linux computer I tried it on, including a Dell Laptop running Ubuntu and an Odroid XU4 (similar to RaspPi) running Armbian.  Same issue on both, namely, I could eventually get it up and running, BUT, if I ever launched WSJT-X, it seemed to freeze the StarTech.

(2) I've also seen the case where you get the "double spectrum" in receive.  I would be highly suspicious that you've probably got a mismatch between your sample rate in Quisk versus the sample rate of the actual sound card.  You probably have Quisk set to 96 KHz for the I/Q RX Audio.  However, you could check to see if that's how the card is actually configured.  If you are using PulseAudio, you can run the following commands, and they should list your audio sources/sinks and their corresponding setup.

pacmd list-sinks
pacmd list-sources

I will note that, for my current RS-HFIQ setup with Raspbian, I do NOT use PulseAudio.  I just have ALSA sound configured, including for the loopback / "virtual audio cable" from Quisk to Fldigi.

Rob KC4UPR

[F5NPV Amateur Radio]

Hello Rob,

Agree with you , the Startech USB is not so bad for Windows but for Raspbian not really suitable. Really tedious to setup and maybe impossible to expect some good performance

I have received 2 days ago my UCA202 and all issues has gone. The UCA is working perfectly with RASPBIAN and nothing really to setup. Tonight i manage to make my first QSO with Kuwait 9K9NLD with a great signal for both side (Only 50watts) , despite the propagation is not really good.

I need know to think about the CW mode

This coming week end we have the French SSB contest and will perform some QSO . I will shoot a video and put in Youtube.

Have a great Week-end

73s Didier / F5NPV

[F5NPV Amateur Radio]

Hello,

A quick video for the first test yesterday during the French National SSB contest.

https://www.youtube.com/watch?v=AQ-64oEaT4o&t

Considering the propagation is not so good , i manage to contact Réunion Island, Martinique Island , Marocco and many more....

-Running only 50watts

-I have adjusted the bias to only 400ma in order to cope with my amplifier.

So far so good and easy peasy really, the RF-HFIQ is really cool and provide great performance.

Working now for CW and i start considering :

-CW IQ keyer

-ad5DZ CW KEYER (Software)

But also the following features since the board will be installed in a bigger enclosure with a Raspberry and a dedicated 50 up to 70w amplifier:

-Rotary encoder 

-push buttons for quick access

-Automatic amplifier LPF switching

73s Didier / F5NPV

[Rob French]

Here's some CW code for Quisk on a Raspberry Pi, as well as a hardware definition for the RS-HFIQ.

http://kc4upr.motd.org/?page_id=32

[F5NPV Amateur Radio]

Hi Rob

Many thanks for the information and sound promising .

I will have some spare time this coming week-end and will perform some test with this Quisk Release.

73's Didier

[F5NPV Amateur Radio]

Hello,

Manage to perform some test tonight after office hours.

I am using a RPI 3B+ and the compile of your Quisk from your repository was without any error.

Like you i use a breadboard with a  quick wiring to the RPI . I did a quick test with a Websdr not so far from my place and everything was fine and working very well.  The only issue i need to check is regarding the sound to monitor the CW output. Currently i have no monitoring sound when i am keying the Keyer. Btw it was a really quick test and will need to wire everything properly.

Tomorrow i will do some additional test

73s Didier

[Rob French]

Didier,

Check that you have "Use sidetone" set to "1".  It should be selectable in Config, under Controls.  However, you can also set it in your $HOME/.quisk_conf.py file.  Ensure you have a line, "use_sidetone = 1".

Rob

[F5NPV Amateur Radio]

Hello Rob,

Thanks , Saturday morning i manage to find out . In fact i did a typo in my .quisk_conf.py.

I manage to perform few additional test and currently i find one main issue.

-When using a Iambic paddle , the latency between the sidetone and Quisk is too high and make it difficult to key the paddle and synchronis with the sound sidetone . The milliseconds gap especially at 25 up to 30 Words per minute is really difficult to handle

-->Saturday  I did some test reducing the latency in the QUISK  "Sound" configuration and it really help a lot.Therefore the minimum i can setup is 50ms . Below the buffering is CPU Consumming and some packets are dropped.

I will perform some additional test sometimes this week and maybe using a sidetone provided by the Raspberry itself through a tiny speaker will be interesting to test in order to drop the latency effect. There is a possibility to use GPIO output to generate a 1000hz which sounds promising.

Otherwise , regarding the paddle i use 2 capacitors for the decoupling with the ground (100nf) and a pulldown circuit with 10K + 1K and 1kohms in serial for each paddle connection. At this stage everything is keying smoothly.

Regarding the RS-HFIQ , during the ARRL contest i use this oportunity to perform some additional test with another homemade amplifier (IRFP250N with 24volts). The main target was to provide about 100 up to 300 or 400 watts and with only 1 up to 2 watts (Which is my current setup). With this input power i have a 150/170 watts which is more than enough.

I am starting to prepare a dedicated enclosure which will be equipped with a 7" Display, an encoder for the VFO . 6 switches for quick access menu, the IRFP250 amplifier, Additional LPF and the power supply.


73s Didier



 

[F5NPV Amateur Radio]

Hello,

After scratching my head regarding the latency issue , i was thinking about few options :

-Using the GPIO in order to generated a tone

-Using the T/R or CW GPIO  output to drive a tiny oscillator

-Use the Bug Key mode (Using only the dash) with an external Keyer.

I setup quickly with an arduino nano a tiny oscillator tap on GPIO 26 --> It is working very well and no more latency issue when keying Iambic mode

Therefore , i still have a K3NG CW keyer i was using with my retired TS520 and since the K3NG is full of options i have decided to go for this solution.

 

I modify the hardware_pihfiq.py and replace the BUG menu with K3NG instead. i just plug it to the GPIO throught debounce and pull up circuit with 10nf decoupling capacitor --> use only the dash --> The result is result great and since my K3NG can provide so many options it will be my choice (The Winkey option with CQRLOG and a keyboard is so amazing to use!). 

Still have a lot of work to achieve , but i am more than happy about this product. The sensitivity is really great . Quisk was a bit tricky at the beginning since i never use it before. all the OM during QSO are reporting a very good audio quality. 

73s Didier

[Rob French]

Didier,

Glad you were able to work it out, though I'm bummed that the iambic mode isn't working out for you (due to latency).  I haven't really been affected by the latency, but I also don't send at 25-30 wpm (because I can't copy that fast!!!).

If you're plugging in an external keyer, you can probably remove the debounce circuit if you want. 

The original N1GP keyer code--which is what I derived the GPIO keyer from--is actually intended to drive either a sidetone through GPIO (i.e. using PWM output), or a sidetone through ALSA and/or Jack.  If I get a chance, I will look at adding one or both of those back into the mix, perhaps as options.  (I will probably use ALSA rather than Jack however, just because that is what I have Fldigi and Quisk setup to use.)

I would think that the latency of the keyer code itself is pretty low, since it's C code that uses interrupts from the GPIO input rising/falling edges.  But, there are also two rounds of polling/updates that occur... at a 1 msec rate inside the keyer code, and then at some other rate (I don't know what it is) inside of Quisk.  So that might contribute to slowing it down too.  I'm surprised it's very high though, because I understood the name "Quisk" came from "QSK", i.e. it was designed for QSK operations...

What kind of sound card are you using, and what kind of latency does it show on the Quisk config menu?  I use an Audio Injector sound card "hat" which connects to the Pi via GPIO pins, and has very low latency... I'm guessing that a USB sound card has higher latency.

Also, what sound system are you using to connect to Quisk internally?  I use ALSA rather than PulseAudio... I know that when I previously used PulseAudio (which was generally easier to setup), it had MUCH higher latency than ALSA.

We might want to move this discussion over to the N2ADR groups.io page, as I think we have moved beyond the scope of the RS-HFIQ Technical Support forum... In fact, I'll post my response over there, here's a link.  https://groups.io/g/n2adr-sdr/topic/71744425

Rob KC4UPR

[ON1OD]

Hello
I have the same problem : a waterfall separated in two similar sides .
Is there an answer to this issue ? ?
Regards
Olivier

[all...@gmail.com]

Hi Olivier

I am not responding in terms of the Quisk discussion above but in general for IQ systems. When the waterfall divides itself into two mirror images of each other it is invariably the result of one of the I or Q channels being missing or the two channels being shorted together. Typically there are phase and amplitude adjustments that will be adjusted to null the unwanted image.

Be sure that both I and Q channels are being presented to your sound card and that the sound card is set for stereo operation.

BTW the principle holds for transmit as well. You must be sure that both I and Q are present and that the balance and phase of the two channels have been set up properly. If you do not then you will have a second transmission on an unwanted frequency.

WA8TOD

[ON1OD]

Hi

Thank you for the answer.

Yes, the problem comes from IQ chanels.

Linux Unbutu 20 does not provide a bar graph VU meter so I installed Audacity. The left channel was missing. In PulseAudio Volume Control I choosed for Internal audio :Analog duplex stereo. Beside, the soundcard (Creative 0204 USB) seems not to be well recognised on Unbutu so I plugged the UCA 222 from Behringer.  Then the waterfall looked fine. Only a slight shadow on strong signal only. It comes probably from the hardware and is not disturbing.

I have to test the TX mode now

73

Olivier

ONOD