QSK with Thetis

User avatar
w9ac
Posts: 290
Joined: Sun Apr 09, 2017 4:01 pm

QSK with Thetis

Postby w9ac » Mon Sep 27, 2021 2:55 pm

For many CW operators, full QSK with Thetis is a pleasure to use. When QSK is engaged, Thetis defaults several operating values (e.g., AGC). However, since implementing full QSK, I noticed that on bands 20m-6m, moderately low AGC hang-time settings resulted in slow QSK recovery.

So, why only 20m-6? Internal RF leakage strength into the 7000DLE receiver during transmit increases with an increase in frequency. This may, or may not be an issue with other ANAN transceivers. Internals of the transceiver, including cable routing can impact leakage. By trial-and-error, I discovered that the slow recovery impact is eliminated by running the AGC (H)ang line up near the top of the GUI display.

If slow QSK recovery is an issue for you, simply grab the hang line with the mouse and slide it up. Bring it up far enough to achieve artifact-free recovery. Keep in mind that output power will have an effect on AGC hang placement.

That's it. QSK is now smooth and artifact-free from from 160m through 6m.

Paul, W9AC
VE6IV
Posts: 6
Joined: Thu Mar 12, 2020 10:07 pm
Location: AB Canada

Re: QSK with Thetis

Postby VE6IV » Tue Mar 01, 2022 1:29 am

Thanks Paul (and others) for your continued work making this radio shine as a real cw dream machine. I just found out today that my 7000 is close to shipping, so I'm madly reading back on all the tips and tricks you've imparted, with the hope this will allow for a relatively smooth transition to QSK-cw mode on this platform. Can't wait to dig into it!
Cheers,

Kirb - VE6IV
RF Unit: ANAN 7000 DLE MKII Black, P2_v2.1
CPU: AMD Ryzen™ 9 7950X Processor, 16 core, 4.5GHz, MB: ASUS ROG STRIX X670E-E, 32 GB DDR5
GPU: ASUS Dual GeForce RTX™ 3060 OC
User avatar
w9ac
Posts: 290
Joined: Sun Apr 09, 2017 4:01 pm

Re: QSK with Thetis

Postby w9ac » Wed Mar 23, 2022 12:11 am

The attached audio .WAV file demonstrates Thetis' CW QSK ability when using low-latency ASIO audio drivers. The audio clip begins with my ID on the 80m band. Shortly thereafter, I inject a carrier to simulate another station on the band. Then, the keyer speed slowly increases from 15 WPM to 50 WPM. Notice that at all speed settings the carrier and band noise is easily heard in between keyed CW elements. Try that test at 50 WPM with any other transceiver including those made by Flex or Elecraft. What does this say about ANAN's "thick wire" disadvantage?

The .WAV file sounds noisy because it's digitally compressed. I didn't take the time to optimize it. The Apache message board wouldn't allow an upload of the original uncompressed file size. Nevertheless, the point is to demonstrate Thetis' fast QSK ability. Also attached are screen captures of the settings that produce this result.

Paul, W9AC
Attachments

CWL 3.550100MHz [32bit 48k] 3-22-2022 7 42 27 PM-1.wav [ 561.41 KiB | Viewed 6569 times ]

W9AC Thetis Audio Setup-DSP.jpg
W9AC Thetis Audio Setup-DSP.jpg (113.8 KiB) Viewed 6569 times
W9AC Thetis Audio Setup.jpg
W9AC Thetis Audio Setup.jpg (137.71 KiB) Viewed 6569 times
PHILIPASCHECHTER
Posts: 19
Joined: Mon Aug 31, 2020 5:24 am
Location: CRESTONE, CO. 81131

Re: QSK with Thetis

Postby PHILIPASCHECHTER » Wed Mar 23, 2022 6:40 am

Hi All,

I am not sure that my CW sidetone issues belong here however I think it can not hurt to share. I am really having trouble using Thetis for any CW due to my sidetone sounding overdriven and sloppy. The first temporary fix was to with RX1 all the down and then turn it back up. alternating between settings for sending and receiving. At least I could use the sidetone to send and hear what I am sending in a way that allows me to send correctly, but I am barely fast enough to have a normal exchange with all the level adjustments. The second fix attempt was to hit mute when sending and immediately unmute to listen. Faster than first attempt and totally inadequate for operating.
Obviously, QSK is not even in the realm of reality for me with this rig. I have an IC-7800, a great QSK rig and it has a useable sidetone. So, despite my intentions to make my ANAN 7000 DLE MKII my primary rig, it mostly sits unused.
Paul, if you see this, I have tried the settings that you have posted and maybe they do help with QSK. I really don't know as I'm stuck at racing back and forth to the mute button to hear an acceptable sidetone. I am hoping that I am missing something simple in the setup, that someone can point me to ending the mute to achieve a hear useable sidetone saga. BTW, I have experimented with lowering AGC, lots of pitch changes and nothing seems to help the quality of the sidetone. I am frustrated to say the least and I get to thinking that this radio was the wrong choice. Any help will be greatly appreciated.

73,

phil, NE0S
User avatar
w9ac
Posts: 290
Joined: Sun Apr 09, 2017 4:01 pm

Re: QSK with Thetis

Postby w9ac » Wed Mar 23, 2022 12:50 pm

Phil,

First, until you have ASIO audio capability, the settings shown will result in horrid CW performance whether in QSK or semi-break-in modes. If you've not read through the forum discussions, that would be a good place to start. In addition to the setup screens shown above, the CW settings under <General><Options><Options-1> also have an impact on CW performance. Since I only operate QSK, it would be worthwhile to ask for Thetis setup screen captures from other CW ops who are currently using Thetis and ANAN models with the Orion MKII board.

Another important caveat to the achieved result is that my ANAN 7000's T/R relay was replaced with a fast RF reed relay with make-break times of 1 ms. That, together with ASIO, a dual-port GPSDO, audio interface with Wordclock, and the Thetis settings shown all contribute to the overall QSK experience. Getting there was an incremental process.

Paul, W9AC
W4WMT
Posts: 325
Joined: Sun Apr 09, 2017 10:12 pm

Re: QSK with Thetis

Postby W4WMT » Sat Mar 26, 2022 10:23 am

w9ac wrote:The attached audio .WAV file demonstrates Thetis' CW QSK ability when using low-latency ASIO audio drivers.


Hi Paul,

Those are very low ring-buffer settings shown in your VAC1 screenshot above. Even with the Studio 192 wordclock synched coherently with the radio, that's a nice trick. How long can you run like that without accumulating ring-buffer events? If you are getting events, are they occurring as singletons or in clusters?

73, Bryan W4WMT
User avatar
w9ac
Posts: 290
Joined: Sun Apr 09, 2017 4:01 pm

Re: QSK with Thetis

Postby w9ac » Sat Mar 26, 2022 11:20 am

W4WMT wrote:Those are very low ring-buffer settings shown in your VAC1 screenshot above. Even with the Studio 192 wordclock synched coherently with the radio, that's a nice trick. How long can you run like that without accumulating ring-buffer events? If you are getting events, are they occurring as singletons or in clusters?

73, Bryan W4WMT

Typically, the over/underflow count is stable until some other PC application starts or terminates. The biggest offender is my web browser. When it opens, a cluster of perhaps 50 flow counts occur. Thereafter, its stable again. At most, I hear a short glitch.

Within Thetis, if the profile is changed on the fly with VAC enabled, several things require rebooting, sometimes the PC. No issues as long as the profile change occurs with VAC disabled, then reenabled. With the PC left running alone without application interrupts, the counters will stay near 0 for a long time.

When nighttime 80m conditions permit, I'll attach another .WAV file that demonstrates an ionospheric delayed echo. The T/R is so fast that on some nights with my amp at 1KW, the echo is as loud as my own received sidetone. Of course, the T/R turnaround time can be delayed if it becomes too obnoxious.

Paul, W9AC
User avatar
w9ac
Posts: 290
Joined: Sun Apr 09, 2017 4:01 pm

Re: QSK with Thetis

Postby w9ac » Sat Mar 26, 2022 12:08 pm

Another important point regarding the browser interrupts: My system has two NIC cards on their own subnets. The second NIC is dedicated to my SDR transceivers.

I followed W1AEX's tutorial and prioritized the SDR NIC. In doing so, web pages opened slowly and became an annoyance. For example, opening a new page took seconds. I then reverted back to prioritizing the first NIC. So, it's come down to what's the least objectionable action and for me it's been the short glitch when opening a web browser.

Paul, W9AC
K9RX
Posts: 414
Joined: Fri Apr 14, 2017 3:47 pm

Re: QSK with Thetis

Postby K9RX » Wed Apr 13, 2022 1:46 pm

Regarding QSK using Thetis. I'm new to Thetis having just installed it for the first time recently - I'm dumbfounded by it, it is absolutely incredible, like a new radio. My fears, or desire to have a reliable radio (I had been using 3.4.9 which worked extremely well re reliability) were unfounded. I put too much emphasis on bug reports here - none of which have been an issue for me.

Now - that stated. I found one anomaly, bug I'd say, the other day. I use SEMI break-in as I've always preferred it. However, I tried QSK. I was amazed at how well it worked. It was as if I was in constant RX. Superb sounding both while transmitting (i.e. the monitor) and RX. Then I noticed that the delay was set to 0... so I changed it to 10. Ok - very little difference - maybe just a slight amount... then I changed it to 100. Now there was this ugly end sound - like.. hard to describe, an echo - louder than the primary sidetone. AND now the AGC was set to CUSTOM I noticed and the RX recovery time was very long. I then went back to 1, same thing - 0, same thing. I was from this point on unable to get it to work as it did initially. That is where I am now. I tried going to a previous DB to no avail. If I hit a single 'dit' I can hear like a pip PIP.

It is not the hang time - just as a note (read here, thanks Paul). If I turn MON off then I hear the sidetone, different tone, and it is no longer doing the end louder tone - but now there is a delay that is enough to cause sending CW to be difficult if not impossible at higher speeds (30+). So MONITOR (sidetone) has to be on - but then the problem is there.

I can try to figure out a way to record this. I have ICE CREAM for video but haven't had a need to figure out how to get it to also record audio - although I suspect that isn't possible since the monitor is IN the radio and not in the computer. Oh - yes - this is with headphones, this is not using audio through a VAC. Also, this is using a K1EL Winkeyer which has a hard wired key/paddle input to the radio (no direct bearing but I mention it as someone will ask).

Now one might say - what's the big deal if you use SEMI. Indeed that would have been the case until I heard that initial action of QSK - it was unlike anything I had experienced before (756PIII or FTdx5K) ... and I think I might want to revisit it.

HELP?!

Gary
K9RX
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: QSK with Thetis

Postby w-u-2-o » Wed Apr 13, 2022 2:33 pm

Gary,

If you mess with the delay, or too much with the custom AGC settings, it will screw up full break in QSK. Also, when in that mode, the RX is NEVER turned off, you are in full duplex TX/RX mode 100% of the time.

Set delay back to 0, set the custom AGC back to Slope 5, Max Gain 100, Decay 1, Hang 12.

Please do read the original release notes for the QSK functionality here, it will explain a lot: viewtopic.php?f=9&t=3049

73,

Scott
User avatar
w9ac
Posts: 290
Joined: Sun Apr 09, 2017 4:01 pm

Re: QSK with Thetis

Postby w9ac » Wed Apr 13, 2022 3:10 pm

K9RX wrote:Now one might say - what's the big deal if you use SEMI. Indeed that would have been the case until I heard that initial action of QSK - it was unlike anything I had experienced before (756PIII or FTdx5K) ... and I think I might want to revisit it.

HELP?!

Gary
K9RX

Gary,

I'm surprised you had any semblance of good QSK without using VAC and an ASIO driver. Generally, the idiosyncrasies you describe go away with ASIO. Also, I am not sure which delay setting you're referring to when set to 0. AGC? QSK delay setting under the <DSP> <CW> tab? Be very careful with 0 Tx settings; it requires looking at the RF envelope with an oscilloscope for signs of hot-switching.

To get back to the original operation, you mentioned a DB reset, but how about a full PC and ANAN power reboot? Anyway, try it.

At this time, I no longer use VAC when QSK is engaged. The entire ANAN audio CODEC is now digitally-replaced with a Presonus 192 sound interface. In fact, my 7000DLE's audio jacks are disabled with digital piping.

As someone once said: "There's QSK and then there's QSK." One man's opinion of good QSK is different than another's." To me, the Holy Grail is full-duplex CW without *any* keying artifacts. To get consistently fast QSK without artifacts, I suggest the following:

1) Use ASIO Drivers;
2) Use an ASIO-capable sound interface with a Word Clock input;
3) Purchase a dual-port Bodnar GPSDO. One port set to 10 MHz for EXT REF in; the second port set to 48 kHz. Feed this to the Word Clock input;
4) Adjust AGC (H)ang time, typically pushing it up just far enough to eliminate AGC artifacts; and
5) Replace the internal T/R relay with a fast RF reed relay that's meant for the job. This will ensure that 0 delay settings won't hot-switch the transmitted RF. I'm using an Aromat RSD-12V that switches in 1 ms. on make/break without the need for contact debouncing. Some datacom relays switch very fast but need an additional few ms. for contact settling time.

For most folks, this will be too much work for consistent QSK results without artifacts. If you're fine with semi-break-In CW, then your best option may be to leave it alone.

Paul, W9AC
K9RX
Posts: 414
Joined: Fri Apr 14, 2017 3:47 pm

Re: QSK with Thetis

Postby K9RX » Wed Apr 13, 2022 3:23 pm

Paul: Not sure where you're coming from, I think you think I'm using VAC?! As mentioned I'm not using any form of VAC. This is strictly the rig's hardware/FPGA, nothing external to it. And as mentioned when I first tried it it worked superbly - or at least I believe it did - I sent (only) 'v' a few times - then noticed the delay was set at 0 which concerned me greatly re RF being present - so I changed it, ultimately to 100 when all of a sudden the action was considerably different.

Scott: I'll read what you presented. Note - I changed the delay from 0 to 1, then I believe 10, then 100 - nothing else was changed by me. When I went back to 1 and 0 it continued with the weird behavior. So any settings other than this were not done by me. That stated I'll check the AGC and see if that corrects it.

Gary
K9RX
Posts: 414
Joined: Fri Apr 14, 2017 3:47 pm

Re: QSK with Thetis

Postby K9RX » Wed Apr 13, 2022 3:39 pm

Ok - read that thread. Although I was following everything prior to installing Thetis (when using 3.4.9) I had no particular reason to focus on one thing or another ... obviously there's very detailed info there. First and foremost I'll need to get to prot 2 as I'm currently on prot1. What is curious however is the fact that it appeared to work until I changed the delay and then wouldn't work again. But I understand that until i get to prot2 all bets are off so let me try that - I'll report back if there continues to be an issue, which I suspect won't be the case.

gary
W4WMT
Posts: 325
Joined: Sun Apr 09, 2017 10:12 pm

Re: QSK with Thetis

Postby W4WMT » Wed Apr 13, 2022 4:32 pm

K9RX wrote:When I went back to 1 and 0 it continued with the weird behavior. So any settings other than this were not done by me.

Could this be attributed to the "latency creep" phenomenon reported earlier (more than once) by W9AC? Paul reported this happens only when using the hardware CODEC in the rig (not when using virtual audio).

73,
Bryan W4WMT

Return to “Keys, Keyers, Paddles & Bugs”