Thetis fork and Update for HL2

417 views
Skip to first unread message

r webby

unread,
Apr 27, 2023, 6:52:50 PM4/27/23
to Hermes-Lite
Hello everyone!

I have been working on a version of Thetis that I forked from Richie, MW0GLE. You can set the radio to HL2 in this version and it will set up the defaults for N2ADR filters and so on.

If you'd like to try it and help me iron out the bugs, you can post issues here:

Download:

Before installing, keep your database safe! Uninstall any version you have, and give mine a try! I am hoping to merge my efforts with Richie's version, but if he does not want to, I will change the name of mine to avoid any future confusion. This is because I have been working for a long time on an HL2 version of Thetis, with a mind to improve mostly VAC1 for SSB. So I ported these changes into the standard version.

Please if you have any comments, try to keep them on github rather than on here, as I do not check in here very often at all,

Steve, KLJ.

Toonces Cat

unread,
Apr 28, 2023, 2:42:33 PM4/28/23
to Hermes-Lite
Steve,

 I liked how 2.9.19 worked with the PS-M . I was able to tell I was in manual PS. The latest (2.9.024) with the check mark in the linearity window. I am back to not knowing if I am in manual, In fact you can turn PS-A on and have the check mark in Manual and it will be in PS-A and give you the (Dalek) randomly.

Thanks for all your hard work...I reverted back to 2.9.19 so I will know for sure I am in PS manual

73's
Mike WA3O

r webby

unread,
Apr 29, 2023, 6:23:36 AM4/29/23
to Hermes-Lite
Yes, it made the ANAN users (or should I say one very important ANAN community guy) scream blue murder. Don't worry, it will return (unless we get to the bottom of the random DALEK/Voice Of the Mysterons/Picket Fencing/ PS Freaks out randomly) on certain ANAN models and the HL2. We need it. Don't let anyone gaslight you into the idea you do not. LOL :-)

r webby

unread,
Apr 29, 2023, 6:25:29 AM4/29/23
to Hermes-Lite
Be careful with SingleCal. It is supposed only to work for that tranmission only. ie.; it is no good as soon as you key up.

On Friday, 28 April 2023 at 19:42:33 UTC+1 toonces...@gmail.com wrote:

Toonces Cat

unread,
Apr 29, 2023, 8:19:20 AM4/29/23
to Hermes-Lite
Steve,

 I have been using the PS-M with great success by going on the frequency I want to be on...using PS-A until I see it lock and then pressing the PS-M. after doing this I can work 100% WITHOUT the DALEK.

Thank you!!

73's

Mike WA3O

RdWing

unread,
Apr 30, 2023, 1:05:51 AM4/30/23
to Hermes-Lite
Single Cal is meant to lock in a correction, i.e. you transmit a two tone then hit single cal, and it will take one correction and apply that to any future transmissions until you either enable PS-A, or click Single Cal again, which will take another "permanent" correction.

I have another question too, I am struggling trying to setup an HL2 with this fork. I am seeing power reported as if there was a 100w PA. I'm sure there's just one button but I'm not that familiar with using Thetis with HL2. Help a fellow out?

WU2O Radio

unread,
Apr 30, 2023, 12:16:16 PM4/30/23
to Hermes-Lite
I am the "blue murder" guy. Or "blue meanie" if you prefer 😄. Steve is doing some really tremendous work, both foundational and in terms of potential feature improvements, and us Apache/Thetis folks are greedy and want all the cool stuff he is accomplishing for the HL2 folks to carry over to us as well. It's the bare truth!

RdWing has it right about how Single Cal works. Assuming that the only work-around for whatever problems HL2 folks are having with PureSignal in Thetis (PS) is to stop the automatic (continuous) PS measurement and correction update process during a transmission and just let a single measurement and resultant correction ride then the already existing Single Cal function will do that. The only problems (really just annoyances) with Single Cal as implemented in Thetis (and PowerSDR mRX PS before that--it's the same) is that a) you have to pull up the Linearity settings window to get a hold of that button and b) when you punch that button the UI doesn't explicitly show that any sort of correction is in use during TX. You have to infer it is working by looking at your transmitted spectra using DUP. If a Single Cal button were to be added to the main console UI, and the behavior of the "Correcting" indicator changed to go green when it has a valid Single Cal correction in effect, this would make it just about perfect. For myself, for when I use my Apache stuff on digi modes and want a Single Cal approach (which is what most Apache folks use it for), I added a Single Cal button to some macro software I have (it's accessible via a CAT command, "ZZUS") but that's just another UI work-around.

I would like to offer help to track down what's going on with PS and HL2 as an exercise in enlightened self-interest. If a common build of Thetis can be maintained for both Apache and HL2 users then it's a win-win. To do this I have a number of questions I'd like to pursue with HL2 users who are having the problem. BUT...this is your thread, Steve. If you don't want me sending it in that direction please let me know and I'll start a separate one.

73!

Scott

WU2O Radio

unread,
Apr 30, 2023, 1:55:22 PM4/30/23
to Hermes-Lite
While I was waiting for moderation on my first post, I had more than ample opportunity to study the group archives on the subject of HL2 PureSignal. I will start a separate topic on this.

73,

Scott

r webby

unread,
Apr 30, 2023, 1:55:26 PM4/30/23
to Hermes-Lite
Hello Blue Murder & friends,

Not too worried what goes on in the thread; I don't claim ownership of it. However, I have some potential bugs that I have found in wdsp code that I will need to take up with Warren directly at some point if he is receptive.

Now I know that a happy Anan user will claim it "can't be the software, I have never once seen picket fence/Dalek/distortion" however there is a notion in C of undefined behaviour. I think I see it in the code, though it needs an expert to check.

When UB occurs, anything can happen. Your rig may work perfectly, forever. Your PS may work 99/100 or even 99/1000 and be fine the rest of time. Next door's cat can get pregnant.
Undefined behavior - Wikipedia


I think this may well be the case here. What makes it so difficult to debug is that there are a couple of threads that are created from scratch not only every time PS is entered, but every time a change is encountered. Since static PS does not spawn these threads, and is stable, this seems to me to be the logical place to look. For those in the know, the time between when you ask for a thread to start, and when it actually starts, is unbounded. It could take 1 microsecond or tens of millis. But it needs a lot of time and effort to deal with it, ideally with a multi-threading expert. And you can count how many of them there are in the world with one hand.

The best thing for now is for everyone to use their favourite revision of my, or anyone else's code; the one that works for your use-case. This, being a github project, you can 'rewind' to your favourite incarnation and compile it, and it will work for you! Every change on every commit is saved and retrieveable. So if you want the PS-M version, just rewing back a few commits and use that one. If you have it already, keep using it.

Bear in mind that this issue may never be resolved (many UB cases are not) and so that's another reason to use what works for you!

Have fun, everyone!

Steve


WU2O Radio

unread,
Apr 30, 2023, 2:20:15 PM4/30/23
to Hermes-Lite
Hi Steve, 

Nobody said it never happened, it's just not happening on the more modern Apache hardware. What's really interesting is the few complaints that are voiced on the Apache side involve the ANAN-10E hardware. The 10E has some very similar issues to the HL2 where RF isolation is concerned, TX to feedback path.

However, more interesting, is the observation that this doesn't happen with PowerSDR but only with Thetis. A big potential clue.

If people are willing to chime in on the topic I'm about to start I'm willing to bet that we can figure out the problem.

You have Warren's email, I'm confident he would not mind hearing from you.

Cheers!

Scott


r webby

unread,
Apr 30, 2023, 3:00:56 PM4/30/23
to Hermes-Lite
Hello,
You put up the source for wdsp, thanks have integrated it.

Steve

WU2O Radio

unread,
Apr 30, 2023, 3:06:14 PM4/30/23
to Hermes-Lite
I saw your commit :-) Just trying to help the best I can, happy to do it!

73!

Scott

--
You received this message because you are subscribed to a topic in the Google Groups "Hermes-Lite" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/hermes-lite/5WWq-BvNVFs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to hermes-lite...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/ff85af85-6c9d-48b7-82b3-a8725163da56n%40googlegroups.com.

Toonces Cat

unread,
Apr 30, 2023, 3:21:46 PM4/30/23
to Hermes-Lite
Steve and Scott,

  I am one of the 10E (protocol1) guys with the "problem" with PS-A. I can confirm that this is a nonissue with OpenHPSDR mRX PSv3.4.9 (3/19/18) and PS-A.
At some point Ramdor was doing upgrades to Thetis and he changed the wdsp.dll to an upgraded version. I know this because I installed a Thetis.exe and it failed to run because of a DLL version conflict. I am not sure if this is the difference between the to software packages .

 "However, more interesting, is the observation that this doesn't happen with PowerSDR but only with Thetis. A big potential clue"

Hope this helps.

73's
Mike WA3O

WU2O Radio

unread,
Apr 30, 2023, 3:29:07 PM4/30/23
to Toonces Cat, Hermes-Lite
Mike,

I don't know if you saw my dedicated topic I created for discussing this issue, but if you would be so kind as to duplicate your inputs there in order to start a focused discussion on the issue I would greatly appreciate it. Please also describe your PS settings in detail. You'll see what I wrote in the other topic.

Thanks,

Scott

Toonces Cat

unread,
Apr 30, 2023, 3:58:48 PM4/30/23
to Hermes-Lite
Scott,

 I Pmed  you (on the Apache Labs forum) my phone number over a week ago  trying to convey the information over telephone call and you choose to not communicate. I think posting the information at text or email is sometimes not the best way for me to explain what is happening. I think a half hour conversation with you may clear some things up on both sides.

Up 2 you.

73's
Mike WA3O

WU2O Radio

unread,
Apr 30, 2023, 5:00:59 PM4/30/23
to Hermes-Lite
I don't think a call is necessary. The problem seems clear. Now we are just trying to find the root cause. And it's important to keep the discussion public so that others can fully appreciate all aspects and, I hope, also contribute to finding the solution.

73,

Scott

Reply all
Reply to author
Forward
0 new messages