ARDOP PTT Hangs

[Ethan Crosby]

Hi, I've been trying to get pat to work with my IC-705 using ARDOP but it seems to hang up after transmitting. I can see it transmit the audio but it doesn't release the PTT. ARDOP also doesn't seem to see the PTT-FALSE command. When the ARDOP program hangs I have to close the terminal window, pressing ^C just outputs "terminating on SIGINT" but the process still runs and has to be terminated using the kill command. I am controlling the CAT (and therefore PTT i assume) using Flrig and rigctld. I can't seem to find any really solid documentation or community around ARDOP so I'm posting this issue here. I tried different configurations of software for abour 2 hours today with no luck and massive frustration. My last ditch effort was to try and find an alternative sound card mode in case its something with the ARDOP program thats giving me grief but it seems like theres no other alternative for HF. Let me know what I should provide to help resolve this issue.

[David R. Andersen]

Ethan - it sounds like RF is getting into your soundcard and keeping
the T/R relay closed. The usual solution to this is to get some
ferrite beads and put them on all leads to/from the soundcard to
filter the RF.

When you select a ferrite bead, you should make sure the material it
is fabricated from works at your intended frequency. Sometimes the
material is called the "mix."

If you google "ferrite beads" there are several good articles on EMI
suppression with these devices, and also several places to purchase
them. You can even get them from Jeff Bezos these days. Not sure if
they should be called BezosBeads in that case or not ;)

73
Dave K0RX

--
int x=90560580,y=32678;main(){putchar(x);x>>=y&7;y>>=3;return y?main():0;}
Contact info: https://k0rx.com/contact/

Quis custodiet ipsos custodes?

[Ethan Crosby]

Hey there Dave. Thanks for the response. My sound card is built into the radio so the only cable connecting them is a USB cable and I have 5 ferrite beads on it to suppress RF and still no dice. I could be wrong but I also assumed that when the audio stops there would be not more RF since the radios using SSB. Other digital applications such as Direwolf, Fldigi, and WSJT work with no problems

Its the same problem as this topic but I don't see a final answer in this thread.

Thanks,

Ethan

[MJ Inabnit]

Good morning Ethan,

I'll have to completely agree with you regarding the lack of definable
documentation. I am not a fan of getting config help from a video and
the process took too many hours to extract the critical information I
needed for a simple config file. I REALLY hope we can change that since
I use ARDOP 90% of the time.

I doubt you are getting hammered by RF. You're using a QRP rig. Like
you, I run the whole gamut of digital software (fldigi, js8call etc)

I also ran into issues early on. It's kind of silly but I was never
able to get the ARDOP_GUI to run. I've posted for help and no help
arrived. So here's how I run pat on my old laptop:

I open *two* terminals. In each terminal, I move to the wl2k directory:
cd wl2k

In terminal one, I start ARDOP: ./ardopc_64 8515 plughw:1,0 plughw:1,0

in terminal two I start pat and rigctl:
rigctld -m 2037 -r /dev/ttyUSB0 -s 115200 & pat http

Then I simply open my web browser and connect to localhost and pat does
it's magic. My rig is the kenwood ts-590sg. Same single USB cable as
your IC-705.

Once I'm done with the session(s) I terminate terminal two with a
control+C, then use the command "fg" and again use control+C. The "fg"
command moved the background process to the ForeGround (active.)

Next I move to terminal one and issue the control+C to stop ARDOP.

Overtly complicated at first, but it's working well for me.

I hope we can attract a hacker(s) to resume ARDOP development, it's
really great code and packed with potential.

One final note. ARDOP signal levels are much higher than any of the
other software I use. I keep a seperate terminal using alsamixer, I
turn the audio level from about 23 down to 11 for ARDOP sessions. Your
ALC meter is critical at finding the proper audio level.

I hope this might help you get it working on your end.

GL, 73
j

On 1/6/22 8:04 AM, Ethan Crosby wrote:
> Hey there Dave. Thanks for the response. My sound card is built into the
> radio so the only cable connecting them is a USB cable and I have 5
> ferrite beads on it to suppress RF and still no dice. I could be wrong
> but I also assumed that when the audio stops there would be not more RF
> since the radios using SSB. Other digital applications such as Direwolf,
> Fldigi, and WSJT work with no problems
>
> Its the same problem as this

> <https://groups.google.com/g/pat-users/c/LXmDb56wyZg/m/denxbCB0BwAJ>

> --
> You received this message because you are subscribed to the Google
> Groups "pat-users" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to pat-users+...@googlegroups.com
> <mailto:pat-users+...@googlegroups.com>.
> To view this discussion on the web, visit
> https://groups.google.com/d/msgid/pat-users/01d84524-1599-4712-af50-854eada8465an%40googlegroups.com
> <https://groups.google.com/d/msgid/pat-users/01d84524-1599-4712-af50-854eada8465an%40googlegroups.com?utm_medium=email&utm_source=footer>.


--

wishing you well
Jaye, ke6sls--via the toshiba w/thunderchicken

[Ethan Crosby]

Thank you so much for you response. I've tested the things you've suggested and it still behaved the same, however I may have found a clue as to where it may be hanging. when the ardop program hangs it won't quit using the normal ^C nor when i try to disconnect in pat not even closing the terminal and using the kill command worked, but when i disconnected the radio's USB cable something did happen. This is the ardop command output:

Host Control Session Connected
Host Data Session Connected
[Command Trace FROM host: INITIALIZE
 Command Trace TO Host BUFFER 0
 Command Trace TO Host INITIALIZE
[Command Trace FROM host: STATE
 Command Trace TO Host STATE DISC
[Command Trace FROM host: PROTOCOLMODE ARQ
 Command Trace TO Host PROTOCOLMODE now ARQ
[Command Trace FROM host: ARQTIMEOUT 90
 Command Trace TO Host ARQTIMEOUT now 90
[Command Trace FROM host: LISTEN FALSE
 Command Trace TO Host LISTEN now FALSE
[Command Trace FROM host: MYCALL KC3ILB
 Command Trace TO Host MYCALL now KC3ILB
[Command Trace FROM host: GRIDSQUARE FN00FG
 Command Trace TO Host GRIDSQUARE now FN00FG
[Command Trace FROM host: ARQBW 500MAX
 Command Trace TO Host ARQBW now 500MAX
[Command Trace FROM host: CWID TRUE
 Command Trace TO Host CWID now TRUE
[Command Trace FROM host: VERSION
 Command Trace TO Host VERSION ARDOP TNC_1.0.4.1-BPQ
[Command Trace FROM host: ARQCALL N3HYM-10 10
 Command Trace TO Host ARQCALL N3HYM-10 10
 Command Trace TO Host NEWSTATE ISS
[SendARQConnectRequest] strMycall=KC3ILB  strTargetCall=N3HYM-10 bytPendingSessionID=d0
Sending Frame Type ConReq500M
 Command Trace TO Host PTT TRUE
[Main.KeyPTT]  PTT-TRUE

This is where it always hangs; The following output was after I disconnected the radio:

snd_pcm_status() failed: No such device
 Command Trace TO Host PTT FALSE
[Main.KeyPTT]  PTT-FALSE
[Command Trace FROM host: DISCONNECT
[Command Trace FROM host: ABORT
 Command Trace TO Host ABORT
 Command Trace TO Host BUFFER 0
 Command Trace TO Host BUFFER 0
 Command Trace TO Host NEWSTATE DISC

So to me it looks like the the problem may have something to do with this snd_pcm_status() command. This leads me to believe that the problem may lie with the ardop code somehow. I think that command has something to do with Linux ALSA and I'm on Fedora Linux so maybe that has something to do with it? I tried to compile the ardopc_64 program from source but it seems the 64 bit source code doesn't exist. To whoever wrote ardop, great software, but please use GitHub.

It's also worth noting that I was able to get AX.25 to work correctly with pat but I have yet to test it with an RMS node because I don't have a suitable base antenna. Plus the whole goal of this project is to get Winlink working on Linux and HF and I won't be buying a PROCTOR modem so I guess I just have to struggle at it until something works.

Thanks for all the help,
Ethan

[James R. Pilgram]

I'd like to share a link with y'all on some great info for sluthing RFI.  On or about page 35 of the .pdf is a meter you can build that has been a godsend for sluthing RFI on cables. It also shows some commercially made ones from MFJ but the home built ones work great. 

--

You received this message because you are subscribed to the Google Groups "pat-users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to pat-users+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/pat-users/01d84524-1599-4712-af50-854eada8465an%40googlegroups.com.

[James R. Pilgram]

Oops,

Hit the wrong button and sent before I was done....

I'd like to share a link with y'all on some great info for sluthing RFI.  The article doesn't just apply to mobile but is a great general read and applies to all installations.....

http://www.3905ccn.com/files/pdf/mobile_installation_clinic.pdf     https://0x9900.com/clamp-on-rf-current-probe/    

On or about page 35 of the .pdf is a meter you can build that has been a godsend for sleuthing RFI on cables for me. It also shows some commercially made ones from MFJ but the home built ones work great and is simple to build. Since the suggested snap-bead in the parts list was from Radio Shack and is no longer available, a Mix 43 snap bead from Fair-Rite of any size works. Choose one for the biggest cable you want to test. I had a 200ua meter and used that for greater sensitivity. The potentiometer adjusts the sensitivity so the meter full scale is not real critical. If you don't have a meter, Amazon has 200ua meters for cheap. I used a die-cast aluminum box for the ones I have made.

Using the box, I have been able to test and find the offender(s). Use 1 pass through the box ferrite. It has been interesting to substitute different cable lengths and I have found that it makes a difference. I have found RF on cables then applied beads and played with the turns on that bead to get the best suppression as indicated on the meter. Just key the transmitter and adjust the sensitivity for a full scale reading. Un-key and add or subtract turns or change to a different length cable then key again. Much guess work and many hours of frustration has been eliminated using this little jewel!  

I have a Winlink server (WH6FG) running HF pactor & VARA HF plus  VHF packet & VARA FM that I had big RFI problems with. It ended up that the mini industrial computer I was using is now sitting on top of my IC-7300 with a Motorola CDM1250 strapped to the right side and a SCS Pactor modem on the left side of the IC-7300 and the longest cable is 18" with no ferrites and no RFI. No long cables coiled up.

Jim NH6HI

--

[David R. Andersen]

Here are a couple of additional resources that may be helpful if you
are interested in working on the ARDOP source.

https://ardop.groups.io/g/users
https://ardop.groups.io/g/developers
https://ardop.groups.io/g/users/message/4793?p=%2C%2C%2C20%2C0%2C0%2C0%3A%3Acreated%2C0%2Ck0rx%2C20%2C2%2C0%2C84127830

Dave K0RX

>To unsubscribe from this group and stop receiving emails from it, send an email to pat-users+...@googlegroups.com.
>To view this discussion on the web, visit https://groups.google.com/d/msgid/pat-users/b5a25c6d-d93c-4a08-8fc1-065f1826883en%40googlegroups.com.

--
int x=90560580,y=32678;main(){putchar(x);x>>=y&7;y>>=3;return y?main():0;}
Contact info: https://k0rx.com/contact/

Quis custodiet ipsos custodes?

-----BEGIN RANDOM OR NOT?-----

7lteacmJxYr73kzBMtk6jRCiPSbs/OaPLqz4DY8CYfM7wI/J+oKLDm7w1noa1Fs8
vOsBh0FvrP24oAySkI1tbTiiPi1/ytrWt4wu+XM9enRLHo0QdFas+DzGYJhi/L6k
MiD3Xbm0l9Id/5CMxPxx5dOWmSeC61IIqwGTpzCGzv6h/0anFfzFnjL2n8m0RRgD
-----END RANDOM OR NOT?-----

[David R. Andersen]

One more....

ardopc source can be found at:

http://www.cantab.net/users/john.wiseman/Downloads/Beta/TeensyProjects.zip

On Thu, Jan 06, 2022 at 01:57:19PM -0600, 'David R. Andersen' via pat-users wrote:
>Here are a couple of additional resources that may be helpful if you
>are interested in working on the ARDOP source.
>
>https://ardop.groups.io/g/users
>https://ardop.groups.io/g/developers
>https://ardop.groups.io/g/users/message/4793?p=%2C%2C%2C20%2C0%2C0%2C0%3A%3Acreated%2C0%2Ck0rx%2C20%2C2%2C0%2C84127830
>
>Dave K0RX

[Ethan Crosby]

Okay so that source code is probably going to be super helpful. I've started trying to build ardopc from source which is proving to be just as challenging as the original problem. I'm going to explore if there could be some sort of conflict with the audio set up that causes it it hang though this doesn't seem likely since it does seem to transmit one frame before it hangs. Of all the programming languages I know C and C++ are not one of them so reading the code would sort of be useless. Anyhow thanks everyone for the help. As frustrating as this is the support here is certainly keeping me motivated.

Thank you all,

Ethan

[Ethan Crosby]

After beating my head off the wall for a while now with no results I FINALLY think I've found something interesting. I set everything up exactly the same including rigctld, ardop, and pat except I changed the audio output to my computer speakers and there it is! It keys the radio correctly and sends all the packets correctly. This means (to my untrained eye) that this issue has something to do with the sound card interfacing with ardop. The sound card in this situation is built into my radio so maybe it has something to do with sample rates. Either way this probably means that this issue is now out of the scope of this group. If no one objects to me cluttering up the group, I'm going to dig deeper and continue to post my progress on here in case someone else runs into this issue.

Thanks,

Ethan

[Dr. Torsten Harenberg]

Am 07.01.22 um 05:26 schrieb Ethan Crosby:

> After beating my head off the wall for a while now with no results I
> FINALLY think I've found something interesting. I set everything up
> exactly the same including rigctld, ardop, and pat except I changed
> the audio output to my computer speakers and there it is! It keys the
> radio correctly and sends all the packets correctly. This means (to my
> untrained eye) that this issue has something to do with the sound card
> interfacing with ardop. The sound card in this situation is built into
> my radio so maybe it has something to do with sample rates. Either way
> this probably means that this issue is now out of the scope of this
> group. If no one objects to me cluttering up the group, I'm going to
> dig deeper and continue to post my progress on here in case someone
> else runs into this issue.
>

Hi Ethan,

stupid question (but it happened to me once): do you have VOX enabled?
Usually the VOX is too slow for data modes and keeps the radio
triggered. I had that once with PACTOR and was looking everywhere (even
attached an oscilloscope to the PTT line) only to find out that I still
had VOX on ^^

73s

  Torsten DL1THM

[MJ Inabnit]

Hi Ethan,

Should have asked, what linux are you running?

Are you running the ARDOP code or the ARDOP_64 code?

I'm using Xubuntu, the most recent point release. I needed to use the
ARDOP_64 code for Pat to function. I used synaptic to pull in most of
my ham radio apps, except pat and ardop which were pulled from the
respective web hosts.

I also wonder what alsa code you have installed. I'll attach my lsmod
and alsa bits I have running here for you to compare against.

jaye@ThinkPad-X61:~$ lsmod > mods_on_thinkpad.txt
jaye@ThinkPad-X61:~$ dpkg -l alsa* > alsa_bits_on_thinkpad.txt

One thing that will mess up my digital connection is when the pulse
system defaults to using my radios audio codec. I make sure that when I
boot up, I move the system over to the default intel audio, then my
digital levels work as intended.

Hope this might be helpful om.

73
j

On 1/6/22 8:15 PM, Ethan Crosby wrote:
> Okay so that source code is probably going to be super helpful. I've
> started trying to build ardopc from source which is proving to be just
> as challenging as the original problem.

[Ethan Crosby]

Thanks everyone for the input. I wrote up a post last night but somehow I deleted it when I went to post it. To answer KE6SLS's question I'm running Fedora Linux 35. I managed to fix the compiling bugs fixed and tested the problem with the new software and still the same behavior. I also got the GUI working not that it makes much of a difference. I also played around with the sample rate conversion methods listed on the homepage with no luck resolving the issue. I don't really have a way to confirm if it actually changed anything except that it didn't give me any errors when I set the audio device to ARDOP.

So to recap here's what I've figured out so far:

1. Something is causing the ardop modem to hang after transmitting the first frame leaving PTT on
2. It has to do with the audio being SENT to the radio because the command ardop 8515 plughw:1,0 plughw:0,0 where device 1 is the radio and device 0 is my computer speakers, works. It keys the radio properly and takes input from the radio as confirmed by the GUI. The command ardop 8515 ARDOP plughw:0,0 yields the same results
3. Arfter ardop hangs, when I disconnect the radio to quit the program the next output is

1. • snd_pcm_status() failed: No such device

2. Sample rate conversion (shouldn't) be the issue

Here's my setup again:

1. Laptop running Fedora Linux 35
2. My rig is an IC-705
3. The two are connected by a single USB cable for CAT control and audio
4. the problem exists
   
   1. When using the ardopc_64 binary
   2. When compiled from the latest source

[Martin Hebnes Pedersen]

Hi Ethan,

Thanks for keeping us updated on the issue. Too bad you are unable to get in contact with the author(s) of ARDOPc for help in debugging this. Let's hope they jump in on the conversation at some point :)

Have you tried looking for clues in the kernel messages (dmesg)? Just a thought.

-- 

Martin

--

You received this message because you are subscribed to the Google Groups "pat-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pat-users+...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/pat-users/8a91357e-837d-4e5a-ac1e-025b6bc27091n%40googlegroups.com.

[KM4ACK]

If I am not mistaken, ARDOPc was written by John Wiseman. You can reach out to him in this forum https://groups.io/g/bpq32/topics

73, de KM4ACK

[Chris Keller]

Closing this out, it looks like this was fixed in ARDOP: https://groups.io/g/bpq32/topic/88545355#34236

Thanks Donnie WD1F for working through this with John G8BPQ, and to MJ KE6SLS for spreading the good word!