Doppler control bug on Icom IC-9700

[Aleziss]

Hello all,

I am trying to get the new Icom IC-9700 doppler control work but I get erratic control.

I did try to use the IC-910 driver from hamlib-w32-3.3 by changing the CIV adress to my IC-9700 and it does the same exact issue.

What is happening is that both Main/Sub VFOs are being swapped each time gpredict send frequecy correction to the radio which make VFOs frequencies swap to Sub/Main and then back to Main/Sub. Both frequencies are being updated though.

On top of that issue, the SPLIT function does turn on on the radio which complicate things even more.

Another issue I noticed is that when I transmit, the frequency stop being updated on the transmitting VFO but the receiving VFOs continues to work for a couple seconds until gpredict disconnect from hamlib, probably due to one of the VFO not responding while transmitting.

See text file of debug trace capture from rigctld hamlib-w64-4.0 (I let the doppler control work for a couple second without touching anything nor transmitting) using these switches: -m 381 -r COM5 -t 4532 -s 19200, remember, the radio behave the same exact way with the IC-910 driver from hamlib-w32-3.3 using these switches -m 344 -r COM5 -c a2 -t 4532 -s 19200. Some user reported that this seem to be a known issue with gpredict and hamlib with Icom radios but I can't verify and confirm for sure that this happend on other models at the moment.

You can SEE HERE how the radio behaves when running gpredict with rigctld hamlib v4.0.

The IC-9700 CI-V reference guide can be found HERE.

Thank you all for your help !

Rigctld Trace diag IC9700.txt

[Aleziss]

I may also add the following notes about the VFOs on the IC-9700.

In NORMAL operation with both VFO on, the TOP VFO (Main) is ALWAYS the TX and BOTTOM VFO (Sub) is ALWAYS RX.

In SATELLITE mode, it is the opposite, the TOP VFO (Main) is ALWAYS the RX and the BOTTOM VFO (Sub) is ALWAYS TX.

In both modes (normal/satellite), VHF/UHF frequencies CAN be swapped between both VFOs but the TX always remain at the same place (depending on being on normal/satellite mode).

It is also NOT possible to run both VFOs on the same band.

[dl7oap]

The topic is, too, discussed here:
Hamlib Developer Mailinglist

Objet : IC9700 hamlib-w64-4.0
https://sourceforge.net/p/hamlib/mailman/message/36783570/

Maybe this information will help, to drill down to the real problem. I will get my IC9700 next week, then i can help to debug.

[dl7oap]

some analysis work:
in the rigctrl-tracing file of Aleziss you see a regular cycle of gpredict calling rigctrl with the following:

--- start of a cycle

rigctl(d): t 'currVFO' '' '' ''  // get_ptt
rigctl(d): f 'currVFO' '' '' ''  // get_freq
rigctl(d): F 'currVFO' '145870639' '' ''  //set_freq
rigctl(d): f 'currVFO' '' '' ''  //get_freq
rigctl(d): i 'currVFO' '' '' ''  //get_split_freq
rigctl(d): I 'currVFO' '435218093' '' ''  //set_split_freq
rigctl(d): i 'currVFO' '' '' '' //get_freq

-- and a new cycle starts
rigctl(d): t 'currVFO' '' '' '' // get_ptt

now switching the focus on what hamlib v4.0 makes (did a fresh compile of the newest master branch of hamlib) for hard commands for -m 381(IC9700) from this, all is fine, but get_split_freq and set_split_freq use following hard Icom CI-V commands:

get_split_freq is separated in 3 steps:
a) fe fe a2 e0 07 b0 fd // exchange MAIN and SUB Bands
b) fe fe a2 e0 03 fd // read operating frequence
c) fe fe a2 e0 07 b0 fd // exchange MAIN and SUB Bands

the same behaviour when using set_split_freq. so hamlib switches short MAIN and SUB, to read or write the frequence of the SUB and then switching all back.

actual i think and guess a) and c) leads to massive switching in the display of IC9700.
i will test this with a small python script, when my IC9700 has arrived, and will give feeback here.

by the way: why is the PTT-Status checked every cycle? did gpredict use this information? Duplex-Transceiver should can switch the frequency during TX. so, IMHO frequence change have not to wait for PTT = off

there must be a solution :-) because on this video it is working fine https://www.youtube.com/watch?v=R1-vnHrARwk. Also frequency change, while TX.

[csete]

Thanks for the info. Not sure why PTT is checked, it shouldn't be necessary if one selects full-duplex mode in gpredict.

[dl7oap]

sorry, PTT is my fault. in the gpredict settings of transceiver setup you can define if you want to check PTT or not. i just switch it off. now the PTT status is not request.

i did some more analysis and hamlib has a subcommand for VFO.
rigctl V VFOA will switch to vfo a
rigctl V VFOB will switch to vfo b
rigctl V MAIN will switch to main
rigctl V SUB will switch to sub

at the moment, i guess, it is a better way not to exchange the MAIN and SUB to set and get the frequency.
But to switch with this vfo subcommand directly to the MAIN or SUB and then change the frequency there.

In the sourcecode of hamlib i see that IC9700 is implemented in sourcecode-file ic7300.c. so they just expand the ic7300.c for some IC9700 commands and the functionality of using the vfo subcomands V MAIN and V SUB are not implemented (the IC7300 just have no MAIN and SUB). They are also not implemented for IC910. so this might be the root of the problems, why many ICOM Transceiver are not working with gpredict.

they are only implemented for IC-9100, but at the moment i do not know if gpredict adress the command V MAIN and V SUB? when my IC9700 has arrived i will do testing with rigctl and IC9100 profil. i have checkout out github of gpredict and hamlib and have compiled the newest version of gpredict and hamlib. so i can test on the newest setup.

[csete]

Thanks for the info, it is very helpful to understand the problems. Lately, I have been thinking that the best way forward might be to write our own rig server for the full-duplex radios since our use case is very specific and only requires implemeting a few CAT commands.

[Aleziss]

@dl7oap

The topic is, too, discussed here:
Hamlib Developer Mailinglist

Objet : IC9700 hamlib-w64-4.0
https://sourceforge.net/p/hamlib/mailman/message/36783570/

That is my post on sourceforge too.

[Aleziss]

@dl7oap

sorry, PTT is my fault. ....

but at the moment i do not know if gpredict adress the command V MAIN and V SUB? ...

So the menu in gpredict where we can chose if we like to use VFO as Main/Sub or A/B doesn't work then ?? I've tried all combinaisons and it always does the same thing. I thought those setting were there to accomodate radios that have different ways to use Main/Sub or A/B VFOs...

[dl7oap]

Hi Aleziss, no, the menu himself is working correct. it is to early, i have to collect more facts.

but IMHO and for the moment my suspicion is, that it is not a bug in hamlib and not in gpredict.
the problem is that
a) IC910H and IC9700 is not completely integrated/implemented in hamlib, because VFO Subcommand MAIN and SUB are lack. [IC9100 it is implemented, so here we can do first test without gpredict]
b) and gpredict MAIN/SUB works ok for most radios, but for ICOM it have to call hamlib in a small other way

at the moment i wait for my ic9700, than i can give detail facts and a plan, how to solve this dilemma. i can develop in freepascal, python, java. C is not my favourit, but i can read it. i will try to extend hamlib v4.0 locally on my plattform, and when it works i will make a pullrequest to hamlib. and maybe there have to be a small adaption in gpredict. for a workaround i just developing a small python tcp server script which can be plugged between gpredict and the ic9700 which corrects this. but at the moment i develop this in a suspicion, because my ic9700 has not arrived.

i will give a signal here in 2-3 weeks, when i got the hardware and can check the facts. :-)
i am using linux, so i am very interested, that ICOM TX works with gpredict. for the last JOTA i have to use IC821H with Windows and SATPC32. But with the IC9700 i definitely will not switch back to windows.

[dl7oap]

Hi, short very first feedback. ic9700 arrived today. i was using my prepared python adapter script between gpredict and ic9700 and dopplershift correction works fine for FM and SSB satellites.
the python script just open a tcp server where gpredict can connect, translate in a very easy first way the downlink and uplinkfrequence and is using fe fe a2 00 07 d 0fd (CI-V command to select MAIN) and fe fe a2 00 07 d1 fd (CI-V command to select SUB) to switch between MAIN and SUB Band and setting the frequence with standard fe fe a2 00 05 * fd via usb/serial.

so there will definitely (in the near future) a way that gpredict will work with ic9700. maybe for now with my pythonscript as workaround.

in the next days/weeks i will try to find a way to send MAIN (fe fe xa2 00 07 d0 fd) and SUB (fe fe xa2 00 07 d1 fd) via hamlib. maybe by implementing this missing feature in hamlib. it should be no big step.

as a last step (when hamlib will have this feature) we have kindly to ask, if gpredict could use a small other way to speak with hamlib while using icom transceiver.

[Aleziss]

@dl7oap thank you so much for the update ! is your python script currently replacing the hamlib driver ? I don't know much about that programming language but can it be ran under windows ?

[dl7oap]

the pythonscript is replacing hamlib. it uses the correct commands. it is very rudimentary but i use it to test out the real civ-v commands and how the ic9700 reacts. i can send you the script. please contact me, via email on qrz.com.

In the last 2 days i monitored the serial communication between ic9700 and satpc32/satpc32iss with a hex-sniffer and have created a complete list of the used civ-v commands. now i compare these commands against hamlib sourcecode (freshest commit from hamlib repository) to find the correct request for rigctl.
Until now i have found 3 relvant civ-v commands which can not be called by hamlib for ic9700.

Next steps:

• complete the necessary command list
• fix hamlib
• write a short concept/userstory how commands have to be used for SSB, FM and simplex (like ISS) satellites in gpredict

[dl7oap]

There is no need of fixing hamlib. All commands are already present in hamlib 4.0 (nightly build).

This is a list of all commands, which will be needed, to control IC9100, IC910H, IC9700 for satellite operation (SSB, FM and Simplex (like ISS)):

concept_command_mapping.pdf

Hint: IC9100 and IC9700 have the identical CIV-V commands here. IC910H has only one difference: rigctl G XCHG has the meaning of exchange VFO A and VFO B (and not MAIN/SUB).

Last step:

• writing a short concept/workflow how to use these commands for SSB, FM, simplex-Satellites for IC9100, IC910H, IC9700

[Aleziss]

@dl7oap wonderful work you've done there, it is getting along pretty good I think ! I am trying to reach you to test your script (on windows ??) but it seem you don't have an email contact on qrz.

I see in your concept command list that there are commands to set tones and modes, that was another suggestion I set to implement in gpredict that it would be nice it could send those information to the radio and set it automatically with the parameters fetch from the internet.

good work, it's nice to see how you work on this and mainly, thank you for sharing the progress on fixing that issue with icom radios and gpredict !