DSM Telemetry support

deviation-devo10-disp_non_zero_v4 and deviation-devo10-disp_all_values_v4
are working.

Edit:
Your last version of deviation-devo10-v4.0.1-4dc1002 does work as expected.
Could not find any bugs jet

Now I let it run over night and see what it does tomorrow.

Thank you very much
Manfred

Hi,

now it comes to the sad and probably difficult part. I try to describe my observations as precise as possible.
Test-environement:
Devo10 and MiniCP which supports TEMP1 amd Volt2
TX powered via 12V AC adapter.
I don't have a power-source for the RX that lasts longer than 4h.

1.) let TX powered on over night ~11h (with telemetry on)
2.) powered on RX
=> no connection to RX, which I think should work.
3.) switched telemetry off
=> connection to RX is working
4.) switched telemetry ON
=> connection to RX is working
- no telemetry values for ~1min (all 0 inverted)
- after ~1min valid telemetry values showed up (as it should)
- but eventually got inverted again (should not, as RX is 3m away)
- Values came back after ~30s -1min
5.) Switched USB mode on
=> is working,
6.) Switched USB mode off
=> telemetry is working as expected!
- valid telemetry values are displayed.
- they don't get inverted anymore.
- everything is working as it should.

All in all everything seems to work as expected, but I am afraid, that
eventually could happen a LOSS OF CONNECTION.
datalog isn't working for me.

Thank you for testing.

Ok, telemetry works but does not update constantly as it should. vlad_vy reported similar problem with Devo telemetry after a particular change was made. Since that change vlad_vy has been using a custom build with the change reversed. I'd like you to test that and see if it works better.

If it is worse, please try v2 of the last one.

Hi,
I couldn't find any telemetry related bugs in deviation-devo10-vlad_vy.zip until now. Need to test longer.

One new bug in mixer page:
(tested with small fonts)
- grap the attached model2.ini
- goto mixer page
- goto to i.e. "ch7 DR&Expo"
- scroll down
=> page becomes empty half way down.

this bug wasn't there in nightly 2015-03-05
"deviation-devo10-v4.0.1-a4e6135"
But it is there in nightly 2015-03-07
"deviation-devo10-v4.0.1-35e768d"

Could you provide more info, what I should look for?
what should be the difference between
"devo10-vlad_vy" and "devo10-v4.0.1-4dc1002_v2"?

greetings
Manfred

v2 is a code tidy up of "devo10-v4.0.1-4dc1002". There should be no change.

"devo10-vlad_vy" is v2 plus reversing the change that vlad_vy said made telemetry worse.

Both are latest nightly merged with telemetry. Sounds like you have found a bug in nightly.

I was hoping Devo telemetry with "devo10-vlad_vy" would be better than v2. If it's the same and it also works for vlad_vy then that is the better version.

Did you have telemetry 'slow to update' problem with either or both?

Indigo wrote: I was hoping Devo telemetry with "devo10-vlad_vy" would be better than v2. If it's the same and it also works for vlad_vy then that is the better version.
Did you have telemetry 'slow to update' problem with either or both?

I only tested devo10-vlad_vy so far and no telemetry issues jet.
Should I test "v2" as well? (I didn't even download V2 jet)

Indigo wrote: Sounds like you have found a bug in nightly.

Should I file a ticket in Bug-Tracker?

No need to test v2.

The bug is caused by my fix for #606. I have found the cause and it will be fixed in next nightly.

I have just uploaded: deviation-devo10-v4.0.1-a0f31b6

It is "devo10-vlad_vy" plus new fix for #606. And a change that might fix telemetry 'slow to update' problem.

Indigo wrote: I have just uploaded: deviation-devo10-v4.0.1-a0f31b6
It is "devo10-vlad_vy" plus new fix for #606. And a change that might fix telemetry 'slow to update' problem.

I could confirm 'slow to update' with nightly 2015-03-05 "deviation-devo10-v4.0.1-a4e6135" witch isn't there in "devo10-vlad_vy"

Indigo wrote: I have just uploaded: deviation-devo10-v4.0.1-a0f31b6

No telemetry values displayed
All telemetry inverted
Mixer page seems to work.

In regard to telemetry IMHO "devo10-vlad_vy" was the best right now.

linux-user wrote: I could confirm 'slow to update' with nightly 2015-03-05 "deviation-devo10-v4.0.1-a4e6135" witch isn't there in "devo10-vlad_vy"

That's great. So Devo telemetry is now working properly without issues?

linux-user wrote: I left the previous version (deviation-devo10-v4.0.1-8d9113c.dfu) powered on over night, and got locked out. (RX not binding automatically)
- switched telemetry off => RX is binding.
- switched telemetry on => RX binds again.
This was done without power-cycling anything.

I want to determine if this is a telemetry bug or protocol bug.
Does it happen if telemetry is switched off and left switched off?

Indigo wrote: Re: The loosing binding problem.
I want to determine if it is a telemetry bug or protocol bug.
Does it happen if Telemetry is switched off and left switched off?

This tends to become difficult to tell, as I am not sure to be able to reproduce this.
But:
1.) cmpang reported Devo protocol being reliable with telemetry switched off #28056 ff
2.) This seems to correlate with my experience a year ago #525

Edit:
I may try to test this but don't expect any result before end of next week. Every test run lasts one night, and I won't trust one single run.

cmpang recently reported it as fixed ON DSM2/X protocol #29304

So I've made the same change to Devo protocol v2 below.

Available for download here :
"v4.0.1-c0b203a" is same as "devo10-vlad_vy" merged with next nightly.
"v4.0.1-c0b203a_v2" is same as above with fix for #525

I'd like to publish v2 as a test version in the downloads section, but first I need confirmation that Devo telemetry is still working. Anyone?

Indigo wrote: "v4.0.1-c0b203a" is same as "devo10-vlad_vy" merged with next nightly.
"v4.0.1-c0b203a_v2" is same as above with fix for #525

Hi Indigo,
quick test of deviation-devo10-v4.0.1-c0b203a_v2.zip looks promising
I will fly it today, and then see what next "over night test" tells us.
- over night test with "devo10-vlad_vy" ended with lost control.

Are you saying, that "devo10-vlad_vy" doesn't include the fix you have done for DSM2/X?
And v4.0.1-c0b203a_v2 should include it?

Yes, only v4.0.1-c0b203a_v2 has the fix and I'm confident that it will fix the problem.

The above issue has been fixed.

A new version is now available:
bitbucket.org/Indigo1/deviation/downloads

Indigo wrote: So I've made the same change to Devo protocol v2 below.
"v4.0.1-c0b203a_v2" is same as above with fix for #525

Hi,
now comes the sad and probably difficult part
1.) I, left devo10-v4.0.1-c0b203a_v2 (Devo protocol fixed-ID with telemetry ON) powered on for some hours. (RX powered off)
=> no connection to RX anymore (when powering on RX)
2.) same test with telemetry switched OFF:
=> RX still connects instantly.

Could anyone confirm this?
cmpang?!

Something within the telemetry code must have the possibility to disrupt transmission. #565 #525 #21247

BTW:
This topic is named "DSM Telemetry support"
Shouldn't we have a separate thread for Devo protocol?

Edit:
follow up -support-for-walkera-telemetry

Indigo wrote: @Thomas.Heiss - If you have the time (this weekend maybe?), can you please test the latest - see other thread. Thank you.

Spektrum DSMx testing deviation-devo10-v4.0.1-c0b203a_v2.zip (Indigo nightly):

Found bugs:
- temp displayed as "18186C" when no temperature sensor is added to TM1000 (AR6210, AR600, AR8000)
- Bat: displays 4.94V / 5.05V
- RxV: displays 12.2V
-> Bat + RxV is switched on TM1000 and connected RX (possibily used to be working on 92e1705 as I can remember)

- fades/frames: will have to further detail test with multiple Spektrum RX:
-> previous collected and shown "fades" (e.g A/L) where erased to 0 after connection was re-established
-> DX8 always showed the last collected fades from RX, even telemetry link has been gone, never dropped fades (e.g AR6210 fades 255)
-> NOT_CONNECTED black inverted fields (B/R) on Devo10 on AR8000 get activated and display numbers once HF-link is weak / lost (framelosses/holds) (as described in previous posts in this thread)
-> A/L on AR8000 still show various fast changing / blending high 5-digit numbers (>9999) once HF-link is bad (framelosses) (as described in previous posts in this thread)

Greetings

Thomas