ERSKY9X Coding

[jhsa]

Ok guys, this is weird..
I was trying to find the problem. I remembered that I also updated the fw yesterday and the problem didn't stop. On startup if i moved a stick, the voice would stop playing but i could hear that static noise of the file being read but the volume is set to silent. Another problem I have noticed and I don't know it is related (will keep an eye on it) is that the "last menu" option in the main popup sometimes takes me to the "model Select" menu even if I was in another menu just before.

Just now, while i was trying to find the problem by changing some settings, the problem suddenly got worse. My radio wouldn't speak at all. Then i think I have rebooted it and it seems that it fixed itself.. Can't make it happen anymore. Can press buttons or move the sticks as much as I like and the radio will start as if nothing happened before
Unfortunately that makes it impossible for me to find the cause of the problem

Weird stuff sometimes.. Eeprom hiccup?

João

[MikeB]

On my 9Xtreme the channel order in DiagAna is RETA, which is what I expect.
On my AR9x channel order is RTEA

Please check your 9Xtreme again. I'm seeing RTEA on DiagAna on all platforms.

Mike.

[jhsa]

I think we should not mix channel order with inputs.. For a moment I got confused.

In the DiagAna menu I see, LH, LV, RV, RH on all my platforms for the analog inputs A1 to A4.. This shows stick to inputs assignments, not channels..

If you call it RTEA, this will change depending on the stick mode you have selected, right? if you select mode 4 you will see ATER because the SOURCES that we call R and A will be assigned to different ANALOG INPUTS/sticks..

João

[Daedalus66]

On my 9Xtreme the channel order in DiagAna is RETA, which is what I expect.
On my AR9x channel order is RTEA

Please check your 9Xtreme again. I'm seeing RTEA on DiagAna on all platforms.

Mike.

João is right. Don't talk RTEA but LH, LV, RV, RH.

Then it makes sense and I see I was confused by different stick modes on the two transmitters.

Forget I said it! No problem.

[willhac]

@Mike :
I'm testing the internal MULTI with the 9Xtreme, my setup is :
- internal MULTI with A7105 (Flysky modified)
- external MULTI with NRF24L01 & CC2500
I noticed that when I switch from a model using external module to a model using internal module the LED of internal MULTI blinks slowly (no valid serial).
Rebooting or switching twice between models using internal MULTI and serial becomes valid. (fixed LED or fast blink & fixed LED if autobind)
There is no problem with switching internal MULTI -> external MULTI.
Can you see the problem ?

[MikeB]

I can see the symptom, the internal output starts inverted. I'll search for the reason. While the output is a serial data stream, it is generated from a timer/counter.

Mike.

[willhac]

OK, thanks

[MikeB]

I think I have this fixed. It will be some time tomorrow before I can post a test version with the fix in it.

Mike.

[MikeB]

I've just posted a specific test version that, I think, fixes this problem.

Mike.

[willhac]

Yes I confirm it's OK, thanks again Mike

[jhsa]

Mike, I have a little situation here Not a real one, but it could very well be, as this setup is quite common on a glider.
I'm helping Nigel with some examples for the manual.

Ch3 100% P4 (My slider - controls the throttle)
Ch6 100% Thr (Throttle stick controls air brakes)

Safety switch programmed as sticky throttle cut on the throttle channel:

Ch3 X !THR -96

The sticky doesn't work as it should as it continues to follow the throttle stick and not the source of the channel the throttle cut is applied to.. It is even dangerous and I explain why.

Imagine neither P4 or the Thr stick are at minimum. P4 is for example at full.
Now you turn off the thr cut protection. Nothing happens. Bring your airbrakes (Thr stick) down, and the model will jump at full power.

The sticky throttle cut should follow the source of the channel it is applied to..

João

[MikeB]

I understand the situation.
We can't really use the mix source, as you may have more than one mix on the channel, so can't decide which source to use.
A sticky throttle cut may be configured (as was done before I added the 'X' option) using logical switches and a mix, and so can apply to any control, not just the throttle.
You can also use a custom check to make sure the pot is at minimum throttle.

I could possibly add an option to specify the throttle control source, for use by the safety switch 'X' option, but since this would also require configuring, it isn't really different to needing to configure a 'sticky' safety operation based on a pot.

The 'X' option is therefore a "throttle stick" sticky option, rather than a "throttle channel" or "throttle control" sticky option.

Mike.

[jhsa]

I understand the situation.
We can't really use the mix source, as you may have more than one mix on the channel, so can't decide which source to use.
A sticky throttle cut may be configured (as was done before I added the 'X' option) using logical switches and a mix, and so can apply to any control, not just the throttle.

Yes, but we have been saying everyone that we should use the sticky throttle safety switch instead..
The other template with the old sticky T-cut doesn't even exist anymore

You can also use a custom check to make sure the pot is at minimum throttle.

Ok, but this only works at startup and changing models?? Also the throttle stick will still control the sticky throttle cut..

I could possibly add an option to specify the throttle control source, for use by the safety switch 'X' option, but since this would also require configuring, it isn't really different to needing to configure a 'sticky' safety operation based on a pot.

This is an idea. Configure what input controls your throttle, and make the sticky safety switch follow it. this could work.
Default to the throttle stick, but allow to change for these special cases.

The 'X' option is therefore a "throttle stick" sticky option, rather than a "throttle channel" or "throttle control" sticky option.

Mike.

I agree this should be based on the channel instead of an input..
But your previous idea of configuring which input controls the throttle would also be a good option..

João

[jhsa]

Or what about instead of selecting X on the safety switches, we selected the input the sticky would follow??

João

[MikeB]

Something may be possible, regarding choosing the input at the safety switch. It may be a while before I can get round to looking into it though.

Mike.

[jhsa]

Thanks for looking into it Mike.
In my opinion this subject really needs to be addressed as I believe there is potential for some dangerous situation.

Thank you again

João

[MikeB]

I may not have mentioned this, but recent test versions include an option on the safety switches to choose the throttle source if safety type 'X' (sticky) is selected.

I'm also just adding an extra curve for the mixer. This is a 6-point curve and is mainly intended for use with a 6-pos switch, as it will map the 6 switch positions exactly to your required outputs.

Mike.

[jhsa]

I may not have mentioned this, but recent test versions include an option on the safety switches to choose the throttle source if safety type 'X' (sticky) is selected.

Mike, quick test and it seems to be working. Thank you, this is great.. Found a little problem though.
I tested it on my Ar9x radio where I have a slider installed as Pot4. The Sticky safety switch works as it should, but displays as HALF instead of P4.

Will check the curve later..

Thanks

João

[MikeB]

I've got the fix for the display of P4, seems OK.
While I'm looking at the curves, I've also added a second XY curve.

Mike.

[jhsa]

Mike, would it be possible to generate a voice alarm for the "Custom Check"?
Or maybe when the safety switch sticky condition is not met at startup or when loading a new model..
In other works, the "Throttle not at idle" alarm should follow the setting on the safety switch menu. When nothing is set it would default to the Thr stick obviously..

João

[MikeB]

I've posted new test versions of ersky9x and eepskye.
eepskye fixes a bug when reading the firmware, adds support for the model name voice and the XY-curve.
Both add a second XY-curve, a 6-point curve and support for individual cell voltages.

Mike.

[jhsa]

Downloading
Thank you..
João

[MikeB]

I'm looking at the possibility of adding more channels (currently 24) and more mixes (currently 48).
Increasing the number of channels involves quite a few changes as thing like lists of sources, where a channel is allowed, need to be considered. Increasing the number of mixes is fairly easy, and I already have most of the code written, just in comments!
I can't see needing more than 32 channels, so if/when I increase the channel limit that's what I shall use.
How many more mixes do we need?
Currently, there are an average of 2 per channel, so at least 16 more (total 64) for 32 channels seems reasonable to me, although I could add a few more than that.

Mike.

[jhsa]

I think it would be better to add more if you can.. There is no flash, eeprom or ram limitation, on these chips right? at least not yet?
Better to think about the future, we never know
Thanks
João

[RCHH]

Yes! 32 channels would be a perfect number.

[MikeB]

I'm updating the editing of the custom telemetry screens. I'm adding a display of the current value of each item, so you may see those sensors that are returning a value.

I think I'll actually release r219 before I add more channels and mixes.

Mike.

[jhsa]

Thanks Mike.. The telemetry menu in general really needs a "Face lift"

João

[bob195558]

Also consider, with the telemetry screen to remove the model picture, so as have more room for larger font
and/or more telemetry data to be displayed. This would be for erSky9x on Taranis display screen.

Bob B.

[MikeB]

OK, the Taranis custom telemetry screens will have 9 items you may select instead of 6.
If I make the font larger, some values won't fit properly anymore.

Mike.

[MikeB]

New test version(s) posted, changes:
Taranis sync PXX with XJT to reduce latency
Taranis allow 9 custom telemetry values per page
Add PXX option for only 8 channels - improves latency
DSM telemetry, add "Use 'A' as RSSI value"
Remove ASSAN protocol
Improve custom telemetry editing menu - displays values
Language updates

To explain about latency with an XJT in the Taranis. The XJT sends a "heartbeat" signal on pin 2 of the module connector. I now use this to time sending the PXX frame to the XJT to minimise the latency between this and the XJT sending the RF frame to the receiver.
Should you use pin 2 in the module bay for a wireless buddy box connection (CPPM or SBUS), the synchronisation is disabled.
Also, the firmware may send 16 channels to the XJT. It sends these in two frames 1-8 and 9-16, but only 1 frame is sent every 9mS. If you only need up to 8 channels, then setting the option to use 8 channels only, means the first 8 channels are sent in every frame.
I may look to see if the heartbeat from an external XJT module could be used on a SKY/AR9X/9XR-PRO, but it does need an unused input to the processor.

The ASSAN protocol was for a DSM module/telemetry development. The module and telemetry worked well, unfortunately I haven't heard from ASSAN for over a year so I'm removing the option. This is done by conditional compilation so can be put back in very easily if needed in the future, however, I'd now simply recommend flashing Orange modules to the MULTI protocol (I've even done one of the ASSAN modules).

When editing a custom teleemtry display, I now show the current value, so you may quickly see items that are being received. For A1 and A2, I also display the raw received value so you may see if a value is being received, but the scaling hasn't been set yet.

Recent Lemon DSM telemetry devices return a RSSI value as the 'A' parameter. There is therefore now an option to use this.

Mike.

Edit: Sources committed to 'next'.