ERSKY9X Coding

[jhsa]

Mike, I think I might have found a little insect
This is intended to remind me that the motor is live (Throttle cut off) everytime the Throttle channel (CH3) is at idle, the THR switch is on (throttle active), every 20 seconds.

C.Switches:
SW1 v~=ofs CH3 -100 THR
SW2 Ntimer OFF 20 ON 1 SW1

Voice Sw:

VS6 SW2 ON 0143

Ok, this works but... Whille I was programming this I noticed that

SW1 v~=ofs CH3 -100 (no AND switch)

Turns OFF if I move the throttle stick even with the throttle cut active. Note that I am using CH3 and this channel is pulled to -100 by the sticky T-Cut..

On the output menus (bars) all works as it should.. It seems that the C.Switches are not recognizing that the channel is being pulled to -100 by the sticky T-cut.

João

EDIT: This is on ersky9x..

[jhsa]

Problem with the latest test eepskye from today. 9XR-PRO selected. Please go to safety switches and try to set a sticky T-Cut. When you try to select a switch there are some taranis buttons and all the other buttons like THR, RUD, etc, are missing..

João

[MikeB]

Your custom switch 'problem' above is 'normal'. The channel values used throughout are the values BEFORE limits and safety switches are applied. This has been the case for 'years'.

Currently limits and safety switches are applied to the output going to the servos only.

Mike.

[jhsa]

Ok.. But then:

sw1 v<ofs THR -100 is equal to:

sw1 v<ofs CH3 -100

which is not the same..

João

[MikeB]

Only if CH3 is set to +100% THR with no curves, switches, DR, offset etc.
It certainly was the same in openTx as well, I don't know if they have changed that either.
It's a long standing question as to whether it is correct or not. Currently, you can make a channel (internally) take on values much larger than +/-100, and this is sometimes useful. Always applying any limits and/or safety could upset this.
Consider:
CH12: +100% THR Offset(100%)
* +20%P1 Offset 80% (Late Offset)
CH03: +100% CH12 Offset -100%

The first mix on CH12 scales the throttle to be 0 to 200, always positive.
The second mix on CH12 produces a mix that goes from 60% to 100%, and multiplies by this, so the 200 gets scaled to 120 to 200.
The mix on CH03 then puts the offset back in.

If limits were applied to the CH12 value this would not work!

Mike.

[jhsa]

ok, but returning to my question and probably me being dumb, but what should I do for in that situation the custom switch to react to the real value of channel 3? because on
sw1 v~=ofs CH3 -100
if a safety sw holds ch3 at -100, but the custom switch turns off when moving the throttle stick, then the custom switch is not doing what I told it to do, which is "stay ON" every time CH3 is at -100%
In this case it's not doing what is supposed to do, which is "watch CH3" it is watching the throttle stick instead..

João

[jhsa]

I just checked the latest version of companion.. It does have the same problem

The custom switch is supposed to be ON when CH3 is forced at -100% by the safety switch.. But it doesn't. also it follows the throttle stick instead of CH3's output.. and on the custom switch I selected CH3 not the throttle stick..
So, the Custom switch, or now recently called logical switch it's not reading the output of the channel as I asked it to when I selected CH3.. As far as I'm concerned we also have an insect here

João

[Kilrah]

As Mike said this is intended, and is how the firmwares have always worked.
Safety switches and limits are NOT taken into account when using the CHx sources. Think of them as a clamp that is external to the radio, which would not know if it is there or not.

You can use a Replace mix instead if you want the "clamp" to be included in the calculations.

[ReSt]

That was the reason why I modified the working diagram last year

Reinhard

[jhsa]

Ok, it is the way it works. But that doesn't mean it is correct.

[jhsa]

CH3 100% THR
R -100% FULL Switch (THR)

CH6 100% THR Switch (THR)

Safety Switch:
CH3 X !THR -100 (this is a sticky thr cut)

Custom sw.
CS1 V<ofs CH3 -98

When I flick the throttle switch I move the thr stick to control channel 6, but I want to keep monitoring that CH3 is at -100. That won't happen because the custom switch will now detect channel 6 instead even if it is set to monitor CH3. it will follow the stick. I think the problem here are the safety switches. try it, program it in eepe and companion and see it for yourself..
So if you guys think this is right, please give me an alternative that does what I want to do..

João

[jhsa]

looking again at the diagram, thank you ReSt, I see where the problem is.. the C switches don't read the safety switches.. Actually nothing is able to read the real output so trying to read a real channel value is pretty much useless..
Could an option be implemented (like the late offset thing) so the switches could read the real channel output, just before it is sent out?

João

[Kilrah]

It's not useless, just things are done differently. In the particular case you have above you simply can't use the sticky throttle cut switch, you have to do it yourself like before using the custom switch and replace mix method.

The current system has a big advantage as you can take decisions on the "logical", intended control value (mixer output). Say you want a custom switch when your ailerons are above 10%, you just use a CS with Ail > 10%. If limits etc were taken into account, you'd need to set your CS with something like Ail > 8.4%, because of the limit. And if you changed the limit or subtrim after adjuting something on the model, you'd have to go back edit your CS with pretty much no way to calculate what you need to put there.

[jhsa]

There you go, some nice feature to implement then. Call it "Late Output" or something similar..

For me that would be more intuitive than having to go back to er9x pre-history and program a hand full of mixes and c.switches again to have a sticky throttle cut, which I have now by just creating one single line..
Ok, In opentx it doesn't matter because as far as I know there isn't a sticky thr cut feature anyway.. you still have to do it the old way..

[MikeB]

I'm thinking about it! I'll need a little time though.

Mike.

[ReSt]

João
exactly the problem I mentioned here
viewtopic.php?f=5&t=1243

Reinhard

btw.
have you ever discovered it here (How it works) ?
http://openrcforums.com/wiki/index.php/Er9x_user_guide

[jhsa]

Yeah, I saw the diagram you posted. But that is not the point. As I demonstrated above I it's not possible to monitor the real output from a channel. The diagram just helped me to find where the problem is.

João

[PNaz]

Mike, I still have problems with latest test eepskye and Taranis. When an eeprom is read from the transmitter the AND switch in Switches is incremented by one if it is a custom switch. CS1 becomes CS2, etc. Physical switches are not affected. SA- stays as SA-. Burning the eeprom back to the transmitter with no changes made returns the custom switches back to the original eeprom values. CS2 in eepskye becomes CS1 in the transmitter again.

Paul

[jhsa]

Problem with the latest test eepskye from today. 9XR-PRO selected. Please go to safety switches and try to set a sticky T-Cut. When you try to select a switch there are some taranis buttons and all the other buttons like THR, RUD, etc, are missing..

João

In case you didn't see it.. we should be posting about eepskye in it's own thread in my opinion

João

[MikeB]

PNaz: OK, got the fix!
On the SKY and 'PRO boards, to keep compatibility with older code, the TRN switch is moved to the end of the lis for AND switches. I had SH doing the same thing on Taranis in eepskye, while it was in its 'correct' place on the radio.

Mike.

[MikeB]

Initial solution:
I'm adding O1 to O24 as sources for the GVARs. These are the actual channel output values AFTER safety switches and limits.
So you will be able to set a GVAR to the channel of interest, and use the GVAR instead of the channel in a custom switch.

I may look to add O1 to O24 in other places later, not replacing CH1 to CH24. The problem is if I add them to an existing list that includes telemetry values, then I want/need to add more telemetry items, the list could become (more) fragmented and difficult to follow. For this, I will probably need to do the same sort of mapping I'm now doing on the switches. This will take more time to implement, so for now, just use a GVAR.

Mike.

[ReSt]

Wouldn't it be possible to process the safety switch as the first step and bypass all other mixes for that channel if the safety switch is active?

Reinhard

[MikeB]

It should be possible, although that would still mean that if the channel output is used as an input to another mix, the safety switch value would override the mixes.
This is one of those situations where 'it's been done this way' for a long time, so any change needs to be carefully thought through.

It does seem like if you apply a safety switch to a channel, then it probably isn't being used as a mix source elsewhere.

Limits are clearly a different thing, and apply to the actual output.

Mike.

[jhsa]

Mike, wouldn't it be possible for the C.switch to just "Verify" how the output is? you don't have to do anything with that value.. just monitor it, nothing else.. then the C.switch would only confirm if the condition is true or not..

Again, this is my logic

João

[ReSt]

Next idea

THIS is, just like having all the preceeding mixes on a virtual channel, but without needing the virtual channel, so saves one channels.

THIS is the value of ALL the mixes on a channel that come before it,

THIS (on ER9x) in the moment seems to work only with slow
If it would also take switches into account, then one could use the THIS like a safety switch

e.g. R 100% THIS SWITCH(anyswitch)

as the last mix of the motor channel and the channel value would correspond with the channel output

Reinhard

[MikeB]

1. THIS does take notice of a switch, I just checked it on both er9x and eepe.
2. THIS is still processed before any safety switches or limits.
3. R 100% THIS SWITCH(anyswitch), I think doesn't change anything. If the switch is OFF, you get the channel output, if ON, the channel output is replaced by the channel output!

Mike.

[jhsa]

as I said before the custom switches just need to read the channel.. it doesn't need to do anything with it. At least it is how I see it..

João

[MikeB]

The problem is the channel values, before safety switches and limits are applied, can exceed +/- 100% (or 125%). These are the values currently used by the custom switches. It may well be, particularly for v1>=v2 type tests, that such values are required. This means we can't just use the channel values after safety and limits as these will be restricted to +/- 100% (or 125%).

I have the O1 to O24 sources to GVARs working on the radio, just need to finish updating eepskye to match.
This provides the functionality needed to test the actual outputs.

Mike.

[jhsa]

thanks Mike. Can't wait to testing it..

Today I went to the field.. On my electric models, I use the trainer switch to make the radio tell me the pack voltage when I want to know it.. Just flick it and the radio speaks all the stuff you programmed.. very nice..
Normally I have the timer set to 7 minutes so I can fly twice with one battery. But the weather was so nice that I decided to fly the complete pack. Of course the timer went to zero and i had to reset it. that was when the idea came to my mind.. What if we could use short and long presses on the trainer switch? (not sure about any other)
for example,a short press would do something like speaking the telemetry, and a long press would do something else like resetting the timer. Double the possibilities. what do you think?

João

João

[MikeB]

Just posted test versions of ersky9x and eepskye.

PNaz: The eepskye should fix the custom switch bug you reported.
João: ersky9x now allows O1-O24 as inputs to the GVARs.

Mike.