ersky9x - what next?

[Rob Thomson]

Hello,

Firstly... let me start by saying. I do not want to kick up a storm. I do not want to nock on anyones hard work. However I think this may cause a few upsets. Please dont take them that way as am simply pointing out some general observations.

I feel that with ersky9x it is probably worth pausing for thought on how to take things forward. The issue I see is that while the ersky9x firmware is working perfectly; functionaly; interface wise etc.. it is still er9x. No problem.

But...

This new board allows us soo much more functionality and power it seems to me that we need to remove the limitations that are in place that we have inherited from er9x. Otherwise whats the point?

The time to do this is now. Before too many people are using the system, and before it becomes a problem to change things.

My thoughts are:

- audio functionality (ditch current functions and pull over the work done my betrand on open9x that is much more capable)
- menu structure (re-work menus to organise data better - potential table of contents/tree structure etc)
- flight phases (the er9x method is a clever hack - but a hack it is. A better system is needed)
- aileron differential (the current er9x method uses far too many mixes for what could be a simple option)
- a complete code cleanup. The current code is full of commented out code; and generally could do with getting properly indented and cleaned up to a standard style format that we all use.

I am sure I had more to add to the list - but it escapes me at present.

Please dont anyone take me the wrong way on this. I love er9x, and I want ersky9x to be better. Just think we need to pause and think about some of these things before the board becomes standard and changes to structure and functionality become trickier!

Rob

[cre8tiveleo]

I've discussed this with someone else.. he knows who he is , and we both thought that a break away from er9x way and open9x is required to fully take advantage of what ersky has to offer, otherwise, it's a patch work of ideas that work , only because the processor can do it soooo much faster. Imagine if it was written for it.

I'm not worried about hurting feelings, not everyone who has a 9x radio will buy ersky, maybe they will, it's evolution, and it has to start evolving, and not worry about backward compatibility.

I love er9x and open9x too, just ersky deserves so much better...

SO in short.. I agree with you!

[gohsthb]

Just a thought I had while reading Rob's post above.
Aileron differential:
How about having 2 mix percentages? Similar to how the dual rate and expo works. There could be a check box to enable the feature on the mix page. So if you don't want to use it the you only need to change one percentage.

Good ideas Rob. I think you are correct, now is the time to pick the direction of the road Ersky9x is paving. It is too hard to turn the paver later. That said Er9x works perfectly well. It is a good base to get things working and for testing the new hardware. We just have to make sure we don't get stuck in the "we started doing it this way, and now too many people would have to change. So we can't change how it works" mentality. Ersky9x is a different more powerful animal. It will be fun to see what it can do.
-Gohst

[MikeB]

From my point of view, I'm still getting the hardware tested and drivers written, SD card and AtTiny167 co-processor are still to do. We also need to decide on a definitive solution to the (possible) regulator overheating problem (I have some testing planned for the weekend).

I must have got the drivers right, because Bertrand seems to have been able to grab them for open9x!

For the future, I think changes can be divided into two types:
a) Those that improve on the current implementation.
b) Those that change the functionality significantly.

a) might include better audio
b) might include a completely new mixer arrangement.

We have enough memory, processing power, and EEPROM storage allow 'old' and 'new' functionality, on a per model basis, to both be present, and be user selectable. For example, we could implement some new flight phase functionality, add the storage to the end of the model structure, but then select between using that, and using the existing version. Existing models would still load and work, but could be changed to the new method if required. Users migrating from er9x would find it familiar, but improved functionality is available at the 'touch of a button' (or two!).
The same applies to the mixers, we could have a new mixer array, with improved functionality, and maybe more of them, and still not get near the 4K bytes available for each model (and that is with no compression for storage).

In due course, very old functionality could be dropped.
This way, hopefully, we can realise the potential of the board, without needing a dedicated team of full time programmers to do the work. At present, there only a very small number of people doing the programming. We could do with a few more. I have done a lot of learning about the processor, and writing low-level drivers over the last 3-4 months, so , of course, I took the quick way of porting the code I knew fairly well to get a working system. Initially I even emulated the 2K EEPROM of er9x as a starting point, but have now written i different way of handling the EEPROM storage, even that may not be the best, but at least it gives us 4K for each model instead of 2K for 16 models.

So I would say, improve things where easily possible, or add in rather like the FrSky internal alarm, there is an option to enable it, that disables the original method, then you get the new menu options and functionality. Clearly, with the ersky board, the audio is likely to be always there, so you would probably now always use the internal alarms, where on er9x, if you didn't do the audio mod, you still needed the beeps.

Mike.

[Rob Thomson]

Mike... I think the route you took is 100% correct. It was the fastest way to proove everything worked.

Moving forward... I can see your thought process - and this will work. However surely a better option would be to start with a clean slate.. and allow a desktop tool like companion to do the conversion of the model settings? Most people converting to the sky board will be happy to convert and resetup this way.

What is the possibility of all the core developers from the different firmwares working happily together to make a new sky9x fimware?

Take either er9sky9x or opensky9x as a base - doesnt bother me.. and work from there? Strip out the unnecessary and add the new etc..?

Either way.. doesnt bother me which way it goes. Just wanted to raise the question

[MikeB]

A clean slate might be nice, but I am concerned at the amount of work needed to do it that way. The way I proposed allows one or two things to be implemented, tested and become available to users without a large effort. If you require a desktop tool to be changed to allow you changes to be used, then you are doubling the amount of work needed to get something done. I know I sometimes think, that would be useful, but I shall have to change eepe as well, and then update the manual to make sure users know what to do. Then I think no I haven't got the time for that, so it doesn't get done.
If we were full time, paid, programmers you could possibly go the clean slate route, but even then it would take quite a bit of time to do.
I reckon users will be happy to have the enhancements as they come, and not all users want/need all the enhancements. For example, I understand why some users like flight phases, but others don't need them at all.

Also, you are relying on Companion, or another tool, keeping up with changes to the main software, this can put pressure on those maintaining it to do so. We all do this because we enjoy doing it. Once it becomes 'must do that because' it can lose the enjoyment and just become a job. Then it can easily become, 'don't want to do this anymore'.

If someone wants to do a major re-write then go ahead, I don't think I've got the time. I should, however have time for significant enhancements. Then, er9x itself needs some updating, I'm still trying to find time to get an audio board mod sorted for it. We shouldn't get carried away with just ersky9x, and ignore er9x, there are a lot of users out there, and they won't all upgrade to the 'sky board.

Mike.

[Rob Thomson]

You are probably right.

My thoughts at present are that the open9x firmware has actually already done 99% of the cleanup required. (much of what I have mentioned in my first post)

Long term however... Open9x will be held back by the fact that it is trying to maintain one firmware across all platforms. This is great - but will never realise the full potential of the board.

The solution may be to pull the open9x core as a base and strip out all the none ersky9x code base. That would shrink the code and result in a good base to grow from?

Regardless - it is a large job!

Bertrand... What are your thoughts on that idea?





Sent from my iPhone using Tapatalk

[Flaps 30]

My concern would be from another direction concerning the future of this project. That would be the availability of the 9X hardware/case. The rumoured arrival of the Turnigy 9XR might prove that the 9X is coming to an end. For now it seems that getting hold of 9X's isn't improving and the price keeps going up much to the frustration of those that are hoping to get on board.

Bearing the above in mind. I am wondering what the thoughts concerning a case that could be produced that works for us, in that it has all the holes for the SD card and other additions that the 9X case doesn't offer. Of course that would also start another round of talks as to where controls should be, do we have sliders or knobs for the other 'analogue' channels, and where the display goes. You can bet that whatever is done, someone isn't going to be happy.

Somewhere in the loft I have an old Skyleader TX. Perhaps that could be modified to suit the new card.

[Clivew]

Mike, without your input here, software and hardware related, I for one would be totally lost!
I think your approach is the right one.
You have my unqualfied support for fwiw.
Thanykou for all you've done and hope you will continue to do for us.
It is much appreciated

A clean slate might be nice, but I am concerned at the amount of work needed to do it that way. The way I proposed allows one or two things to be implemented, tested and become available to users without a large effort. If you require a desktop tool to be changed to allow you changes to be used, then you are doubling the amount of work needed to get something done. I know I sometimes think, that would be useful, but I shall have to change eepe as well, and then update the manual to make sure users know what to do. Then I think no I haven't got the time for that, so it doesn't get done.
If we were full time, paid, programmers you could possibly go the clean slate route, but even then it would take quite a bit of time to do.
I reckon users will be happy to have the enhancements as they come, and not all users want/need all the enhancements. For example, I understand why some users like flight phases, but others don't need them at all.

Also, you are relying on Companion, or another tool, keeping up with changes to the main software, this can put pressure on those maintaining it to do so. We all do this because we enjoy doing it. Once it becomes 'must do that because' it can lose the enjoyment and just become a job. Then it can easily become, 'don't want to do this anymore'.

If someone wants to do a major re-write then go ahead, I don't think I've got the time. I should, however have time for significant enhancements. Then, er9x itself needs some updating, I'm still trying to find time to get an audio board mod sorted for it. We shouldn't get carried away with just ersky9x, and ignore er9x, there are a lot of users out there, and they won't all upgrade to the 'sky board.

Mike.

[Rob Thomson]

I think that what is key to realise.

I am not talking about writing everything from the ground up. That would be way too much work.

My primary point of post 1 in this thread is.

'Now is the time to remove some of the limitations imposed by the earlier hardware contraints'

I believe 99% of the code would remain the same. But ditch systems that are 'workarounds' and replace them with better solutions - ones that are not held back by the memory/firmware space issues.

Quite possibly - I am wrong; but to me it seems that the time to do this work is before too many people jump on board and compatability becomes an issue

Either way... I personally am not going to make a big fight about this - just wanted to point out the observations and lay down a thread for thoughts on the matter.

Rob

[bertrand35]

Rob, I am not sure that 99% of the cleanup is done in open9x Well ... I don't know. If Mike wants to take parts of open9x, I will be really pleased to have contributed. As previously said, without his work on the low layer thing of the ersky9x board, I would not have succeeded to have almost everything working so quickly ... As discussed in chat, my next priority will be the SD card.

Bertrand.

[jhsa]

Hi guys. Let me start by telling you what attracts me most on er9x, and I'll use only one word.. FREEDOM.
I think I don't need to remind anyone of how all this project started.. You just have to go to the RCG er9x thread and read the first few pages..
It's funny that this project started based on how free a radio could be, I mean all the free mixes, no channel order, etc.. I thought, wooow, I gotta buy one and change it to er9x.. the main goal was to make a great radio out of a not so good chinese radio, and keep it as cheap as possible.. ok, some mods require some soldering skills and not everybody have them.. but in one way or the other people managed to get things done.. many learnt to solder by doing the mods on the 9x.. Then, along came the smartiepaRts, a solderless option to flash the tx.. wonderful.. it helpped a lot of people..
Recentlly I've noticed that people have been asking for features that are available on many normal radios.. stuff like flight phases, differential, etc..
One day we will have to change the name of the fw to FutabaJrSpektrumHitec9X

Now, we have a new board for this cheap radio. very powerfull, no doubts about it, but the only relationship as far as I can see with the good old 9x project is that it fits inside the tx's case and it is compatible with the LCD, and the plugs from the pots and switches..
I don't think we can say that we are modding the 9x anymore.. that's a complete new brain, heart and soul.. well the soul (firmware) is still the same for now but I think it should be something specially made for the new board..
I think it should be a complete different project, separated from the one called er9x..
I know everybody is excited with this new project, me too (shame that I can't afford it at the moment) but I would like the cheap er9x project to continue and not to be left in the corner,
So, the $1000 question is:

What is going to happen to the er9x project as we know it? is it the end of it? I really hope not...

Thank you Mike for all the excellent work on this firmware and for showing some interest in continuing with the original project. I just don't know how you can possibly find time for all this stuff..
Also thank you for not giving up on the audio mod project for the er9x
It's difficult to explain in English what I mean as it isn't my mother language, sorry..
João

[bertrand35]

Jhsa, believe me or not, the Flight Phases feature adds freedom. And you keep the freedom not to use it
I even have no doubt that with this new ersky9x board, the implementation will be even more exhaustive that what we have succeeded to do on the stock board now.

Bertrand.

[Kilrah]

FutabaJrSpektrumHitec9X

Well, *9X indeed does everything those do, once added together, and much more conveniently



I'm personally also for continued development on the stock board. I love the freedom of the alternative *9x firmwares, and the different developers have already achieved something awesome there. The development of the ERSKY9x board is a great initiative too, but... looking at it now after a few weeks and flight hours with the stock board I personally can't really find an interest in it, expect maybe for one thing, which is model memories. As I'm now going for the 9X (2 of them) as only radios, I will definitely be tight with 16 models. But apart from that, all I need and use for everyday flying is already available on the stock board.
Yes ERSKY9x can have a BT module, potentially connected to a phone etc, but... IMO those are mainly "cool features" that require lots of development, are fun a couple of times, but don't get much use in the long run. Mostly because they require cooperation between several devices (the radio, the phone, the PC software, sensors in the model), which complicate the development a lot and force deciding to "lock" a specific set of features as it's just impossible to maintain the current level of freedom across 4 different projects that all need to be updated everytime something changes. *9x is so generic that anybody can do from the simplest to the most complicated set of mixes to meet the requirement of any model whatever how twisted, which is the whole interest. Missing something? Add a mix.
If you offer phone display, but that only works with android phones, requires you to have (insert name of non-existing bidirectional RF-module here) and not another, same for the sensors, lets you display a set of 6 parameters, then if you want to add one you have to reprogram the sensor board, adapt the RF protocol, add support to the 9X software, then recompile the phone application to add a display field, it becomes the complete opposite. The user can't do it himself, it has to go through 4 potentially different development teams that have to decide on how to do it, then implement it... long and complicated process. And then of course the next guy needed that other sensor, but can't as it would mean repeating the whole thing.
So, as usual when you start having a lot of "toys" (powerful hardware) to play with, it becomes more and more difficult to use it wisely, keeping the original "spirit" and not getting into just another radio that is full of single-use features.

Jhsa, flight phases are a little more specific indeed, but only just a little. You can use the mix fade in/out for about anything, that was one thing that was missing in ER9X IMO, whether implemented as flight phases or a parameter in the mix itself. But if you don't want them, maybe Bertrand can implement a model-specific hiding of screens... so for example in a Spektrum BNF plane, you can hide the heli, flight phases, curves and telemetry screens so you've got nothing out of context disturbing you

[jhsa]

ok, I do understand what you mean about the phases, the fade in/out was something we requested sometime ago on the er9x side if I remember well.. Can also understand a little bit the differential, due to the mixes being limited to 32, But would prefer not having it.. why not doing a v-tail, elevon, whatever mix ready available then?? I would go back to my graupner radio that has better pots and switches.. it can do all that..
The firmware was the reason I've changed to er9x.. I'm not saying that open 9x is not good.. no, no, no..
Very far from it.. it seems that it is excellent according to it's users.. I couldn't possible say such a thing because I didn't try it yet.. but I will definitely try it as soon as I have my other 9x..

My concern was/is: Are these projects going to end due to this board arrival? Well, I can only speak for myself and I can't afford it at the moment.. I think there are many people in the same situation.. As far as my needs are concerned, as soon as er9x/open9x have support for an audio board so It can insult my "Big Headed" club mates, I'm sattisfied..

I wouldn't mind seeing them both kind of merged.. But all this is just my opinion. Hands up to you guys that work on these firmwares..

P.S. Let's see what 9XR brings as well..

[Rob Thomson]

Something to remember with open9x.

It currently supports:

stock board
gruvin board
ersky9x board

So... I am not convinced support will drop for the stock board. All that happens currently is you have to choose which features you want - because not all fit into the tx's memory!

[jhsa]

Oh, I had someone posting a nice comment about er9x on one of my tricopter videos on you tube. This one about the strobe light.

http://www.youtube.com/watch?v=bQPCJlfYTh4

And here is the comment

"I took my er9x and bashed it with aluminum bat over and over bigest mistame i ever did was changing it to er9x after i flashed it the mixing becomes so complex that if your not a rocket sientist with life time experiance flying you will mever get the helicopyet to mix right in er9x"

Well I gotta say, this is not something you buy.. It's free If you're not willing to learn how to do the mods and also how to program your models, then this is not the right thing for you. Stay away from it and keep the stock firmware. and instead of bashing it, you could have just put the original FW back on it.. what a waste...

[Romolo]

Something to remember with open9x.

It currently supports:

stock board
gruvin board
ersky9x board

So... I am not convinced support will drop for the stock board. All that happens currently is you have to choose which features you want - because not all fit into the tx's memory!

I'm quoting Rob but I'm also answering to Jhsa,
Everybody may be sure, that nobody will abandon stock MB support,
Speaking for me and Bertrand not for sure, one of our biggest concerns is not to broken stock board support and to add as much features the remaining flash allows.
About added feature: with 5 flight phases 32 mix are really not so much, differential was an easy, logical way to save curves and mixers.
They are there, but nobody force anyone to use them if they are not needed, but in complex gliders they are useful, and as Rob says, open9x likes gliders (glider9x ?? )

[jhsa]

What is good for gliders, is also good for other planes.. The gliders always need more complex mixing..
it's good to know you will continue to develop on the stock board.. thanks for that..

João

[Reacher10]

Oh, I had someone posting a nice comment about er9x on one of my tricopter videos on you tube. This one about the strobe light.

And here is the comment

"I took my er9x and bashed it with aluminum bat over and over bigest mistame i ever did was changing it to er9x after i flashed it the mixing becomes so complex that if your not a rocket sientist with life time experiance flying you will mever get the helicopyet to mix right in er9x"

Well I gotta say, this is not something you buy.. It's free If you're not willing to learn how to do the mods and also how to program your models, then this is not the right thing for you. Stay away from it and keep the stock firmware. and instead of bashing it, you could have just put the original FW back on it.. what a waste...

I had to comment on his comment:

"Did you video tape it...there are sooo many excellent ways to bash a 9X to peices.... I ran one over with a pay loader but it was anti climactic. I will experiment with fire works when the 4th of July arrives...celebrate the freedom of blowing of 9x to smithereens. However...I will leave it with the stock formware...ER9 is just too good to waste."

[jhsa]

[erazz]

Change for change's sake is not something I personally peruse.

er9x, or any kind of FW is a bunch of compromises. The nice thing about open source is that you can choose what you want and you have a lot of options. While I agree that starting with a clean slate is great it's also impractical. Would you throw away all the work Thomas, Mike, Rob, Shadi and others (and I) have done? It's quite a lot of work....


I suggest to start off and spell out what is missing/wrong/should be changed in er9x. I'd also spell out what is right/good/should be kept.
That will at least help us see if there is a better way to go.

[Rob Thomson]

100% agreed. My original post never started out by saying chuck it all out. The point was to say... change firmware formats and settings now before it gets adopted and it is hard to change.

I would say..... (and may have become noticed already)... I have come to love the open9x approach. The extra configurability.. features etc.. are all there. And all wrapped in a 99% er9x format.

I see there being two options for the new board.

1 - strip out the none sky9x board features from open9x and use that as a base. (I already use this on my sky9x board - it is rock solid)
2 - pull in loads of the functionality from open9x/er9x and do something new based on the code already in play from the current ersky9x build.

Personally.. I like option 1. It will IMHO be a quicker route to getting the foundation right.

But... I am not the decision maker

[Kilrah]

IMO, open9x has a comprehensive and cleaner approach on everything it does. The consequence is that every "advanced" function takes a bit more code space than ER9X's simpler implementation, which means not everything can fit at the same time in the stock AVR's flash, leading to lots of compilation options and the need to make choices.

So, on the ersky board that can accomodate it, just compile everything in and turn the options that are "incompatible" or relate to user preference (frsky/nmea/ardupilot/jeti, potscroll/audio/haptic) into menu settings. Easy. Everything one could want and has been implemented to date is at hand. Pretty sure it's the cleanest AND more complete approach that could be done simply

Then, build upon that.

[Hexperience]

FWIW, all I really want is er9x on the ersky hardware. I can't solder for @#$ and that was my #1 reason for wanting ersky9x. I'm excited about the possiblites and want to see what happens, but er9x is great and I'm happy with it and eePe. I can't believe Mike got the drivers and base FW running so quickly and with very little errors from what I see in these forums. Fantastic work! Cheers!