Dear fellows and developers.
I've been in copter use/testing for a long time (paparazzi, open pilot, dji, some multiwii clones own developed, px4, apm) and I have my own minor understanding of what is going on.
Nowdays APM 2.* controller series become very popular and even gain fair name among copter enthusiasts and professionals. Opposing to proprietary and "hardcoded" DJI systems or "undergrown" MultiWii products, APM platform was offering what you could call "value for money".
It has GPS, Baro, IMU, mag support with virtually endless tuning potential.
Right now it's beeing replaced by Pixhawk and PX4 systems (I own full set of PX4 hardware, so don't think that this post is because I have no means to migrate to other platform), which have great STM32 architecture and are not limited by arduino memory size.
I personally like my PX4 board, but last two weeks I got chance to meet APM 2.6 based aircraft (in fact I built it from scratch). I'll be honest, it's not OpenPilot, when you can set up aircraft in 3 clicks and 2 hours time, but if you read manual and follow everything that is written, the result will excel your expectations.
On fairly cheap APM 2.6 (original, not clone) board I built FPV/Video GPS/mag/baro assisted quad with RSSI, Voltage control (by the way, Multirotor mania (yeah, bad guys) didn't ship APM power module to me, so I had to find alternative "resistor divider" solution for voltage monitoring). On top of that I have OSD that receives all data from controller via MavLink which makes driving UAV pleasant and convenient. There are not many systems that can offer such functionality, and especially for that price.
It would be very unwise to kill everything now.
Bottom line:
I want to ask developers not to abandon APM controller line and not to close development. There are too many people are using this autopilots and rely on you.
APM users, if you want to see further development of the board, please vote.
Here is mine +1
With regards to comunity, with respect to developers.
M.P.
Replies
Once again on this topic.
I don't really understand why can't APM have Z axis autotune. Well, maybe it have no resource for kalmann filter, but some simple features and some minor improvements could be done.
I agree with resume development of ArduCopter 3.2* branch for apm2.5.2 or 2.6 without new feature but just more stabilization . It is really good plateform for drones. It's so pity to drop it.
+1
Considering that the development of APM has been underwritten by 3DR it is unreasonable of them to even have to consider further development from a simple commercial perspective when they have stopped development of the APM 2.6 and just about everyone buy the cloned versions including the HK 2.7. And anyway the current version of AC 3.2 is really solid, mature platform. You could use it for years to come as is. It has been pushed as far as it can go hardware wise. Strikes me that it is about as good as an 8 bit controller can be right now. I will use my APM 2.6s for some time. I actually like the fact that it is not now being continually overhauled and is mature.
If you even look at DJI you see many pros just want stablilty and reliablity and that seems to be why so many still swear by tried and tested 8 bit Wookong FC instead of the newer A2 (that had teething issues). (Not a DJI user so that is what I read).
Just my 2 cents.
Agreed, but AFAIK all pixhawk products have open schematics (pixhawk, px4, apm) so it's a matter of time for HK to make clone of this "next generation" boards.
But it does provide one interesting speculation: At the moment even pixhawk (not saying about px4) firmware is considered under heavy development, and is in Beta state at best. It matches with my usage experience (but I have to admit they are making huge progress every release).
APM board was positioned as ready to use/release product. Why don't make the best of it? At least bring 3.2.* firmware to it's stable state.
Is 3.2 8-bit firmware not already at its stable state? I understood that it was only being touched to fix known bugs that turn up. I thought 3.2 was stable but maybe I am wrong on that. And yes Pixhawk is widely cloned already and available for $99. I use a VR uBrain - not a clone but a 32 bit board, but only 99 Euro.
Here's topic that concludes that the only way to get satisfactory performance of althold/loiter modes is to flash 3.1.5 stable firmware. And there are at least three similar topics only on this forum.
Well, as I remember you can get new Pixhawk as low as 99USD on ebay, but as I mentioned above, that's not the point.
I must confess that my APM 2.6s fly very well on 3.2. There are a lot of people successfully using 3.2 on APM 2.6s. Sometimes it is hard to nail these things down and figure out what is causing this. I have seen this issue referenced to in 3.2
I may be experiencing the same issue. I am having Alt Hold issues with a Mini APM on 3.2. It tends to vary its altitude by 3-4 meters or more in Alt Hold and it behaves quite differently to my other machines. It is not vibes and it cannot be tuned out. Packing even more foam around helps - in as much as it buffered the barometer somewhat but the issue is still there. I even purchased another board - same result. I have not tried flashing 3.15 though. Not sure it is related to these issues linked but one wonders if 3.2 does less averaging of the baro output and these sensors vary in performance from one to another. On my mini apms the baro performance on the bench is worse than my APM 2.6 clones.
APM2 is short on RAM and flash, and today's development goes in the direction of more redundancy, and better EKF, which APM does not even have sufficient processing power to run.
All the new code is still available for people to try to squeeze in, but there is no way it can take it all.
I understand, but it would be better to leave 3.2 firmware branch on some stable release, without critical bugs in latest frimware.