APM 2.5

[salience]

Well there certainly seem to be a MULTITUDE of them, and so few of us traditionalists...
The main problem is that all the wikis and even the terminology in MP and elsewhere are geared for the multi-ists.

[salience]

Much improved. Changed Loiter PIDs thus

LOITER_LAT_D,0.2
LOITER_LAT_I,0.4
LOITER_LAT_IMAX,400
LOITER_LAT_P,0.8
LOITER_LON_D,0.2
LOITER_LON_I,0.4
LOITER_LON_IMAX,400
LOITER_LON_P,0.8

Now the rock'n'rollin is quite subdued, but still present. Flew quite well today in spite of some stiff gusts from the Mistral. Loiter stayed right in place. The Roll/Pitch graph still shows the R&P sine-wave trace, but the amplitude is now about ± 6, rather than ± 20 when the cycling led to runaway uncontrolability a few flights ago.

Next will try the Ps at 0.6 and Ds at 0.3 with Is staying at 0.4

Any observations? If I had to attribute a cause for my difficulties with BH & the comets aka HulaHoop aka toilet bowl I'd say it was the unusual weight and weight distribution of my TRex500 - with a heavy tail motor and lots of battery upfront to balance. It's not your classic TB with circling of a certain radius, but rather a stay-in-place gyration on both axes, as if the heli is trying to swing a HulaHoop. Sexy on certain birds but not airborne ones.

Btw, you have to change these from the complete parameters list or the new parameters tree. If you change in the extended tuning window only the LAT settings are changed. Thus you cant use the Ch6 slider method to adjust "live". A BUG in MP? Heavens...

[salience]

A QUESTION:
What do the RATE_FF parameters do?
I notice MP does not specify the range of possible values, why so?
and what wd they be?

[jolyboy]

Good to read you have made some progress mate.

Rate FF is the feed-forward term. Its essentially a direct proportional connection between your desired rate and the rate controller output without manipulation by the PID controller. It gives you a better feeling of control.

Rob will fill you in on the details but that's essentially it. He has done some fantastic work getting this to work better on our latest code

[jolyboy]

So its gyrating without moving in circles?

Can you post the latest log?

[salience]

Log and parameters attached. See what you can deduce, I'll post my own analysis tomorrow - too late to do so now...
Yes, the position holds well, but the heli gyrates in place.

[jolyboy]

Haven't looked at your log yet. According to your param file:

RATE_PIT_D,0.007
RATE_PIT_FF,0.01
RATE_PIT_I,0.2
RATE_PIT_IMAX,450
RATE_PIT_P,0.08
RATE_RLL_D,0.005
RATE_RLL_FF,0.01
RATE_RLL_I,0.2
RATE_RLL_IMAX,450
RATE_RLL_P,0.08

Your IMAX looks way too low. Try 4500 instead of 450.

Also, did you lower your rate_p down again? If so, I'd try keeping it there and increasing rate_i and see if you have better results.

How is using rate_d going for you?

[salience]

A couple of observations on the log:
Look at just the R&P at lines 17-24. At the beginning we are still in Loiter, with the typical quite regular sine wave output. Due to recent tweaks this is now down to an amplitude of ±5 and the resulting HHing remains small, more like a bit of bobbing mostly fore and aft, and the heli is stable in place, with a little random drifting of a meter or two. Awhile back the amplitude was ±20 and the HHing was gyrational and got uncontrollable quite quickly.
The tweaking consisted of - as I mention in above posts - changes to the Loiter PIDs as well as Rate PIDs. When I first tried using Rate D-terms awhile back, my P-terms were quite a bit higher and even small D-terms resulted in severe shaking. Biut now with P-terms @ .08 I have been increasing the D-terms slowly, but hard to tell what effect they might be having. I was hoping increased D-terms would damp down this strange sine-wave oscillation, because it seems to be present not only when engaging Loiter, but also in Stabilize and Alt Hold. See the log. In these latter modes, however, the oscillation is not regular as in Loiter, probably because I am constantly "correcting" it with stick inputs. Still, it is there and I suspect my aim should be to find its source.
GyrX and Y also show the sine-wave pattern, but I presume it is a consequence of the heli movement and not the cause.
MagX and Y ditto.
I've nary a clue what the NTUN traces mean, but there too we see the sine waves, but ONLY during Loiter. A clue, maybe?

Back later with more, when I go up to the shop.

[jolyboy]

You really should increase the IMAX values to 4500 from 450 though. Rob has also stated that low IMAX causes problems with loiter

I run 4500, so does rob. In fact, everyone should IMO.

[salience]

Will test that tomorrow. Maybe I'm mistaken, but I just DL'd the latest MP a few days ago, and yes, now IMAX range goes to 4500 - if I remember correctly, the previous version showed max range only to 450 but units were x10. I think that's how I changed to 450, believing it was 4500...

[R_Lefebvre]

In log files, and in the Full Parameters List, MP does not manipulate any values, including IMAX. If the log file says 450, then it's 450. Which is too low.

[jolyboy]

Any updates salience?

I've been busy learning some 3D so I can go nuts on this KDS Chase 360 I have sitting here I'm using a V977 BL micro heli to learn on. Man, this thing is fun. So far I've gone from basic circuits to hovering all directions, flips, rolls, inverted hovering, inverted circuits and now tick-tocks. I still crash very often (once every few flights), but its far less than before! Most crashes are from bogging or tail blowouts, but they happen less often now that collective management is getting easier.

[R_Lefebvre]

Salience, have you checked your swash plate leveling? Apologies if already stated.

Joly, so is that little heli any good? I hadn't heard of it, but it looks like it could be a fun thing for me for the winter. Is the tail motor brushless too?

[jolyboy]

Its pretty damn decent if you ask me. For the price its way better than my old mcpx v1. Its faster, more powerful, tail's more solid, more durable. Others in the same price range might be as good, but without a BL motor. The flight times are longer than my mcpx v1 flights were. These are around 5 mins going hard out, never hit LVC yet either.

The tail motor isn't brushless but it is larger than the mcpx's one. The main blades are the same size as the mcpx so it still has tail blowouts if you aren't careful, but its way better than the mcpx v1 was.

It has self levelling which is useless once you learn to fly in all orientations (you know this). Its actually annoying and no fun to fly with SL enabled, but is good for indoors I guess. Otherwise outdoors, SL is a hindrance.

Its durable. I've crashed >30 times, and only broken a main gear. It split in half. I bought the white main gears and have since never broken a part. Now I have a box of spares that I'm unsure I'll ever use hahaha. I only ever fly over grass and hit TH early which makes a difference to crash severity for sure.

There is one annoying thing about this heli which is a deal breaker for many. The radio protocol isn't compatible with any other remote other than the stock one, or a devo TX. So if you intend on flying it, make sure you get the combo with TX, not the BnF! I made this mistake because they do not tell you this on the website. If I had known I would have gone for a different heli.

[salience]

The wiki:
"
Overview

When RTL mode is selected, the copter will return to the home location. The copter will first rise to RTL_ALT before returning home or maintain the current altitude if the current altitude is higher than RTL_ALT. The default value for RTL_ALT is 15m.



This isn't very good for mountain flying! I'm at 950m altitude, and may well fly down into a valley or two from here, i.e., down to maybe even 600m. RTL failsafe will not save the bacon!


What's needed is an absolute RTL_ALT, in my case 950+m.

[jolyboy]

I don't fully understand what you are saying. The RTL_ALT is relative to where you are standing. So if you fly into a valley and engage RTL, it will rise in alt until it is 15m higher than where you are standing. As long as you don't stand in a valley and go behind a hill you should be OK right?

Selecting the right location for your launch pad is important as you can see.

So is your heli flying better?

[salience]

http://copter.ardupilot.com/wiki/rtl-mode/
The wiki is very badly worded then. It should say not
"The copter will first rise to RTL_ALT before returning home"
but rather
"The copter will first rise to RTL_ALT above the launch altitude before returning home"
or something like that.
Are you sure that if I take off at 900m, fly down into the valley at 600m, then switch to RTL with RTL_ALT set to 15m, the heli will then rise 315m before returning?

[jolyboy]

You're right, it is badly worded.

And yes, I'm sure -- that is, unless something has changed in the code which I highly doubt.

I've flown the same heli with the same settings (bar magnetic declination) at sea level, then the very next week on the top of a mountain at 2200m ASL without any RTL troubles. My most regular flying spot is 600m ASL. I have regularly flown at different locations with no RTL troubles whatsoever, and without needing to change RTL_ALT either.

Hope that helps mate

[salience]

OK then, I guess I can depend on failsafe RTL to take it up to Launch Altitude PLUS RTL_ALT no matter what the actual altitude at failsafe. Meanwhile. here's my NQRTFAA

[jolyboy]

!!!!!!!!!!!!
OMG that looks freaking awesome!!

FYI, looks like these ESCS have a built in RPM sensor:
http://www.banggood.com/Hobbywing-Pl...-p-921279.html

RE RTL, make some safe tests from short distances to start with. If it looks like its going to do wierd stuff when you engage RTL, just switch back to manual