You mean with the new FW?

You can use the original firmware. The problem was not caused by what I thought.

_________________
Spirit System developer

I've just tested the style. The problem occurs with both style mode 1 and 8.

I can confirm what Björn says.
It seems to me that the effect is more pronounced when there is more wind.

Hello

now it's getting a little strange!?

a few days ago, i flew with my Alien 600 and was able to confirm this 15 ° behavior.
but today I flew my Logo 600SE and tried to provoke this behavior, but it wasn't possible.

_________________
official hacker distributor USA
www.hackermotors.us

Yes it is strange.
I think it depends on the self learning mode, which was described before.
If you fly only this candle, it is better. If you fly much movements it forgets the optimization.
The problem is, that we need for F3C an absolutely reproducible flight pattern. So currently the Spirit can not be used
for competitions.

@Tomas: Have you tried it yourself and have you been able to reduce it yourself?
Maybe it also helps if you come to us and we try it together.

Bjoern

Hello,

in supplied firmware there is no learning algorithm enabled.
So it is very interesting.

We have several f3c pilots that never suffered from this issue. Some were attending in the biggest heli events. Also we cant reproduce it.

We are unfortunately very busy to focus on your issue, but I can recommend to verify if swashplate is level at the lowest and highest positions too.

_________________
Spirit System developer

Hello Tomas,

I aligned my swashplate with a level tool.
Furthermore, my helicopter flew through all figures before switching to the Spirit 2.
Only the FBL system was exchanged.
I hereby rule out a mechanical problem.

As I have already written, the problem occurs in a reproducible manner with high speed and a small radius before the vertical ascent.
On the other hand, if you fly the radius large and take pitch and elevator slowly back it goes.
In my opinion, that shouldn't matter and it always has to work reliably.

We currently know another user of the Spirit 2 (with another helicopter) who has the same problem.

We (Frieder and me) are not aware of any competition pilots in Germany who use a spirit.
Can you please tell us the names of the F3C pilots who fly a Spirit 2 for you? Maybe we can get in touch with these pilots and solve the problem.

I would be happy if we could somehow solve the problem and I am also happy to do more tests for you.

Best wishes
Bjoern

Hello Tomáš,

I checked my swashplate-levelling too and for my believe its quite good. So the issue might have another reason.
But I don‘t want to give up already because I really like the Spirit in principle. I‘m confident that there is a way to solve our issue, and I would be ready to do more tests for you if you have any suggestions what we could try…
Maybe you don‘t have that much work that way and we can make it perfect together.

Greetings,
Frieder

I've just watched the two videos that were uploaded earlier in the thread, and it's interesting.

I have a Spirit Pro on a Trex 700, and a Spirit 2 on a Goblin 700, and neither suffer in the same manner as shown - where what I might describe as a fast, tight elevator maneuver such as used to enter the initial part of a stall turn, or similar, results in a nose up/collective jump type behaviour, when the helicopter is vertical.

However, what both my helicopters do suffer from, but inconsistently and rarely, is a nose up/collective jump type behaviour - similar to that seen in the video, but when the helicopter is horizontal rather than vertical - and only when performing a fast, tight loop. Perhaps that is in some way related? Perhaps something to do with a fast, tight radius type maneuver?

My settings are somewhat different to those shared here, so I shall spend some time trying variations, once the weather allows.

The units are mounted entirely differently in terms of location on helicopter, and orientation of unit, and settings are similar, but different. I believe the sensors are somewhat different between the two units, too? If so, then it does start to sound like a hardware independent characteristic...