Beta75X Washout Tuning TIP - trunglp/PA-Readme GitHub Wiki

What’s new in Version 2?

Enable “I Term Relax” (a new feature in Betaflight 3.5). Leave the “I Term Relax” options at their defaults. This helps to reduce i-term blowouts!

Lower the Yaw P to 65 and increase Feed Forward (another BF 3.5 feature) to 96/96/100 for roll/pitch/yaw (respectively). This preserves the same responsive feel but reduces oscillations!

Added a note about how bent props make this problem much worse.

These changes have been incorporated in the CLI listing (below). Huge thanks to Patrick Clarke for these updated suggestions!

Introduction:

I love 2S whoops like the Beta75X. They already fly great in most circumstances, especially if you come to understand their strengths and limitations. The biggest struggles come when pulling out of long dives, or loops. Whoops will sometimes “wash out” in these conditions, meaning they lose their grip on the air and tip, spin, or wobble unexpectedly. As far as I can tell, this has always been an issue with whoops and the Beta75X is no exception. In fact, the combination of weight, speed, KV, and prop area really seem to bring it out. Lately, I have been using a lot of my free time to investigate this topic, and learn as much as I can. I believe there are some specific things that CAN be improved, but we need your help with testing and feedback!

Causes of Washouts:

It is difficult to explain in writing why whoops wash out. It would be easier to show you. I hope to make some videos about it in the future, but that will take time and I want you to have as much information as possible now. The first thing to understand is that washouts have more than one cause. In my experience, the two most common are:

Low RPM Motor Stalls - Micro motors have high KV ratings. They are great at spinning fast, but they are actually relatively weak at low end torque. They are especially weak when first starting up. You will be in trouble any time a motor stops, even for a fraction of a second. There are a lot of forces acting on these props during high speed maneuvers. If the RPMs ever drop too low, then the motor is in danger of a complete stall. When this happens, Betaflight will try to compensate, but the attempt will fail because the motor will not be able to spin back up fast enough. If you see the camera tip up and to the side, then I believe that this is the most likely cause. Don’t forget that active braking can slow a prop almost instantly, but if the RPMs drop too low, then it may not be able to recover as quickly. This problem can be mitigated by increasing the startup power and idle speed, but most of the burden falls on the pilot (see below).

I-Term Blow Out - There is another issue that happens when the PID values (and especially the i-term) escalate too high and get clipped by limits inside Betaflight.

This usually happens late in the maneuver while increasing throttle. For example, it might happen in a long arching dive, or near the end of a tall split-s maneuver. When it happens, it feels like everything is holding strong at first and then the quad just “lets go”. If the quad performs an unexpected yaw spin, then this is almost certainly the cause. To avoid this, we suggest maxing out the “pidsum” and “iterm” limits, reducing “antigravity”, and enabling “I Term Relax” (see blelow).

BLHeli Settings To Test:

Startup Power = 1.50

This setting affects how much power the ESC can use when first starting the motor. It’s the kick that gets it going. Increasing this value can help the motor recover more quickly from a stall. It can also help ensure that all your motors spin up consistently when you arm.

WARNING: Increasing this value also increases the severity of power spikes. Use these settings at your own risk! In some cases, it can trigger a “death roll” (quad spinning rapidly out of control). It will be most likely to happen when entering or exiting a snappy roll (especially with very high rates). If you experience anything like that, then decrease this value and let us know.

Demag Compensation = High

Helps prevent desyncs, death spirals, and the like. Setting this value to “high” totally fixed it for me (despite the high “startup power”).

Motor Timing = Medium High

My understanding is the “Medium High” is a pretty good default for most quads. I believe the 75X ships with this value set to “High”. I have no concrete evidence as to which is better. But “Medium High” is what I have been running lately, so please use it for your testing as well.

Betaflight Settings To Test:

Copy and paste the following commands to your Betaflight CLI window. Don’t forget to save and reboot.

set anti_gravity_mode = SMOOTH
set anti_gravity_threshold = 250
set anti_gravity_gain = 2000
set p_pitch = 65
set i_pitch = 80
set d_pitch = 50
set f_pitch = 96
set p_roll = 65
set i_roll = 80
set d_roll = 50
set f_roll = 96
set p_yaw = 65
set i_yaw = 75
set d_yaw = 0
set f_yaw = 100
set iterm_relax = RP
set vbat_pid_gain = ON
set iterm_limit = 500
set pidsum_limit = 1000
set pidsum_limit_yaw = 1000
set runaway_takeoff_prevention = ON
set osd_throttle_pos = 2401
save

Betaflight Notes:

Motor Protocol = DSHOT300 This one is up for debate. DSHOT600 also works well. There is some evidence that DSHOT300 might be a bit more reliable in extreme washout conditions, but the evidence is inconclusive. More testing and feedback needed.
MOTOR_STOP = OFF Never ever ever let your props stop spinning in the air. Ever!
Motor Idle = 6.5% Increasing the idle speed helps to prevent low RPM stalls. If you continue to have problems, you could try increasing it even more. Just be aware that it could have undesired side effects, like making the quad feel too “floaty”
Antigravity Gain = 2x This feature helps the quad to hold its orientation when you rapidly increase the throttle. That’s normally a good thing, but in this case, it was causing the i-term to blow out. We had to dial it way back. This is the reason the quad would wash out as you would increase the throttle at the end of a maneuver.
PIDs The P terms have been reduced, and replaced with Feed Forward. This helps to reduce oscillations and produces a smoother overall feel, while preserving responsive controls. The other values have also been tweaked. Please try it out and give feedback, then tune it to your liking.
pidsum_limit, pidsum_limit_yaw, and iterm_limit These need to be totally maxed out, and it can only be done in the CLI window.
I-Term Relax This is another new feature in Betaflight 3.5. In larger quads, it helps to reduce bounce back after a snappy roll. In these micros, it also helps to prevent the i-term blowouts!

Props Out:

The motor direction has a significant impact on these washouts. The Beta75X comes pre-configured for “Props Out”, meaning that the front props spin out away from the camera (the opposite of the Betaflight default). That’s a good thing in general, but there have been reports that switching back to “Props In” significantly reduces the yaw washouts experienced during dives. That has been my experience as well. However, switching back to “Props In” causes a different set of problems! With “Props In” you tend to get washouts during high speed turns. When it happens, the camera will suddenly tip up and to the side. The problem can be quite severe, especially since high speed turns are something you tend to do a lot. During early testing of the Beta75X, we tried it both ways and determined that “Props Out” significantly improved handling in the turns, and resulted in a better overall experience. That said, you are welcome to experiment with prop direction. Please share your results in the “2S POWER-WHOOP” group on Facebook.

I believe I can explain why whoops have problems in high speed turns and why switching to “Props Out” makes it so much better. It has to do with the relative speed of the motors and the airflow around the props. The yaw spin, however, is more difficult to explain in terms of physical forces. As far as I can tell, it is caused by problems in the Betaflight software. The PID controller becomes numerically unstable and the values get blown out. Can it be fixed through tuning alone, or is it just a bug? Hopefully we can continue to learn more about this issue, and continue to make improvements. If you are a Betaflight dev, then we could sure use your help!

Things You Can Do as a Pilot:

Use higher throttle input, and start increasing it sooner.
Never let throttle input drop too low when the quad is moving fast (including falling fast).
Avoid dropping into your own propwash. It is rarely the best option anyway.
If you need to pull out of a very fast dive, increase to high throttle BEFORE you pull up.

Other Factors That Might Affect Washouts:

Weight - Reducing the weight always helps. My original 2S whoop was 7g lighter than the Beta75X!
Prop Load - In theory, props with fewer blades or less pitch should be less likely to stall. That’s because they are easier for the motor to spin, and because you will naturally run them at higher RPMs. It’s a theory, but I have not had time to test it.
Prop Condition - We have observed that you are much more likely to get washouts if the quad is running rough because of a bad motor or damaged prop. Be sure to check those things if you ever experience an increase in washouts.
Battery Sag - If you do a huge move and your voltage sags way down, then the effectiveness of your motors will be reduced. Try to give your battery a few seconds to recover, or just get better batteries.

How To Test:

Enable AIR MODE on the Configuration tab, or on a switch. Please fly with AIR MODE turned ON for this test (even if you don’t normally fly that way).
Make sure your OSD shows both voltage and throttle input (throttle is included in the CLI commands above)
Fly the way you normally do, but keep these suggestions in mind.
Let us know if you are able to avoid washouts in situations that were problematic before.
Let us know if you experience any new problems (like death spirals)
Any other feedback is welcome. It will be best if accompanied by short DVR clips.