[JustAPilot]

I FOUND THE PROBLEM

While testing, Motor1 has no issues. Only Motor2,3,4 have the "sync" issue.
I connect Motor4 to the output of Motor1 and Motor4 now runs fine with No issues.
The difference between output1 and output2,3,4 ? = GROUND
Only output1 has a Ground wire on the MiniREVO. Apparently my Afro30A ESCs do need that ground connection to work as expected. (servo tester has ground)

Sigh,..finally figured this out. Thanks for shouting some directions.

[JustAPilot]

The wind was calm today so I tried the AutoTune. The tuning went fine but the results are strange.
Without stick movement, the copter hovers fine. When I move the stick (roll/pitch) the copter oscillates.
When I looked at the stored autotune values in GCS, they are high and even red.  Integral(R,P,Y): 256, 294, 462   (picture attached)

I will use my default mushy settings again. And I will try what you suggested: "StabilizationSettingsBank1 -> YawPI -> P.  Cut it in half for a test"
Hopefully this will remove that aggressive correction when I let go of the stick.

[TheOtherCliff]

Sounds like a bad tune, which does happen sometimes.

If you would post System -> Settings -> SystemIdentSettings
I might be able to tell you what number there looks wrong.

After that, I would run AutoTune again. 

[JustAPilot]

I've attached a screenshot of the SystemIdentSettings.  (after the bad tuning I did reset bank3 to defaults)

Please have a look at these two videos?  and listen to the sudden boost at the moment I start moving the stick (roll or pitch or yaw) and also at the moment of centering the stick. I seems like at the moment of transitioning from/to another "subroutine" there is this twitch/boost.
This is in PH+GPSAssist.

youtu.be/rQs-mWN-C3E
youtu.be/cuvGy3NN7-8

The copter loses altitude when I bank. Tips?

Thanks in advance

[TheOtherCliff]

(Your screen shot) Those are SIS default values.  You must plug in the FC to see the current values.  They are only stored on the FC.  But I don't see a problem that needs fixing in the videos.

Those sudden boosts are perfectly normal.  If all those twitches were caused by transmitter stick movements, I would say that it was flying very well.

Let me give you a yaw example.  Yaw is a "rate control" which means the rotation rate is a simple proportion of where the stick is.  In the center you have 0 degrees rotation per second.  At full stick you may have configured 300 degrees per second.  So at half stick right, you get a right rotation of 150 degrees per second.  What happens if you move the stick from 0 to 150 degrees per second "instantly".  You have commanded it to accelerate from 0 degrees per second to 150 degrees per second instantly so it does the best it can to get it to 150 as quickly as possible.  Once it gets to 150, the motors slow down because it takes much much less force to hold it at 150 rotation rate.

This is just like a car.  You hit "resume" on the cruise control and so you have commanded it to get to 110 quickly, but once you get to that speed, it takes a much smaller amount of power to keep it there.

If you don't like the revving up, you can reduce it, but not eliminate it.  This will make the quad smoother, but slower to react to your commands.  Personally, I like to be able to tell it to do something quickly if I want, and just move the sticks slower if I want it slower.

One way to slow it / smooth it is to reduce "Maximum Rate (all modes)" (I forget the exact name) toward the top of the Stabilization page (bank 1, 2, or 3) Advanced tab.

In GPS (VelocityRoam) mode, the roll and pitch sticks command horizontal speed.  Say that you have set 8 m/s configured as max speed.  Instantly moving the stick to 1/2 stick is asking for it to instantly be moving 4 m/s.  It tries to instantly bank the maximum configured amount in order to accelerate horizontally as quickly as possible.  As it gets to 4 m/s it levels back out to almost level, just enough to maintain 4 m/s.

[JustAPilot]

The SIS values can't be read via OPLink? I was connected to the fc via OPLink. I am at this moment too to double check and those values are greyed out. (other values are not greyed out).  Maybe this is the problem?

The sudden boosts occur without moving the stick a lot. Even if you extremely slowly move the stick just a tiny little bit, it will do that sudden boost. And also when going back to center extremely slowly with the stick, you have that sudden boost. It does this only when in PH mode.   
In attitude mode this does not happen at all.

[TheOtherCliff]

Ahh yes, SIS doesn't show up if AutoTune module is disabled.  AT module gets enabled if you have AT on your FM switch or if you manually enable it in System->Settings->HWSettings->OptionalModules

The sudden boosts occur without moving the stick a lot. Even if you extremely slowly move the stick just a tiny little bit, it will do that sudden boost. And also when going back to center extremely slowly with the stick, you have that sudden boost. It does this only when in PH mode.   
In attitude mode this does not happen at all.

Does full yaw make it spin very fast (top speed, not acceleration)?  How fast would you guess?  If so, reduce the Stabilization->Advanced->bank1.2.3->MaxRateLimit(and RateMode)->Yaw

For yaw, one difference may be that in Attitude mode, the yaw is actually in Rate mode, but in PH/VR it is Attitude mode (I haven't looked at that code to verify).  If that is the case then reducing System->Settings->StabilizationBank1,2,3   YawPI->Kp  or  YawMax might help.

Because of the way it works, yaw control is fairly ineffective, so it takes a lot of motor RPM change to make a small amount of yaw force.  To make a yaw, it say reduces the NW and SE motors and increases the NE and SW motors by a fairly large amount.

Again, to me those videos looked like good examples of a GPS quad working well.

[TheOtherCliff]

I just noticed that you wrote PH+GPSAssist.

I didn't think that would even allow you to arm that way.  You shouldn't use GPSAssist with any GPS mode.  GPSAssist is meant to give you a temporary PH when you release the sticks when you are flying a non-GPS mode.

Maybe you meant AH AltitudeHold.  In that case let me suggest you use AltitudeVario instead of AH.  Center throttle stick is the same as AH and there is deadband there so it doesn't even drift (because of the stick anyway).

What mode were you flying when you didn't like the roll/pitch?  Again, in some modes the stick is commanding speed and any quick stick motion is commanding basically an infinite acceleration to get from 0 to some speed very quickly.  To do that, it must bank steeply and quickly.

[JustAPilot]

Sorry.. I meant VelocityRoam+GPSAssist.  I used PH a while till I was confident that works and then I start using VelocityRoam instead of PH.
Now I'm flying VelocityRoam+GPSAssist (see screenshot)

VelocityRoam is also a GPS mode I should not use with GPSAssist ?  The system allows this. If I'm not incorrect I even get a config error if I do not select GPSAssist with VelocityRoam. I have to check this

I fly (1)Attitude, (2)VelocityRoam+GPSAssist and (3)RTB+Autolanding. Only 3 flight modes.
The banking angle and speed is fine in all modes.
Only in VelocityRoam+GPSAssist there is this sudden boost when the stick is moved from or to center, no matter how slow and how little you move the stick.

youtu.be/rQs-mWN-C3E   @ 0:18   here I move the yaw just a tiny little and let go right away. The time between the two boost is so small that they seem to amplify each other. (In the other video I show the same effect on the roll axis)

The boost occurs at the transition from CenterStick to non-CenterStick and vice-versa. The mode used is VelocityRoam+GPSAssist.

I don't know what else to try. 

[TheOtherCliff]

Sorry, I find it hard to see how even the tiniest amount of yaw the slowest possible would jump. 

I gave you several options for turning this down.  How did they help?

It looks like you have negative expo set somewhere in your yaw.  Transmitter?  FC setup?

[JustAPilot]

This is hard to believe and explain and to see on video. The copter flies fine. That jump is just unnatural. I will figure it out.
I have tried everything you suggested.
One thing I will try on Monday is to select None for assisted mode of VR. I just checked and it does not say Config Error. But the system allows both PH and VR with GPS assist.

From your reaction I understand that this combination is not common. Maybe here there is something "double" going on and is also the reason for that "boost".
I will try the following tests:
1. set None for assisted mode of the VR.
2. Tune quad in Rate to perfection and then tune in Attitude.

As I said before, I have the exact same quad with OmnibusF4+Betaflight. Everything else is exactly the same (except the props) The props for both quads is 8x4.5 but the Betaflight copter has bidirectional props.
The betaflight quad is tuned perfectly and one peculiar thing during tuning is that P value was constantly increased but there was no oscillation. The max for P is 255 and even when I tested with 250 there was no oscillation at all. I fly it with P set to 200.
What I'm trying to say here is that this copter probably needs high values and still doesn't oscillate. That's why the AutoTune showed red values twice.

The hunt for the bug goes on 

[JustAPilot]

There was a dry period so I went to test and tune the quad today.
1. I set None for assisted mode of the VR and didn't notice any overall difference compared to with GPSAssist selected. So I will leave it to None.
2. Tuning in Rate mode (beyond default values for P) makes the copter less mushy and more responsive but also increases that Boost/jump when centering the stick.
I like the more responsive/locked in reaction but that aggressive short oscillation at the end of a movement is frightening. I wish I could only tame how hard it stops a movement. I tried many different settings, (very high P with much higher I and with much lower I, etc) but that aggressive stop with the boost/jump can only be decreased by lowering the P.

I guess I have to live with this.  The copter flies fine. 

[JustAPilot]

another question...

I set my ReturnToBaseAltitudeOffset to 4m. Does this mean that the copter will fly at 4m above its (Armed)take off altitude on its way to base?
If I'm flying at 12m height when activating RTB, will it descent to 4m when returning to base?
I'm seeing inconsistent behavior so I'm trying to understand how it's programmed to function.

When I arm or disarm the copter there is about 1 second delay. Can this be shortened via (Dis)ArmingSequenceTime ?

[TheOtherCliff]

A while back I listed a VtolPathFollowerSetting that may help and I don't know if you tried it.

You can at least reduce Stabilization page Advanced tab, Rate and Max Rate and not adjust the PID.

I also suggested that you may be running reverse expo.  I can tell you for sure that some correct expo would make it where you like it.

[JustAPilot]

I have tried everything you suggested. (the YawPI->Kp had no effect).
My rates are low and I set the max rate to 3 deg/s higher than my rates. The copter turns and banks very gently.
Expos are not reversed. I have no expos anywhere.
This boost thing I'll learn to live with it. It's a minor issue.
Thanks.

Maybe you missed my other questions...please if you have time:
I set my ReturnToBaseAltitudeOffset to 4m. Does this mean that the copter will fly at 4m above its (Armed)take off altitude on its way to base?
If I'm flying at 12m height when activating RTB, will it descent to 4m when returning to base?
I'm seeing inconsistent behavior so I'm trying to understand how it's programmed to function.

When I arm or disarm the copter there is about 1 second delay. Can this be shortened via (Dis)ArmingSequenceTime ?