LibrePilot Forum
Users => Vehicles - Helicopters => Topic started by: jcg1541 on October 27, 2017, 10:38:27 pm
-
Has any one done the urban proximity diving with a traditional heli, like youtube RotorRiot team?
I think most helis with fpv video onboard are too expensive to do those dare devil acts where only
50% chance a craft can get out alive. But, how about purely direct drive helis, where the main
shaft is the motor shaft itself? And whole craft is as cheap as quads.
And it is carbon fiber shaft, so no bending, either crack and die or live. And the entire craft has
only the same number of moving parts as a quad with folding props.
-
OK, I got the pitch and throttle curve for the gearless heli. Video of the onboard FPV camera DVR is here
Liftoff happens at 0:25 , and the 1 second dive happens at 0:56
https://www.youtube.com/watch?v=1ss5iPBlxnk
. And the ground camera video is here
https://www.youtube.com/watch?v=jDdS0xVSPgA
I was afraid that my high P gain will put the airfoil in the stall region at the bottom of the dive when
it needs no stalls. But it looks OK, and the audio at the pulloff does not sound like stall turbulence
happened. The CC3D config and building discussion is
https://nocomputerbutphone.blogspot.com/2017/12/converged-drone-developers-platform.html
. I am still trying to dive it at straight 90 degrees. Right now I fly this in attitude mode, and the dive is
80 degrees downward when throttle is lowest and pitch is at extreme forward. I don't know if I want to
fly it in rate mode.
-
Be careful when heli forward speed becomes comparable to blade tip speed. Even 50% of blade tip speed.
Of course the trailing blade will loose lift (worse effect in high G maneuvers), but the leading blade will have more effect and you may see oscillations. This happens in fixed wing aircraft. The faster you fly the faster the roll (/pitch/yaw) rate and so you must reduce the PID at high speed.
-
Interesting. I have not thought about the forward speed problem.
-
In Heli mumbo this is called "Retreating Blade Stall".
Its related to the inevitable effect from the unbalance of lift with a leading and trailing rotor blade.
When forward speed increases it will put a theoretical limit to how fast forward a Heli can go (not very fast).
The retreating blade will be unable to create any lift since forward speed is equal or higher, and nose of Heli will be forced upward.
For the purpose of your project (as I have understood it) it will not be a concern.
Let me get some docs for you to indicate the speed at which your heli will experience the retreating blade stall.
...
Dont let it stop you :)
-
Ok, my throttle curve and pitch cure here in the pictures.
During slow descent, at the 1:13 and at 1:30 - 1:40 mark of the youtube video, the cyclic
compensation has oscillations that didn't occur during ascending because the collective lowers
down while throttle is still high at 55% when collective stick is 70% position and blade collective
pitch is 80%. I though about lowering the 55% to 50% , but to keep the high rpm during diving,
the 55% seems unavoidable. So, I am planning to overhaul the entire curve scheme for such a
converged drone.
I want to eliminate any negative collective pitch all together. Normal people shouldn't do any
inverted flying. Dedicate the entire 5 tuning points for the pitch curve should give me much
finer tuning. And the entire craft will fly just like a quad, except that the zero collective pitch
point will be a true zero lift-drift. Most stunt quads have a non-zero "idle up" lift-drift to maintain
the rotor's spinning momentum.
-
Some guesses. :)
Making a guess for 250mm diameter rotor turning 4000RPM (I think that is probably very fast even for 250mm).
Guessing that the blade stalls completely out if 75% of the way out is stalled (farther out is not stalled, but farther in actually has negative lift) and assuming that you start to have problems at half that speed (trailing blade has only half the lift)...
250mm diameter x .75 x PI = about .6 meters of 75% tip distance per revolution
4000RPM is 4000/60=67RPS
67 x 0.6 = 40 meters per second is completely stalled out
For 4000RPM I would guess that, unstabilized, you would have to hold a lot of cyclic at 20m/s to fly level. And don't try to pull a lot of G's.
For 3000RPM head speed that happens at 15m/s.
Diving straight down with no vertical thrust (since blade is vertical) and the acceleration due to gravity is 10m/s/s you could dive for no more than 2 seconds with no initial velocity (2 seconds at 10m/s/s is 20m/s). 1.5 seconds would get you 15m/s. 1.0 seconds would get you 10m/s. But I bet you are already flying forward fairly fast when you start the dive.
Larger heli is better, larger diameter means higher tip speed for a given RPM
Higher RPM is better, more tip speed
No negative collective? I guess you don't want to fly any autorotations. OK with electric, but for nitro, I would want to allow for motor failure that requires an autorotation. :)
-
250mm is indeed the radius of the rotor. The blades are oxy2 210mm, and the feather shaft and grip add another 40mm. Diameter is 500mm.
4000 rpm is close to it. The motor is 400kv, and voltage is 12 volt.
Not bad guesses.
The speed will tip the nose up to help pulloff the dive, so, the P gain should be small. That makes sense.
The youtube "Trump tower building dive" free falls for 6 seconds. Will be interesting to see how it goes.
So it will be 60 m/s at the high balcony palm tree tips. So, the blue trump tower is about 6s * 60m / 3 = 120m from the tree tips to penthouse , about 400 feet.
But I can't find the tower on google. The las vagas one is golden yellow. The new york one is white.
-
Twice the diameter that I described, so twice the speed before you have problems and you can dive twice as long time before you have problems.
80m/s is completely stalled out
40m/s is half stalled out and might be OK for max speed (certainly slower would be safer :) ), but that is very fast (144kph / 89mph) and is probably only possible in a dive (4 seconds when starting from 0 speed).
-
Wow, now I really do think retreating blade stall is something to worry about here.
At least to be aware of and take pilot measures at early symptoms,
or it could potentially get permanently out of control.
Since the high speed is the problem, recovery can be difficult when the heli is in a vertical dive. Thats exactly what you want to do. Its said that in horisontal flight, the symptom of a retreating blade stall is a pitch up of the nose and it may roll as well, mostly to the right for CW rotor. The recommended recovery is not to move cyclic forward or aft (pitch) since that will only worsen the situation, but to first lower the collective and then use aft cyclic to slow down the speed (pitch nose up). But you will have no collective angle to reduce when in the vertical dive.
Have a second look at the video we discussed before.
I noticed how quickly he pulls out of the dives, even though he could go much further.
It may have to do with this risk.
https://www.youtube.com/watch?v=N-AqVbbfIoI
-
I want to eliminate any negative collective pitch all together. Normal people shouldn't do any
inverted flying.
No need to do inverted flying just because having some degrees of negative pitch :)
Just thinking it may be easier to control it and have more options to handle any drift.
-
For me, I think it makes more sense to use a powerful fixed wing with a thick airfoil.
Powerful enough to easily fly straight up. Thick enough wing to slow it down enough so that zero throttle terminal velocity straight down is reasonable.
Then you can fly straight up and straight down for as long as you want. You could even have an FC to assist in hovering.
There are many 3D airplanes that can do this from small sized to large sized.
I always think about this when I see a video of somebody doing mountain / cliff diving with a quad where the mountain is very very steep
https://www.google.com/search?q=drone+mountain+diving
-
The youtube "Trump tower building dive" free falls for 6 seconds. Will be interesting to see how it goes.
That was done with a quad right?
I am not sure about this but, I think a quad will also experience trailing blade stalls but have a better chance than a heli to get away with it since they have 2 CW and 2 CCW propellers?
-
There is certainly a terminal velocity issue too. The rough calculations I did to calculate number of seconds to reach a certain velocity assume there is no drag. It looks to me that the quads doing long cliff dives are hitting terminal velocity and don't get any faster.
-
All the quads crash 10+ times for each youtube take. We can't be too concerned about the physics.
My servos are 3 dollars each, and it takes at least 20 crashes to destroy it. They want us to crash, and I will crash it for them. See the video how resilient the servos are in simulated crashes,
https://www.youtube.com/watch?v=nKVo3AxR_8I
-
I am having doubt about using attitude mode to do straight vertical diving. In my testing
when I hold the heli with my hand, straight downward position causes roll oscillation vibration, in
this video, at 0:37 mark and again at 0:52 mark,
<video removed due to private information involved>
.
It appears that attitude code loop is trying to find the gravity acceleration on the bottom (belly)
side of the heli , but the ground is pointed to by the nose when in straight downward dive.
The attitude mode diving is only stable when I dive at 80 degrees.
-
OK, I just tried rate mode with the same exercise. Rate mode has not oscillation when nose points straight down.
-
I can imagine that. That doesn't help you much.
I can imagine it does that because with nose pointed straight down, then any direction of roll is as level as any other. So it does not know where to put the roll axis.
From that I would think that if you tried to do more than 90 degrees of pitch, where the roll is actually upside down that it would try to roll out. Imagine holding 180 degrees of pitch so it was perfectly upside down in pitch. Then the roll would also think it was upside down even though the roll stick was neutral. It would (I guess) try to roll back to upright.
I don't think this is related to gimbal lock, but it's possible. It makes me wonder what INS13 would do in this case on Revo.
-
Thanks for confirmation.
I had a bliss with my first ever FPV flight today with this gearless setup with low P gain (36) and in
attitude mode.I didn't turn on the onboard DVR recording. So, no video recorded to share. It was
so intense leading up to today's flight that I forgot to attach the FPV goggle's antenna and still
managed to fly it. The low P gain makes it very stable in open field, but the drift makes it unsuitable
for testing hover in office cubicles.
I figured out why FPV videos on youtube are always action packed. Today in my first FPV flight, I
instinctively had to do non-stop forward flight because the FPV video feed gave no perspective of
depth of the scene. The FPV pilot needs to continuously turning and moving to get the visual ques
to know where the craft is. It is the opposite of line-of-sight flying, where we stop all actions and
try to hover still when we lose track of the heli's movements.
I am still bewildered by what happened today. The feel of insecurity about what is on my sides
and what is behind the craft made me want to land the heli on a landing strip while in forward
flight to avoid "danger". I used to land the craft next to my parking spot in between the fence
and the trees, but now I landed it on the open grass. It was not planed. I didn't plan a landing
approach. But it was a safe touch down without a scratch.
-
Yes :) I can certainly relate to your fpv description.
I flew my first fpv with a heli that did not even have a flight controller, just a tail gyro.
It was exhilarating. So different from los flying.
-
The first one is much harder than the second or third. :) You quickly get used to seeing it from the air. I hope your camera does well when looking into the sun, else it is all silhouette and you don't know how close the tree is.
My first FPV was fixed wing, but you would probably say that I cheated. I used a laptop display and took off LOS. It was still hard to do. Fixed wing it is flying faster than a quad and will fly away sooner. If you get lost FPV you will have a hard time looking up to find it. ;) After about a half a flight I got used to seeing things from the air. Second flight I even landed FPV.
-
OK, Did my first rate flight today with a hard landing because I didn't realized that I could/should have switched to attitude mode for the landing approach.
Next is to book a flight to Las Vegas.
-
I have been struggling with rate mode tuning for 2 weeks. Whenever I fly rate mode toward some direction for
more than 2 seconds, there will always be a "rebound" to stop my action on track. CC3D seems to take on
a life of its own.
I still have not managed to dive straight down for more than 2 seconds.
-
Just some thoughts:
You might try increasing pitch/roll PID "I" term. Quads "blow out" (unstable pitch) at high speed if the "I" term is too low.
You could try speeding up / slowing down the main rotor to see if the problem happens later / sooner. If so, it may be you are hitting the copter top speed.
Have you tried from a dead stop so that you know you are not flying faster than 20m/s after diving for 2 seconds? Free fall is about 10m/s ... for 2 seconds is 20m/s. Is it powerful enough (and you brave enough :) ) to fly horizontally at faster than 20m/s? I bet it has the same problem when flying fast horizontally.
Rate mode is similar to unstabilized. Can it be flown unstabilized to see if it is a copter issue or a FC issue? If it always fails in pitch direction, maybe just set pitch to manual for a test??
-
... Whenever I fly rate mode toward some direction for more than 2 seconds, there will always be a "rebound" to stop my action on track. CC3D seems to take on a life of its own.
John, can you please elaborate a bit what you mean here? What are you doing and what is happening with heli?
Rate is hardest to fly but still the most predictable stab mode, in my experience.
I think it will not be difficult to fly just with yaw set to axis lock and all others manual. To test what Cliff suggests.
-
Good suggestion. I will try the axis lock of the tail and manual cyclic with my gearless direct drive.
If anything goes wrong, I will just switch back to attitude mode to regain control.
I was too cocky today and didn't turn on the fpv recording for tracking flight path in case of lost direction. The geared setup flew too far. My brain froze and didn't switch back to attitude mode to regain control before I lost sight of it. It flew away. Lost somewhere in the snow. 250 dollars.
I will have a good story to tell the Dunkin Donut girl tomorrow. She is curious about my heli on my dash board when I drive throu for lunch every day. I put both the geared and gearless builds on the dash board and tape them down.
-
I tried both manual cyclic and high gain rate cyclic today. Both with tail axis lock.
Both gave me toilet bowl effect. But I don't know if it was pilot induced oscillation of if the gyro physics makes the wobble. Manual cyclic is unflyable . High gain rate cyclic is barely flyable when I try to counter the flushing action.
-
It should have been pilot induced oscillation. I had the roll stick with zero exponent but pitch stick with very large 75% exponent.
I was not aware of the situation until just now. No wonder it behaves so out of wack. When I try to correct roll, it is very sensitive, so when I try to correct the pitch it is unresponsive, so I force it hard, but then when I try to correct roll , I over correct it. Then I soften my finger, and the pitch becomes unresponsive, and go back to over correct roll, and then it just spirals. Roll is over corrected and pitch is under corrected. I have been flying like drunk for 2 weeks.
-
You are talking about stick exponential settings of pitch and roll in the transmitter right?
If so, yes then I can imagine its difficult to fly, regardless of rate, attitude or PID settings of the FC.
Sorry about your fly-away. I have just gotten a pair of these TrackR bluetooth units I am trying out.
-
OK, Got my first rate mode "110" degree diving . Apologies for the protection film smudge.
The dive lasted 2.5 seconds. I was so tunnel-visioned on this dive, I flew it with the smudge on FPV,
totally unaware of the blurred feed,
https://www.youtube.com/watch?v=UuLjEBTcxl4 .
I flew another round laid-back attitude mode after taking out the smudged film. And it was a beautiful day,
https://www.youtube.com/watch?v=MYZc57_7Byw .
-
Cool.
Glad to see you got it sorted out and got it this far.
So your speed at the end of that dive, when you bailed out, should have been like 25 m/s right?
Interesting to see next step!
Best of luck
-
Yes, At the bottom of the dive, the speed was about 25m/s . The top of the dive was about 30 m above ground.
25 * 2.5 / 2 = 31.25 .
The nose pitch up was very sensitive at the pull up. I will gradually reduce the pitch up so that it will be a pull out to a horizontal level flight instead of a stand still hover.
I have ordered a 1080p cam runcam split mini.
-
Interesting, the sensitive pitch up may be an early symptom of a 'retreating blade stall'.
As discussed before, maybe 4 sec vertical dive is the maximum speed and you did already 2.5 sec.
Hopefully when the heli reaches its max speed the nose will pitch up by itself and interrupt dive and slow down in a controlled way.
-
I assume that somehow you are running an "idle up" or similar. I would guess that you go to neutral collective pitch during the dive. It is important that you keep your main rotor speed high during the dive, so your neutral collective should still have full RPM.
-
Yes, it is idle up, fully acrobatic mode.
OK. Got a 5 second dive. Dive starts at 0:03 and pulls up at 0:09 . It could have been longer, but I chickened out. The zero pitch was not clean zero pitch because I was frightened and left 3 degrees on the collective during free fall. It was FPV. It felt like I was the one falling from sky.
https://www.youtube.com/watch?v=wVyx11qVg-Y
The craft was only 240 grams because the 10-gram raspberry PI was removed before taking off.
-
OK, Got a slightly over 6 seconds dive. The collective pitch is cut to zero at 0:01 , then the nose is pointed down in the next 1-2 seconds. So, technically the dive starts at 0:01 . The video feed does not look like diving until 0:03 .
https://www.youtube.com/watch?v=XZJF7gACwoc
My rate mode PI gains are on the high region.
https://drive.google.com/file/d/1sZWrp5uXmfFvvFws00yjMYA2tGCyeg8t
-
So are you using higher head speed during dive that now allows you to dive longer?
-
I have not changed head speed. I am doing the 110 degree dive instead of 90 degree straight down dive. 90 degrees is very unstable.
I am thinking about cheating by tilting the camera lens up 20 degrees, so that when I am doing 110 degrees dive, the video feed will look like a straight 90 degrees down.
-
The apparent singularity when pointing straight up / down very probably only exists in Attitude mode. You might consider Rate mode test.
Even easier, you might consider trying Rattitude mode on Roll / Pitch. Holding about half stick should give you about 90 degrees. Beware that the stick gets very sensitive when away from center stick, 79% of stick is 180 bank (upside down), 81%+ of stick is continuous flips. It basically feels like Attitude with so much expo that full stick does flips.
I would also guess that the issue doesn't exist in Revo INS13 even in Attitude mode.
-
I have been in pure rate mode for the last 10 dives.
Attitude mode could not have a single second of stable straight down dive. Attitude mode does not allow any "110" degree dive because the singularity problem yanks the heli out of any dive when it approaches 90 degree.
Rate mode does have stable zero collective pitch straight down dives for the first 2 seconds until the velocity is high.
The rate mode does not have singularity problem. The rate mode problem is the pitch up/down sensitivities with single rotor at high velocity.
At high enough velocity, like 50m/s when air moves parallel to the rotor disk plane , 0.01 degree positive collective pitch will generate the pitch up force equivalent of full elevator stick down; 0.01 degree negative collective pitch will generate the pitch down force equivalent of full elevator stick up.
It is worse than "complete stalled out" of the retreating blade at 50m/s. The air blows from the rear edge of the retreating blade to the front edge of the retreating blade. The retreating blade's collective pitch has a reversed lift force.
-
Karla, yes, the 'retreating blade stall' gets worse and becomes a lift reversal when air moves much faster than the retreating speed, when I go beyond 5 seconds.
Interesting, the sensitive pitch up may be an early symptom of a 'retreating blade stall'.
As discussed before, maybe 4 sec vertical dive is the maximum speed and you did already 2.5 sec.
Hopefully when the heli reaches its max speed the nose will pitch up by itself and interrupt dive and slow down in a controlled way.
-
Wow John, I see you are really pushing the limits here :)
I had a look at your PIDs. The only thing that sticks out is your I terms, never seen so high in 250 or 450 size heli. But I would not worry, if its working.
B t w your outer loop PIDs are very high but you only do rate mode so it is not used.
Totally see your point of 'falling out of the sky' when dive using fpv :o
When you did the 6 sec dive, heli reached 60 m/s over 210 km/h? Or even faster since you do 110 degree dive and not just falling, but pushing the dive.
So how does it behave when approaching stall, nose pitches upward? any roll?
-
John, just an idea.
You could consider doing the dive in a different way.
Do it fully upright, horizontal and apply negative collective.
You now have a fully 3D heli setup right.
You will not have any issues with 'retreating blade stall', not even for the tail rotor since it spins typically 4 * faster than main rotor.
However, you will likely run it to another issue, Vortex Ring State.
But, I was just thinking this should just be initially, and then you can push through it, descending much faster.
This style of dive will give same speed and sensation as 40-50 m/s but you can control the descending speed so it keeps under control.
This will allow you to use Attitude stabilization for simplicity but more interestingly, a much longer dive, like any distance you prefer, really.
Move the camera to a downward position, or have two cameras, or maybe a camera that can tilt.
This is the problem of Vortex Ring State and how to recover from it. But you don't want to recover :)
https://www.youtube.com/watch?v=HjeRSDsy-nE&t=95s (https://www.youtube.com/watch?v=HjeRSDsy-nE&t=95s)
But very little or no research has been done if you want to push it faster downward.
No one have ever had an interest doing that.
Fascinating John!
-
The nose pitch is either up or down. There is no involuntary roll. Roll is very stable other than a twitch when elevator stick is applied suddenly to start the dive. My attitude mode gains are high, and it is very stable and precise maneuver when I fly it in the office aisles. Thanks for looking at my UAV file.
I have not used the heli's own thrust to dive faster than free fall. It is all free fall. Yes, would be close to 120 mph or 200 km per hour if air had no resistance to slow it.
My zero collective pitch is not 0.00 degree. It is maybe 0.3 degrees positive. My camera lens is mounted horizontally. In the last dive video, when it was in 150 degrees dive, it was stable. When it was 90 degrees straight down, it had pitch up. But, I can only imagine that if my zero collective pitch was 0.3 degrees minus, the pitch up sensitivity problem would reverse and became a pitch down.
Wish I can dive next to a building to get a visual of the speed.
When you did the 6 sec dive, heli reached 60 m/s over 210 km/h? Or even faster since you do 110 degree dive and not just falling, but pushing the dive.
So how does it behave when approaching stall, nose pitches upward? any roll?
-
Got it.
Maybe good to wait diving close to building until very familiar with heli behavior?
-
Yes, good idea.
Looking at the picture, the vortex ring seems to be some kind of inverted ground effect that comes out of nowhere.
-
Haha yes that's a way to put it I guess.
Its the number one dangerous flight conditionfor a full size heli and a common cause of crashes. It may happen when descending vertically (90, tops 70 degrees) and rate of descent is relatively high for a normal heli, 2-4 m/s. I don't know how this translates to our smaller helis, the tip of a rotor moves typically at 200 m/s, we have higher rpms but much shorter rotors. I just have a gut feeling our helis will need to have a higher rate of descent to get in to it. Anyway, the heli basically gets trapped in its own main rotor downwash. Descending faster or slower will move it out of the predicament.
I think you could expect to move through this state as you dive.
Other than that it could be useful to your ends.
-
Karla, My setup is closer to 200 size than 250. The rotor head is from the Blade 180 , and blades are Oxy2 210 blades that are barely over 200mm from bolt screw to the beginning of the sloped cut.
The rotor diameter is close to 490 mm . The weight is less than 250 grams including Raspberry Pi.
I do use the attitude mode to launch/land from/on hand . The outer loop high gain is very helpful for catching.
https://www.youtube.com/watch?v=TrHm1BYM3fo
I had a look at your PIDs. The only thing that sticks out is your I terms, never seen so high in 250 or 450 size heli. But I would not worry, if its working.
B t w your outer loop PIDs are very high but you only do rate mode so it is not used.
-
Got it John,
Will try later to calculate at what decline rate such heli would expect a vortex ring state.
Meanwhile, I think you can really just experiment and try it out.
Now you know what to look for and how to recover from it.
symptoms
. heli vibration
. random yawing, rolling and pitching
. increasing rate of descend
. less cyclic stick respons
recover by rolling
What do you think?
-
I have been asking around a bit but seems no readily available method to do this calculation.
Answers mostly like this:
- VRS is a very turbulent state and therefor stochastic and hard to predict/calculate.
It happens, roughly, when rate of descend of downwash is equal to heli descend. But how to figure out difference between full scale heli compared to rc size heli - not clear.
So no real help for you.
-
My experience with this is with quads. I never had much issue with my unstabilized .30 to .50 sized nitro helis.
Falling straight down is part of the issue. Descending at a 45 degree angle does not have the problem in my experience.
Also if you are falling quickly (straight down) there is no problem.
At least with a quad, well tuned PIDs (such as from AutoTune) change the problem almost into a non-problem.
Best to experiment in a controlled way to know where the issue is, to avoid it when needed.
-
Karla, I have cut 40% of the gains with much larger swash travel range to make the heli agile enough to flip.
Previously I tried to optimize the collective for long duration by stipulating the collective range in Librepilot PWM output tab.
Now the swash travel range is almost doubled, and the gains are all different now. I am still yet to cut the gains further down, maybe half of my previous.
I had a look at your PIDs. The only thing that sticks out is your I terms, never seen so high in 250 or 450 size heli. But I would not worry, if its working.
-
Just some thoughts:
You might try increasing pitch/roll PID "I" term. Quads "blow out" (unstable pitch) at high speed if the "I" term is too low.
That was a valid idea. I had a Proprotional 35, high Integral 150 dive today. And it felt more controlled and not on the whim of heli physics pitch-up-down.
The dive is at the end of the video, 1:06 to 1:15 , sorry it took a minute to climb to the diving height,
https://www.youtube.com/watch?v=2InEBgHKfI0
I was able to do 120 degrees from 1:06 to 1:08 , then to aim for 90 degrees dive at 1:09 - 1:11 , then again to go at 120 degrees from 1:12 to 1:15 . I was in control.
Tank you my friend.
-
:)
-
OK, I have the leisure to do a aileron flip while diving straight down now with a even higher Integral .
The first 15 seconds is climbing, then the diving immediately follows at 0:16 ,
https://www.youtube.com/watch?v=R1DbTP3CIHc
. I still have the bad habit of elevator pitch up trying to pull up the imaginary airplane at the end of
a dive that looks stupid. In helicopter, elevator pitch up doesn't really help any thing, and elevator
pitch up should be substituted with elevator pitch down for forward pull out.
-
I think it will not be difficult to fly just with yaw set to axis lock and all others manual. To test what Cliff suggests.
Karla, I have tested manual mode with this 210 sized heli now. The result is that it tilts and twists all directions.
(the heli is too small to be seen most of the time, and police cars zipping by not noticing what goes on there)
https://www.youtube.com/watch?v=CsPT7BiAaVE
The rotor is too small to have stabilization by itself without flybar. This is a FBL setup. This is a FPV session.
0:01 tilt right and backward
0:02 attitude mode to rescue
0:03 tilt right and forward
0:04 to attitude mode
0:05 tilt left straight
0:06 attitude mode
0:08 tilt backward
0:09 attitude mode
0:10 tilt forward
0:11 attitude mode and stay in attitude mode until safe landing
0:25 landing approach too high and close to trees
0:35 backing off far and fast from the botched landing
0:40 good landing approach
0:53 touch down
Every second is an adventure of twists and turns. It needs a computer to stabilize it every second of the way.
-
Smoothest and stable dive so far,
https://www.youtube.com/watch?v=C6aAeSFlfw4
-
That is looking acceptable! What, about 4-5 seconds?
Just a question: are you using TPS (Stabilization->Advanced->ThrustPIDScaling)? Mathematically at least, whenever you double the head speed, you should halve the PIDs, etc. So if you are using a throttle controlled variable head speed, this will keep the PIDs in sync with the head speed.
-
Interesting. I am not using the thrustpidscaling. My thrust ranges from 78% to 87%. And P is 15 now. The scaling will reduce 1.5 of P.
I will have my pid around this number for a while until my proximity dance skill leap to the next level.
https://m.youtube.com/watch?v=3imtW_ho2zU
The initial half flip overshot made it dive on the side, still straight down dive though. And then elevator erroneously used to correct it. The elevator pitch up was intended to turn the nose down after a half flip. But now the elevator only made it swing on the horizontal plane. And it ended up with piro correction. All this need to be fixed on the pilot's brain. And work on the brain starts now.
-
Got leveled pull off at the bottom of the dive now thanks to Sikorsky's research for turbulence solution with a canted tail rotor,
https://www.youtube.com/watch?v=64EaENqfc2s
. The heli now has the same 20 degree slanted tail as the Blackhawk to fix previous ugly high-G-pullback at the closeup of the dive videos.
The swinging during diving is due to the high wind today at 15 MPH on the ground and even higher in the air at 120m altitude. You can see how the wind blows at the tree at the end of the second video,. It is not PIDs problem, this is the view from the FPV goggles, transmitter is a generic tiny whoop 200mw VTX,
https://www.youtube.com/watch?v=gAIVv7u5EAA
. No changes to PIDs since last 2 posts.
-
Interesting.
Now you are doing 10s dives and pull out in good order.
How does that tail rotor looks from the side view?
Is it canted like this
https://www.google.se/search?ei=ZbQIW6GxD8exsQHu_rLYBg&q=canted+tail+rotor&oq=canted+tail+rotor&gs_l=psy-ab.3..35i39k1j0i22i30k1.19401.21345.0.21760.6.6.0.0.0.0.414.414.4-1.1.0....0...1c.1.64.psy-ab..5.1.414....0.PvzdiYXefxI
-
Sorry I have been distracted by the search for abandoned tall buildings. I am sure I would have dived many decaying skyscrapers of Detroit if I lived there. Here in the north east, I found the rusting old Ramington Arms munition factory with a tall smokestack. See attached photo. I passed by it over the highway just this memorial weekend. Guess I will dive it in the next few weeks.
-
:P
-
Just a question: are you using TPS (Stabilization->Advanced->ThrustPIDScaling)? Mathematically at least, whenever you double the head speed, you should halve the PIDs, etc. So if you are using a throttle controlled variable head speed, this will keep the PIDs in sync with the head speed.
That was an excellent question. I have tried it today, and the result is a dramatic improvement. Previously either the ascending was very slow or too much vibration to be shown on video. Now the punch up is forceful, yet no oscillation anywhere in flipping or diving. The ascending is 28 seconds, then the 9 second dive follows,
https://www.youtube.com/watch?v=ZRezTcZxFwM
. And the extreme scaling down is between thrust 75% toward 100% . The flipping and diving thrust is below and right adjacent to 75%.
The curves are attached in pictures.
-
This is a tribute to Boston Marathon spirit. We lost citizens, and we lost limbs. But our spirit is not hurt one bit.
The takeoff weight of this drone is under 250 grams. I have the PIDs for quadcopter-like diving now,
http://nocomputerbutphone.blogspot.com/2018/09/converged-iot-drone-build-notes.html
. FAA rule is that if you are within 400 feet next to a structure, your altitude limit is based off the top tip of the structure. The dive is about 600 feet from the top of the tower to the church spire. The following video will be deleted after being viewed by this forum.
https://www.youtube.com/watch?v=SjaLnku8ueQ
-
Well done
-
I might have run into the Vortex Ring State now. I have been trying to push the limit of my tail motor ESC to the point of failure by forcing full throttle torque at the bottom of my dives.
2 incidences . The first dive starts at 1:00 mark. In either case, maximum throttle-pitch was applied at least 2 seconds before the touch down crash and the craft crashed with full throttle-pitch still in effect. The pull-up deceleration appeared weaker than with "regular" pull up moderate collective pitch.
https://www.youtube.com/watch?v=AP6Rui1JB-k
The solution is to just go back to the more moderate pull-up that we instinctively develop.
I have a diagram my reasoning of the vortex , which is a stall vortex due to wind direction change with the dive itself, in attached picture.
I believe this happens to quadcopters too, just that no one tries to push the limit of the tail ESC because there is no tail motor there.
I have found my tail ESC that allows the the vortex ring fault to occur before the ESC fails.
John
John, just an idea.
You could consider doing the dive in a different way.
Do it fully upright, horizontal and apply negative collective.
You now have a fully 3D heli setup right.
You will not have any issues with 'retreating blade stall', not even for the tail rotor since it spins typically 4 * faster than main rotor.
However, you will likely run it to another issue, Vortex Ring State.
But, I was just thinking this should just be initially, and then you can push through it, descending much faster.
This style of dive will give same speed and sensation as 40-50 m/s but you can control the descending speed so it keeps under control.
This will allow you to use Attitude stabilization for simplicity but more interestingly, a much longer dive, like any distance you prefer, really.
Move the camera to a downward position, or have two cameras, or maybe a camera that can tilt.
This is the problem of Vortex Ring State and how to recover from it. But you don't want to recover :)
https://www.youtube.com/watch?v=HjeRSDsy-nE&t=95s (https://www.youtube.com/watch?v=HjeRSDsy-nE&t=95s)
But very little or no research has been done if you want to push it faster downward.
No one have ever had an interest doing that.
Fascinating John!
-
It could be a simple matter of the main blades stalling out if you pull out too sharply; just like a fixed wing. High angle of attack associated with backwards flow through the main disk / high G's.
-
I am not sure, but I think this looks like another trap for helis called 'settling with power' (main rotor stalls).
It looks like your heli is slowing down the descend but just do not have time enough to stop before it hits the ground.
That is different from Vortex, since you will just keep falling without slow down.
Anyways, it do seem you really reached the limit now :)
This guy from Canada is trying to separate the two things.
https://www.youtube.com/watch?v=ehV9vLnBICE
-
OK. I have been testing a 3-second collective hold-off rule. I keep the collective pitch at zero for the high-G pull-up at the transition out of the dive for 3 seconds, and I have no more crashes.
During the 3 seconds, the craft can do all cyclic banking and piro maneuver without any problem.
https://www.youtube.com/watch?v=7X1T3yMSLt4
I have to admit that, in my previous crash tests, I did feel that the craft was going to keep sinking indefinitely with the hard full-collective pitch, especially in the first crash test. To the point that I doubted that there was some gear-strip, even though the build is gear-less.
The operating note about this trick and all other tips are at
https://nocomputerbutphone.blogspot.com/2018/09/converted-iot-drone-operation-notes.html
https://nocomputerbutphone.blogspot.com/2018/09/converged-iot-drone-build-notes.html
-
I've wondered about that. At zero collective, I would guess that the rotor acts more like a wing flying out of the high G maneuver, rather than a prop (with air blowing backwards and probably stalled out) trying to thrust out.
-
latest test with 3150 rpm , 5 degree hovering collective angle , 1 degree during diving, PIDs tuned down 2680/3150
https://youtu.be/VXKGSyEdRI4
all previous tests before today had rpm 2680 for 6 degree hovering collective angle maximum fuel economy, 0.5 degree during diving.
-
Looks very stable and controlled.
The only thing I would wonder is how well it responds to the 3 stick axes in the dive.
-
I have been trying a new servo that is marketed as "high precision helicopter servo" , but the high integral gain in the servo is interfering with CC3D's PIDs.
In low-speed, line-of-sight flights, it looks nice, but the PIDs gains becomes inadequate at terminal speeds unless I dedicate a CC3D mode with 180% proportional gains for the dives.
The following video is with the 194% "high-gain mode" during the dive. And immediately after diving pulloff, CC3D PIDs are changed back to normal acrobatic mode.
https://www.youtube.com/watch?v=eBw200KLlp4
. The servo is E-Flight S60, 6-gram servo. The parts of the video of switching-back-and-forth of the flight modes have been trimmed away.
Prior to this month, I used the Emax 9051 that likely reacted to split-millisecond adjustments from CC3D at terminal speeds so that the low PIDs numbers for cruising/acrobatics was adequate for diving. The emax 9051 servo likely had no integral circuitry. A youtube reviewer compared it with a more expensive servo and showed that the Emax 9051 servo had larger error angles.
Further more, the Emax 9051 servo likely used large proportional gain to compensate the omitted integral circuitry, serendipitously giving it excellent sub-millisecond reaction time.
-
What does it do if you leave the normal PID active in the dive? Oscillate? Wobble around as if the servo isn't strong enough when you know it is strong enough? Come straight down and require more pull out pitch than normal? Is the new servo rated to have at least as much torque as the old servo (maybe the strength is marginal)?
I wish we had telemetry to look at to see what the servo is being asked to do. It could be something like a very jittery servo signal that is being smoothed too much by the new servo, or a PID oscillation with one (slower) servo and not the other (faster one) like an oscillation when you run quad ESC protocols that are too slow for your PIDs.
Just an observation: Measuring servo error with a fixed servo signal is not the same as measuring the heli leveling error with a stabilizing FC running. Any slop in the servo or even the linkage is removed by the FC. Imagine a roll in manual mode with the stick released when it is level. Ignore inertia and assume that the only issue is slop. Imagine that with the slop the swash plate never comes all the way back to neutral and so you get a continuing slow roll. Now imagine what the servo does with the FC in Rate mode when you release the stick when the copter is level. It continues to drift past level and the FC sees that the rate is not zero and drives the servo far enough past neutral to both counteract the slop and the inertia. Servo and linkage slop doesn't really matter because the FC takes the slop out.
I think I saw a slight circular roll/pitch oscillation in forward flight. I wonder if the FC is not well aligned to the frame or if the heli roll and pitch response is not well aligned to the frame. I would fly in Manual mode and test to see if pitch and roll flips are well aligned to the frame. If not, I would adjust whatever is needed to get the manual flips aligned well.
I might also be inclined to measure vibration with GCS and a USB cable attached and try to get a good rotor balance if you think it could use it.
Hope these ramblings help. :)
-
What does it do if you leave the normal PID active in the dive? Oscillate? Wobble around as if the servo isn't strong enough when you know it is strong enough? Come straight down and require more pull out pitch than normal? Is the new servo rated to have at least as much torque as the old servo (maybe the strength is marginal)?
I have a video footage of using normal PIDs in the dive using the new servo. I will post it next. It wobbles around.
The new servo, eflight s60, is more beefy than my prior servo, emax 9051. I have developed a "servo saver" modification to prevent gear strips in crashes. The mod weakens the maximum torque of the servo arms, but I am pretty confident that the mod does not impair the precision of the servos. See attached pictures. I use this mod in all my videos, old and new.
-
What does it do if you leave the normal PID active in the dive? Oscillate? Wobble around as if the servo isn't strong enough when you know it is strong enough? Come straight down and require more pull out pitch than normal? Is the new servo rated to have at least as much torque as the old servo (maybe the strength is marginal)?
https://www.youtube.com/watch?v=O4JJHkmpZHg
The first dive was with the normal PIDs. The second dive was with 194% high gain mode. The 2 dives were in the same flight back-to-back.
I spun the rotor with consistent constant RPM by using BLHeli MULTI software with PI governor. Both dives had exact same RPM.
The first 5 seconds of the first dive was OK, but I had to abort it after 5 seconds when it swung so much I was not sure what it would crash into.
-
I saw that same circular oscillation (pitch/yaw with roll too) in fast normal flight with normal PIDs. I am guessing that the fact that the oscillation has a yaw component (with different servo, linkage, servo loading, PID) is saying something?
I would start with Manual mode, making copter adjustments so that it flips (roll and pitch) and pirouettes in alignment with the frame or at very least in alignment with the FC. Then check that Rate mode flips are also in alignment.
I might try tuning with ZN method like this: Set I and D to zero and adjust P up till it oscillates. Cut P in half and adjust I up till it oscillates and reduce I a little. Leave D set to zero.
-
Just an observation: Measuring servo error with a fixed servo signal is not the same as measuring the heli leveling error with a stabilizing FC running. Any slop in the servo or even the linkage is removed by the FC. Imagine a roll in manual mode with the stick released when it is level. Ignore inertia and assume that the only issue is slop. Imagine that with the slop the swash plate never comes all the way back to neutral and so you get a continuing slow roll. Now imagine what the servo does with the FC in Rate mode when you release the stick when the copter is level. It continues to drift past level and the FC sees that the rate is not zero and drives the servo far enough past neutral to both counteract the slop and the inertia. Servo and linkage slop doesn't really matter because the FC takes the slop out.
I see your concerns with the slop play. I want to rest your mind that the problem was investigated and resolved with my 250-gram craft as attached first picture with the 0.1mm patching films (tape). The slop play was a very serious destabilizing problem even for simple hovering without any acrobatics because the airfoil CG shifted as explained in the second and third pictures. I have built the same 250-gram craft 10 times, and I can consistently build stable crafts recently.
I know that symmetrical airfoils have a stable center of pressure CP (but not zero CP shift), while cambered airfoils have a positive feedback CP shift, meaning that a CP shift resulting in more angle of attack change resulting in even more CP shift - an inherent unstable system.
So, what I learned was
helicopter + small slop = stable
helicopter + large slop = unstable
airplane + any slop = unstable
.
-
I might try tuning with ZN method like this: Set I and D to zero and adjust P up till it oscillates. Cut P in half and adjust I up till it oscillates and reduce I a little. Leave D set to zero.
The ZN method is used by the hobby industry to make the expensive "cyclic servo". They cut the instantaneous output from CC3D for swash in half and reallocate it for I gain in the servo's digital board. I-integration needs time to close the error gap, and for common helicopter flying the half gap is closed very tightly by integration over time, but there is no time when the craft dives at terminal speed. Every millisecond spent on integrating the error is a millisecond the craft spend on swinging like my last video's first dive. That explains why I needed 194% P gain, almost exact doubling, in CC3D as the "dive mode" PID when I used the precision cyclic servos. We are in "nested" PID loops when it comes to the swash. See this video, starting at 1:30 for the nesting control,
https://www.youtube.com/watch?v=tbgV6caAVcs
The ZN method is absent in the, also expensive, "helicopter tail servo". I installed 3 tail servos on the cyclic swashplate today. And my "normal mode" worked for both diving and cruising. As this video,
https://www.youtube.com/watch?v=F9I_hD_9evc
The "tail servo" needs the exact same PIDs for the swashplate as the lowest cost, simplest F3P servos, EMAX 9051, that I started this project with.
I just now realized that when I dive vertically, my main rotor stands up straight, exactly the same as a tail rotor. Maybe that is why they mean to sell me tail servos for my main rotor.
-
I have been trying out a new set of heavier rotor blades. And somehow the PIDs ratios between the 2 sets need to be the square power of the blade weight ratio.
The lighter set weighs 14.6 grams,
https://www.youtube.com/watch?v=RMf1LfdpxVY
, and the heavier set weighs 17 grams,
https://www.youtube.com/watch?v=JVFXiyKf_lg
.
The 2 sets of blades have the exact same geometry, shape and dimensions.
-
Some wild guesses. :)
I can imagine that pitch flexibility (twisting) or loose blade grips or blade grip pivot hole farther back in airfoil or blade airfoil being more tail heavy could cause the blades to overreact to (cyclic) blade pitch changes... and thus to need smaller PIDs. Are either set of blades weighted? Weight is usually added close to blade leading edge and that would be more stable and need more throw and more PID to move it?
-
Everything geometry wise is identical between the 2 sets. Pivot holes, bolts, etc. The airfoil is the popular NASA0015 .
The airfoil is pictured on this page,
https://nocomputerbutphone.blogspot.com/2018/10/oxy-210-blade-characteristics.html (https://nocomputerbutphone.blogspot.com/2018/10/oxy-210-blade-characteristics.html)
The lighter set is simple nylon. The heavier set is nylon reinforced with graphite.
The only difference is the material.
Grip is tuned by the same procedure as this video,
https://www.youtube.com/watch?v=Rm-M_b-C1XA
Some wild guesses. :)
I can imagine that pitch flexibility (twisting) or loose blade grips or blade grip pivot hole farther back in airfoil or blade airfoil being more tail heavy could cause the blades to overreact to (cyclic) blade pitch changes... and thus to need smaller PIDs. Are either set of blades weighted? Weight is usually added close to blade leading edge and that would be more stable and need more throw and more PID to move it?
-
Mystery solved. It was my grip setting procedure allowing too much error margin.
I allowed the blades to drape between 45 degrees and 90 degrees, and the 2 extreme ends had very different rotor stiffness. The looser the grip, the higher the PIDs were needed.
Now my new procedure is to set the grip to allow the blades to drape at 80 to 90 degrees,
https://www.youtube.com/watch?v=O2LQsu6K7kk
, and the PIDs for diving is almost the same for either the 17-gram blades or the 14.6-gram blades.
https://www.youtube.com/watch?v=1oO6FHYu-hU
.
-
Glad you got it sorted. I always enjoy watching your flying straight down.
On my .30 size helis I just went by feel. Equal and fairly stiff. I figure the force pulling out on the blades will straighten almost anything out. :)
-
Appreciate that. The last video was an example of slight rotor imbalance. Today I got the well balanced rotor with the same PIDs and everything.
The 2-blade system is actually equivalent of a string music wire tensed up. And my RPM happens to be close to the natural resonance frequency.
I am balancing it with pieces of 1 square cm scotch tape, less than 0.01 grams each. And each piece counts. Yesterday's flying was 1 piece off balance.
Today's dives are here,
https://www.youtube.com/watch?v=upqED-hF6A8
https://www.youtube.com/watch?v=Fz3vTEh6z2o
https://www.youtube.com/watch?v=y2iRtDIjkW8
-
Really admire your perseverance 8) You been working on it for a long time not giving up.
-
I am balancing it with pieces of 1 square cm scotch tape, less than 0.01 grams each. And each piece counts.
For my low RPM (1000KV@3S) plastic quadcopter props (after trimming on magnetic balancer) I use black plastic electrical tape of about that size to balance them running. Also the short way (thinner strips of tape) of the prop vs. the long way.
I also balance motors with several layers of the tape pieces. One out of ten might need it. I might add a little CA around the edge of the tape square, especially for motors that get warm and the tape flies off.
Even glued, the black electrical tape sometimes comes off. I have always just trimmed a blade in the long direction (research one bladed props if you think trimming a blade tip is bad). I now use hot melt glue on the hub for the short prop balancing direction. It works well, at least on my 10000RPM motors. None has come off.