cant keep a perfect turning

[Skkuda]

Hi!

The point is, when I start turning the shark I can maintain a perfect coordinated turn in level flight at the same speed but passing lets say 180 degrees the chopper makes an agressive jump, dont know how to tell in English, something like entering in a hole.

This situation got me intrigued.:huh:

Anyone else getting the same effect?

 

Skkuda

[Frazer]

Do you have a track maybe? Would be easier to analyze ;)

[ScEBlack1]

Could be collective brake? are you setting collective brake after each collective adjustment?

[EinsteinEP]

Russian airway management is notoriously lackluster. Sounds like you're running into air-potholes. ;)

 

Seriously, a track would be helpful here, but here's a wild guess: you could be in an autopilot mode (route? auto turn to target?) and starting your turn by providing cyclic inputs that counter and swamp out the AP's feedback. Once you get 180° around, the AP switches the direction of input, so it seems like a sudden jump in cyclic controls.

 

Just a guess.

[AlphaOneSix]

It's the autopilot, though exactly why it does that is still a mystery to me. At 180 degrees from your original heading, you'll get a spike in rudder input from the autopilot in the opposite direction of the turn, which will look like the nose jumping up. If you go the entire 360 degrees, you will notice that the opposite occurs, and you'll get a spike in rudder input in the same direction as the turn from the autopilot, resulting in a downward pitching motion. This only occurs when the heading mode of the autopilot is on, and you are not holding the trim button during the turn.

[takamba]

Try to turn off the heading hold during the turn, or fly the hole turn in Flight Director mode.

[sobek]

Or trim while you already have a positive yaw rate. If you trim while your yaw rate is >3deg/sec (methinks), the autopilot will try to maintain the turn instead of trying to force hold your old heading.

 

Although i am an avid AP fan, i must say that coordinated turns are easier if you hold down trim during the turn. On the other hand, if you are used to the procedure, you have less workload during the turn, although a little higher on entering and when ending the turn.

[Skkuda]

Originally posted by AlphaOneSix

It's the autopilot, though exactly why it does that is still a mystery to me. At 180 degrees from your original heading, you'll get a spike in rudder input from the autopilot in the opposite direction of the turn, which will look like the nose jumping up. If you go the entire 360 degrees, you will notice that the opposite occurs, and you'll get a spike in rudder input in the same direction as the turn from the autopilot, resulting in a downward pitching motion. This only occurs when the heading mode of the autopilot is on, and you are not holding the trim button during the turn.

 

Thats exactly the description of the problem that Im having!

The only difference is in the first 180 degrees it makes your second description Alpha!

need to try the entire 360 degrees, to test a little bit more!

 

Originally posted by Takamba

Try to turn off the heading hold during the turn, or fly the hole turn in Flight Director mode

 

Seems a little bit waste of time if you need to react fast! But Im going to try!

 

Originally posted by Sobek

Or trim while you already have a positive yaw rate. If you trim while your yaw rate is >3deg/sec (methinks), the autopilot will try to maintain the turn instead of trying to force hold your old heading.

 

 

Seems to be the solution, gonna give it a try even I dont like to hold the trim button while maneuvering because of the aircraft behaviour.

 

Real Thanks!!

 

Skkuda

[Haril]

It's the autopilot trying to point you back to the heading you were at when you last trimmed. For the first 180 degrees, it'll be fighting you, trying to turn the 'copter back to where it started. When you reach and pass 180, it'll switch to helping you in the turn, as that's now a shorter way to get back to the trimmed direction. The jolt is caused by the autopilot going full displacement in one direction to full displacement in the other.

[beers]

Seems a little bit waste of time if you need to react fast!

 

I have trim and FD mapped to the joystick in the most convenient of places. I think of them as;

Trim = I want to point her in a different direction/attitude

FD = :helpsmilie: drop to the deck

[Frederf]

The jolt is caused by the autopilot going full displacement in one direction to full displacement in the other.

 

The confusing thing is that shouldn't be the case. The AP does not exercise its authority to fight the pilot's input. Easy example, you start perfectly trimmed heading North so the AP is using +/-0% control. You push in +19% (right pedal) as the pilot. You would think that the AP would add in -19% to keep you facing North, right? Well it doesn't. The AP is not always using up its authority to realize its hold... sometimes it just waits at its current input until the pilot lets go of the controls.

 

That's not to say that there isn't some weird bump at 180° from the hold heading, just that the AP shouldn't be going form +20% to -20% there.

[ARM505]

Well, I'm sure you'll know what I'm going to say already :)

 

I've said it before, if you plan on flinging the helo around the sky (ie generally a lot of heading changes), then just turn the heading AP channel OFF completely. Yes, yes, Kamov says you're a naughty boy if you do that (apparently) but hey, I'm the pilot and it makes my job easier then. No funny spikes, no bouncing and fighting. Just smooth turniness :) Ball in the middle, nice smooth pedal inputs, no problems....the pedal trim will still work fine. Again, I have a sneaking suspicion that all is not 100% correct in the AP inputs, but we'll have to see what happens.

[Frederf]

Alternatively you can keep the channel on and only turn the hold off (FD, trim button held). No reason to throw out the baby (stabilization) with the bathwater (hold).

[Haril]

The confusing thing is that shouldn't be the case. The AP does not exercise its authority to fight the pilot's input.

 

Sure it does. It just doesn't do it immediately on pilot input. If you fed in 19% rudder, the helicopter would start turning away from the selected heading. As the heading difference increased, the autopilot would start feeding in more and more opposite rudder to try to bring the nose back again until it's at 19% and balanced with you with the nose offset from the selected heading.

 

The amount the autopilot puts in is proportional to the difference between the selected attitude and the current attitude. This creates the jolt at 180 degrees off the selected heading as the difference goes from 100% in one direction to 100% in the other, making the autopilot reverse itself as quickly as it can.

[sobek]

The amount the autopilot puts in is proportional to the difference between the selected attitude and the current attitude.

 

Well, there has to be some connection between deviation from trimmed attitude and AP input, but what you say here would result in an infinite impulse response (IIR) control circuit, meaning that the autopilot would oscillate infinitely around the trimmed heading. It's a little bit more complex than that. :)

[Haril]

Well, there has to be some connection between deviation from trimmed attitude and AP input, but what you say here would result in an infinite impulse response (IIR) control circuit, meaning that the autopilot would oscillate infinitely around the trimmed heading. It's a little bit more complex than that. :)

It's probably handy then that the helicopter has pitch, roll and yaw damping built in and is flying through a medium which provides a nice pile of friction.

[EtherealN]

This creates the jolt at 180 degrees off the selected heading as the difference goes from 100% in one direction to 100% in the other, making the autopilot reverse itself as quickly as it can.

 

Nah, what creates this is the fact that people don't trim... :P

 

If you are going to turn around, initiate your turn and find a stable turning attitude, trim (and note how the AP now does not fight you anymore), then end the turn when reaching desired heading, trim. Alternatively hold the trim button down while maneuvering, as previously mentioned.

 

Like Frederf said, don't throw the baby out with the bathwater. Just trim for the turn. ;)

[ARM505]

Nah. That baby is useless. I chucked it, sorry. :)

 

Turn to target - that's what my pedals are for. I mean, 'look at tgt cursor, turn towards it!' That's hard man, we really need an entire function just for that one - NOT.

 

Keeping ball in middle - yup, pedals again!

 

Heading hold - just trim (ACTUAL trim, ie applied force, not wierd damping and fumbling about chasing some arb heading) properly, and use pedals in the conventional way (step on ball, etc)

 

Hover hold, holding a specific heading - thats what my pedals and trim are for.

 

Sorry, I can still not think of ANY situation that I can't do just fine by simply using the pedals for what they were intended. If God had intended man to fly with the AP yaw channel engaged, He wouldn't have given him pedals! :) Actually, this is all really directed at Kamov, not you guys (you've heard me before).....Kamov, are you listening?! ;)

[EtherealN]

Actually, the fact that Russian Army Aviation is going for two-seaters in future and canceled production of the single-seat Shark indicates that even with those functions they still feel that there's a bit too much workload for the pilot in a helicopter like the Ka-50. Those functions are there to ensure that the pilot doesn't have to bother with minute house-keeping while his head (and mind) should be outside the cockpit scanning for threats.

 

Pilot Saturation is a serious problem, especially in helicopters.

 

So instead of having to maneuver completely on your own to get pipper on the target, you just have ToT on and can focus entirely on your threats and targets.

Heading hold - the same. Trim up to where you want to fly, and then it'll keep you heading over there while you are handling things like planning your ingress, looking for threats, scouting possible egress routes etcetera. In the case of hover hold - you're usually there because you are about to employ weapons, which means the enemy can see you too. Your mind needs to be fully on your targets and on spotting any threats that might pop up. The little mental effort spent on managing your position and direction can very well be the difference between spotting that Stinger and not spotting that Stinger.

 

For regular flying about - you are right, you don't "need" it. Airline pilots don't "need" all their fancy AP kit either. But the kit decreases fatalities... Or, in the case of Kamov's product, decreases friendly fatalities and ensures that more time can be spent increasing hostile fatalities. ;)

Edited February 13, 2010 by EtherealN

[Frederf]

Sure it does. It just doesn't do it immediately on pilot input. If you fed in 19% rudder, the helicopter would start turning away from the selected heading. As the heading difference increased, the autopilot would start feeding in more and more opposite rudder to try to bring the nose back again until it's at 19% and balanced with you with the nose offset from the selected heading.

 

The amount the autopilot puts in is proportional to the difference between the selected attitude and the current attitude. This creates the jolt at 180 degrees off the selected heading as the difference goes from 100% in one direction to 100% in the other, making the autopilot reverse itself as quickly as it can.

 

That's a fine theory and one I should easily be able to test. According to your theory if I were to limit my pedal input to +/-20% (via the joystick setup page), then I could stomp on my pedal from a perfect trim situation with a heading hold and the Ka-50 would slew off that heading. As the heading difference built the AP would kick in opposite rudder up to 20% and the rudder would be in the same position it was (from external view) before I did anything. Right?

 

Well, there has to be some connection between deviation from trimmed attitude and AP input, but what you say here would result in an infinite impulse response (IIR) control circuit, meaning that the autopilot would oscillate infinitely around the trimmed heading. It's a little bit more complex than that. :)

 

It's probably handy then that the helicopter has pitch, roll and yaw damping built in and is flying through a medium which provides a nice pile of friction.

 

Haril's right, osculation is not guaranteed so long as there are the channel dampers, fluid air mass, and maybe some tiny deadzone built into the system as well.

 

Nah, what creates this is the fact that people don't trim... :P

 

Don't trivialize the effect. Trimming doesn't solve the 180 degree bump unless you are actually disabling the hold altogether. It's nonsense speak to counter "the hold exhibits this weird behavior" with "the hold doesn't do that if you disable it" well, Duuuuuuuuuuuuuuuh.

 

Nah. That baby is useless. I chucked it, sorry. :)

 

Turn to target - that's what my pedals are for. I mean, 'look at tgt cursor, turn towards it!' That's hard man, we really need an entire function just for that one - NOT.

 

Keeping ball in middle - yup, pedals again!

 

Heading hold - just trim (ACTUAL trim, ie applied force, not wierd damping and fumbling about chasing some arb heading) properly, and use pedals in the conventional way (step on ball, etc)

 

Hover hold, holding a specific heading - thats what my pedals and trim are for.

 

Sorry, I can still not think of ANY situation that I can't do just fine by simply using the pedals for what they were intended. If God had intended man to fly with the AP yaw channel engaged, He wouldn't have given him pedals! :) Actually, this is all really directed at Kamov, not you guys (you've heard me before).....Kamov, are you listening?! ;)

 

Do not confuse AP HOLD with AP STABILIZATION!!! You don't like the AP hold? OK, then turn off the hold, not the entire channel!!!!

[EtherealN]

Don't trivialize the effect. Trimming doesn't solve the 180 degree bump unless you are actually disabling the hold altogether. It's nonsense speak to counter "the hold exhibits this weird behavior" with "the hold doesn't do that if you disable it" well, Duuuuuuuuuuuuuuuh.

 

In my experience it does - if you trim for the turn (that is, trim while at >3 degrees/s yaw rate) instead of trimming for course (trim before and after turning).

 

EDIT: Very interesting. I just made a flight specific to testing that, and even when trimming for turn rate there still appears to be a bump as described. My apologies. I'll have to do some more testing of this when I have more time available (heading out on a business trip in a few hours so it'll have to wait). Looking forward to your results though.

Edited February 14, 2010 by EtherealN

[Frederf]

It makes sense that the turn rate hold wouldn't have a bump. What's 180° opposite +4°-to-the-right? :P It is even more confusing if there is a bump for turn-rate! Are you sure you exceeded the °per-second threshold to achieve a turn rate hold? Maybe it's all down to how ED coded the whole mess where it hiccups when swapping directions on the backside of the circle, nothing Kamov about it.

 

Small niggle but the hold for the AP involves heading as distinct from course. Course being the path over the ground while heading is the direction one points. The closest thing the Ka-50 has to a course hold is DT while in ROUTE.

 

I wish I could get a raw data output of flight control position at various stages in the procress. It would make realizing how the stab, AP, etc all work.

[EtherealN]

I'm quite sure I was trimmed further than the requirement for turn rate trimming (I was around 10 degrees per second). I would have to eliminate any momentum buildups and such though, but sadly that will have to wait until I get back home.

 

And yeah, I'd love it if it was possible to see all that info in real-time. Probably doable with LUA export ofc.

[sobek]

Haril's right, osculation is not guaranteed so long as there are the channel dampers, fluid air mass, and maybe some tiny deadzone built into the system as well.

 

Duh. Yes i know that, i just wanted to emphasize that the autopilot does not work that simple, of course there is backcoupling and damping in the control system, also full AP authority occurs well below 180° deviation from the trimmed course. There's a differential equation of at least 1st order involved into such a circuit, to counter overshoot.

 

I wouldn't count on friction though, the Shark weighs in at roughly 8 to 10 tons, which gives it quite the fair amount of moment of inertia. Considering the low tangential velocities that would occur in a yaw oscillation, i'd say if it was only for friction, it would go on for hours on end. Dunno what you mean about the deadzone, or how that would counter oscillation.

Edited February 14, 2010 by sobek

[Skkuda]

Originally posted by EtherelN

trim while at >3 degrees/s yaw rate

How do you all calculate this 3 degrees/s yaw rate per second?

Are you estimating it? or are you watching some instrument? because I cant find it.

 

Skkuda