Can anyone explain why we would want hysteresis and slower response as offered by the Futaba 14MZ?

Is it something helicopter guys want?

Jim O

Jim,

What is hysteresis in radios? I know that you have been discussing in the NSRCA group but I didn't have chance to ask.

What is the equivalent in JR 10X radios?

Thanks,

Vicente "Vince" Bortone

Apparently WE asked for it!

Hysteresis adjustment was added at the request of some of the F3A flyers, who had become used to systems with a bit more hysteresis in their systems. This can help eliminate aircraft movement with small inadvertent movements of the sticks, such as when you're holding aileron in a roll and have to add elevator at different points. A bit of hysteresis keeps you from changing the aileron.

[link=http://www.rcuniverse.com/forum/m_5864828/anchors_5864828/mpage_1/key_hysteresis/anchor/tm.htm#5864828]Original post[/link]

I use it to turn on a condition one point of stick movement and turn it off at another. If that is the same hysterisis your meaning.

Bill

Hmmmmmm . . must be a Mode 2 thing . .

Hey, can't you accomplish the same thing by switching to decaf?

Sorry for not quite understanding this. Does this mean that if you're doing a slow roll and your finger moves the ailerons slightly while introducing the elevator or rudder, the ailerons would HOLD that position? In effect, your direct stick movements will not affect the aircraft's flight...in other words, you're not directly flying the plane during the stick positions where hysterisis is in effect for that control input?

My other question is a little more fundamental but also a little rhetorical: Does this and any other "control augmenting" programs like dual rates, and exponential make us fly better or does it actually just hide our "mistakes"?

...or is this another gimmick like the words "organic" and "free range" so we end up having to buy the latest and greatest to gain that "edge"?

Jeff,

You're not gonna start picking on all us mode 2 fliers now are ya??? It would be interesting to know the number of people that fly mode 1 and mode 2... Here in Australia it's mainly mode 1. I must be a rebel...

Back onto hysterisis, I think a better solution would be to use flight phases whereby the aileron is set so that full stick movement is used for slow rolls. It would give me the Sxxxs having wishwashy controls. Like having too much expo.

Cheers
Jason.

Naahhhh . . There is a lot of nice F3A planes here being flown Mode 2 . . what does that mean? . . a lot of nice planes I can't borrow []

Hysterisis . . . hmmm ! I use a lot of Expo on ailerons (about 80%) and about half that on elevators and rudder. I like "soft" ailerons for gentle corrections in a schedule (hate to see jerking of the wings trying to find the horizontal) BUT I like a more "solid" feel on elevators and I set up the rudder to feel the same as elevators so rolling through KE, etc. has the same stick feel (if ya know what I mean )

I am not sure I would like hysterisis. I think it would introduce an element of vagueness that would bother me. From what I understand, it is kind of like expo but operates around stick position where-ever it is . . I think??

Couldn't be too sure until I tried it, I guess.

http://en.wikipedia.org/wiki/Hysteresis

Probably will give you guys more than you ever want to know about hysteresis.

I really cannot understand why you would want it in a control system like ours as it would introduce a level of uncertainty with regards to stick position and corresponding servo position (ie: that relationship would be in constant change)!?.

Thanks for that, Chad . .

I have hard enough time with schedules now . . . imagine doing it with a "magnetic furnace" . .

Hi Vince,

I could tell from your post on the NSRCA list you understand hysteresis, but I thought I'd try to upload a sketch that might make it clear how it works in the 14MZ. Without hysteresis we get a linear plot of servo position vs. stick position that is a single straight line that goes through zero-zero on an x-y plot. With hysteresis we get a loop as shown in the sketch. There are two parallel lines both of which are displaced from zero-zero. To see what happens let's start with the stick moved to the left with the servo positioned at the left defined by point 6. As we move the stick to the right. towards neutral, the servo holds the left position until we get to the right hand line and then it starts moving towards neutral but notice it is still to the left when the stick gets to neutral at point 1. the servo doesn't get to neutral until the stick is at point 2. As we continue to go right the servo follows right but when we reverse stick direction at point 3 the servo holds that position until we get to point 4. Once again when the stick gets back to neutral the servo is still lagging and is now right of neutral at point 5. So we can have two servo positions for the same stick position depending on which way we approached neutral. In real life it is near impossible to keep going in one direction unless you move the stick fast and you end up toggling back and forth between the two lines as you move the stick slowly.

With the hysteresis set to max (32) I measured about 8 microseconds (change in pulse width to the servo) between the lines so the neutrals at point 1 and 5 would be 8 microseconds apart. How this would affect your elevator trim depends on you set-up but it could be significant. I'd like it to be zero so I won't be adding hysteresis to get good neutrals. I may try it to see if I like the feel but it seems illogical to me.

Hope this helps and hope the sketch loads. If it doesn't please tell me how to do it.

Jim O

Hi Jim,

I knew that you will do an excellent job explaining the funcion. For sure, I understand now. It appears that we want to minimize the hysteris in pattern. Now, I have another question or comment. I wonder if this function is created to minimize the effect of some hysteresis already present in the system by introducing some amount of hysteresis in opposite direction. It won't be perfect at all points (ver good around center) but it is possible that the new technology using 2048 is so good that the effect of hysteresis in the control system becomes important.

With my best regards,

Vicente "Vince" Bortone

GEEZ, what is the hysterisis of the a person? Between eyes seeing, brain interpreting, and fingers commanding sticks, the hysterisis in a human is pretty large. Yet I know that I can feel the lag in my JR 10x. That is, when I fly alot (which hasn't happened lately).

But I also don't think I can tell an 8 microsecond lag, no matter how much I fly

MattK

[/quote]


I'm not talking about a lag. I'm talking about a neutral position error as measured at the receiver output. This can be translated to a control surface position error depending on how you've set the ATV and the mechanical advantage in your linkage. It is probable that your plane would climb and dive between the two positions. I think you'd detect it. I know if you had a good servo you would see it as a position error at the servo.

Jim O

I may be absolutely wrong, but I always understood hysteresis was some sort of "moving" exponential. Normal expo tends to soften responde around center. Hysteresis does the same at all stick positions (i.e. "filters" small non deliberate movements requiring a larger deliberate movement)

Jim O,

So for F3a we should set hysteresis to 0 for better servo centering and response should be to 1 for the fastest response?

When setting your deadband are you still using spline with +/- 0.5 either side of center or have you found an easier method ?

Shane

Correct. To convince yourself on the hysteresis, set it on 32 on a channel that has your best servo on it. In my testbed I've got an old JR 4000 Super Servo that is still probably the best in terms of centering and low deadband. When I set the Hysteresis to 32 I could see the servo move in small steps and see the poor centering with the naked eye. No other measurements are needed. We spent a lot of money to get better resolution and faster response so I can't figure out why anyone would want to degrade it.

I haven't been able to fly the system yet but I do plan to do the stick deadband thing. With the +/- 5% I can't feel a problem on the bench but we'll have to see if it is a problem in the air. It is very easy to put in or out so I can try it both ways. I'll try to test the centering both ways with Jerry Budd's meter to see if it is measurable.

Jim O

Hi
I performed a test the other day, involving some new JR DS9411SA servos, using the servo-pulse generating option of the Jeti programming box for the SPIN brushless regulators. With this function it is possible to change the pulse-width in 1µs steps. By holding the servo close to the ear one can hear a distinctive buzz from the motor for every 1µs change in pulse-width. A 10µs change in pulse-width makes for a visible change in output position. From this I am convinced that the servo's of today is very precises and I state that the hysteresis defenitely should be set to 0.
Ola

PS The Jeti-box also has the option to measure the pulse-width coming out of the receiver, measuring servo speed etc. it's a good tool even though you don't use a Jeti SPIN regulator.

Pure hypothetical - A plane requires -2% elevator (push) to hold straight and level inverted flight. Set the elevator hysteresis loop to a total height of plus and minus 1% each side of zero. Now trim the elevator down -1% and go fly. For straight and level flight, pull very slightly and then release - the hysteresis loop of +1% offsets the down trim of -1% and the plane holds straight and level. Now roll 180, push slightly, and then release. The hysteresis loop of -1% adds to -1% down trim for a total of -2% to give hands off straight and level inverted flight.

I'm not proposing this and I would not fly this set up, but it is the only scenario I can imagine where hysteresis might be hypothetically useful. As a practical matter, I would think having non-centering servos would be a real pain in the backside.

Bob