Servos for Heli

[detach8]

Hi guys

I have soem questions regarding servos.

Basically what we've been seeing in specs is the speed and torque. What about precision? I believe every servo has a precision factor. Here's my encounter.

I have BeltCP with el'cheapo Esky servos. I saw Airtronics AT-80A on sale at Jet so I bought it. The Airtronics was rated 0.9s while esky at 0.12s.

I put the airtronics on and even with same servo throw, I set my GY240 gain up to almost 80% and I don't see wagging. I put my esky back and it started wagging a little.

My guess is that the faster servo should require lower gain (because the speed of the servo compensates for it?) or am I wrong?

How do we determine the servo's precision as well?

In the mini Titan thread I've asked several questions about servos and it seems that people who are using servos with speed 0.12s are having their tail hold well. See http://www.daddyhobby.com/forum/show...4&page=4&pp=15

So what exactly makes a good tail servo? What makes a good ccpm servo?

[detach8]

After some research, I wrote in my blog... hope to share:

My el'cheapo servos have taught me some valuable lessons. In my previous post I mentioned about "upgrading" my tail servo to an Airtecnic AT-80A. I had this believe that a servo with higher speed rating would be better for the tail. I mean, who doesn't think so? I was so wrong. It was more of a downgrade. Here's why.

I was doing a test flight of my Belt CP in my room and had to set my gyro gain way above 80%, and still no sign of a tail wag. (Bad, dog!)

Then I swapped it for the stock Esky servos. My tail started to wag at 75% gain. That was a sign of improvement, i.e. the tail servo is doing its job and now the gyro's over-compensating the movement instead.

I stared closely as the servos move and realized that the Airtecnic servo had much coarser movements than the Esky's.

A little research afterwards, I realized the speed specifications (in seconds per 60 degrees) was for the fastest movement for a servo arm around a complete 60 degrees turn. It's not stated anywhere if it includes the delay from electrical signal to mechanical movement, neither does it tell if the servo moves as accurately as we wanted it to go.

I do have a believe, for example, that a servo rated 0.12 seconds for a 60 degrees turn might not move as fast for a 0.5 degree turn as another rated slower.

Let's just put it somewhat like driving cars. Every different make has different acceleration curves. Although the BMW can hit 100kph in 7 seconds but a Honda probably beat it within 30kph due to weight and other factors.

In a heli's tail servo, it hardly moves much to keep your heli stable. If it went 30 degrees clockwise, you'll be in a crazy pirouette.

So, how small a movement can your servo make and how accurately can it move from point A to B?

I read a little deeper and stumbled upon this link on Futaba's website. It describes how digital servos are made to respond faster and more accurately. The graph made a lot of sense as well. It tells me that a good servo provides very consistent torque throughout it's movement until it stops.

A futaba digital servo

With a more consistent torque, changes to your tail will become more linear and thus easier for your gyro to compensate its movement. If your servo had an exponential-ish movement curve, it would become very difficult for the gyro to correct its motion.

I may be wrong. But these are based on my first impressions and thoughts. The physics of it makes some sense as well. What do you think?

[detach8]

Ah mistake in my first post.. it's AIRTECHNIC not AIRTRONICS.

[baby_zyklon]

Your belief was correct, higher speed servo works better for tail. Why leh? Because lets say your tail drift left, your servo is very slow, of course it will take longer to compensate right? So the faster the better. However, its also dependant on many other factors.

The part about wagging at lower gain value= doing its job is correct, but its not an improvement. Wagging at lower value will mean the servo already reach it max gain possible for it. To understand why you must see why is it over-compensating. I try my best to explain but I think I will not do a good job. Anyway, here goes. When your tail drifts left, your gyro sense it and will apply left rudder to correct it. The gyro compensates a little by a a little and when tail reach center again, it will ask the servo to move back to neutral so you don't over compensate. However, lets say the servo is very slow to respond and move back to center, it will overcompensate. So now your tail is in the right. So gyro ask servo to compensate and give servo right rudder, your tail reach center, the gyro sense it, but alas, too late, when servo return to neutral position your tail already at left. This constant correction will give wagging.

While I maybe wrong, I think the higher the gain will make the tail servo gives more movement to stabilise the tail faster, too much gain=too much movement=wagging? The servo horn also play a part. Furthur the distance of link to center screw of servo arm=lower gain possible.

As load increase, your servo might slow down a little, especially if it is analog servo because analog servo don't produce the max rated torque immediately off center, it takes a few degree of movement. Digital servo is around 1degree of movement for max torque.

Try about the 30degree part, but I still think it is possible to use quite a bit of angle if you do 3d.

Yup, how small a movement make is resolution of servo. Coreless servo usually have better resolution than cored assuming both non digital.

[baby_zyklon]

Try about the 30degree part, but I still think it is possible to use quite a bit of angle if you do 3d.

I mean true not try.

[detach8]

well i said i was wrong because speed isn't the only factor. so in summary a good servo is both fast AND precise.

[junyang]

yep..precision is impt too.. thats why some people choose to go digital servos.. as their resolution is much higher.. than some analog servos..