Gyro Gain

[Steve]

I've always used CSM gyros up until recently, and am now trying a Futaba 401 for the first time. I've been using it for a few weeks in a reasonably new, quality heli that has had no prangs to date, so everything is still still in good shape.

I'm getting tail wobble in high speed forward flight which I believe is a sign of too much gain on the gyro - please correct me if I'm wrong.

However, I've been tweaking the gain down and down, and it's now down to 30%. She still climbs dead straight under full pitch, but I'm just getting a bit nervous about how low I should allow the gain to get. I've also got the 'delay' turned up to 100% having followed the instructions for stopping the tail from wagging at the end of a pirouette.

My CSMs are usually set around the 60-70% mark.

Tail servo on this particular model is Futaba 9250.

Any advice would be gratefully received.

[antsis]

The 9250 is a digital servo and delay is generally used for standard servos. Here's the description from Futaba;

"Delay affects how rapidly the gyro responds when input is given, and when the stick is returned to neutral. If the delay is too low on return and you have a slow servo, it will think the servo should have been there already and give more input, then overshoot, then give more input to correct, and so on, causing a wag. Faster servos can handle smaller and smaller delays."

I'd start by putting the delay back to 0. Set up the movement on the servo and then connect the gyro. Adjust the limit on the gyro to avoid binding. What tx are you using? If you have the 9C, the CSM gyros are usually setup in the end point menu whereas the 401 is setup in the gyro sens menu. I start with AVCS and NOR set to 50% and the end points at 100%. The end point setting will determine the speed. I've had tail wag in forward flight, but adjusting the gain by a couple of % usually cures this.

Also check there's no binding in the rudder linkage.

[Ade_Law]

generally i recomend using the 9254 on the tail as it was designed for the job.

9250 seems to be a medium torque highish speed servo so its going to be a little bit more tricky to get setup nicely.

what model is this on?
how long is the servo horn?
is the horn at 90degs in normal mode?
does the tail drift in normal mode?
what is the limit pot set to?

delay pot should only be used on non digitals. It is sounding like your servo horn is too long, which reduces the value of the limit pot and gain value.

Ade

[Steve]

You're right about the servo speed and torque. I originally bought it for my Shuttle which had a notoriously stiff linkage through to the tail blades, so I went for one that was reasonably fast with a bit more oomph as well than a 9254. I use 9254s and 9253s in my other helis.

This model is my JR Voyager 30. The tail mechanism is perfectly alligned and at 90 deg for stable hoverring in normal gyro mode. The output ball on the servo arm is 13mm from the servo centre (as indicated in the heli instructions) - on the last hole of one of the standard Futaba 6 arm horns. My Tx is a JR 9Xii. The rudder limits are set to 95% each way and the gyro limit pot is set at 100%, which is just within the binding points on the tail pitch slider.

I've increased the delay from 0 to 100% in 25% steps. The initial increase from 0 to 25% made a very slight improvement in the amount of bounce at the end of a pirouette, but from 25 to 100% has made no further improvements.

[Steve]

Anybody help here would be gratefully received.

[Raptorite]

The next step is probably to bring the ball in one hole on the servo arm, this will mean the gyro wont be quite so sensitive so you should be able to increase the gain a little. If the tail is pretty well locked in now, there is no point changing the gain just yet, but it will certainly help out with the hunting problem.

Just to clarify one or two things, with the gyro in normal mode, the servo arm is at 90 degrees and there is somewhere between 3 and 5 degrees on the tail blades? An easy way of telling is if the model hovers holding its nose straight without drifting?

Cheers,
Pete.

[Steve]

The tail is spot on at 90 degrees with no drift in normal mode - I made a point of setting this a couple of weekends ago.

Also did the same with my other 2 helis.

I'm learning....... slowly!

[Raptorite]

OK, thats great, so the next step is to bring that ball in one hole on the servo arm, this will have the effect of desensitising the gyro a little which should eliminate your hunting problem.

Cheers,
Pete.

[Steve]

That would make sense Raptorite.

In science, they say that you should never accept the accuracy of readings of an instrument below one third of its nominal operating range - hence the uncertainty about the current state of affairs with this gyro at 30%

I think I understand your point about bringing the ball in one hole - this will (correct me if I'm wrong) reduce the output of the servo, and therefore make the gyro work harder for it's living, hence requiring a higher gain value to get it back in line and bring it up above my scientific 33% limit.

[Ade_Law]

normal gain setting is around 50-60% for the 401

Ade

[Steve]

Thanks Ade.

Do you agree with the synopsis so far?

[Raptorite]

Yes, you are sort of there, the hunting problem is caused because by the gyro is sending a signal to the servo to move the tail left or right, but because the servo isnt as fast as the gyro would like, it takes longer than it expects to make the correction, therefore the gyro then says, oh thats too much and then tries to move the tail back the other way. This process then repeats over and over again, hence the hunting.

By moving the ball in one hole on the servo arm, you are moving it closer to the centre of the arc of movement of the arm, therefore it will give the impression the servo is operating faster than it actually is, due to the fact that the ball has less distance to cover the further down the arc of movement it is.

This will mean the gyro has less work to do and should therefore settle it down, thus eliminating the hunting problem.

I hope this makes sense!

Cheers,
Pete.

[Steve]

Does this account for the hunting in fast forward flight as well as at pirouette stop?

If so, why?

[Raptorite]

Yes it will, the reasons are that in fast forward flight the gyro is working quite hard to keep the model in a straight line, as there is a fair amount of torque being generated by the rotor head which is also constantly changing with every pitch and cyclic input. This means the gyro is constantly modifying the tail rotor thrust by changing the pitch of the tail blades, but here again as the servo doesnt react as fast as the gyro would like, by the time it has moved to where the gyro has told it to go and paused there briefly (we are talking milliseconds here), by then the gyro has already worked out that the tail is moving too much one way as it "thinks" much quicker than the servo can react. It therefore sends a signal to the servo to move the tail back the other way to compensate for what it already see's as an over correction. This process then repeats over and over, therefore you see the hunting in forward flight.

The bounce you will see after a stop of a piro is again because the gyro is putting in the opposite input to stop the piro, but by the time the servo has reacted the gyro is one step ahead and thus determines that it has over corrected and therefore needs a correction the opposite way. This gives the impression of a tail bounce.

The delay pot on the gyro is for exactly this reason, as it works by slowing down the speed the gyro "thinks", effectively tuning the gyro and servo to operate at the same speed. This is most useful if you are using a standard coreless type servo, typically these days only used in micro electric heli installations such as I have in my 2 zoom 400's which run 401's.

By moving the ball in on the servo arm you are effectively speeding up the servo, which should then match the speed the gyro "thinks" more closely, therefore eliminating the hunting issue.

I hope this helps.

Cheers,
Pete.

[Steve]

Ok Raptorite - I'll move the ball in one hole in the morning and set the delay back to 25% (for starters) and report back.

Cheers.