Need help understanding gain

[chouserchop]

Hello - I've read about everything I can get my hands on regarding "gain" and I'm still not sure exactly what it does, e.g., 70% gain. What does "70% gain" mean as opposed to, say, "30% gain?" I've been experimenting with different gain settings on my Futaba 6EX and have to say that I don't really notice the heli behaving much differently from 30 to, say, 70.
Now under some conditions I notice that when I throttle up hard the nose will sometimes swerve to the left. If I increase the gain will that "swerving" be corrected?
Thank you.

[DebianDog]

Man I wish I was a better searcher. Someone wrote up a nice "exactly what gain is" here somewhere...

[DebianDog]

From: http://www.raptortechnique.com/gyroconfusion.htm

What is gyro gain?
==============
This refers to the sensitivity of the gyro. When the gyro senses an unwanted movement it commands the tail servo to move in the opposite direction to compensate. How much it tells the servo to move is the 'gain'. Ideally the amount of gain should match how much the helicopter was rotated so that it stays pointed in the same direction and does not move. If the gain is too high then the helicopter over compensates. The effect you will see is the tail will bounce back and forth (aka wag). If the gain is not enough then you will notice the tail does not hold very well. When setting the gain, you want to turn it up until you see the tail 'wag' (bounce back and forth) then turn the gain back down until it stops.

[fogger]

Here is an analogy that might help. Think of your gyro as a record player, amp, and speaker.

The turntable (specifically the needle) is the ring sensor: it picks up vibrations induced in the ring (record) by rotating the heli, but also it picks up vibration noise from the environment.

Gain is the volume control on your amp, it controls how amplified the signal from the turntable (ring sensor) gets.

The speaker is the output to your servo.

So if you have a low gain, only the loudest vibrations are "heard" and acted upon by the Amp output section to drive the servo. If it's too low, even vibrations caused by heli rotation will not be picked up (hence tail drift or kicking out on power changes). On the other hand, if the volume is too high even vibrations from external sources will be heard by the servo (hence wagging).

The trick is to select the highest gain you can without picking up noise, IE without inducing the wagging.

-Fog

[DavidH]

I will try and give you a brief description.
Think of gain as being sensitive to movement. So for example, 70% gain is more sensitive to movement than 30% gain.

Don't know if you have the Futaba GY series gyros. But they hold very well with what is considered very little gain. The GY gyros will hold well with the gain in the high 30's if the tail is set up correctly. It will also hold just as well with the gain in the high 70's.
Now here is where the problem starts with the higher gain setting. With the higher gain setting the servo is being worked more than with the lower gain setting. So this at times heats up the servo and it will fail prematurely. So just my opinion, it is always best to use the lower gain setting that will still hold the tail.

As for the nose moving to the left with a quick movement of the throttle and pitch. I really don't know of any gyros that will compensate for it. This is a result of the throttle and pitch curve being set up incorrectly. Or if they are setup correctly, the motor is not tuned properly.
What is happening when you raise the throttle/collective stick quickly. The pitch is ahead of the throttle and it is creating a lot of torque. The tail is not turning fast enough to over come this torque. So the nose will go left, till the tail can compensate for the torque from the main rotor. You want the throttle to lead the collective pitch added, With a governor on the throttle, this will keep the rotor rpm constant and torque changes will no occur when the pitch is changed. Because the tail is turning the constant speed also and does not have to catch up to overcome the torque.
Hope you can understand this.
David

[HFG]

Here is an analogy that might help. Think of your gyro as a record player, amp, and speaker.

The turntable (specifically the needle) is the ring sensor: it picks up vibrations induced in the ring (record) by rotating the heli, but also it picks up vibration noise from the environment.

Gain is the volume control on your amp, it controls how amplified the signal from the turntable (ring sensor) gets.

The speaker is the output to your servo.

So if you have a low gain, only the loudest vibrations are "heard" and acted upon by the Amp output section to drive the servo. If it's too low, even vibrations caused by heli rotation will not be picked up (hence tail drift or kicking out on power changes). On the other hand, if the volume is too high even vibrations from external sources will be heard by the servo (hence wagging).

The trick is to select the highest gain you can without picking up noise, IE without inducing the wagging.

-Fog

Sorry Fogger but this is all wrong. The gyros we use are control feedback systems, I will attempt to explain the basics.

The mechanical gyro (the sensor itself) outputs the rotational rate of the heli. The control system takes the difference of this signal and the pilot's signal. The difference is the undesired rotational rate and is the error signal. The error signal is amplifed by your "gain gain" before being used to control the servo.

Lets take a case as an example. The tail begins to drift to the left, the gyro outputs this small drift. Assuming the pilot has given no control to move to the left the error signal will be the opposite of the gyro output. This is then amplified and fed to the servo, the servo changes the pitch of the tail blades to counter the drift to the left. If the gain was set right the resulting rotational will cancel out and the tail will stop moving. If the gain was too high the tail will start to drift to the right now, again the gyro will try to compensate. But with the gain too high it will start to drift to the left again, this produces a waggin in the tail.

Lets take another example. Your pushing the tail really hard and it starts to drift to the left but you dont want it to move. But your tail pitch is already at the maximum so your gyro has no way of correcting. The drift will continue to grow until the tail blows out.

A heading hold gyro adds an additional stage in parallel with the amplifier. The additional stage is an integrator. Because the amplifiers job is only to stop the tail from rotating it doesnt care where the tail stops moving. The integrator is able to keep track of how far the tail moved and will move it back.

Hope this is clear.

[fogger]

you appear to be saying they change sensitivity on the output side of the circuit, but I believe it makes more sense to change sensitivity on the input section. Obviously the ring (or piezo) sensor outputs an extremely small ac signal which must be amplified, and presumably converted to a digital signal. Why would you set up a complex algorithm to adjust how much the servo is driven when you could just adjust the amplification of the sensor output signal and achieve the same result?

as in your signature, I may be wrong, but I doubt it.

-Fog

[HFG]

Don't confuse signal conditioning the output of the gyro sensor and the control feedback amplifier. The signal output of the gyro itself isn't that small as the pickup electronics for the gyro is built into the gyro integrated circuit itself (usually includes built in test functions and temperature correction). This signal is easily converted with a ADC so it can be compared to the PPM signal (this also has to be converted into digital first). The result is the error signal which will be amplified by the "gyro gain" ( a simple multiplication in the digital domain). The gyros in modern "gyros" (we need better names) are MEMS based not peizo-electric. There is a mass thats forced into oscillation in the "drive axis" (a large rotating mass can't be used as traditional gyro as no MEMS based bearings exist). Any rotation about that drive axis will result in a coriolis force (virtual force). The coriolis force will produce an oscillation in the "sense axis" (90degrees to the drive axis). To drive the drive axis into vibration they use electrostatic force of alot of comb fingers. These same comb fingers are also used to measure the sense axis. As the comb fingers move closer together their capacitance increases (and vice versa), this capacitance is converted to a voltage and demodulated. The output of the demodulator is amplified before being fed outside the chip. Hope that clears my previous post up.

[fogger]

Interesting, but why would they amplify the delta signal after the comparison, rather than amplify / attenuate the sensor output before? It seems to me it's just as easy to alter the magnitude of the error signal (sensor output) before the addition step to let one vary the sensitivity.

Also I believe that logictech are still using piezo. Only the futaba and the latest generation of JR's are SMM / MEMS afaik. Also - also the sense vibration axis is 45 degrees out, not 90. I've read several white papers on the mems. Here is an interesting new one being developed, or at least proposed :) :

http://users.ece.gatech.edu/~ayazi/pub/JMEMS_Gyro.pdf

-Fog

[BarracudaHockey]

To the original poser of the question. Read DavidH answer. That is what you need to know without the confusion of a masters level discussion on it.

More is not always better and turning up to tail wag and backing off a hair is old school, run the gain you need to hold the tail and no more and your servo will thank you with long and faithful service.

[HFG]

Interesting, but why would they amplify the delta signal after the comparison, rather than amplify / attenuate the sensor output before? It seems to me it's just as easy to alter the magnitude of the error signal (sensor output) before the addition step to let one vary the sensitivity.

Also I believe that logictech are still using piezo. Only the futaba and the latest generation of JR's are SMM / MEMS afaik. Also - also the sense vibration axis is 45 degrees out, not 90. I've read several white papers on the mems. Here is an interesting new one being developed, or at least proposed :) :

http://users.ece.gatech.edu/~ayazi/pub/JMEMS_Gyro.pdf (http://users.ece.gatech.edu/%7Eayazi/pub/JMEMS_Gyro.pdf)

-Fog

The logitechs also use AD gyros which are MEMS based. Traditional MEMS gyros the sense and drive axis are 90degrees to each other unlike the ring gyro. That ring gyro is a few years old but has some interesting properties. I've since left the MEMS game so i haven't followed up on anything in the last 2 years.

Amplifying the gyro signal before the error amp isn't the same as amplifying the error signal. I'd rather not get too deep into control systems as this what i do for my day job (i design phaselocked loops) and i'd rather not take my work home.

[fogger]

Cool. I'm sure I saw a white paper from Futaba describing their use of the ring type mems. I'd bet this is the one they (and JR) use... Japanese companies tend to stick together and specify other Japanese components.

http://www.sssj.co.jp/en/products/gyro/crs07.html

http://www.sssj.co.jp/en/products/gyro/principle/index2.html

-Fog

[teeson]

so to help me the newb, who is totally lost by all this, i say what i think is the way to setup gain...
on gain theres a sweet spot, too high the tail wags, to low and you lose tail control. you want to try and find the sweet spot when doing so which changes for each heli based on any number of variables.

so.. am i right or way off base?

[BarracudaHockey]

Run as much gain as you need to hold for what you can do. If the best you can do is hover you wont need as much gain as someone flying backwards hurricanes and inverted tail slides.

Some will tell you to crank it up till the tail wags then back off a hair but thats old school and hard on the servo.