|
The Gyro
I happen to really like the design approach of the CSM 540. Its an extremely simple unit, with little more than a microcontroller, piezo sensor, and associated signal conditioning and voltage regulation circuitry. Through the PC interface, the gyro offers a great deal of flexibility for tailoring the gyro to various setups, allowing it to extract premium performance from any machine its installed in. Furthermore, the CSM 540 is a true rate-demand gyro, unlike the Futaba gyros. This means that, when working properly, the helicopter will rotate at a constant rate for a given control input, regardless of speed or direction. The Futaba gyros use a system that automatically trims for a constant heading at low rotation rates, but once the rate goes up, it essentially becomes a conventional gyro, and rotation rate for various stick inputs will vary with speed and direction.
One of the most polarizing features of the CSM 540 is its ability to be adjusted through a PC interface. Some people will say that they don't want to spend hours adjusting the settings through the PC interface. Others, like myself, like the fact that you can cram a great deal of versatility into a unit, without carrying the overhead of things like LCD screens and the necessary user interface software. I believe that if CSM had explained the PC interface adjustment procedure as thoroughly as I have in this article, there would be fewer people disheartened with the CSM PC interface, because the process is really pretty quick and painless if approached methodically.
I don't want to compare the CSM and Futaba gyros for the sake of determining the better of the two, nor do I want to convert Futaba users to CSM users. Both gyro designs have their advantages, the CSM's being better flight quality (when properly adjusted), and the Futaba's being drastically simplified setup and better performance out of the box. Given the overwhelming popularity of the Futaba gyros, my priorities are definitely in the minority. But, if this article is helpful to anyone at all, it has served its purpose.
Setup Foreword
In this article, I will assume you have read the directions that came with the CSM. I expect that you can use the auto setup and adjust gyro settings via PC, and have followed the basic mechanical setup instructions (some of which will be thrown out the window straight away). The helicopter should be flying, and you should be able to set gains and change between modes. Its also invaluable to have access to a laptop that either boots DOS, or has a non-NT windows installed, ie. 95, 98, or Millenium. Without a laptop, adjustments can't be made inline at the field, and it could take many flying sessions to converge on the proper settings.
Initial Mechanical Setup
It is a gyro's job to extract maximum performance from the mechanical setup it has been given. Give a gyro a lousy mechanical setup to begin with, and it will never perform well. Give the gyro a great setup and, in the case of the CSM 540, all you have to do is adjust gyro parameters to match. This may sound a bit counter-intuitive, eliciting the question, "If I've already got a great mechanical setup, why should I have to play with the PC interface at all?".
I can only answer that by saying, "That's just how the CSM works." This is a primary tenet to CSM setup, and the primary reason people have flocked to the Futaba gyros in droves. The CSM works by trying to predict the response of your helicopter to a given servo movement, based on its gamut of settings in the PC interface. Regardless of how well refined your helicopter is, chances are it won't respond to a servo movement the same as any other helicopter. Granted, you can get acceptable performance by adjusting your helicopter mechanically to respond similarly to the helicopter the gyro was designed for (rumored to be a Robbe Millenium), but rarely will you get the response to be identical. And, as such, you'll rarely get peak performance.
Nonetheless, there are certain aspects of mechanical setup that must be paid attention if you want great gyro performance. These aspects can be put into one of three categories: linkage geometry, mechanical gain, and system stiffness.
To ensure good linkage geometry, first make sure the servo arm is nearly perpendicular to the tail boom when the tail pitch slider and servo are at their neutral positions. Adjust this by changing pushrod length and the position of the servo arm on its servo splines. Next, if your tail rotor has adjustable pitch (like some Century and XCell designs), adjust the tail rotor pitch so that the tail rotor blades are pointed with their leading edges either straight forward or slightly towards the rotation direction of the main rotor when the pitch slider is centered in its travel. For example, on a clockwise-rotation helicopter (which includes 95% of the helicopters sold in the US), the tail rotor blades would be pointed with their leading edges slightly left with the pitch slider at neutral. If your helicopter doesn't have this adjustment, don't worry, the designers have probably taken it into account. These two adjustments attempt to make the tail rotor response symmetric both left and right. The CSM 540 can accomodate some pretty extreme asymmetry, so this adjustment just takes care of some of the more severe cases that can cause problems.
Next is the issue of mechanical gain. There are two factors that affect mechanical gain, the servo arm length and the tail rotor blades. Servo arm length seems to be pretty easily generalized to a wide variety of setups, and the ball link should be positioned nominally about 18-20mm from the servo center. Go out a little further if you have a slower non-coreless servo on a high performance setup, and slightly further in for a conservative setup on a high performance servo. For a high speed servo on a high performance setup, or a slower servo on a conservative setup, 18-20mm works just fine. Tail blades have a similar effect on mechanical gain, with the exception that changing tail blades changes the overall control power of the tail, where servo arm length does not. Long, thick tail blades will give the best holding power, but may require finer tuning to make the heli stop well and avoid nervousness in tail response. Stock plastic tail blades are fine for more conservative flying, and generally give slightly smoother response.
Gyro performance potential increases as the link between the gyro and the helicopter's response becomes more direct, which is a property of the system I like to call its "stiffness". Stiffness is affected by the rigidity of the helicopter's tail boom, the rigidity of the gyro mounting scheme, and the servo speed and associated control slop and binding. You want to increase this stiffness as much as possible by making sure boom supports are rigidly tied to the frames and tail boom, and that the gyro is mounted on a rigid platform. Furthermore, it often helps to mount the gyro using a firmer tape than that supplied by CSM, though firmer tapes will lead to more vibration at the gyro, which tends to upset the internal drift cancellation on many units. It might, at first, seem like a stretch to call servo speed and control slop a "stiffness", but it has the same detrimental effect that tail boom rigidity and gyro mounting rigidity have, it increases the delay between what the gyro's commands and the corresponding reactions. Therefore, it is also extremely useful to have a quick tail rotor servo with a free-running, slop-free linkage to the tail rotor. The addition of a digital high speed servo will greatly improve gyro performance. There should also be ball links on both ends of the link to reduce slop. Carbon fiber tail pushrods aren't necessary, just make sure that whatever pushrod you have is supported well enough to keep it from bending.
In summary, here are the things you can do mechanically to make the CSM 540 work better:
| 1. | Get a high speed tail rotor servo. A coreless, digital servo made explicitly for gyro use works best. |
| 2. | Be sure that your tail rotor controls are smooth and slop-free. |
| 3. | Try to make your linkage setup as symmetric as possible about neutral. |
| 4. | Set the length of your tail servo arm appropriately. |
| 5. | Make sure your tail boom and servo mounting tray are as rigid as they can be. |
| 6. | Consider using a firmer mounting tape, especially if other aspects of your setup are less than ideal. |
PC Interface Parameters
I assume that if you're reading this, you tried the default CSM setup and have found it unacceptable. Therefore, after initial setup, I think its best to dive head first into adjusting paramters with the PC interface. This process can be done quickly and systematically if you understand what all the parameters do and have a method for setting each one.
Many parameters have been poorly or incompletely explained in CSM's documentation. I'll try to clarify the ones that need clarification, but I still assume tht you have read the manual and know what most of them do. We'll begin with the parameters on the "General Parameters" screen.
Throw Limits 1 & 2-The limits of physical travel of your servo. Interestingly, these have an undocumented feature, in that they are also the GAINS each side of center. They can be used effectively to balance out any asymmetries in tail rotor setup by changing the gain each side of center.
Look Ahead On Motion/Stick-Determines how the gyro tries to predict future motion of the helicopter based on how its moving at any instant. The default setup does not enable look-ahead on stick, because doing so would require the use of the PC interface to adjust look-ahead gain, whereas the look ahead on motion is self adjusting based on gyro gain. Look ahead on stick becomes an invaluable tool for making the heli stop crisply.
Glitch Limiter Window-This smooths out control input. It does the same thing as turning up the servo delay on some radios, but has the added advantage of reducing the effect of glitches in PPM radios. You want to have this enabled, because it keeps the heli from accepting stick movements that change faster than the heli's rotation rate can.
There are two flight mode parameters screens, labeled 0 and 1. Mode 0 is set up as a conventional mode by default, and mode 1 is heading hold. The parameters are as follows:
Look-Ahead Gain-Controls how powerful the yaw rate prediction controls the gyro. This setting is most useful for adjusting for crisp stops when combined with look-ahead on stick.
Conventional and Heading Lock Gains-These two gains must be balanced for good tail response. If one is increased, the other should be decreased. Heading hold gain tends to make the tail feel locked in, while conventional gain keeps response smooth and prevents hunting. There will be an optimum value where the heli holds well while still responding smoothly.
Exponential Gain-Adjusts the amount the gains increase when large tail commands are issued. Higher values tend to settle down the tail during small disturbances, such as collective changes, while preserving stop quality. This parameter doesn't interact much with the others, and doesn't do much, so it can safely be "played with" until a best value is found.
Max Yaw Acceleration of Helicopter-Similar to the way heading lock range keeps the heli from trying to hold harder than it can, this keeps the gyro from trying to accelerate faster than it actually can. When things are working properly, this parameter shouldn't be necessary, as long as it is set higher than the actual max acceleration you demand in flight.
PC Interface Adjustment Procedure
This is the part of the process with the most importance and the least documentation, and one that CSM could have made much clearer and more straightforward. I will present a procedure for setting all of the gyro settings in a logical manner, and you will find that you can set up all the gyro settings almost as quickly as you can set your pitch and throttle curves. To start out, reset all values to their defaults in the PC interface. This will provide a clean slate to work from.
We will start with the parameters in the general parameters screen. To begin, you should do the following settings:
| 1. | Take the average of throw limit 1 and throw limit 2, and set both throw limits to this value. You may then need to adjust tail pushrod length to avoid binding at one of the extremes, so be attentive when turning on the gyro again. |
| 2. | Make sure look-ahead time is "short". |
| 3. | Set look-ahead to be on both motion and stick. Look-ahead on stick can dramatically improve stops, but requires a careful setting of look-ahead gain. |
| 4. | Set the glitch limiter window to 75%. This is really the parameter that adjust how quickly your helicopter will change direction, and you can lower it for faster stops if you find later that your hardware is capable of them. Conversely, if you can't seem to get rid of bouncy stops, increase this parameter. |
Next, choose "OK" and move to the Flight Mode 1 menu (I assume that all you will use is flight mode 1, because once you go to heading hold, you'll never go back).
| 1. | Set look-ahead gain to 0%. |
| 2. | Set conventional gain to 85%. |
| 3. | Set heading lock gain to 60%. |
| 4. | Set heading lock relaxation rate to 0%. |
| 5. | Set stick exponential and stick linear sensitivity as desired. |
Now, shut down the PC interface, fly around a bit, and notice how lousy this setup flies! If this setup happened to work for you, its miraculous, because it should be extremely poorly tuned. What we can glean from this setup, however, is a baseline configuration and a gain setting.
Set the gain in mode 1 by requesting some stops from a moderate speed piro. If the helicopter oscillates more than 3-4 times when stopping, turn down the gain. If it doesn't oscillate at all, turn it up. What we're looking for is a setting that provides a fair bit of hunting when stopping, but not prolonged hunting.
Next, try stopping from both directions. The helicopter may hunt more from one direction than the other, indicating that the gain is higher in one direction than the other. To fix this, use the PC interface to set the throw limits (General Parameters Screen) different for the two directions. Try to offset any increase in one travel with a decrease in the other travel to maintain the same overall travel. Try making adjustments at about 5% at a time, and stop when the hunting is pretty much the same when stopping from both directions. Be sure to adjst the tail pushrod length after each throw adjustment to prevent binding.
Now that asymmetries are taken care of, it is time to set the heading hold and conventional gains for flight mode 1. These gains are easiest to adjust if you keep their sum constant, ie. if you increase one, you decrease the other. You try to find a balance between the two that gives just the amount of heading hold gain needed to do the job. Too much heading hold gain, and the tail will never be smooth, and too much conventional gain, and the tail will never hold well. The amount of heading hold gain in my default setup is almost certainly too low, which is good, because its easier to see changes when increasing heading hold gain as opposed to reducing it.
Start by flying in fast forward flight, and do a slow pirouette (about 3 seconds per revolution). If the helicopter rotation slows down when the tail goes upwind, and speeds up when it goes downwind, the heading hold gain needs to be increased and the conventional gain needs to be decreased. I typically will make a 3% increase in heading hold gain and 3% decrease in conventional gain for each trial. I continue this process until I see no discernable change in rotation speed as the helicopter rotates in fast forward flight, being sure to check rotation in both directions. I then give an extra 3% change to both gains for good measure. If you get a constant pirouette rate with my default setup, be glad, because your mechanical setup is better than most.
At this point, the helicopter should be holding like a rock, but stops will probably not be crisp. To improve stop quality, increase the look-ahead gain in flight mode 1 in 3% increments. After a few trials, you should find a value that gives the fastest stops without any hunting.
If you're ambitious, you can adjust exponential gain to try to get both fast piros and slow piros to stop quickly. Its a trial and error process, and more exponential gain tends to make fast piros stop faster. You will probably need to re-adjust look-ahead gain a bit after doing this.
That should be it. With a laptop, all of this should be achievable in a day at the field.
In summary, to set the throw limits, heading hold, conventional, and look ahead gains, do the following:
| 1. | Set transmitter gain to give about 3-4 oscillations when stopping from a moderate speed pirouette. |
| 2. | Check stops from both directions. If one stop direction oscillates more than the other, adjust throw limits 1 and 2 by 5% at a time, keeping their sum constant. Readjust the pushrod length to prevent binding. Repeat this step until you get a similar amount of oscillation stopping from both directions. |
| 3. | Fly in fast forward flight, and do a slow pirouette, looking for changes in rotation rate as the tail goes upwind and downwind. Increase the heading hold gain and decrease the conventional gain in 3% increments until the pirouette rate is constant, then add another 3% to your heading hold gain and remove another 3% from conventional gain. |
| 4. | Increase look-ahead gain in 3% increments until stops from a pirouette are crisp. Be sure to check both directions. |
| 5. | Adjust exponential gain to try and balance both fast and slow stops. Look-ahead will probably need readjustment after changing exponential gain. |
|
