I forgot to add one cause of hunting that I solved on a particularly recalcitrant Goodyear racing plane made from the SIG Buster kit. It is a happily resolved case, and a good one to think about if you're having issues.
10. Transient separated airflow caused by poor shapes. My SIG Buster has been a real workhorse now for years, but it hunted badly for a long time. I tried all the usual fixes like shifting the CG forward (and then aft), bigger HS/elevators, tighter pushrod guide, engine thrust line, and so on. The controls were always quite free. None of this had a positive effect. So I sawed off the ¼” thick wing and replaced it with a 3/8” thick semi-symmetric wing. This actually helped some. Note that there are actually three different things happening, or potentially happening, at the same time with this rework. The airfoil is different; the wing/tail decalage may be different (I tried to keep the zero lift line parallel to everything); and ….? This improvement was not simply a fluke. My racing partner liked what he saw and he scabbed another ¼” plank on the bottom of the wing of his Buster and reshaped the airfoil. While his Buster never hunted quite as bad as mine, his improved also. But I still was not happy. What made a major improvement was putting a much larger radius on the front edges of the cheek cowl. This was a clean experiment, since no other changes were made at the same time. Note that this cheek cowl was a modification of the SIG kit added to stiffen the nose for improved engine performance. The basic kit would never have had this issue to begin with. This aero buffeting source is one reason I am not a fan of bluff features such as the canopy on the Nobler. Another feature that makes me leery is a large cooling inlet with an exhaust ramp that has a large angle. This creates a large turbulence zone and may blanket any surface that is in its wake. What you want is something closer to parallel flow ejection. Go look at cowl exhausts for light planes to get some ideas. A particular feature on a particular plane may not affect things at all—but when you go fly it is trial by error. In a kit you presume your risk is low because any problems should have been solved by then. But the design may not have much margin relative to stability (freedom from hunting) and the mods you make might not be aerodynamic improvements.
Other comments on the discussion so far:
There is some pretty good anecdotal data from some pretty experienced guys that putting slop into the elevator controls tamed hunting. What is hard to confirm is that the controls were equally free before the slop was added. If the controls were slop-free but had significant friction initially, then they might be attributing the improvement to “slop” and not “sufficiently friction free.” So the experiment is not really conclusive or complete.
Relative to the discussion on control friction and ball links and rod ends: If the plane has plenty of margin on controls performance, then a bit of friction shouldn’t be noticed. A good, straight design with no other likely causes may be immune. Now compare size. If you used the same exact ball link setup on a Ringmaster at 28 oz. running slow lap times, the line tension needed to overcome friction is a much larger percentage of available tension than on a 60 oz. ship running ProStunt everything and using the exact same controls. The use of smaller bellcranks on the smaller plane makes this even worse. So if (put your favorite ProStunt name here) uses XYZ controls, they must be the best, and better than you need for your 28 oz. Ringmaster—well, maybe not. What did the same ProStunt guy use on his Ringmaster?
Something getting lost here is the stick/slip action that is likely occurring as one root cause. It takes a certain control input force to overcome the stiction. (Also known as breakaway friction, which is always higher than running friction.) When the contacting surfaces break free the friction goes down. It is highly non-linear. So the input from the pilot goes past what he intended, and he immediately has to try to compensate in the other direction. No pilot can do this. (Think of our setup as “servo between the ears” which is pretty accurate. But because of this, the bandwidth of the system is really low!) As Tim alluded, even creating software to simulate this plant response is problematic because it has so many dependencies like temperature, vibration level, humidity (nylon is hygroscopic), whether the rubbing surfaces are “contaminated” by engine exhaust or brand new clean, and so on. So the solution is mechanical, not in software, as was the root cause—you have to reduce the stiction. I have had pretty good luck with using a kerosene/turbine oil mix on ball links and rod ends. You can always try it to see if it helps. I would not agree that “…a little bit of friction…is [adding] damping and…assists in stopping hunting.” If there is no control slop, then all friction can be doing is turning the pilot into a bang-bang controller with the range large enough to visibly see the result of the plane going up and down. This overshoot characteristic is similar to the heater control in your house—well, at least in my old house. If you don’t want to alternate between being hot and being cold, then you have to reduce the deadband—and you can’t unless you can measure smaller temperature differences and you can make smaller adjustments. Just like our pilot with servo between the ears. The only way he can make smaller control inputs is to have less friction in the controls and thereby reduce overshoot.
Note that the sticky controls are on a differ leg of the root cause tree than the aerodynamic disturbances than might be root cause like my Buster example. I spent a lot of time “improving” the controls and it had no effect, because the controls weren’t the cause.
Relative to Ken’s replacement of controls and his experience that hunting was reduced (eliminated?) even though the ball links (rod ends?) may not have allowed a “free dropping” low friction condition: note that you like changed multiple relevant things at the same time, whether you intended to do this or not. With a new pushrod or horns or bellcrank the rigging has been changed. Just as a minimum example, if it was a 4-40 ball link, your increment to get things to line up is one full turn. That is .025” or an angle of about 1.4 degrees for a 1" horn height, and would scale down with increasing horn height. This sounds small and is difficult to measure without fixtures. But for reference, some of the ProStunt guys argued a lot about using 0.5 degrees positive on the stab during build. Not 1 degree or 1-1/2 degrees. So even if you did not intend to change the rigging, there is a limit on your ability to measure, and the prior and new setups won’t be the same. It may be very close if both measurements were highly precise. Then the question is how much margin did the plane have to this adjustment? If a lot of margin, then it suggests this is not an additional cause. So we reject it out of hand, or discount it. But it is real and in some cases of hunting, maybe it is the cause.
I don’t understand Ken’s comment about one cause being “pilot light.” Can you explain so we understand your thoughts on this?
I hope we get some good reports from the Motorman. I’m looking forward to hearing whatever fix he comes up with! Sending some good weather your way for testing….
Dave Hull