In my case, installing four analogue JR 537s cured the problem completely. Even with the poor mechanical advantage of a close clevis to surface set up. I thought this was due to the 537s not trying so hard to find the exact center point of the surface throw. In testing to figure this problem out I used two 4.8 V 1850 NiMh receiver packs. One at full charge and one that had four 14 min. flights on it, my second JR R700 reciever and a second powefr switch. No difference was noted at the time untill I installed the 537s. All with the sam HD JR extentions.

The DS811s were fine in a scale aerobatic EDGE 540 and a CAP232 with less surface deflection and area and more resistance to pushrod movement. They seemed to find and hold their center without passing it back and forth just fine in that aplication.