Torque and Speed. Servos have two specifications by which they are rated (and priced…): torque and speed. Torque, usually given in ounce-inches, is a measure of rotational power. Speed, usually given in seconds, is a bit more specific–how fast will the center shaft rotate through a given distance (measured in degrees)^*?

Let’s examine torque a bit. A typical rating is 43 oz-in (this for a JR 517 Standard servo). What this means is that if we suspend a 43 ounce weight on a servo arm that extends 1 inch from the center of the shaft, the servo should be able to lift it over the entire distance of rotation. We can increase this weight, but we will have to shorten the arm in order to remain within the capabilities of the servo. (For instance, if we have an 86 oz weight, the arm might be only be ½" long.)

To show what power can mean, suppose we have an airplane weighing 8# 7 oz and a wing area of 800 sq in. Further, assume all the weight is spread evenly over the wing. Ailerons would be, say, about 12% of that area (for a sport ship, NOT a 3D monster!). The wing loading is thus 24.3 oz/sq ft. The aileron share is then 16.2 oz ((0.12x800/144)x24.3). This is how much static force, or ‘1-G’ the servo must overcome for level flight. Most servos could handle this load, right? Not exactly…when coming out of a dive, for example, the G-forces could be more than 3-5 times higher! Suppose the G-force is 4; the aileron load then is 64.8 oz–quite a bit higher than what a ‘standard’ servo can take.

The assumption here is that the servo arm is 1 inch long, of course. But suppose it’s shorter–say, 5/8". Then the force available is increased to this value: 43/0.625=68.8 oz. Here you can see that the servo is (apparently) capable of handling the load. What happens, of course, is that the servo is being pushed to its limits, gets hot, and is more liable to fail–especially during a prolonged demanding flight! There’s always a price…

Now look at speed. As stated, the (rotational) speed of a servo is measured in degrees per second, under no-load conditions. If you have a servo arm of 1 inch trying to pull a load of 43 oz through an arc of 60º, the servo will NOT be able to do it in the stated time–remember, the specification is "under no load conditions". Speed is therefore directly proportional to the mass of the load–the heavier the load, the slower the servo.

Tangential velocity is a function of the length of the servo arm. A longer arm (with the connector at the end of the arm) will produce a higher tangential velocity:

(V = [angular]Distance x Time),

so the control surface will move faster. But this is deceptive, due to the load on the surface resisting the movement, and the geometry of the linkage (see above). There’s always a price…

Rotational Distance. If you connect a servo to your receiver, plug in the receiver battery, turn on your transmitter and exercise that servo, you’ll probably find that the servo arm goes about 90º total from one extreme to the other^†. This seems, also, to be a standard setup for computer radios; the Travel Adjust/ATV/EPA values are defaulted to 100% in either direction.

Most servos, however, allow for ±60º rotation. What’s happening here? Why don’t the servos move the entire 60º when the TA/ATV/EPA value is 100%?

The manufacturers are being conservative, that’s why– the manufacturer really can’t know the specific use for which the radio was bought, so he supplied good, inexpensive servos that would serve the sport flyer quite well. Since sport flyers make up a large portion of the flying public, this makes economic sense. And to accommodate the varied skills (or lack thereof) needed to set up an airplane, he erred on the conservative side of total travel allowed, as a default. (Well, ya gotta start somewhere!) So, you get about 3/4 the total throw, right outta the box.

If you accept the default values (i.e., 100% gives 45º rotational distance), you can easily set up your airplane, but you’re wasting some servo power.

More Power, More Worries. There is a way to utilize all that your servo has to offer–set your (high rate) TV/ATV/EPA values to the maximum percentage amount that your radio manufacturer allows. But be careful–make sure that your servos don’t buzz (an indication of being overdriven) at their extremes. As has been stated many times, an overdriven servo leads to rapid battery depletion and possible servo failures, neither of which is desirable (!).

If you do this, be sure you do it at the beginning of your setup procedures for your aircraft–attempting to do this after you’ve set up the proper geometry will certainly change it. This will lead to non-linearities and frustration.

Increasing the total throw also has another benefit–an increased servo response from idle–by reducing the deadband^‡.

Hmmm–it’s beginning to look as if increasing the total throw allowed is A Good Thing, and we should do it as a matter of course. But, there’s always a price…and in this case, it’s a bit insidious. It has to do with dual rates–if your radio allows you to set your dual rates to some value greater than 100%, AND you’ve set your throws to the maximum allowable value, you WILL overdrive your servos! For instance, suppose your radio manufacturer stated that the total allowable throw is 125% for a particular travel direction. If you set your travel to this amount, the servo will duly travel 60º, as desired. That same manufacturer, however, ALSO allows you to set the maximum dual rate to 125%. If you did this, you are trying to drive the servo to 75º! This is guaranteed to be BAD!

If you want to set your dual rate to more than 100%, then in most cases, you’ll have to be sure that your total travel is less than the maximum allowed. Since manufacturers give different specs, it’s left as an exercise for the student (you!) to figure out what that number is!

* Some radio manufacturers don’t even give you the distance, just the time. Since there are a few standards, I usually make the assumption that the distance is 60º (half the servo’s maximum travel).

† Most computer radios are this way. I don’t own an analog radio (non-adjustable, that is) any more, so I can’t say what the travel is/should be for those radios–you’ll have to measure for yourself.
‡ Deadband is usually defined as the time it takes to discern a control surface movement after a control input is given. It is comprised of two parts–electronic response, and slop in the linkages.