In light of the S-Tec and GPSS discussion on the Grumman Gang, I was asked in private email about why I installed GPSS. The implication was that I had "wasted" my money on that feature. I thought it might be helpful to others considering installation of an autopilot in their Grummans. Take the technical details with a grain of salt.
>>> What exactly does it [GPSS] do for you?
I'm not exactly sure of your question, so I'll give you a good run-down of the technology.
The regular autopilot, whether it's an S-Tec or a Century, is a simple wing-leveler. The turn coordinator (TC) head note any change in turn and/or roll rate and sends a plus or minus voltage signal to the aileron autopilot servo. That servo, based on the TC output voltage, actuates the ailerons. The autopilot's goal is to keep the TC from showing any turn, thus keeping the airplane upright and straight ahead. It's called a "rate based" system because the farther the airplane (actually, the TC) deviates from the center, the higher the plus/minus voltage sent to the servo, and therefore the faster the servo turns the appropriate direction. The max voltage output causes a standard-rate turn.
Note that the autopilot has no clue "where it is" in the world, it's just trying to keep the airplane upright (actually, it's trying to zero the TC, but the end result is the same.)
The next step is to add an input for navigation, usually from a VOR or GPS. That is accomplished by having the nav head (the CDI) send out a plus/minus voltage based on its deviation from the center. 1 dot off to the left is xx positive volts, 2 dots off to the right is more than xx volts negative, and so forth. The signals go through the TC head, and now the head is doing both: combining the "wing leveler function" in addition to the nav function. It's still rate-based, as the amount of CDI deviation from center causes a larger movement of the servo and a larger turn rate. Also note that it still does not know where in the world it is, it's only trying to center the CDI, and it will never do more than standard rate turns (max CDI deviation).
One really odd result of that: consider if you were in Nashua and you dialed in the Hartford VOR (assuming you can get it.) Let's say you dialed in the 345 degree radial of Hartford and told the autopilot to capture that radial. Your VOR is telling you to "go right" and it's sending out a negative voltage signal to your autopilot turn coordinator. Your autopilot, diligently obeying that "go right" signal and sensing a very large deviation, will send signals to the servo to make a standard rate turn to the right to capture that radial. Problem is, you're way too far away from the radial to be able to capture it and you'll fly standard-rate circles over Nashua until you run out of fuel. Because a standard rate turn will never allow it to capture that radial, you get dizzy.
The next step: consider if you added an autopilot heading indicator (HI). All you've done then is added an instrument that compares your current (compass) heading to the position of the heading indicator knob/needle (the compass direction you want to go). Same thing applies as for the other features: If you want to go to the right, the HI sends out a negative voltage to the TC, which then combines that with the wing-lever and turns you to the right. Maximum deviation from the current heading gives you a standard-rate turn, and that rate of turn is reduced until the current heading and the desired heading indicator matches. Again, no intelligence, just "go left, go right."
One more step: altitude hold. I know your airplane has the S-Tec altitude hold function installed. That system is, effectively, a completely isolated system from your turn coordinator. The altitude hold (which can be installed onto an airplane with the Century autopilot) adds a separate air pressure sensor and a separate servo connected to the pitch control. When you get to the altitude at which you want to be (assuming S-Tec 30) you select the ALT HOLD button. The system senses the air pressure and uses the pitch control to attempt to keep the airplane at that air pressure. Any deviation from the selected air pressure (note it has NO CLUE what your numerical altitude is) it will send plus/minus voltage to the pitch servo and pitch the airplane up or down to maintain that air pressure. It also senses forces on the motor, and since we do not have automatic electric trim it will notify you that the forces are excessive and indicate to you which way to adjust the pitch trim manually. Yes, the airplane's actual MSL altitude will change with changes in barometric pressure, and yes it will continue to pitch up, to a point, if you lose your engine. Will the airplane stall? Dunno, haven't tried it, but I suspect it will only hold so much force against the pitch trim until it allows the cable clutch to slip.
Anyway, back to our GPSS story. Notice that on everything above, there's no navigational intelligence at all. It's just a dumb system, saying, "Duh, go right, or "Duh, go left."
Your GPS has a lot of intelligence built in. It not only knows where it is, it knows where you want to go. This has allowed us to do really neat things like program in a complete flight plan, center up the autopilot, and let it make all our turns for us. Just like the VOR, the GPS CDI sends out the plus/minus signals to the turn coordinator based on your relative position of the desired track. When you hit a waypoint and it's time to make a turn, the GPS says to the CDI "Whoops! We're off track to the right (assuming a left turn). Hard to port, let's capture that new course!" It will make the turn to the left, swing through the desired course a couple of times, but eventually it will core it and you're on your new course. Really neat.
The only real problem here is that until the CDI deviates from the course at the waypoint, the autopilot is flying along fat, dumb, and happy saying "Duh, I'm on course." At the waypoint it goes "Duh, we're off course. Go left." The funny part is during all this, the GPS is saying "Hey, George, we're coming up to a turn here, buddy, why don't you start turning?" while George is watching the CDI saying, "Duh, we're on course."
That's where GPSS comes in.
All GPS have a data stream of position, vectors, deviations, and so forth, usually in a standard format. Since our autopilot only knows how to translate plus/minus voltages, we add a small GPSS navigational computer. The GPSS computer ignores the "plus/minus" signals coming from the CDI, instead looking at the digital data stream from the GPS itself; it then makes calculations based on that data. It can calculate a lot of neat stuff, like how far away from the Hartford VOR 345 radial we are, and how far away from the next waypoint we are, and what's the best way to get there. Based on these calculations, it sends an intelligent plus-minus voltage to the TC, which dutifully turns based on that voltage.
For example, with GPSS we don't have to worry about circling over Nashua. With GPSS it will say "OK, we want to capture that radial inbound, intercepting on a 45-degree course. We need to first turn to course 210 to do that". It commands the TC to turn right ("Duh, turn right.") When the inbound course is approaching 210, it begins to reduce the voltage to the TC ("Duh, don't turn as hard.") When it reaches 210, the voltage goes to zero ("Duh, stop turning.") and maintains the appropriate voltage to stay on that 210 course. As we approach the 345 radial, it will ANTICIPATE the turn inbound, and the next thing you know we're flying inbound to Hartford without having touched the yoke.
Same thing goes with our enroute flight plan. Whereas before we had to fly past the waypoint before the airplane would turn, with GPSS the computer figures out the required turn rate and ANTICIPATES the turn, smoothly transitioning us to the next leg. I *hated* it when I was enroute and the flight plan called for a turn of more than 45 degrees; the airplane would always overshoot the course several times. Whereas before George would suddenly kick in some aileron and wig-wag to find the next course, with GPSS it makes a nice smooth transition. Plus, it's a value even on regular enroute "straight lines." Since the deviation from course is based on much finer data than a plus/minus voltage from the CDI head, with GPSS the airplane stays much better on course and minimizes deviation. Now I never fly with the autopilot connected to the GPS CDI output any more, it's all done via GPSS.
It's pretty sweet stuff.
With GPSS, GPS approaches become a snap. Program in the approach, turn on the GPSS and the computer figures it all out. It will turn towards the first waypoint in the approach and head right for it. It will then anticipate the turn to the next waypoint, make a smooth anticipated turn, and look for the next waypoint after that. It will literally fly you all the way to the missed approach point and, based on your autopilot, all the way to the holding fix. If you have altitude hold, all you have to do (besides monitoring George) is work the altitude hold to be at the right altitude. Get to minimums, set the ALT HOLD to 400 feet AGL or so, and now you have a lot more flexibility to look outside for your airport. See the airport, hit the autopilot kill switch, and land the dang thing.
Like I said, Sweet Stuff.
So, I would disagree with the comment that folks are "wasting their money" on GPSS. Can I do without it? ABSOLUTELY! My S-Tec 30 autopilot was fully functional before I installed a heading indicator and GPSS and more than enough to improve my safety in IFR. However, do I really like the GPSS and the convenience and accuracy features it gives me and find it a good value?