I do know that variac controlled platters are very nit-picky with the timing. Mess with the speed of one deck and you affect them all. But I have seen these things payout slower and slower over the course of my shift so I am sure that no one is tampering with them when I am not there. The platters need to be retimed once every 5 to 7 days it seems.
Any suggestions on what I can do to make these things run smooth 100% of the time? I would like to stress again that these are brand new.
I am ready to answer questions if you all need more info about the set up.
Also, check the control rod from the elevator to the variac. If the jam nuts are loose, it could be doing some nasty adjustments to itself.
Another thing to look at....where is the elevator riding when the pay-out falls behind, and where is it riding when the platter is running normally? You will see a change over the course of the movie. If it is favoring the timing hole in the mast, it is probably riding too high. I would recommend you verify the variac setting before you start to time your platters. Remember - you have two holes in the mast. A Threading hole, and a Timing hole. The Timing hole is the upper hole.
I could be wrong on this, but when the elevator is resting on the bottom, the wiper on the variac should be about 50%. Would someone verify that setting, please?
Be careful of your motor tracking. If it is set too heavy, the motors will require more current. This was also a common failure with the Christie Autowind I's and II's. It caused alot of module failures on those machines. The SCR in the A-3 phase controls might not like that, either.
Never raise the elevator without turning off drive motor switches first. The surge current could damage the SCR's in the phase controls.
I believe that if Strong's engineers did something about the inductive spike that could occur across the SCR's, the phase controls would last a very long time.
I might increase the size of the ballast resistor inside the motor can by a couple of Ohms. The resistor is a 25-watt 2 Ohms. This resistor is switched in series with the motor during the take-up mode. The elevator will ride lower if this resistor is electrically larger, hereby allowing more power to be applied to the pay-out motor so it won't fall behind. After it finally blows all the SCR's and is converted to micros as Mark describes, one can put 10 ohm 25-watt shunt resistor across the micro switch so the platter will not whip so badly.
I tried that already, and it works to a degree. However, the resistor I used was not large enough in resistive value. It has to be large enough so the platter will not creap in the pay-out mode while the plunger is engaged in the threading hole. It will be a trade-off between variac settings, the resistor value and motor tracking pressure for it to work right.
On a second thought, I think I will take a phase control unit and install it on a retro fitted micro-switch pay-out head on a CFS platter. The 10 ohm shunt resistor inside the motor, but I am not sure how the phase control will like that resistor. I'm gonna find out!!!!
So, off I go to Oak Harbor to snitch a brain for modification.
The old orange POTTS and blue CFS platters used a set-up very similar to this. Check the schematic, and you will see.....
As far as the Strong Digital platters are concerned, the jury is in. They have been out in the market for about 5 or 6 years now, and according to the feedback I have received, it is a very reliable machine. I will consider them in new construction. 
Perhaps this will help with your experiments.
Notice the 2-Ohm 25 watt resistor is in series with the motor on the rewind (take-up) mode. I have a tendency to believe if a 4-ohm or maybe an 8 ohm resistor was used, it would give the phase controls a little better head room to play with.
I installed a phase control on a CFS pay-out head. (it fit perfectly.) I have not tested it yet, hopefully I'll be able to do that in a few minutes.
It looks like my experiment is going to work with the phase control on the CFS platters. It ran nice and smooth, but I could not complete the test of running a movie with it. I was jazzing around with the little potentiomenter, and I might have popped the LED. In addition to that, the cheap crimp jobs of the wires on the pin plugs broke. Those were factory crimps - only about two three strands of the wire were actually crimped and the rest of the strands were busted. Nice quality
....but it didn't come from Strong International.
That specific brain died while running a movie once....then it was put in the brain repository. Now I know why.... 
I used that trick on desoldering vacuum pumps and other equipment using ad/dc series wound motors while I was in the navy.
I do believe this is perhaps a payout/brain problem... more specifically the spring that returns the payout arm. Perhaps it is just too stiff since it is so new. The film needs to tug just a bit harder to make the platter turn at the right speed, and when it does (after wrapping about 3/4 of the way around the brain) it unwraps and the payout arm moves back a touch. Then the cycle repeats again and again. The elevator does not move at all during this. All I know is that the Strong phase control A-3's at my old theater that is now over 10 years old never had problems like this. In fact I think we only had to replace one brain due to a bad LED the entire time I was there.
I'd rather spin both decks by hand myself than have a microswitch system. Hey waitaminute that's actually happened with a microswitch system!
Incidentally, make sure the platters are level. If they are not, the swing arms do not like it. They should practically stay wherever you set them.
In my phase control experience I have found it nearly impossible to time them so that they pay out perfectly without touching any of the feed plate bumpers for the entire show. SO, as I have said in previous threads, I time them to spec (as close as possible), then raise the motors a bit more so they are timed slow. Most of my Strongs and Potts platters will wrap 3/4 or more around the brain. I have platters that have been doing that for YEARS and I have NEVER seen any film damage from it....and no brainwraps with phase control or micro since they are timed slow.
After spending a small fortune to convert some of my platters to phase control a few years ago I am converting them, one stack at a time, BACK to microswitch as the phase control modules crap out.
I have found a workaround and will post it in a few days if it seems to be a rock-solid method, which I think it will be.
------------------
Samual Hunter Sr.
KC5ZSL
1. Build up all film reels onto one of three decks.
2. Feed film out from center of roll using fancy-schmancy centerpiece.
3. Film goes into projector.
4. Film is projected and audience is amused.
5. Film exits projector and returns to platter system, taking up on different deck.
6. Show ends.
7. Repeat from step 2.
73,
de KJ5KX
------------------
Better Projection Pays!
I think the problem was that the springs were very sensitive to temperature variations. If you suddenly had a very hot day the spring would expand, which would make the arm reluctant to return to 'neutral', effectively running the feed plate too fast. On a cold day the opposite would happen, increasing the resistance on the arm and running the platter too slowly. I remember one very cold day on which I'd forgotten to switch the air handling unit on in one of the boxes: the first show virtually wrapped because the feeder arm was so stiff. If the temperature in the box remained steady at around 20-21, then no problem.
------------------
Samual Hunter Sr.
KC5ZSL
The projector pulls the film from the center of a platter.
That film as it comes from that platter passes through rollers to create a slack segment, then through a widget called a "brain".
The brain detects by changes in that segment if the platter is moving fast enough to pay out the film at the speed the projector is pulling it. The speed of the payout platter is controlled by the brain.
Potts platters rule.
As with any electrical/electronic device, it is necessary to establish a reference when you reverse engineer it for troubleshooting purposes.
It makes me angry when some manufactures won't release a schematic on some of their equipment. Any good technician will be able to reverse engineer practically anything on a simple PC board such as the Phase Control of the A3's. Ken ran into that snag.
Well, Mr. Engineer, We are not idiots! We will figure it out. If your design is patented, pony up, please. We have to keep some of your junk running in the field. At least make it easier for us to do just that! We are not interested in how "part happy" you are. We want to help you make it as good as it can be through our suggestons.
OK - Flame me....see if I care. I get tired of some of the
I have to contend with in fixing some of your un-documented garbage. 
You know it is true. 
End of rant.....
, the takeup is kinda like a reel takeup system ie its speed is regulated by nothing more than the sound projector itself. ------------------
Samual Hunter Sr.
KC5ZSL
On the other hand, the IWERKS and IMAX platters are just the opposite. IMAX and IWERKS pull from the edge of the platter, and taken up on the center. However, there is a center ring in place even on the feed platter. Those platters can be looked at as a horizontally mounted reel-to-reel system. Even so, there is undoubtedly some kind of motor assist on feed-out because of the tremendous weight of the print. At the end of the presentation, 70mm IMAX and IWERKS must be rewound.
You can't really look at a conventional 35mm platter as a reel-to-reel system because the center ring is removed on feed-out. The 35mm system is said to be automatically rewinding as the movie runs. At the end of the show, just remove the center ring, re-thread, and start the next show.
With most reel-to-reel systems, the takeup reel is connected to the soundhead by means of a belt which drives the takeup clutch on the lower reel arm. Tension is adjustable, but does not remain constant throughout the entire reel. Some machines (e.g. the electronic Kinoton models) do have motor-driven takeup and provide constant takeup tension, but that is the exception rather than the norm.
Paul: the newer IMAX setups use platters very similar in concept to conventional 35mm setups and films do not need to be rewound. The 2-strip 3D setups, in particular, are quite impressive to watch when running.
Then there is the MK2 reel unit that placed the 4 disks one above the other on two different angles to the post (looks more like a platter) and again it rewinds and uses a servo control
(Pictures of the MK2 under Ontario Science Centre)
Finaly there is the QT which stands for Quick Turnaround unit.
It is a beefed up custom platter from Kinoton and operates like a conventional platter and is available in a 5 deck and 3 deck version with up to a 7' disk