So basically, here's the rundown. I'm sharing because I have seen other CGE users run into similar issues - although, I'm the first case of a complete failure that I've yet come across.
In the beginning...
Looking back through my log, it's clear that the real issue started early on, I just didn't recognize it for what it was yet. I got the mount in February 2016, and it wasn't long after that I started complaining of it "not being able to handle the cold." I mention having very bad initial guesses for alignment stars as early as late February, and definitely in April. Also in April was the first time I recorded that it "moved on its own," probably an instance of it "jumping" when I either touched a cable or the cold shrank the metal in the shielding contact in the port, causing bad ground contact and the motor to jump.
Long story short, here is what was happening: So the Celestron CGE is two parts, the electronics pier, and the mount head itself. The two are connected by...wait for it...ethernet cables. Worse, they used shielded ethernet cables - and worse yet - they use the shielding as a ninth pin to conduct the ground signal to the RA axis. (Whenever I tell my electrical engineering friends about this, they are as flabbergasted as I am. If you are a Celestron engineer and reading this...please please please never do something so dumb again.) To top it all off, if the ground signal to the motor is lost, the motor starts slewing out of control, at a speed higher than the highest normal slewing speed, and that motion is not encoded (meaning, if the mount slews this way for just a short period, it doesn't know that it moved, and I have to restart the entire mount and re-align all over again). I hit the internet over this issue, and eventually it came down to those blasted ethernet cables.
The first thing I tried, back in July 2016, at the recommendation of many other CGE users, was to replace the shielded ethernet cables with flat, more-flexible ones. This appeared to work, and since the weather was warm, they were pliable enough and the metal expanded enough that I had several nights of success without any issues. I also got cable guides that I attached to the mount to let me create some slack on the cable where it goes into the mount housing so that it wouldn't get pulled. This worked for a while, but then as the cold weather of fall came again, it became clear that a more drastic solution was needed.
The Bennett Mod
Many CGE users online recommended the Bennett Mod. There were so many CGE users who had this exact same issue that a Canadian, Gary Bennett, finally created a mod kit to fix it. Basically, it replaces the ethernet ports, both in the electronics pier and the mount housing, with an 8-pin screw-on connector that was much more robust. The chassis was still used as the RA ninth pin, but a good connection was ensured because the cable screws on at both ends. After reading about all the success that many other CGE users on forums had, and saw how many people swore by it and said it solved all of their CGE issues, in December 2016 I bit the bullet and paid the $400 for the "easy" version of the kit, which doesn't require you to drill new holes into the mount housing. I received the kit, and it was indeed rather easy to install - no soldering required, just unplugging headers and plugging in the new ones, which even came labeled. The instructions were quite clear and had helpful pictures. It took me about two hours (I meticulously checked the tightness of the connections, triple-checked that I was plugging the right cable into the right header, etc), but then it was done.
So I plugged int back in, turned it on, and after getting some connection errors and tightening up some connections, got both axes talking to the hand controller. I hit Enter to find switch position...and while dec worked on finding the switch position no problem, RA started slewing out of control! Pressing keys on the HC yielded error codes about no response, and finally I just had to cut the power before the mount tipped itself over (it's rather top-heavy).
I got in contact with Gary Bennett, and he was very helpful at providing me with ideas of things to check to try and figure out what went wrong. He even sent me the schematic for the mount, and I dug out my multimeter and meticulously went through every connection, testing for proper voltage and connectivity (or grounding). Everything looked by the books, at least as far as I can tell, being a physicist and not an electrical engineer. The dec axis worked just fine, and it continued to work even when I swapped the RA and dec cables to test. I've got a whole bunch of notes on all of the strange behavior that was going on, but I won't exhaust you by putting it all here.
Calling In the Pros
After struggling with it for a few weeks, I finally got in touch with Dr. P. Clay Sherrod, who does a "supercharge" service for several kinds of mounts, hoping that his familiarity with the inner workings of my kind of mount would be able to help me figure out what was wrong, and how to fix it (his website is here). He gave me a great deal of helpful advice, and I re-did a number of things I'd already tried, and also tried some new things to test it. I looked for burned out resistors, checked the function of the limit switches, swapped out hand controllers, used different power sources (AC wall power vs DC battery), checked resistance across the optical encoder, plugged the ethernet components back in (essentially "un-doing" the mod), all sorts of stuff. But, two months later, I was still getting nowhere. So I had a choice: throw in the towel and get a new mount, or take a several-hundred-dollar bet and ship it to Dr. Clay for diagnosis and, hopefully, repair. I just had a feeling (and my hunches are usually right) and there was just a single bad connection somewhere, some ground wire that needed to be re-soldered or replaced, something cheap and easy that would just unravel the whole mystery.
Dr. Clay spent several weeks on it, even taking it into his electronics lab at the University of Arkansas to test all the circuitry. He was just as baffled as I was. After much effort, I finally had him send it back to me. I started shopping for an alternative mount, although I was unable to find anything that matched the capability and payload of the CGE. Everything was either too light (55 lb payload or less) or too big (75 lb payload or bigger). The CGE had a 65-lb payload capacity, and only weighed about 40 lbs. In astrophotography, you want your mount to be able to carry twice your actual payload in order to have good tracking; between my 11-inch telescope, the guide scope, and the camera, I am at about 32 lbs, so I really do need about a 60 lb capacity. But anything with that capacity was beyond my pricepoint, so I was going to have to settle for something like the Celestron CGX, which has a 55-lb payload capacity and cost $2200.
Another miracle, of a sort
So far in my astrophotography journey, I've had a great number of small miracles. Nearly all of my equipment has been given to me, or at least sold at a steep discount. Literally the only things I have purchased on my own at full price are my DSLR and some miscellaneous parts and pieces, like dovetail bars from ADM or cables. My C8, C11, and Borg refractor were all gifts; my Orion ST-80 was sold by a friend on the cheap; my QHY5 was also sold to me on the cheap.
So when Dr. Clay emailed me and told me that he had a Celestron CGE Pro, just back from Celestron refurbishment (and Hypertuned!) that he would sell me for $2k, I about fell out of my chair. The Pro is the successor to the CGE, with a payload capacity of 90 lbs, and - wouldn't you know it - screw-on 8-pin cables (and even a screw-on DC power port). It sells new for $4,500. My only reservation that I had to sleep on for a few nights about it was that it weighs a lot more than the CGE did - the EQ mount head alone weighs in at a hefty 75 lbs, and then there's the counterweight bar and electronics pier in addition to that. Luckily for me, the mount head comes apart into two sections, but they are not very evenly distributed - the upper portion still weighs some 50 or 60 lbs. So, of course, I bought it! And, bonus points, it arrived just in time for the 2017 Texas Star Party. Blessings all around.
As far as Dr. Clay and I can tell, something got fried when I installed the mod kit. Dr. Clay thinks it's a juxtaposed cable in the mod kit, but Gary Bennett says he tests each kit on his own CGE before shipping. I followed his instructions to a T, and I am no foreigner to circuitry - I am an experimental physicist, after all, and have soldered my fair share of both low- and high-voltage circuits and have thus far not blown anything up. And this didn't even require soldering, just plugging stuff in, it's very nearly fool-proof. So, who knows. It may have even been completely unrelated to the Bennett mod. The last night I used it, in November, it did the slew-like-crazy thing, and then it under-slewed in dec several times in a row, even with power cycling. Maybe it broke that very night, before I even bought the Bennett Mod. We will never know.
As far as the CGE itself goes, I would like to sell the parts to other CGE owners, since they are hard to come by now that Celestron has discontinued it. But I'll have to sell them with some kind of money-back guarantee, since I don't know which part is the fried one. I haven't gotten around to it yet, so we'll see.
As far as the new CGE Pro goes, it was a champ at Texas! I guided as long as 7 minutes, and the mount wasn't even level (my tripod has a gimpy leg, and of course that always ends up on the lower side). It is, however, prohibitively difficult for a single person to set up and use, so it is going to become my long weekend and star party scope, situations where I can set it up and leave it up for several days (and have some help getting the heavy mount head on top of the rather tall electronics pier and tripod combo).