FWIW, I have my company's products run through Highly Accelerated Life Testing (HALT) where the primary tests include applying strong vibrations at different frequencies while concurrently heat cycling the item being tested as a means to get it to fail quickly. Kinda like what happens on the handlebars of our FJR. From what we learn during HALT testing we refine our product design so that weaknesses are mitigated and the product becomes very reliable. Motorcycle specific GPS should be more robust than most car type GPS due to vibration testing the design. Old school circuit boards with thru-hole components tend to be more robust than modern surface mount boards where the components have no leads making the components rigid and more prone to fracturing. Also, surface mount boards tend to have more latent defects which won't show up until much later in the product's life unless the manufacturer has good screening processes. Today's circuit boards made for the public no longer have lead in the parts or in the solder, however, for some military electronics and implanted medical devices lead is still used for improved reliability.
Anything you can do to soften vibrations will vastly improve the chances of your GPS having a long life.
Ion gives a great explanation of the root of many failures. My take in your original situation would be that your mount was not able to provide consistent contact between the unit and the power source from the mounting cradle due to vibration or whatever. That's all a mechanical thing, and mechanical things will fail as we all know.
I switch my Zumo 550 back and forth between my 2015 FJR and my HD Ultra Limited. Both cradles are mounted dead center, not out on the ends of the bars where vibrations are usually magnified. I have yet to ever have a failure with the 550 related to intermittent power to the unit on either bike. I think your move to the middle will give you better more reliable service assuming the unit is 100% reliable within it's enclosure.
Both correct. Mostly. Some surface mount (SMT) parts do have leads that extend from the part to a surface pad on the PCB, some do not. But in either case, cyclic flexing of the PCB itself can gradually cause cracks in the solder joints. A crack will lead to an intermittent open circuit, then finally an always open circuit when a small amount of corrosion builds up on the two sides of the crack. Heat cycling makes this worse, though with a typical case of many hours on then many hours off, the number of heat cycles is probably less of a factor. It is more common in cases where the unit is started and stopped many times per day.
Some of the failures are also internal to the PCB, most commonly found in the barrels of inter-layer plated through-hole connections, called vias. What happens there is that if the plating in the barrel is inconsistent, thermal cycling can cause a thin area to crack and separate, ending in the same kind of open circuit failure as with a failed solder joint.
What can be done about this?
1. Hot-wire your power connection. I run a cheap car GPS that has a USB power connection. The battery does not last long, and the USB connection in the back of the unit stopped working consistently a long time ago due to vibration. So I carefully opened it up, found where the 5V and ground from the USB connector gets onto the PCB, and soldered the two power wires in a USB cable directly to those spots. Then I used hot glue to hold down the wire to a small exit hole I drilled in the side of the unit. At the exit point, I left an approximately 1" service loop in the wire, then hot-glued it to the back of the GPS in such a way that it departs from the GPS near the bottom. This hot glue provides a strain relief for the wire. I have a 5V switching power supply installed on my bike, wired to a 4-up USB strip, so the larger USB plug goes in there. The larger USB plugs are not nearly as susceptible to vibration as the smaller ones like in phones and GPS's.
2. Ruggedize the GPS by filling the inside of the unit with hot glue. This will dramatically reduce the vibration of all the parts in there. The way to do it is open it up and find several places where you can drill a small hole in the outer casing. Drill the holes, then put the unit back together. Use the hot glue gun tip to inject the hot glue into the unit through these holes. You may not be able to fill it completely, but anything you can do will help. The less anything can move around in there, even if the movement is very small like with high-frequency vibrations, the better. Hot glue will not adversely affect any electronics.
3. Mount somewhere other than on the bars, or on the ends of long attachment devices. The longer an attachment device is, the more any vibration present will be amplified. If you could glue it right on the gas tank, that would be best. Usually that's not practical, but make any mount as close to the frame as possible and as short as possible.
By the way, I'm a PCB designer by profession. I've worked mostly for defense companies like Honeywell, Raytheon Missile Systems, and BAE Systems.
You might be asking yourself at this point how they make PCBs for things like fighter jets and missiles. The methods are similar, but all the failure modes are designed out. Every single aspect of every feature on the PCBs, and the soldering methods, are done in such a way that these failure modes are reduced. The cost for doing this is a big part of why everything for aircraft is so expensive, and would be prohibitive for the vast majority of consumer electronics.
