It is not surprising that Garmin produce the same crappy software as everybody else 
The difference is that they are still in business and thus able to fix it, compared with say HBK (Honeywell Bendix King) who lost all their tech staff in 2003 and are still listing the KFC225 (with its smokin’ servos) as available for sale…
There are some appallingly personal comments under that article.
This is a software problem because the servos use stepping motors which are not capable of moving unless fed with the right pulses in the right sequence. What I don’t know, without diving into the manuals, is whether the servo has its own CPU + ADC, or some equivalent means of pulse generation (and thus accepts an analog control voltage) or whether the autopilot computer generates the pulse train. @wigglyamp may know.
Just do add something. Whether one can actually hold against trim depends vastly on the speed. I can by no circumstances hold against trim at cruise speed. But I can at approach speed. So next thing after disconnecting the AP on trim runaway could be throttle cut-off to get into a manageable range of force.
In the GA context, the manual trim should always work, no?
As stated above, not all GA aircraft (and many large aircraft) have a manual trim.
Most GA autopilots will send a torque command to the servo, and the servo will self-generate whatever it needs to drive the motor. In the analog DC brushed motors I know, the torque signal itself drives the motor via a servo-local amplifier, sometimes with a separate or integrated clutch command for some (usually trim) servos. On digital DC-motored ones (KFC225?) I don’t know.
With brushless motors, there is no reason to send to a servo motor the drive signals all the way from the computer, all it needs to be fed is DC power and a digital torque command signal (as part of a data bus, CAN or otherwise), plus whatever additional safety or redundancy measures are designed into the system. The motor drive signals can be locally generated within the servo: it is a much more robust arrangement. Also typically undriven engine torque is low enough that, unlike DC motors, no clutch is needed: this plus no brushes to wear mean it should be more reliable too. I don’t see why one would want to send the AC power drive signals all through an airframe to each servo, but worse mistakes have been done in aviation….
Jodel trim was a lever connected by a solid flexible cable to the trim tab. I once had it jam in climb, and returned and landed with no difficulty.
I flew with no elevator trim tab. Only a problem with a rear pax – I had sore muscles after a ~1.5 hour flight.
If the trim jammed at full extent, I would have difficulty controlling at full power, even at below cruise speed.
