When we are using a mount that does not have absolute encoders, the question is not if, but when. At some point, for whatever reason, it's inevitable you sit there with the roof open on the observatory, and have lost sync with the mount. Maybe a power surge, or a brownout, or any number of reasons, it's possible. You know the roof is open, and the telescope is pointed _somewhere_ but dont know where. Not an issue using a mount that has absolute encoders.

So the next problem is, we need to park the telescope in order to close the roof.

A secondary part of the same issue, before closing the roof, one would like to have some sort of confirmation the mount is indeed in the correct park positon, separate from the stepper motor feedback, before the button gets pushed which will roll a few hundred pounds of roof over it all, just to be absolutely sure that roof is not going to hit a telescope and possibly do damage.

I've been pondering this question for our build, and a few things come to mind.

a) Some folks use some sort of beam detect, with a reflector mounted on the telescope that only reflects back when the scope is physically in the park position. This works well to confirm the park position. Doesn't help in the 'I am open, and lost' condition.
b) Some folks spend a lot of money on a new mount, with absolute encoders. This is (my opinion) the ideal solution, but, it hits the budget rather hard, and is more than I would like to spend. We have 4 mounts, three of which are perfectly good for the telescopes we have, but all 5 are stepper driven with no absolute position feedback. I'd prefer to NOT spend another 8 to 10K per telescope to upgrade mounts.

One idea I have been toying with, is to mount a 3 axis accelerometer onto each telescope. With feedback from that, we get a 'rough idea' of telescope orientation, irrelavent of where the stepper motors say it is pointed. This data would be good enough to do a cold start with no data, and move the telescope into a safe position. It'll also give us a feedback we can use indendant of the mount data, to act as a second confirmation that 'yes, it's safe to close the roof now'.

I'm curious how other folks have dealt with this issue, or if you have even considered it ? I know the photos of the heq5 sitting in the box look slick, but, we have one of those mounts, and it doesn't have absolute position reports, just positions based on stepper motor counts. If for whatever reason it loses sync, then you end up in the situation of 'dont know where mount is pointed anymore'. One solution is to take a starfield shot and solve it, then sync on that position, but, that doesn't work in daylight. If the worst case scenario is a total power loss for some period of time, then power is restored after sunup, using a starfield to resyn wont work.

I've pondered the self contained 'fail safe' system for our installation a lot, and, so far, the 3 axis accelerometer seems like the most realistic solution I've come up with. Has anybody else dealt with this issue, and if so, how did you do it ?

I have added a gyro sensor to the mount, L3GD20 Gyroscope
It will allow me to know exactly how the mount is oriented.

&

&

I am using the code from the link as is, downside is that you must run the script before moving the mount as the script re-calibrate the sensor on each execution.
But I guess you get the idea, I'm sure there is room for development on my idea.

The problem of 'needs initialization' is exactly the reason I'm hedging toward accelerometers instead of gyros. This way, even from a full cold start, we get a hint toward orientation without any 'known' conditions. It allows for an intelligent set of inputs to a full cold start, from any random orientation.

I've been experimenting this morning. A fresh vivid install with the xwiimote stuff installed, does indeed read the accelerometers from a wiimote over a bluetooth connection. I will probably spend a little time fussing with this on my test setup, and see if I can get intelligent transformations in place.

In our final configuration, there will be one computer managing the observatory, and one for each telescope. I would really like the 'safe to close' indication to be independant of the computer managing the telescopes, and be a completely separate input that goes to the roof manager setup. If I can develop orientation data from accelerometers, then that can become my second indication, and if it's via bluetooth from a gadget mounted on the telescope, that's even better me thinks.

Have you made any progress with the accelerometer?

I have not got to it yet. I got a lot done a couple months ago, but life has got in the way again, very little time to work on any of this right now.

What about adding some encoder rings, like the following ones:
www.jimsmobile.com/buy_ngc.htm