Why Level

curtismacc

We know that the Seestar uses an accelerometer to feedback to the user information when trying to level it. The Seestar then uses the level information along with the internal compass for the 2 data points it needs to find, or at least, come close to finding the sun and the moon during the day. But, why do we need to level the Seestar anyway since the accelerometer measures the Seestar's orientation relative to the horizon, level or not. Let's say the Seestar is 2 degrees out of level pitched downward. This should be sufficient information along with the internal compass data to do just as good a job of finding the sun as when the Seestar is level. After all, we do not have to point the Seestar to north for the compass reading to be used. We can point the Seestar in any direction and the compass will still figure out which way is north. I would expect the actual internal accelerometer orientation, level or not, to be sufficient to provide the second data point so that they Seestar has enough information to make a good first guess where to point to find the sun or the moon.

What am I missing?

Support@Seestar

Although the accelerometer can detect the current tilt, the initial GOTO algorithm is based on the assumption that the Seestar is level.
Any tilt detected before leveling is treated as an error, not as an actual pointing reference.
Once the Seestar is leveled, the current orientation is accepted as the valid reference frame, and only then does it use the accelerometer + compass to calculate the direction of the Sun or Moon.
Leveling tells the Seestar “this is the correct reference plane”, without that, it will not use the tilt information in the pointing model.

Lilo

Support@Seestar
What about Horizontal calibration switch?
I assume it's able to measure a tilt in AZ mode and calculate needed correction if is see stars.

curtismacc

Support@Seestar
You clearly explained how the algorithm works, but missed my point. Based upon how you explained it, the algorithm could be designed to use the actual accelerometer reading regardless of whether the Seestar is level or tilted by x degrees up or down and still add the compass data to create an accurate pointing model. There is no mathematical reason why being a 0 degrees than at -2 degrees or +2 degrees, for example, would make the pointing model any better or worse. If the algorithm were written this way it would be even simpler for the end user. Maybe ask the software guy about this?

In the end it is not difficult to level the Seestar but if I were designing the algorithm I would try to make it as simple as possible.

Support@Seestar

Lilo
We calibrate the horizontal sensor during factory production, so normally there is no need for users to recalibrate.
However, during use, if the horizontal sensor experiences a zero-point drift, calibration may become abnormal, and a horizontal sensor recalibration will be required.
In addition, due to differences in tripod stability and ground levelness, it is still necessary for the user to adjust the Seestar to be level.
Once the device is properly leveled, enabling the Horizontal Calibration function allows the algorithm to automatically compensate for any remaining horizontal errors.

Daemon42

curtismacc I suspect it's already trying to do what you're suggesting. The problem is the compass is really never that accurate, so when you combine variability in the accuracy of the compass with it being off level then as it swings around to find the sun or moon, it can easily miss it a bit high or low. Levelling the tripod base first, makes the error in the compass irrelevant, and you can see that in the way it searches it just starts roughly 15-20 degrees off angle to the target, and then sweeps past it until it sees some bit of the sun/moon pass through frame. If the compass were accurate it wouldn't even have to search that way.

curtismacc

Daemon42 I fully understand what you are saying, but unless you think the accelerometer is also not accurate to within a degree, then it makes no difference whether the Seestar is level or not as the accelerometer tells it where it is pointing in Altitude (o deg, +1 deg, +2deg, -1 deg, etc.). Perhaps the problem is that the accelerometer is not all that accurate. A Google search suggests the iPhone accelerometers are not accurate to more than 5deg. That would be a problem, but then it would also be a problem for the App's leveling function since the feedback is coming from the internal accelerometer.

Support@Seestar

Daemon42
We require the device to be physically level because in Alt-Az mode, Seestar needs to find targets based on a true horizontal reference. The purpose of the horizontal sensor is to help the device maintain physical level.
As long as the device is physically level (checked with a spirit level), the horizontal value shown in the app will not affect target finding.
If the device is physically level but the app shows an incorrect horizontal value, this indicates that the horizontal sensor itself is off and needs to be recalibrated.
In this case, if you do not calibrate the sensor but instead adjust the tripod manually, it may cause the device to become physically unlevel, preventing it from finding targets.
However, if the device is not physically level, and the horizontal value is incorrect, you do need to adjust the tripod.

LA3QMA

Regarding the compass could it turn 180' back/forth to "calibrate" and then use the accelerometer to use whatever degree its of as "0" ?Like @curtismacc explains?

Daemon42

curtismacc A properly calibrated accelerometer can have accuracy much better than 5 degrees, and I'd say, a little less than 1 degree. The only time I've seen the accelerometer drift a lot is when they experience a large temperature change, but generally that can be mitigated by putting it somewhere on the board inside that maintains a fairly consistent temperature. But the compass tends to be all over the place, for a variety of reasons.

curtismacc

Support@Seestar
OK, one more comment from me and then I will drop the issue because it seems that I am not getting my point across.

Fundamentally telescopes do not need to be level to be polar aligned. All leveling does is eliminate cross talk between the Az and Alt adjustments. I have forgotten a few times to level my MyT mount when in the field and all that did was make it take a few more back and forth adjustments of both the Alt and Az knobs to get an excellent final PA. Telescopes in Alt-Az mode like the Seestar, Celestron Nexstar line and Skywatcher AZ GTi, for example, direct the user to level the mount but this is only necessary because the software assumes that the mount is level when it solves for a GoTo model. In the case of the Seestar, it has the advantage of having an internal accelerometer and an internal compass which the Celestron and SW do not. So the Celestron and SW need to be set up in a configuration called "Level North" which means the optical tube is both level and pointing north. Otherwise these mounts would be lost. But the Seestar has both internal sensors and could use the data from both (so long as it is accurate enough) to avoid making the user level the mount and point the telescope to north. Apparently the ZWO software engineer decided to use the data from the compass but not the accelerometer to create the first pass pointing model (which is refined by plate solving, i.e. Horizontal Calibration). I see only one reason for not writing the code to use the data from the accelerometer to compensate for whatever angle the Seestar is relative to the local horizon and that would be if the accelerometer is not accurate enough for this. But then it apparently is accurate enough to tell us when the Seestar is level, so there you go.

I suggest you ask the software folks about this and maybe they will get it and then you can have a code update that makes the Seestar even more user friendly.

Thanks for listening.
Curtis

Support@Seestar

curtismacc
Got it, I will pass your feedback to the software team so they can consider using the accelerometer data to further optimize Seestar’s pointing model.