Hi Alacant,

Thanks for the reply. I hadn't thought of that. The button on the logging page opens a directory in a completely different location but when I think about it, the logs for this are coming from an external program so perhaps that button is unrelated. Sadly however, despite ticking verbose and file there are no useful or relevant files in /tmp. The 2 that come closest are one called internalSextractorSolver_4.cancel which contains the word "cancel" and nothing else. The other one is called indififoa34d22f6 and is completely empty. There is also a directory called kstars but that is empty.

Something else I just noticed: With a camera gain of 0.0 it claims to find 4 stars even though I can see more than that in the image. Setting the gain to 10.0 and it then claims that it cannot find any stars.

I'll maybe give it another go by building from sources on a Linux distribution of my choice rather than using Raspbian stuff that I don't like but then again I am not sure it is worth the effort.

Thanks again anyway.

It should work if you decrease the scope focal length from 900 to say 400. Go to mount tab, change focal length of primary scope to 400, then go back and solve and it should work. Problem is with narrow FOVs it can fail or take a long time to solve.

Hi Jasem,

Thanks for the suggestion. 400 still didn't work but I made it 100 and that worked in that it took three images and plate solved each of them. However, it never produced the RA and DEC vectors which I have seen on Youtube videos and in the documentation. Well I say that it worked but it is hard to be sure since the image on the screen is so small. Even full screen and with maximum zooming it still only occupies a square about 5cm by 5cm in the middle of the much larger screen. Why it will not zoom out further than that will remain a mystery to me.

I have given up now. The only reason I was interested was because it was stated by several people that it supported the camera I have but as I can't get it to work with the simulated camera I suspect I am not going to get it to work with my old Nikon.

Thanks for the suggestion anyway.

I was able to solve in just 1.3 seconds with a focal length of 700.

  

Hi,

Thanks for the reply. I had another go. Starting with a formatted SD card and reinstalling everything all over again including all the index files. It finished at around 5AM this morning. Having sat through the entire night mainly watching it downloading index files you can imagine just how pleased I was when it didn't work again. It completely fails to solve and it cannot do polar alignment using the simulated devices. it makes little or no difference what focal length I use and it makes even less difference what exposure settings I use.

Note this is a straight install from the recommended download site (astroberry). I get the same results whether I use the astroberryos or a vanilla Raspbian with the ekos stuff added. It just can't solve any image no matter where I point the camera/telescope. In the attached screenshot I have copied the settings as far as possible from the image you attached and the result is the same: "Solver failed" as you can see in the attached screenshot.

There aren't enough stars there, put the gain to 90 or something.

Hi Again,

Thanks for all your help with this. Setting the gain to 90 has made it work. Would I be right in thinking that simulated mount is already pointing at true north? I wondered about this as when I do a polar align the green/yellow/purple triangle is tiny and the alt az error is 10 or 20 arc seconds which is why the triangle is tiny I suppose. Anyway, thanks for all your help with this and for getting it working.
 

Also, DO NOT try and solve on Polaris, you are asking for it not to work, move to another part of the sky….mine never works on Polaris either…..

Hi Astronerd,

Thanks for the info. I'll give it a go.

@astronerd: Solving, in general, should work near Polaris--people wouldn't be able to polar-align near the northern pole if this was not the case. 
I both just successfully tried with the simulator, and also polar aligned on my telescope the other night. In both cased I was able to solve repeatedly.
I tested with ASTAP and internal Solver on the simulator, and used ASTAP with my scope the other night. [All that said, yes, solving is one of those
magic technologies that also frequently causes headaches.]

Debugging with the simulator can be tricky, for sure, and there are many parameters.  I know I had issues when debugging my polar-alignment code with it.
In this case, make sure your focal lengths are consistent (simulator and align need to be using the same--simulator focal length is set in the
Indi Telescope Simulator --> Options tab) and that the gain is high enough (I set the gain to 100 in the align tab and exposure to 4s). 

When setting the Telescope focal length on the Mount Tab, you must click on the Save Telescope Info button for it to be applied. Verify it is correct with expected FOV in the alignment tab. It should update the telescope info in the INDI driver at the same time, so save the INDI driver configuration for next time.
Simulators aren't particularly useful for testing the polar align assistant either because by default they don't have much, if any, error.
Do, take the time to get your real mount, camera and guide camera connected and working and try it with real stars. 90% of the time, failed solves have to do with missing indexes, incorrect telescope and camera information, wrong scope/camera combo selected, underexposed frames or bad focus. There are also different solver profile defaults, try them all to see which works best for your equipment.
Sometimes it helps to uncheck Use Scale in the options. This will be slower, but more reliable if your telescope info is inaccurate. Once it tells you what the focal length truly is, you can save the updated telescope info and re-enable Use Scale.

Hi, Thanks for the info. Yes, I thought it should be possible to solve near Polaris. I have an iOptron iPolar system which does plate solving near Polaris and that works really well. Of course I suspected that part of the problem is that the simulated mount is already pretty close to perfect alignment and, I think, Andrew in the next post has confirmed that.

Anyway, thanks for the info.