Mesu mount modeling

[dan_adi]

Hello guys. I managed to get a good mount model for the Mesu 200 mount and I thought I should share my experience. I had great advice on the sitech yahoo groups from mr. Dan G. and mr. Don. The sitech controller uses active pointxp modelind and platesolve 3 engine, the same ones planewave intruments uses for their mounts. At first, searching the forums, I did not find any mesu user to have succesfully made an acceptable model, the reason beeing some mount slippage with this friction drive. I don’t know how much slippage there is but that was the reason put forward. The steps I took are:

1. For accurate tracking and modeling time synchronization between sitech controller and pc clock is essential. The pc clocks are usually bad at keeping time with an accuracy of miliseconds. In order to fix this there are 2 methods: a. A less precise one is using software to syncronize the pc clock with an online time server. b. A more precise method is to buy a small time server. I tested both methods and the dedicated time server  was much more accurate and stable. I set my pc time update from the time server every 15 seconds. There is a powerpoint presentstion by mr. Dan Grey as to why time sync is essential. You can find it on the sitech yahoo groups.

2. Reduce the acceleration rate and slew speed from servoconfig in order to minimize slippage. Also allow a delay after slew for settling.

3. Set encoders to polite mode. 

4. Balance the telescope as best as you can. I had a hard time to balance the mesu in every direction, because of the friction drive. There is no precise way to know. But I know for sure an incorrect balance will make a model with poor performance. If there is a way to know how to balance a friction drive precisely please let me know :). 

5. Polar alingn as best as you can

I hope I did not forget any detail. With these tips I made a model with 36 points with a total RMS value of 42 arcseconds. All my targets were near center after the model. I as understand from mr dan G a good model also improves tracking. 

Also the best part is that a 36 point model is built pretty quickly because platesolve 3 engine is really fast. It takes 1-2 seconds per image to platesolve.

 

[steppenwolf]

Excellent write-up - thank you for posting. Item number 2. was of particular interest as I haven't considered that aspect before.

[Singlin]

Thanks for this as I have never managed to get an acceptable model.

Which Time server bundle di you opt for?

[dan_adi]

Sidereal technologies makes a small time server (sitech time server). It’s easy to configure, and it connects to about 12 gps sattelites. It has a precision of about half a milisecond. It even sends the gps coordonates to sitech.exe. I set mine to update the pc clock at 15 sec intervals. This is the link from mr Dan Grey http://siderealtechnology.com/IncreasingTheAccuracyOfYourTelescope.ppt 

I must admit I was pretty stuborn regarding modeling the mesu, I did not give up easily, and from what feedback I got from the sitech yahoo group, 40 arc sec seems pretty good. Direct drive mounts with absolutes encoders have below 20 arcsec or below 10, but who has the cash :))), one can only dream, maybe at retirement :). The limit I think in the case of mesu 200 is some small small slip which is not detected by the scope encoder that has a resolution of 2.1 arcmin. With a better scope encoder the pointing could improve. 

[Singlin]

Thanks for the link.

How do I open the file in the above link?

[Singlin]

I have just had a quick look at the instructions.

My problem is that my router is far away from my observatory.

It says that you have to connect to your router with a cat 5 cable aswell as a usb cable to your pc.

[dan_adi]

The usb cable connection with the pc that runs sitech.exe is just a one time operation. If you have just one utp cable in your observatory you can buy a small 4 port switch. If you only have wifi acces, you can connect the time server to the router in your house. As long as the wifi is on the same LAN you can connect to the time server. My time server is in the house not the observatory. Is it clear? My english is rusty.

the link with the ppt presentation takes a while to load, if not you can find it in the sitech yahoo groups

[dan_adi]

The above link works, it just takes 2 minutes to load, they are lots of pictures so it takes time.

 

[dan_adi]

do you guys know any precise way to balance the mesu? I tried to look at the current amps in sitech.exe while manualy slewing but the current varies to much to get a consistent answer for comparison. I was thinking of a hand scale to measure the force applied while moving in dec or ra. Any other ideas?

[tomato]

With regard to the balancing question, I have considered the use of a sensitive spring balance to manually measure the forces being applied to move the mount but confess that with a portable set up only until now, I have not pursued this idea any further.

I was aware of the clock issue with the Mesu which was identified some time ago but was not aware of the fix implemented by ST, so thanks for this.

Steve

[dan_adi]

I will try the spring ideea and see if it adds some objectivity to balancing. The clock issue with the sitech controller is not because of sitech. As I understand the internal controller clock is much more accurate than the pc clocks. Because of these differences, in a short time there will be a disagrememt between the 2 clocks and the tracking will suffer. A while ago mr Dan Gray updated the sitech software to automaticly compensate for the time difference between the 2 clocks, but you still have the main issue -> an innacurate pc clock. So in order to benefit from the best tracking your mount can achive it would be best to have an accurate pc clock

[dan_adi]

I have made some progress with the mesu modeling. After a long and painful :)) drift align in phd2 and a modification of the supervisory threshold in sitech config, I rerun the model with only 20 cal points cause I was sleepy and behold I've got a total rms value of 27.3 " (thats arcseconds :) ) with only 1 outlier eliminated.

So as a conclusion the Mesu 200 can be modeled and has great pointing accuracy. In a  few days I will write a tutorial with some pictures of sitech configuration for fellow mesu users.

Clear skies everybody!