ATIK OSX Drivers R1.00 & Example App

[NickK]

So.. sorting out the ATIK Titan.. I spent about 45 minutes chatting through the camera in depth with ATIK, this is now looking very good. 

The camera is happy as larry with ExampleApp again - so now I have that working, I need to rebuild and retest on the C++ drivers TSX and Nebulosity4.

So some sets of files for people.

TheSkyX Professional OS X R1.49 Plugin update: ATIK_TheSkyX_Plugin_R149.zip

Nebulosity 4 application framework library update for R1.49: Neb4_Frameworks_update_r149.zip

Note - the fine focus should be working shortly, just need Craig to update Neb4. Just make sure you backup before changing the files over.

I'll sort out a general driver release for OS X tomorrow.

[NickK]

5-05-01 19:53:25.316 ExampleApplication[38232:1589606] FPS: 17

2015-05-01 19:53:26.317 ExampleApplication[38232:1589606] FPS: 13

2015-05-01 19:53:27.314 ExampleApplication[38232:1589606] FPS: 9

2015-05-01 19:53:28.312 ExampleApplication[38232:1589606] FPS: 8

2015-05-01 19:53:29.315 ExampleApplication[38232:1589606] FPS: 9

2015-05-01 19:53:30.312 ExampleApplication[38232:1589606] FPS: 14

2015-05-01 19:53:31.315 ExampleApplication[38232:1589606] FPS: 17

2015-05-01 19:53:32.314 ExampleApplication[38232:1589606] FPS: 17

2015-05-01 19:53:33.313 ExampleApplication[38232:1589606] FPS: 17

2015-05-01 19:53:34.312 ExampleApplication[38232:1589606] FPS: 17

2015-05-01 19:53:35.312 ExampleApplication[38232:1589606] FPS: 16

2015-05-01 19:53:36.313 ExampleApplication[38232:1589606] FPS: 17

2015-05-01 19:53:37.312 ExampleApplication[38232:1589606] FPS: 17

 

Just playing with the new titan code.. and optimising ExampleApp - this is the OpenCL realtime registration pipeline.. spot when I opened the window (I need to work out how to speed up OS X OpenGL windowing to display at 20fps whilst hammering it for OpenCL).. the Titan and pipeline is giving me near 20fps

[NickK]

Ah ha! Found it .. it appears I was making the window for redraw after redrawing it - causing the open window to double draw.. hence the drop to 9-8 fps when it was open.

2015-05-01 21:14:13.290 ExampleApplication[38910:1616162] FPS: 15

2015-05-01 21:14:14.287 ExampleApplication[38910:1616162] FPS: 16

2015-05-01 21:14:15.290 ExampleApplication[38910:1616162] FPS: 15

2015-05-01 21:14:16.287 ExampleApplication[38910:1616162] FPS: 16

2015-05-01 21:14:17.287 ExampleApplication[38910:1616162] FPS: 15

2015-05-01 21:14:18.290 ExampleApplication[38910:1616162] FPS: 16

 

Much better

[NickK]

I just noted SB's Daniel Bisque has indicated that ATIK and SB have green lighted the inclusion of the Plugin into the standard build of TheSkyX - so you'll not need to install it separately

[manu57]

Hi nick,

Thanks for this great work !

I was wondering why the legacy EFW 1 is not covered in your drivers ?

In my quest toward a 100% mac for astro capture I am only missing this. I wrote a simple app to control it from my mac so I know it is actually possible to pilot this filter wheel but I would love to have it integrated in nebulosity for example.

Any chance to have it in the future ??

thanks

manu

[NickK]

Hi Manu - I arranged to borrow an EFW1 at the SGL star party but in the end it was all a bit rushed and I didn't manage to get back to the person that volunteered :/

I have the code, so in theory I could control it, however I don't know it's behaviour (i.e. when you switch or if you press the separate key pad IIRC it had one!). So it makes it very difficult to develop for something I can't reference.

Sorry to say that priority at the moment is getting a job and then I can worry about enhancing. The GP guider is another device that needs adding..

[manu57]

I understand. getting a job is an obvious priority.  

I have a working sample with the major commands this filter wheel accepts. I can give it to you or even help you if needed. don't hesitate !

[rumbleripper]

All the best with the job situation Nick. You deserve it! 

[NickK]

I had some 'free' time last night so I set about improving the ExampleApp - allowing the system to execute the pipeline on CPU only or a combination of CPU/GPU or GPU only.

I'm still attempting to get down to the bottom of the clFFT issues (it provides FFT on GPU, alongside my code inside the GPU). Literally, for some sizes of camera the library simply aborts processing and returns blank. Now there was a bug with OpenCL array offsets, i.e. myArray[index+1] where adding a constant (i.e. +1 or +C) caused the runtime to return zero and not the contents of the location. Annoyingly I can't run clFFT on CPU on the mac due an Apple limitation (workgroups {1,1,1}) in an attempt to track down the bug - I'm going to have to switch to Linux in a VM with GPU support to track that one down further.. more time lost on something that's distracting to somewhere else..

Having this done GPU/CPU hybridising means I'm not blocked developing the new asymmetrical deconvolution reconstruction process I've been working on as I can switch to CPU with some performance impact. 

Switching to CPU also allows for easier debugging although the Mac is notorious for having really poor support from AMD etc - including the lack of development tools. /rant

Incidentally the company interview was great - not sure yet if I make round 3.. the location is perfect - the office sits in an old mill, with the mill stream running underneath and a pub next door! Product great, people good too.. fingers crossed.

Anyway.. today is more DIY..

[NickK]

Well after getting the MBP to boot into ubuntu on an external drive - not an easy task. It appears that installing the linux radeon drivers causes the system to fail on startup (even with nomodeset etc). It's working but the display is blank. From the web this looks to be a problem with having two GPUs switchable - another difficult situation as once the drivers are installed you can't get to a prompt to uninstall them (also the uninstall instructions are wrong/outdated)... I love linux.. (sarcasm overload)

So once I have the issue of the GPU drivers fixed.. this format is good in providing a Linux install for development. I'll target ubuntu 14.04 64bit as it seems a stable build.

[NickK]

Long time without an update

* Still job hunting - priority that is taking all my time..

* MBP GPU finally died - this will need to go to Apple for a fix.

* ATIK have now adopted the drivers although when possible.

I will get around to sorting out the coding for the pipeline at some point but everything's a moot point until I have a job! (stinks when you're made redundant due to restructuring.. nobody's fault :/)

[carkinzo]

Hello everyone, I'm sorry for my not good English.

I just joined and I need the following help. I would use my CCD ATIK TITAN COLOR as guider and not knowing English, I downloaded from this forum PHD2_241c_R145 and I tried to use it with the version 10.10.4. It does not work very well. The CCD is detected, images are shot and then choose the guide star. But there is a communication issue with my mount VIXEN SXP with STARBOOK TEN. When to do the calibration, the first step (WEST) never ends and does not pass to the next steps and occasionally on STARBOOK TEN lights slightly the led of the button WEST and soon PHD shows an error message. Not being computer savvy would like to know if I'm doing something wrong in the process or if I have to install additional software because I have only installed PHD.

If anyone can help me I would be grateful.

Thank you

[NickK]

Hi Carkinzo,

Appologies for being slow in responding - I am currently working to find a job which takes my time away from the forum.

The OpenPHD2 work integrated the drivers to the existing OpenPHD2 software. If the images are being taken - this means the drivers are working for image capture. I would ask how are you connecting the SXP to OpenPHD2?

Is it using the ST4 port on the Titan?

[carkinzo]

Thanks for your help.

the problem is that as I said in my post,

It is that Atik is connected and takes pictures,

but no calibration.

The Titan is connected to the Mac via the USB cable and the

STARBOOK with cable ST4.

In the settings of PHD2 I selected:

Camera: Camera ATIK

Mount: On-camera, because it does not work with the other settings.

Aux Mount: None

I wanted to point out that if you use Parallel Desktop with win7

and PHD1, all function well.

Thanks again.

[NickK]

Thanks for your help.

the problem is that as I said in my post,

It is that Atik is connected and takes pictures,

but no calibration.

The Titan is connected to the Mac via the USB cable and the

STARBOOK with cable ST4.

In the settings of PHD2 I selected:

Camera: Camera ATIK

Mount: On-camera, because it does not work with the other settings.

Aux Mount: None

I wanted to point out that if you use Parallel Desktop with win7

and PHD1, all function well.

Thanks again.

Hi,

I just wanted to check if the mount connected as: Computer --usb--> Titan --ST4-->STARBOOK-->Mount ?

The reason is this is that the drivers provide the ST4 support for the Titan. I have tested the ST4 with a voltmeter, however I use a USB-to-EQ6 box.

The drivers may differ from the windows drivers in that they will not perform changes to the ST4 port whilst a camera image is being taken (or downloaded). It could be this you are seeing - are the Titan images long exposure?

These drivers are purely OS X, the standard OpenPHD2 on Windows uses either the ATIK ASCOM or ATIK SDK drivers directly. The Windows based drivers are not related with these drivers.

Nick

[carkinzo]

Yes, the SXP STARBOOK is connected to one cable D-SUB9PIN.

The Titan is connected with USB cable to your Mac and cable ST4 at STARBOOK.

-

Images taken from PHD2 that I see, I tried with its exposure time available, that if I'm not mistaken there's a maximum of 10 seconds.

Bye

[carkinzo]

I did a test at this time.

The same thing happens. It does the calibration.

It lights up slightly 3/2 times the indicator DX (West) of STARBOOK.

Start with the West and after 61 steps shows an error message:

RA Calibration Failed: star did not move enough.

I did several tests, but the result is always the same.

Now I have to dismantle everything because is coming

a thunderstorm.

Bye

[NickK]

Ok, that's interesting - OpenPHD2 is indicating that it's not seeing the star move enough during it's calibration. This could be down to one, or more, things:

a) calibration steps - the OpenPHD2 calibration need the focal length etc to understand the pixel size. This gives PHD a sense of scale and allows it to know how much to move.

OpenPHD2 integration - this may not be passing the signals through properly. Although the drivers have been voltmeter tested, the OpenPHD2 integration has only been tried on the NEQ6 with EQMOD using EQMac.

c) OSX ATIK Drivers - this has been voltmeter tested however I need to get an ST4 cable. As my preference is to use the EQMOD - this means the guiding can be sent over a different USB channel to the images (regardless of the current state of the Titan or 16IC image download - the old16IC is a 4 second download!).

I don't have the time at the moment todo any astronomy or computing unfortunately. However once I have a job and things have settled down I will continue on this. I use OpenPHD and the Titan on my garden pier.

[carkinzo]

Hi, I would like to ask another question:

these tests I'm doing with the Mac at home, a Mac Pro 8 Core.

The portetile I still have to buy it and to have a longer life

battery, I would buy a MacBook Air, which has 12 hours of battery life.

I would try using Parallels Desktop or Bootcamp and in this

so, I can use the various applications with window and I solved

the problem. But I talked to several people who have told me that

in this way, the battery has a much shorter duration.

What do you advise me?

Thank.

[NickK]

Hi, I would like to ask another question:

these tests I'm doing with the Mac at home, a Mac Pro 8 Core.

The portetile I still have to buy it and to have a longer life

battery, I would buy a MacBook Air, which has 12 hours of battery life.

I would try using Parallels Desktop or Bootcamp and in this

so, I can use the various applications with window and I solved

the problem. But I talked to several people who have told me that

in this way, the battery has a much shorter duration.

What do you advise me?

Thank.

I've used Parallels a few years ago, virtual box and I've used bootcamp.

I found that if the laptop has a discrete graphics chip (like the MBP with the ATI Radeon 6700M and the inbuilt Intel 3000HD series) then using windows tends to make the drivers switch to use the discrete GPU under windows. This causes the increased battery drain.

If the laptop doesn't have two GPUs then it's likely to be just the additional CPU time if you're running under parallels but I found Windows on bootcamp to be less battery efficient but not really too bad.

In the end I have always used an mains electrical hook up. It's just too much hassle to go from battery on night-long sessions.

Using the OS X software is more efficient (my reason for doing this) - however some OS X software is very graphics & CPU hungry and will flatten batteries very fast.

[carkinzo]

Ok, but for the working method that I would like to do,

Parallels would use only for PHD, then with OSX, Red Screen and Canon utilities.

So for me it is enough for the battery to last at least 6/8 hour.

What do you think about it?

Thank.

[NickK]

Ok, but for the working method that I would like to do,

Parallels would use only for PHD, then with OSX, Red Screen and Canon utilities.

So for me it is enough for the battery to last at least 6/8 hour.

What do you think about it?

Thank.

Parallels used to have the issue where the switchable GPU would be running even in idle burning up battery. I'm unsure if they've improved that and how much impact that has depends on the age of the battery with everything else running.

With PHD and red screen (turn down low brightness for less battery use) it should be fine for hours - but it all depends on the laptop and the battery life varies. I found the best thing is test it out at home before attempting to use it in anger (such as at a star party).

My laptop died a death so I'm stuck with using the little mac mini and long cables!

[carkinzo]

Hello everyone, I wanted to know if there is news
for PHD2 and Atik. I have upgraded my Mac to
version 10.11.4 and the program will not start properly,
but when I connect the CCD Atik Titan Color, the program
It closes and then does not work anymore.

Thank

[NickK]

Hi - I think the best option would be to use INDI and OpenPHD. I know that Peter in Cloudmakers still keeps his ATIK drivers up to date with use in INDI.

I'm still on 10.11.3 and won't be able to focus on development of these drivers for the foreseeable future (I will be starting a new job faster than originally planned).