7344 Telescope control question

[jhudler]

Is there a way to connect more than 1 trajectory generators to single axis?

For Example: The RA axis on a telescope must constantly counteract the rotation of earth while at the same time accept moves to a position in the sky. Ideally this is accomplished using 2 trajectory generators one clocking the sky and the other moving to commanded positions. Both generators are responsible for maintaining there respective positions.

TG1: ~0.13 RPM Constant Velocity
|
+----> Motion Output
|
TG2: Move to a new position

Hopefully I�ve sufficiently described the problem.

However, if this is not possible is there alternate solution?

Note: I don�t have LabV
iew.

This is currently implemented using a Compumotor PC-23 in streaming mode where I command the motion in a Windows device driver every 50ms.

Thanks,
Jack Hudler
Lone Star Observatory
http://www.lonestarobservatory.org

[Scott_R]

Jack,

What type of motors are you using (servo or stepper)?

This application can be developed using some programming environment (LabVIEW, Visual Basic, Visual C++, etc) in which you combine the constant velocity move for the earth's rotation with user-defined position moves. You could do this by using the motion controller in velocity mode, where a constant velocity is applied to the axis for the earth's rotation. Then, when adjustments are made based on user-defined position commands, your code would need to compute the velocity move parameters (velocity and time) in order to reach that particular position. You could then combine the two moves mathematically and use this new velocity control for the axis. This code would result in a constant polling of the
user control that is being used in order to determine if the velocity parameter needs to be adjusted.

Let me know if you have any additional questions on this issue.

Regards,
Scott R.
Applications Engineer
National Instruments

[jhudler]

To answer your question: Steppers

What you proposed is what I'm currently doing with the Compumotor setup. Which frankly I'm loath to do again. I was hoping to have the NI card do all the tedious work for me.

While I was waiting on a response I started to look into short blended moves with the occasional long move to the target object. The observed RA position of an object doesn't change however, the Hour Angle (HA) does. HA = RA � Sidereal Time.

So in my test; I loop waiting on completion of the last blended move. I then recalculate the new HA angle in steps and start another blended move. This results in proper tracking but... a very big BUT, it's not very smooth and not acceptable. This would result in elongated stars on the CCD. (See attached
picture)

With the Compumotor setup I wrote my own NT Device driver that sequences the velocity commands to the controller every 50ms on an interrupt from a timer card. This approach is very accurate and over the course of 24 hours only missed 2 interrupts when the system is under stress and results in a very smooth clock rate.

The whole point of all this to loose the need for a device driver and ISA bus. It's getting hard to find motherboards that have an ISA slot in them.

I'm beginning to think I chose the wrong card however; I�m not going to give up just yet.


Jack Hudler

[Scott_R]

Jack,

To improve the smoothness, perhaps you could try polling the user control more often in order to perform the user-defined movements. Instead of performing occasional long moves, it may be better to perform a greater number of small moves that are combined with the constant velocity move.

So perhaps every 50-100ms you could check the user control to see if it has changed and if so, add the position changes to the velocity move which would occur anyway if no position change had been made.

Regards,
Scott R.
Applications Engineer
National Instruments

[jhudler]

I tried that last night.

The issue is I have a demanded position that I MUST keep the scope at. Also the tracking rate is not constant, it changes based upon the angle of the scope with respect to the horizon. Each step is 0.054 arcseconds which leaves me 4 or 5 steps to move around with before the tracking error becomes noticable on the CCD. (still playing with it)

Looked into contoured moves which is similar to the Compumotors streaming mode. But this highlights another issue of reaction time. If the CCD requires a correction during tracking it must do so within 50 ms with the total response time not to exceed 150ms.

I'm starting to get this feeling that I'm stuck with my experience on moving telescopes around and need someone that h
as different exerience and perspective to get me of my rut. Most of the stuff I read online is... well... kids stuff. I'm not spending a $100, I'm spending $10,000.

Can you look around NI to see if there are any telescopes (optical or radio) that use your motion cards. If so can you relay me some contact information or an NI Engineer that helped them with the project?

Thanks again for your support,
Jack Hudler

[Scott_R]

Jack,

I have spoken with our R&D department, and our boards are not able to handle simultaneous control loops. To my knowledge, there are not any boards out there that do.

In order to use the PCI 7344, you would need to make use of a contour move whose datapoints change depending on the angle from the horizon. You could then add the user-defined motion in increments small enough to avoid the significant streaking that you mentioned previously. This would require some programming prior to sending the move parameters to the controller. You could implement a loop for which each iteration performs a move based on the angle from the horizon and any user-inputs.

I checked around, and was unable to locate anyone who had worke
d on a project similar to the one that you have described. I suggest contacting one of NI's alliance members. These are companies that are NI certified to act as consultants, system integrators, developers, etc. You can search through our alliance member database by navigating to:

Alliance Program Search
http://sine.ni.com/apps/we/niaa.search_display

You can search based on location, industry, etc. That page will provide you with contact information for the alliance members who have worked in those areas. Our alliance members have consulted on or developed a variety of applications, and may be able to offer a fresh perspective.

Regards,
Scott R.
Applications Engineer
National Instruments