by Frank Freestar8n
TABLE OF CONTENTS
MG used to connect to mounts many different ways – but it now only connects via an ASCOM device. Most telescopes have an ASCOM interface, and ST4 guider interfaces also usually have an ASCOM interface – so if you need to guide please install the appropriate ASCOM driver.
Note that collimation can be done with no connection to the mount at all – but if you want mount control for collimation or autoguiding, you must set up one of these connections.
ASCOM stands for Astronomy Common Object Model and allows plug-and-play connection between astronomical applications and devices. For more details, and to download the latest version and drivers, see http://ascom-standards.org/.
Be sure to install the latest ASCOM environment, plus the latest ASCOM driver for your mount
Open Setup and select your mount in the ASCOM Scope Chooser
Select your mount’s driver. Then select Properties and set the COM port. If you have done this before in a different application, it may already be set for you
Then press OK and the mount should be connected
Try pressing the N/E/W/S buttons and see if the mount responds by reading its RA/Dec. values. Note that each press of a button will move the mount for 1 second at autoguide rate, which may not be noticeable without reading the display of the mount’s RA/Dec coordinates
You can practice all this with the ASCOM telescope simulator, without any connection to a telescope. You can view the simulator display to see changes in RA and Dec.
MetaGuide requires an ASCOM driver that supports PulseGuide. If your driver does not support PulseGuide (MetaGuide should indicate this when it tries to connect) see if there is a newer or different driver for your mount that does support PulseGuide. MetaGuide cannot guide via ASCOM without PulseGuide
You should enter the desired autoguide rates for RA and Dec. in Setup. Typical values are 0.5 in each direction. After you enter them, MetaGuide will try to apply those rates to the mount, and the mount will use the best available values that match those rates and return them. If the values in Setup don’t match your request exactly, it is because the ASCOM driver has chosen the best match instead.
MetaGuide shown connecting to the telescope simulator directly via ASCOM
Once you have a connection to the mount that appears functional, enter the declination of the star you are observing, and the RA/Dec autoguide rates. (ASCOM will set the declination automatically, and will override the guide rates with values from the driver.) Make sure the star isn't too close to the edge of the view and press FullCalibrate to calibrate the view orientation and scale. A dialog and progress bar will appear in the upper left. Do not interact with the telescope or the application until the calibration completes. The star should never leave the view; if it does, cancel the calibration. Make sure the telescope is not bumped or leaned on during this procedure, and that it goes to completion. The calibration should take 1-3 minutes or so and you should see indications in the graph that the star is moving. If nothing appears to happen then check the ASCOM connection to the mount.
If you have just one bright star in the view, the crosshairs should stay on it steadily – but if there are other bright stars that pull the crosshairs away, simply press Lock and the crosshairs will turn red, indicating they are locked on to that one star. The star may drift about and the crosshairs will track – but abrupt motion may lose the lock. In most cases, it is best to lock the star during calibration.
If you want to choose a different star, such as one that is more centered or one that is not saturated, simply ctrl-click on the star and it will be locked.
The calibration process moves the mount in a back and forth pattern to determine both the scale and orientation of the guide camera. Toward the end of calibration it moves a small amount perpendicular to the E/W direction, but only briefly. The guide camera need not be aligned in any particular way. There is a quicker calibration, QuickCal, described below – but full calibration should take about 1-5 minutes on a typical mount. QuickCal avoids this delay and can be faster. QuickCal can also be used if the mount has no N/S motion. Remember – with MG the mount never moves faster than the autoguide rate. So the motion may be slow but you should still see the mount is moving by changes in the XY plots in the upper right. If you don't see any changes in those plots and calibration takes a long time – please confirm your mount is connected properly.
The motion in RA should be fairly rapid since there is little or no backlash in the constantly-moving RA drive, but the calibration may stall for some time toward the end when it moves in declination. This is entirely dependent on the behavior of the declination gears, so don’t be alarmed if the calibration seems to stall when it begins to move perpendicularly. Note that the final part of the calibration motion will leave the star a bit off center at the completion.
The calibration procedure is important because it sets the calibration factor so the image scale is exact. Furthermore, it determines the orientation of the E/W direction in the field, and whether the N/S direction is inverted. Although the E/W and N/S alignment of the view does not matter for the calibration, it is sometimes convenient to have RA motion nearly along the horizontal axis so that drift in Y corresponds to N/S drift. If you want to orient the guide camera N/S E/W, use the crosshair feature (checkbox at the bottom of the main dialog) and move the star back and forth (with the telescope motion buttons) as you rotate the camera until it is aligned to the crosshair.
ASCOM is able to deduce much more information from the calibration process than other connections because it can tell which whether guide commands move the telescope north or south. This is described in more detail later.
Once the telescope is calibrated, the guiding-related features such as Guide and Center will be enabled.
If you ever exit MetaGuide and return to it with no changes at all to the telescope and camera, you will retain the calibration but you must manually SAVE the setup info (File->Save or SaveAs). If you save the setup file after calibration, MG will connect to the mount and reload that calibration the next time you start MG. If something has changed and the calibration is no good anymore – simply re-run the calibration and save it.
To center a star approximately in the field, just press the Center button (assuming the view is calibrated). This may take some time if the declination backlash is large. This behaves differently when guiding, as described below.
To begin guiding, just select Guide and the star location will be stored and maintained – and displayed in the GuideXY dialog boxes. The location of the star in the field does not matter at the time Guide is pressed. You should probably Lock the star also to avoid another bright object conflicting with the guidestar.
The Center button takes on a different meaning during guiding. Instead of simply bringing the star near the center, during guiding Center will make the target location be the center of the screen. This will slowly bring the star to the exact center of the screen and keep it there.
Note that once you press Guide, the guide error plot starts scrolling – showing errors in E/W (white) and N/S (red) on a +/- 4” scale. This plot is updated every 0.5 seconds, independent of the video rate. The ability to provide a steady update of guide errors at this rate is unique to MetaGuide and its video guiding methods, and is very helpful in tuning the guide parameters because you have prompt, continuous feedback on the guide errors.
There are many ways to tune guiding, including the usual RA and Dec. aggressiveness (here on a scale to 1.0 rather than 10), along with StackTime, which is the amount of video time to use when calculating a star location, and GuidePeriod, which is the time between corrections. These are the main things to focus on to get good guiding.
These and more parameters are described in detail below – but the basic idea is to set aggressiveness so that corrections are made quickly, but do not overshoot and cause oscillations. GuidePeriod is related, and should neither be too fast nor too slow. For my CGE and CGE-Pro I correct every second and typically use a frame rate of 5-10 fps and a StackTime of 0.6s. This amounts to approximately a 0.6s “exposure” and a correction every second. If you are guiding you should make sure the StackTime is less than the GuidePeriod to avoid oscillations.
In addition to the plot that updates twice per second, MetaGuide allows GuideView during guiding, which gives a close up of the live video behavior of the star in relation to the guide location target. This can yield important insight into what limits autoguding with a given setup, because it shows all the fast and slow motion of the star as it tries to center on the guide target. Overshoot, oscillations, and slow backlash recovery are all evident in this view.
GuideView is an extremely powerful and unique feature of MetaGuide because it lets you see what the mount is doing, in realtime, as it chases the star to keep it on target. You may see fast oscillations and other things that you would never have known if you only had long exposures guide images to go by.
Since MetaGuide relies on small corrections, it may be best to set the RA guide rate on the mount to a small value around 0.5x sidereal. On the other hand, Dec. motion may be limited by backlash, so its guide rate may be best set at 0.9x sidereal – along with tuning Dec. backlash so that it is somewhat responsive in both directions, but does not overshoot. But starting values of 0.5 and 0.5 should work.
To summarize, autoguiding just requires connecting to the mount, finding and focusing a guidestar, pressing FullCalibrate, then pressing Guide. You can Lock on the star with Ctrl-Click to avoid being confused by other stars.
Collimation involves adjusting the mirrors of a telescope so that everything is aligned well and stars become nice, round dots rather than flared, comatic smudges. A very good telescope, when slightly out of collimation, can give horrible views - but after a slight adjustment can show a dramatic improvement. This is particularly true with high power views of planets, where details of Jupiter’s bands or Saturn’s rings will suddenly pop out with greater clarity.
Collimation is not something to be feared. It is important to do regularly, and it is much easier with a web-cam and MetaGuide.
The first step in collimation is to view a bright star overhead at prime focus with the telescope – i.e. with no Barlows to magnify the image. I recommend working with the actual configuration you intend to use, and not make changes to the components for the sake of collimation. So – if you intend to view with a star diagonal in place, then leave it there and insert the web-cam in the diagonal where the eyepiece would go. If you intend to image with no diagonal – then insert the web-cam with no diagonal.
With an SCT or Newtonian, the star will appear as an out of focus blobby donut. The first step in collimation is to get this donut round and uniformly illuminated. As you change focus and the donut gets larger or smaller, you will need to change the gain/exposure of the web-cam to keep it from saturating. You may want to put “star-stacking,” in the menu->settings, on Centroid to keep the zoomed in view from jumping around. (Peak stacking, as described below, will center on the bright part of the donut – which may move around.)
One of the most common beginner errors is to use a star that is too bright, and flattens out the radial profile plot at the top. This will lose precision in collimation, and should be avoided. To help avoid this, a red bar appears over the star image whenever one of the pixels becomes saturated. If the red bar appears, go to VidProps and reduce either the gain or exposure until the star peak is reduced. A bright star may require an exposure less than 1/1000 second when focused tightly, so you may need to use a fainter star if the camera cannot go faint enough. But there is no problem using a bright star for collimation – as long as your camera has low enough gain and exposure to handle it.
The old advice to avoid bright stars for collimation does not apply to video – as long as you can set the exposure very short, and the gain low enough, so that none of the pixels is saturated.
Try adjusting the collimation screws to make the donut fairly concentric. It need not be perfect because, as described below, best collimation may in fact occur when the donut is off center. Once the donut is fairly centered, you want to move to high power and focus on the Airy pattern itself – perhaps with auto-centering of the star by the mount.
A detailed view of the Airy pattern requires a high effective f/number, which usually means the field of the camera is very small and the mount must be very stable. Typical f/numbers for a good view of the Airy pattern are in the 25-50 range, and may require a Barlow. If you have some experience with planetary imaging and your mount is stable at high power, this may pose no difficulties. Otherwise, you may want to start at lower magnification.
Find a bright star near the zenith – the brighter the better, but it should also be very high up to improve seeing. Center it with an eyepiece, then substitute the web-cam and focus. For imaging the diffraction pattern the star must not be saturated, which means extremely short exposures are not only possible, but required. For an 11” aperture, the exposure may be 1/200 to 1/1000 second. Once the star is in the field of the web-cam, its zoomed in view should be immediately visible in the lower right, along with its radial profile in the lower left. You do not need to select the star with a bounding box or click on it – it is found automatically. Carefully adjust focus and exposure so that the profile height is about 2/3 maximum and as narrow as possible. Note the FWHM plot on the lower right for reference. Note that for best results with the in-focus star appearance, you should set Star-Stacking to ‘Peak’ instead of ‘Centroid.’
Other video settings may be adjusted, but frames per second should be 10-20 and gamma should be 1.0 so that the radial profile corresponds to linear intensity. Use the gain setting to place the peak height about 2/3 of maximum.
Superimposed on the image of the star in the lower right is a red dot. If the star is flared at all due to coma or other aberrations of miscollimation, the dot will be offset in the direction of the flare. To collimate, carefully adjust the collimation screws so that the red dot moves toward the center of the star. If the star easily goes out of the field, try working without a Barlow – which is fine as long as the magnification is large enough that you can still see the elongation of the star. Note that the direction of the flare in the zoom view matches it in the wide view. With an SCT, you typically want to move the star in the direction of the flare itself – and re-center the star to see if the flare is reduced. Repeated motions like this should center the dot on the star when collimation is achieved.
One of the goals of MetaGuide is the auto-centering of the star during high-power collimation, which requires that guiding be enabled as described in the next section.
View of comatic star with coma dot in the upper left, in the direction of the flare. To correct collimation, simply adjust the collimation screws to make the star spot move in the direction of the coma dot – i.e. to the upper left. You need to keep re-centering the star in the middle of the field as you make adjustments – either manually by moving the telescope, or by auto-centering (see below).
View of coma dot when in collimation – also showing the Airy pattern. This is only possible at high f/ratio with a Barlow – in this case a 12.5” cdk with 3x Barlow. Note that the first Airy ring has appeared, and is in roughly the correct location based on diffraction theory.
For an excellent write-up on the strange appearance of stars away from the zenith, see http://www.paquettefamily.ca/astro/star_study/. This is a wonderful example of the power in combining empirical results with simulation to document a phenomenon that is largely unknown, regarding the wedge-shape appearance of stars. This effect could greatly confuse collimation efforts if not taken into consideration.
MetaGuide not only lets you see the diffraction-limited performance of your telescope in a live view, but the autoguiding features can automatically keep the star centered as you adjust collimation. This means that as long as you can see the computer screen, you can focus on adjusting the collimation screws and not have to worry about recentering the star. This requires you have set up guiding and have MG connected to the telescope as described in Chapter 4, and calibrate the camera orientation.
To collimate precisely, first make sure you are roughly collimated using standard procedures – then aim at a bright star high overhead as described above. Adjust the brightness and gamma so that you can see the shape of the star – particularly any comatic or oblong appearance.
Now, center the star and press Guide to begin guiding. As long as you don’t let the star leave the field, you may now make adjustments to collimation and the mount will recenter the star in response to changes. With practice this feels natural, and allows you to get immediate feedback on the changes made by a turn of a collimation screw. If the red dot, indicating coma, is consistently off center of the star, adjust until it is centered. It is much easier to keep track of the changes and their effects since the star is automatically centering itself after changes.
MetaGuide works with video cameras that output any size, and the user can even select alternate output sizes supported by the camera. For example, a 1280x960 camera may provide additional formats at 640x480 and 320x240. To choose the desired size, just enter it approximately (it need not be exact) in the CamWidth and CamHeight entries at the upper right of the Setup dialog. MG will use the closest size that the camera actually provides. Note that the user enters the DESIRED camera size, but once the camera is connected, if you go back to the Setup dialog you will see the ACTUAL camera size in use, which may be different.
No matter what the camera size, the view will be sized to fit exactly in the MetaGuide display, with a possible gap along the top and bottom or sides if the aspect ratio is different from 640x480. Although the video is resized for display, the raw data from the camera is always used for centroiding and for the zoomed-in star view at the lower right. Keeping the view the same size is particularly convenient for small computer displays such as on netbooks.
If you aren’t sure of the native size of the camera, in pixels, you can always request a huge size, such as 8000x8000, and the largest camera size available will be selected for use.
Large format cameras may allow you to select a smaller size and then move an ROI (Region Of Interest) around the sensor to select different parts of the scene. This is particularly useful when exploring edge stars because it allows you to have a high frame rate, with the smaller ROI window, while still seeing how the entire sensor behaves.
For the ASI1600 cmos sensor it is over 4000 pixels wide but you can select a much smaller size in the MG setup dialog – and then use the VidProps proper page – and the ROI Misc. tab – to move the ROI region around – as shown here:
You can use that dialog to move theo ROI as you like – e.g.
This lets you have a high video frame rate and good interactivity – while also having small pixels to get a good view of the star shapes around the field.
Normally MetaGuide will lock onto stars directly with no special settings by the user, but sometimes there is a need to refine the sensitivity of the star detection algorithm. Examples are when the star is faint against a noisy background, or when thin clouds pass through resulting in changing star brightness. The two main parameters to adjust are the Star Threshold and Lock Radius. Star Threshold controls how faint a star can be relative to the background noise so that it is detected as a star. For a faint star against a noisy background this threshold should be low, while for a bright star it can be high.
Normally the Threshold and Lock Radius don’t matter, but if a cloud goes by and blocks out a bright star, MG may then find a noisy pixel nearby to lock onto, causing the mount to drive away from the guide position while the cloud is present. To prevent this, set the threshold high enough so that the noise level in the area of the star does not cause false star detection, and set the Lock Radius to a small enough value that it can track the star as guide corrections are made, but reduces the search region wherein false stars may be detected. The Lock Radius must, at the same time, be large enough that when the cloud passes and the star returns, it will still be within the search radius – otherwise it will be out of range unless it happens to re-enter the original target area. Although you may lose a sub-exposure while the cloud is present, due to a loss of guiding with no guidestar, the goal is for the star to be recovered when it reappears within the lock radius, so guiding can then resume.
MetaGuide has several checkboxes to control the ‘look’ of the main screen: Enhance, Crosshair, Cleanview, BigView, GuideView, and SeeingView. Enhance is intended for faint scenes and especially for integrated video scenes of deep sky objects. Enhance will bring out details in the faint information and make faint stars more visible. It has no effect on the star centroiding or locking, which is based on the raw video data.
Crosshair simply places a crosshair on the screen to assist in alignment and centering of objects.
Cleanview removes the plot from the lower left, and the star view from the lower right so you only see the view of the camera. When you combine this with Integration (from the Settings menu) and Enhance, you can turn a simple video camera into a deep sky viewing camera.
BigView makes a much larger image of the star spot and coma dot for easier collimation when the screen is far away.
GuideView shows the motion of the guidestar relative to the guide location during guiding. It won't show anything useful if you are not guiding, but when guiding it will show a tilted crosshair indicating the N-S/E-W directions and show fine motion of the star around the guide target as the mount receives guide corrections.
MetaGuide has a Focus button that opens a dialog with a running plot to show the size and brightness of the star. Just choose the size that matches the expected range of the star size during focusing, and monitor the intensity and fwhm values as you change focus. There is also an audio button that gives a higher pitched beeping sound as the fwhm decreases. Note that the fwhm can be misleading if the star spot is not Gaussian because sometimes the spot can be very flat and wide, but with a small, narrow peak in the middle. This would have a small fwhm even though it is far from focus. To avoid this issue, make sure that as the fwhm decreases, the peak intensity increases – indicating a true focus.
The audio feature is somewhat experimental and should be used with caution, to avoid bothering people or even animals – particularly outside at night. It also has a problem with some computers in that the audio stops, particularly running on battery. So it may or may not be helpful to people.
“Seeing” causes a star in a telescope to blur slightly and change shape rapidly, while moving around a bit at high magnification. The main cause is attributed to shearing layers of atmosphere moving at different velocities, particularly near the jet stream, but more local effects near the ground or even in the telescope can contribute. There are many ways to measure or estimate seeing and they all have their quirks and ambiguities, and the one in MetaGuide is no different. I personally prefer to go by things that can be directly measured – but they don't directly correspond to some absolute measure of the atmospheric conditions. You can measure the size of stars in long exposure images – but that includes errors from guiding and quality of focus. You can use a particular device such as a DIMM, but even they have their own forms of bias that make the result less clear cut.
The seeing measurement in MetaGuide is simple: it is the fwhm of an integrated star spot that has been exposed for two seconds. As simple as this sounds, even it is full of ambiguities and issues of interpretation because there are heuristics involved in defining the center of the star and the background level – both of which are needed to determine the fwhm. Furthermore, the imaging system must be well focused and very high magnification – perhaps 0.5” per pixel, to get a good measurement of a 2” fwhm.
Nonetheless, if MG is used with a video camera at the focus of a long focus telescope with reasonable aperture, perhaps 3” or greater, it should at least capture the trend in seeing conditions during a night and in comparing different imaging sessions. And if you can attach the camera directly to your actual imaging system and focus/calibrate it accurately, you should be able to measure the approximate limit of fwhm that you can expect with that system on that evening with good guiding.
Note that this will not translate directly if you are measuring the seeing through OAG or in a smaller guidescope – since neither will have star spots as small, in arc-seconds, as can be reached with a larger, full aperture system. So it is important to recognize the potential benefit of this seeing feature, while at the same time be aware of its limitations. Despite these limitations, it is much better defined than simply going by the measured fwhm in your images as “the seeing” that night. That value will place an upper limit on the seeing, but the contributions from imperfect focus and guiding remain.
Unlike the other plots in MetaGuide, the seeing is not smoothly updated every 0.5 seconds, but instead is updated every 2 seconds. It is shown in the fwhm plot in the lower right, and numerically in the lower left. SeeingView lets you see an image of the 2s star, along with its radial profile. If the star is moving around a lot, the image may be very bumpy and the fwhm value will not be very meaningful. The seeing value is also broadcast in the messages received by MetaMonitor, so live plots of seeing can be viewed on a different computer on the network.
It's important to distinguish this “Seeing” fwhm, or SFWHM, from the normal aligned and stacked fwhm, or AFWHM, that MetaGuide normally shows for the star spot. The whole point of collimation with MetaGuide is to remove the effects of seeing so that you can study the diffraction pattern with less impact from the atmosphere. So, the fwhm reported by MetaGuide (the aligned fwhm, or AFWHM) is normally much less than the current seeing value and should not be confused with “Seeing.” Now that MG provides both measures, it's important to keep them separate and use them appropriately.
MetaGuide can create very realistic simulations of the expected star appearance in Simulate mode, under Menu->Settings->Simulate/Integrate. The results can be surprising because they reveal just how tiny the diffraction pattern can be unless you use very high f/number, as with a Barlow lens. The simulation mode requires a camera to be connected and runs at the current frame rate of the camera. A simulated Poisson background noise with mean=10 provides a realistic version of video noise – and it shows more clearly in Enhance mode. Try changing the f/ratio and secondary obstruction to see not only the changing appearance – but how it begins to stand out against the pixels.
MetaGuide can be controlled by other applications using Windows remote messages, as described below. This also allows you to control it from scripting languages such as Python or Visual Basic. Unfortunately Maxim does not allow direct invocation of these Windows messages, but you can wrap such invocations in a script invoked by Maxim. To make this easier for Dithering, an executable, MGDither.exe, and a script, MGDitherVB.vbs, are included in the MG directory. To perform dithering between exposures in Maxim, simply invoke the MGDitherVB.vbs script between exposures. You may need to adjust the delay time in that script so there is time to settle after dithering, and you may need to place the script and the executable in a directory where they can be invoked.
Also included in the MetaGuide directory is a sample Python script that shows how to access a telescope via ASCOM and control MG using remote messages. The script is called MGControl.py, and it requires the installation of the Win32 Python package.
MetaGuide doesn’t have a direct connection to Canon or Nikon cameras via LiveView, but there is an app that can provide a DirectShow interface that works with MG called SparkoCam - https://sparkosoft.com/sparkocam It is a low-cost but powerful app that allows your dslr to be used as a video camera with a very large sensor for collimation. The large sensor is important because you can study evidence of coma in the stars near the edge to get it balanced. This is somewhat different from collimation per se and amounts to optimizing the overall field performance – but that is what is needed for deep sky imaging with a large sensor.
MetaGuide can output guide information onto the local network for monitoring with a remote computer. This is very convenient, especially when conditions may be changing due to thin clouds, etc. The Setup dialog has an Extra tab that shows a Broadcast Mask and a Broadcast Port. If your local network allows UDP broadcasts, set the mask appropriately and all computers in the range defined by the mask can receive guide status messages from MG. Since this is UDP, there is no handshaking of MetaGuide with the monitoring computers, and they may come and go as monitors unbeknownst to MetaGuide. The default mask is 192.168.1.255 and the default port is 1277. That mask happens to work with my Linksys wireless network.
MetaMonitor is a separate application with a graphical display of the star intensity in one graph, the guide error in another, and the star location in another. There if a fourth graph showing the sky temperature, but currently that only works with the MGUSB protocol under development.
If the star is lost, the star location screen goes red and serve as an obvious sign of trouble. An alarm can be set to go off if the star disappears for some time. If clouds pass through, the star brightness will show reduction even if there is no cloud detector installed.
The UDP packets are very lightweight and should be no burden on the network. Note that the router must enable such UDP broadcasts, and the computer firewall settings must allow MetaGuide to make such broadcasts. You may get a warning from the antivirus or firewall software indicating MetaGuide is trying to broadcast. If you are unsure, simply disable the setting at the top of the Extra dialog page and the broadcast will not happen.
MetaMonitor currently uses Port 1277 for guide messages.
MetaGuide has many features to learn, but some are more advanced than others. In order to focus new users on the key features, MetaGuide has an Advanced and Beginner mode. Beginners should start in Beginner mode, but can switch to Advanced mode through Menu->Settings. In either mode, every GUI item has flyover text to provide helpful hints. Just pause the mouse over an item to get a hint of its purpose and usage. To read this help document, go to Help->Help or press the F1 key.
Setup Access the control panel for describing the optics, camera pixel size, and for connection to the mount. Also allows specifying the desired camera frame rate
VidProps Control the camera brightness, gamma, etc.
FullCalibrate Do a full calibration of the guide camera, with a ‘+’ pattern
QuickCal Do a quick calibration with motion only in RA. This requires proper setting of NSInvert and ViewParity, as described below
Calibrated This is a readonly indication that the mount has been calibrated and now Guide can be pressed to begin autoguiding. Note that you can recalibrate any time – even though the mount is already calibrated
LockStar Lock onto the current selected star so that other stars are not selected if they come into view
GuidePeriod The time, in seconds, between corrections sent to the mount. Too fast can lead to oscillations, and too slow can result in poor tracking. This tends to be faster for mounts with imperfect gears, but adequate response time to corrections
FrameRate This is the actual frame rate of the camera. The desired frame rate is entered in the Setup dialog, but the true frame rate provided by the camera is shown here
StackTime The time to gather frames used in calculating the star centroid. This acts as an effective exposure time, but it is based on analyzing each individual video frame. StackTime should not exceed the GuidePeriod when guiding
StarThresh This value helps select stars in challenging conditions. The star detection algorithm works very well, so this may never need changing – but if stars are visible yet not detected, try changing this
GuideView When guiding, this shows a live video view of the guidestar relative to the guide target – along with axes showing RA (long) and Dec (short) directions. This is very useful for seeing how the mount behaves as it chases the guidestar
SeeingView This shows a view of the star integrated over 2 seconds, without aligning. This gives an idea of the seeing conditions – but it is mainly just a relative guide
Enhance This will performa image processing on the main view and let you see much fainter stars – but it is computationally demanding and may slow down the frame rate. It has no effect on guiding or centroiding and is just a visual aid for the user
SaveImage Save annotated images of the zoomed in star, plus the main screen. These are PNG and FITS files in the output directory specified in Setup. This will document the appearance, profile, and fwhm of the star, both to document your ‘scope’s performance, and to share with others. The main screen will be saved at the full camera resolution currently in use, and with Enhance, Cleanview, and Crosshair as set by user. The FITS files will include the RA/Dec information if an ASCOM mount is connected. The FITS files are always raw images – either straight from the camera or integrated with Integration or LockStack60.
Focus Opens the focus dialog
Flexure Opens the flexure dialog for measuring flexure with another instance of MetaGuide – as described below
ZeroDelta The drift of the star is constantly being measured, and the motion is shown in dEW, dNS, Distance, Rate, and Drift PA. ZeroDelta lets you reset the calculation of this drift. dEW, dNS, and Distance indicate the current displacement from the start, while Rate and PA indicate the rate of drift and its direction, respectively. This can be helpful when drift aligning. Note that since it uses the EW/NS directions, the mount must be calibrated for these to operate.
Upper/Lower These control the appearance of the zoomed in star, to help show the shape and structure of the comatic outer region. Reset will restore them to default values
MetaGuide MetaGuiding is different from normal guiding because it locks onto a specific gear period, measures its amplitude and phase, and corrects for it in realtime. This is useful for fast gearbox terms and terms that are not harmonics of the worm period and therefore cannot be fixed by PEC. The desired period is entered in Setup, and the MetaGuide checkbox causes corrections to be made. As you guide, MetaGuide tracks the corrections and determines the amplitude and phase of the specified period. If something changes about the guiding, press Reset to start a fresh calculation of the phase. Once you see the amplitude and phase stabilize, and while already guiding, press MetaGuide to begin preemptive corrections for the specified term
Shift This enables “shift guiding,” which lets you track slowly moving objects such as comets. Simply enter the E/N rates into the setup dialog and begin guiding. If Shift is enabled, the guidestar will slowly be offset to track the object – but you must also be guiding on a star at the time
Center This button does two different things depending on whether or not you are guiding. If you are not guiding, but you have calibrated, Center will just move the star to near the center of the screen and stop. If you are guiding, Center will set the target location in the center of the screen permanently. This is useful during collimation to force re-centering of the star as you adjust
Dither Every time you press Dither while guiding, the target x/y location of the star will be moved to a random Gaussian spot around the guide location. This is useful to reduce background noise and avoid hot pixels. Note that it should be done between exposures and not during. This can be controlled remotely, as described below. The current sigma of the Gaussian is 6 arc-seconds or 6 pixels, whichever is greater. Note that this is as-measured by the guide camera, not the imaging camera. Normally this button is not pressed manually – and instead is controlled by a separate app that sends remote messages to MG to dither between exposures
GuideXY GuideXY allows you to enter a specific x/y coordinate for the guidestar, and guiding will keep it near there. The actual location may be different by +/- 1 pixel or so. Whenever you press Guide, the x/y location are placed there and stored with the .mg file so the exact same camera framing is repeated. This is particularly useful when continuing an exposure run of the same object on multiple nights. Be careful to press GuideXY if you want to use the specified x/y values, because pressing Guide will overwrite those values with the current star position
x, y Pixel coordinates of the currently selected star – also shown in the adjacent plot
East/North East/North coordinates of the star relative to the center of the screen, in arc-seconds. Note that any value involving East/North requires a prior calibration
RA/Dec Drift Drift rate as measured over previous several minutes, in “/minute. This is useful for drift aligning the mount. Note that periodic error can make the RA rate very misleading, since it tends not to be a linear motion but somewhat oscillatory
FWHM This is a very important item indicating the Diffraction Full-Width at Half-Maximum based on the optics, including the secondary obstruction. This can be quite different from the True FWHM, which is measured directly from the live, stacked star profile. Note that the measured FWHM does not represent a measurement of “seeing” because the video images are aligned and stacked prior to the measurement. Furthermore, small f/ratio systems on large pixels will tend to have much larger FWHM’s than expected by diffraction. This is why a view of the Airy pattern usually requires a Barlow to get to around f/30 for the Airy pattern to appear. But when guiding, FWHM should be as small as possible to guarantee good centroid accuracy. Ideally it should be in the 2-4” range if possible
Log Log starts multiple logs of the star location and guide corrections. If the mount is not calibrated, a simple x,y log of the star location is output. If the mount is calibrated, many more quantities are output for use in studying the guiding quality and the periodic error of the mount. More details of the log format are below. The log prompts for a comment each time, but this can be turned off by directly editing the .mg file. Comments can be very helpful to understand what the log captured when viewed much later
File-Save Save the current configuration and guiding parameters in a .mg file. This .mg file will be used by default the next time MetaGuide starts
StarStacking StarStacking determines the type of centroid used for the star position. Centroid is the usual center-of-gravity calculation, while Peak uses a windowed centroid around the point of peak brightness. This avoids biasing the centroid by noise and turbulence in the outer region of the star spot. Peak is the recommended mode, but there is a more extreme version, Point, that uses a very small window around the brightest pixel of the star. In most cases Peak will work best, but when shift-guiding on a comet, Centroid may work best. Collimation is not possible in Centroid mode, though, and the coma dot is disabled.
Color This mode sets the color of the zoomed in star view in the lower right. Some users may prefer a simple gray scale vs. the colored version
IsoLines This is a simple way to show Isolines in the zoomed in star, to help reveal its shape
Gamma This alters the view of the zoomed in star and can also help reveal its shape. Note that when measuring the fwhm and comparing to the diffraction plot, use Gamma=1
Simulate If you have a video camera hooked up, Simulate will show the theoretical appearance of the star along with its diffraction pattern. Note that this will tend to be MUCH smaller than the actual star spot unless you are at high f/ratio through Barlowing
Help Help should bring up this document in pdf form, but it requires the Adobe Acrobat reader to be installed
About Shows the MetaGuide version number
Username, Scope Description, and Location are provided for documentation purposes. Log files include the user name and scope description in the file name for convenience
Aperture, Prime FNumber, and Barlow/Reducer magnification help define the image scale and predict how quickly calibration should succeed.
CalFactor can be entered by the user, but its value is normally set by the calibration process after the true image scale is determined. This factor results in the calibrated FNumber and the calibrated focal length
Sec. Obs Secondary Obstruction should be entered as percent of diameter. For an SCT it is typically around 34%, while for a refractor it is 0. This value affects the diffraction pattern – in particular the brightness of the surrounding rings
Declination With an ASCOM connection this is filled in automatically, but with other connections you must enter the declination of the current object during calibration and guiding. This value is not very critical, but should be accurate to at least 5 degrees or so. You should set it every time you change declination significantly to calibrate or guide
RA/Dec Rate These are the current autoguide rates set in the mount. For ASCOM, enter desired values and the true values of the mount will be determined automatically. For non-ASCOM connections, you must enter the actual values set in the mount. Typical values are 0.5x sidereal. The RA guide rate should always be <1 to avoid backlash, but the Dec guide rate could be more than 1 if it seems to work better. These values require tuning and experimentation to determine optimal values
WAngleXAxis This value indicates the direction of West relative to the X axis. This is determined by the calibration process and normally the user does not modify it
RA/Dec Agg Aggression determines how much of a correction should be made based on the measured guide error of the star. A responsive mount under good seeing may benefit from a high aggression value of 0.9 or so, while poor seeing may benefit from lower values around 0.5. These values require tuning, and optimal values may vary from night to night
Dec Lash This is a form of software backlash correction. When the guide error of the star reverses direction from N to S, a long pulse of this duration, in milliseconds, is added in to help kick the mount the other direction, taking up backlash. With imbalance in the dec. axis, the optimal pulse time may be greater in one direction than the other. Typical values are 200 ms. This may be used in concert with the mount’s built-in backlash correction
Dec. Reverse This specifies how far the star must drift in the opposite direction in Dec. before corrections are sent, with corresponding backlash pulses. If the value is 0”, then every time the star changes from N to S the mount will react with backlash correction. A value of 0.2-0.5” will reduce how often this kick occurs, and avoid oscillations
Block Corr Corrections can be blocked in any of the directions with these checkboxes. Some people like to offset the polar alignment so a star drifts only one direction, allowing them to disable corrections in the other direction. You can also disable corrections in E/W – for whatever reason
ViewParity This indicates whether the camera view is correct in terms of handedness – i.e. rotation doesn’t matter, but the E direction must be counter-clockwise of the N direction for ViewParity to be false. If it is a mirror image, ViewParity will be checked
NSReversed This indicates whether guide corrections to the mount in the N direction actually move the mount north. Some mounts, such as the CGE, will change the direction in which N moves the mount on a meridian flip. This is described in more detail below
KillHotPixels If the camera view has hot pixels that distract the star detection algorithm, cover the scope and press KillHotPixels. The hot pixels will be identified and removed. This can be very effective both in finding stars well and removing distracting spots from the view
Occultation This is an experimental mode in which the selected star is measured for each frame and dumped to a log with approximate timings. It is not intended for actual occultation timing, but for general experimental needs where a star and/or seeing are monitored at high speed
PixelSize Enter the pixel size in microns, for proper image calibration. Typical values are 5.6um for a Toucam Pro, 7.4um for a SKYnyx 2-0m, etc.
Pixels/Arc-sec, Arc-Sec/Pixel, Field Width, Airy FWHM: These values are all based on the calibrated focal length and on the theoretical diffraction pattern, including the effect of the secondary obstruction
Manual Control These buttons move the mount for one second at guide rate in each of the directions. Holding a button down does not maintain motion – each press is only for one second. This is largely for diagnostics and testing the connection to the mount
UsePrevCal This is an important button that lets you recover a previous calibration as long as nothing has changed in the camera or mount. After you calibrate, press Save or SaveAs and the calibration info will be stored with the mount. On restarting MetaGuide, connect to the mount and press UsePrevCal and the mount will immediately be calibrated and ready to guide on a star. If the camera has rotated at all since the calibration was saved, a fresh calibration should be performed
MeridFlip This is an important button for equatorial mounts after a meridian flip, where the guide camera view may rotate 180 degrees. See below for more details on how to use this feature. You may also need to press NSReversed on a meridian flip
ASCOM/USB Make the correct selection for your mount as described above
ASCOM Chooser If guiding by ASCOM, press this button to select your mount driver and connect to the mount
ScopeName If using ASCOM, this is the ASCOM name of the mount driver, including connections via Hub. This value is saved in the .mg file and reloaded automatically
OutputDir This is the directory in which logs and images are saved automatically. It may be desirable to store a .mg file and logs along with images in the same directory to keep track of the session information, and to compare results and guide parameters later
MetaMonitor This enables broadcast of status information from MG over a local network using UDP. Your network must allow UDP broadcasts for this to work. Make sure you set the broadcast mask and port for your network.
Shift Rate Shift rates for comets when shift guiding. This usually comes from a planetarium software
DitherRadius Radius of a random pattern used for dithering, in arc-seconds
Min and Max Move These set the min and max pulses sent to the mount when guiding
Rotator Angle If you use a rotator or manually adjust an OAG angle, you can maintain calibration as long as this angle is set properly during calibration – and every time you change the angle you also change this value.
Reverse Rotator Calibration This allows reversing the meaning of the rotation angle if it happens to be backwards
Remote Guide →> GuideXY If a separate app tells MG to begin guiding, this forces MG to place the guidestar at the previous x,y location to avoid drift between guide sessions
Hot Pixel Sigma Reject This sets the sensitivity to hot pixel rejection. A typical value is 3. If you find there are still hot pixels remaining, try a lower value
Wavelength This is useful when collimating and using a special filter, such as IR. An IR filter is handy for collimation because it makes the Airy pattern much larger and easier to see. Entering the wavelength here will make sure the theoretical size of the diffraction pattern is a good match
Video has an undeserved poor reputation for autoguiding, but it has many advantages both for advanced users and beginners. The key advantage is that everything appears in real time, with no latency or pauses. This lets you catch guide stars when they appear in the view, and lets you see first hand the fine motion of the mount that lead to guiding problems. Video at 8 fps with a SKYnyx 2-0m camera allows my C11 to see guidestars down to about mag. 10. There is no need for the user to squint at the screen to see guidestars since MetaGuide finds even faint stars automatically and places crosshairs on them.
Even focusing is easier with video because it is so interactive and responsive. This makes it possible to use a simple thumbscrew to focus the guide camera, since slight adjustments of the camera are immediately evident in the live view of the star.
A key problem for beginners is simply focusing the guide camera on a star – any star. A good trick is to adjust gamma so that even a faint “doughnut” of an out-of-focus star appears clearly as a large speckled washer in the view. With gamma=1, nothing may appear – but with higher gamma it is evident. It is good to experiment with all the parameters of your video guide camera to help detect out of focus stars, and to help faint stars show when in focus.
Note that collimation with a bright star may require very short exposures – perhaps 1/5000 second, depending on the f/ratio. The profile of the star, in the lower left view, should not be flat at the top, and should only go up to ½ or ¾ of the plot scale.
The plots show two curves, and the first item shown is white; the second is red. For example, in the X/Y plot, X is white and Y is red. The main fwhm plot also shows seeing as purple.
The X/Y plot is normalized in each direction, so it goes from left to right and bottom to top for X and Y respectively.
The next plot shows X/Y drift, and is simply a zoomed in view of the first plot. This takes some getting used to, but it is extremely informative. In order to give a sense of the motion of the star in x,y while still at a magnified scale, the plot “wraps” on itself when it goes off scale. Thus a star with increasing X value will rise up from the middle of the plot to the top of the scale, and then drop down abruptly – then increase again creating a sawtooth pattern. It is essential to realize that the sawtooth has nothing to do with the actual motion of the star, except that it is steadily increasing. The frequency of the sawtooth then corresponds to the rate of motion. This is an unusual way to show drift, but once you get used to it, it is very informative.
The Error E/N plot only applies when guiding, and shows the current guide error in arc-seconds in the east direction (white) and north direction (red). By studying this plot as you adjust parameters such as aggressiveness, you can get a feel for how to keep the errors as small as possible, and to avoid oscillations and overshoot.
The FWHM/Intensity plot shows FWHM (white) and Intensity (red) for the star. For best focus, the white line should be low and the red line should be high.
Since MetaGuide uses a rolling set of frames for its centroiding, these plots are always updated twice per second, regardless of the effective exposure for the centroid. This gives much more consistent feedback on focus and on guide tuning.
There is much talk of the sub-pixel accuracy of centroiding algorithms, but it is usually based on assumptions of nice, round, Gaussian stars and ccd’s suffering only from read noise. In fact, for short exposure guide images, the guidestars tend to be misshapen blobs with ill-defined centers. A typical center-of-gravity algorithm will find the center of mass of that blob and be heavily biased by whatever strange things are happening near the edge. This can result in completely overlooked errors in the centroid that directly increase the resulting error in the guide corrections. In addition, ccd’s are susceptible to other forms of noise, including pixel crosstalk, that cause guide stars to bloat far beyond the size expected based on diffraction and seeing conditions alone.
MetaGuide uses a novel form of “Lucky Guiding” that directly targets the problems of these misshapen stars. Instead of using the entire star in a 1 second exposure to determine the centroid, which would be biased by edge distortions, MetaGuide finds a windowed centroid around the peak of each video image of the star. These stars are sorted based on quality, and the best fraction, specified by AccepFrac, of those stars are used to determine the full centroid. Each of the accepted frames has its own windowed centroid value, and the centroid of those centroids becomes the final centroid used for guiding.
This improved accuracy and avoidance of star edge distortions directly improves the centroid used for corrections. This, coupled with rapid corrections with low latency video allows tight and aggressive autoguiding even with a mid-range mount.
One example of the benefit of this algorithm is to look at the FWHM of stars stacked by the normal centroiding algorithm, and one stacked using the Peak (recommended) centroid of MetaGuide. The version stacked using Peak will have a smaller FWHM. This is essential to resolving the Airy pattern during collimation also.
It used to be that astro-imagers would focus on the optics, with less concern for the mount and camera. This changed recently with better autoguiding when people realized how important the mount is in getting small, round stars. Much of the difference is associated with reduced periodic error – but more important may be better bearings and smoother gearboxes – both of which require much more expense to manufacture to the required tolerances and high quality. The situation went from a $300 film camera on a $500 mount with $1000 optics – for example – to a $1000 camera on a $9000 mount with $2000 optics. Without a good mount, better optics were not a win since the results were limited by noise in the mount.
Although autoguiding software and technology have improved over the years, high-end mounts tend to give much smaller star fwhm’s than mid-range mounts using typical autoguiding setups. Adaptive optics can be a cost-effective way to get high-end results from a mid-range mount, but it is limited by the availability of bright guidestars and added complexity of the autoguding setup. What struck me, though, is that mid-range mounts worked noticeably better even when the adaptive optics were operating as slowly as 1 Hz. To me, this meant that tighter autoguiding at the 1 Hz range without adaptive optics might still show improvement, and the two key issues were to reduce latency, and to improve centroid accuracy. With tight corrections every 1 second, using a video centroid, a mid-range mount can now achieve 2” fwhm or less and yield results that rival high-end mounts. This inverts the priority of cost in an imaging setup - reducing the need for an expensive mount, and instead putting more emphasis on the camera and optics.
There are numerous ways to collimate a telescope, but many concentrate on the appearance of the out-of-focus star and ignore the in-focus shape. In fact, complex optical designs with several components may show the best stars when the out-of-focus appearance looks misaligned. Although the out-of-focus appearance (e.g. “centering the donut”) is a good first-step in collimating, it should be followed by careful in-focus collimation at high power. Normally this requires a night of very steady seeing, but with MetaGuide the diffraction pattern is more readily visible and collimation based on the in-focus diffraction pattern is now easier.
Understanding the nature of the diffraction pattern and how it relates to star size and resolution involves distinguishing the physical size of the diffraction pattern on the image plane from the angular size it corresponds to in the sky. The former is determined entirely by the f/number of the telescope, while the latter includes the magnifying effect of the telescope’s focal length.
The size of the Airy pattern is proportional to the f/number of the telescope, and independent of the focal length. Thus the size and appearance of the spot formed on a Toucam Pro will be the same for a 5” f/10 as it would be for a 50” f/10 telescope. Only the irradiance would be different, with the 50” concentrating 100 times as much power into the same size disk.
The angular size of these disks will be quite different, however, and will be 10x smaller, in arc-seconds, for the 50” than for the 5”.
In short, the physical size of the Airy pattern depends only on the f/number, while its angular size depends only on the diameter of the telescope.
For some concrete examples:
The angular resolution of a 6” telescope is twice as good as for a 3”, regardless of focal length.
A 3” f/10 will have the same size (in micrometers) Airy pattern as a 6” f/10.
If you have a 3” f/10 and you replace it with a 6” f/5, which has the same focal length, you will quadruple the light-gathering, and you will halve the size of the Airy pattern. This means that for the same detector, the gain in ADU of the central pixel will be roughly a factor of 4x4 = 16!
Of course, this all assumes the seeing and imaging are “diffraction limited” when in fact the measured FWHM of deep sky images is typically much larger due to seeing.
MetaGuide lets you see how your telescope is performing at the diffraction level so that you can collimate it optimally, reduce guiding errors, and keep the net FWHM as low as possible in your final image.
NOTE! You must use a high enough f/ratio for the Airy pattern to be large enough to be resolved with a web-cam’s resolution. This may require a Barlow, and an effective f/ratio of 20-50. At lower f/ratios, the star spot is blurred by the bleeding and crosstalk of the pixels and will be artificially much larger than the diffraction limit. See the examples at the end of this document.
The SaveImage button creates two .png image files with annotations. One is a small image that shows the raw view of the star, along with the stacked view and the radial profile; the other is the wide view of the camera itself. These images are annotated with many of the optical parameters and capture a great deal of information about your telescope, collimation quality, and seeing.
The images are also annotated with UserName+ScopeDescription on one line, and Location on the next. These lines are truncated to 30 characters, so keep them succinct so they will appear in the PNG dump. The image files are named and indexed automatically based on the date.
People often talk about how great their optics are, and how they saw several rings in the Airy pattern, but until MetaGuide there has been no easy way to document this performance and it is left as a subjective anecdote. MetaGuide allows people to document their seeing with the push of a button, as shown in the examples at the end of this document. This is the first tool that makes it easy to document and compare their star images, with much of the atmospheric effect removed. If someone is boasting about their optics and the diffraction rings – why not push a button and share the result with the world?
The MetaGuide installation package associates the setup files with .mg extension with MetaGuide, so that launching from an icon will restore the full session settings of that setup. It may be beneficial to store a .mg setup file with each imaging session to keep track of what settings you used each time. Alternatively you can have a single folder with several setup files in it – corresponding to different OTA’s, different cameras, and different reducer/Barlow combinations. You can also launch MetaGuide from a command line, with the .mg file as a parameter.
Have imbalance both in RA and Dec. for better response
Polar align well so there is minimal dec. drift, and chase dec. errors aggressively
Study the plots to tune the guide parameters so there are no oscillations and no long periods with constant error
Have a good, well-aligned finder scope with crosshairs oriented NS and EW to help center on stars and focus well
Have good automatic focus setup – preferably using mult-star curves that find focus with a parabolic fit.
I much prefer off-axis guiding with stars at long focal length, rather than a small guidescope. But MetaGuide should work fine at guiding either way
For Off-Axis guiding, get a compact OAG and keep the distance from the pickoff mirror/prism to the guide chip as short as possible. This gives the guide chip a wider view of the light cone and increases the guide star intensity – while also minimizing aberrations. If you have spacers both in the guide path and in the imaging path – try to remove them so spacers are only required in one path
I use a driveway setup, and one thing that improved my productivity enormously is a TeleGizmos 365 series Scope Cover. This allows me to leave equipment set up and aligned even in bad weather, so I can start imaging effortlessly compared to lugging equipment out and setting up each time. Just beware of spiders and wasp nests.
I often see people talk about how ‘great’ their autoguiding is with particular equipment – but without quantitative information, it is hard to know just what that means. Some people cite the displayed error of their guidestar while autoguiding, perhaps with values like 0.1” rmsd - but what does this have to do with the size of stars in the image, particularly when flexure may be producing oblong stars?
To me, the best measure of autoguide quality is the roundness and smallness, in arc-seconds, of stars in long exposure images – at least as long as one full worm period. An autoguide setup may be keeping the star roughly centered and even display excellent values for the guide error – but this can all be misleading if the centroid is not accurate or if there is flexure between the guide and imaging chips.
Stars may appear round and small, but when measured they may actually be 6” in diameter, instead of a more desirable 2”. Without measuring the FWHM of stars in 10-20 minute raw exposures (no processing other than calibration with bias, flats, darks) it is hard to know just from appearance just how good the guiding really is.
Widish field imaging with refractors can produce impressive, sharp stars across the field – but due to the short f.l. of the imager, those stars tend to bloat and mask underlying guide errors. This makes for a much easier setup with which to autoguide and image, particularly since refractors can be guided well with guidescopes, but the resulting stars may have FWHM in the 4-6” range, which is much easier to achieve than 2” FWHM with a long focal length SCT. The latter really does require good guiding. For this reason, I recommend noting and tracking the FWHM in your images. This will not only allow you to document and improve your autoguiding results, but you can share them with others in a quantitative way that can be understood and compared. When you get long exposure images in the 2” FWHM realm, you really are starting to guide well.
There are several very sensitive NTSC/PAL analog video cameras used for astronomical applications such as meteor imaging and occultation timing. Although they are sensitive, they suffer a loss of resolution due to the analog nature of the signal, compression used to encode the signal, and the conversion of the raster scan to a rectangular pixel array in the image. Nonetheless, they can be used with MetaGuide through an inexpensive video2USB converter.
Given these concerns, analog video cameras may not be ideal for collimation or autoguding – but they may work allright. One way to reduce the effect of pixel bloating and compression artifacts is to guide at long focal length – so the angular size of pixel-scale effects is reduced.
See more details on the usage of these cameras in the above section on Enhance, Crosshair, and Cleanview modes.
NexRemote and MetaGuide work well together with or without ASCOM. First connect NexRemote to the mount, and specify a virtual port for use by other applications – say COM 10. NexRemote should connect to the telescope over a real COM port, nowadays usually a USB2Serial converter. Let’s call that “real” port COM 5.
Once NR is connected to the scope and the mount is aligned, connect MetaGuide to the mount via ASCOM.
Drift alignment is a standard method for polar alignment of a mount. It has a bad reputation for being time consuming and difficult, but is well known for the resulting accuracy. MetaGuide makes drift alignment much easier by providing quick feedback on the declination drift of a star as you adjust the mount.
Details of drift alignment can be found at many sites on the web, so I will summarize here how to incorporate MetaGuide in the process.
First, find a brightish star near the meridian and near the equator, and calibrate the mount with it. (You should only have to calibrate the guider once during this whole process.) Do not press guide, but let the mount stabilize for several seconds so the drift plot of the star looks steady. Now press ZeroDelta to reset the measurement of the drift rate of the star. After several seconds, you can read the Dec. Drift rate just left of the plot showing X/Y drift rate. Although the RA drift rate may oscillate due to periodic error, the dec. drift rate should be fairly steady. (I recommend using Periodic Error Correction, PEC, for this process if it is available on your mount – but it is not essential). Note the sign and magnitude of the dec. drift and adjust the mount azimuth control a bit (by physically rotating the mount head – not using direction controls) while keeping the star in view. You may not need to recenter the star as long as it stays on screen, and you definitely do not need to recalibrate the guider (only the mount has changed – not the relationship of the guider to the mount). Press ZeroDelta again and watch the new value of the dec. drift. If it is worse, you know you went the wrong way in your azimuth adjustment. If it is better, you are now on track to get the drift close to zero.
Once the drift of the low star near the meridian is minimal, move to a star low in the east or west and near the equator but probably on the same side of the meridian so you don’t have to meridian flip. Repeat the dec. drift measurement above, but this time adjust the mount in altitude, changing the up/down angle of the polar axis to zero out the dec. drift.
You may iterate this if you like, but just one iteration with an East/West star and a meridian star should get you well polar aligned with much less effort than squinting in an eyepiece trying to gauge star drift.
FullCalibration will always work to find the orientation of the guide camera, but the added time it takes to move in declination is undesirable. QuickCal does a calibration based on motion only in RA, but you must have ViewParity and NSReverse set correctly or else the N/S corrections will be backwards. You could do QuickCal and press guide – and see if the N/S guiding is centering the star or pushing it away. If it pushes the star away, just press NSReverse and it should then work fine.
A more systematic way to do this is to use ASCOM to do a full calibration, and then make note of the values it finds for ViewParity and NSReverse. An equatorial mount may have different values of NSReverse on either side of the meridian. ViewParity, however, should depend only on the camera and number of mirrors. Normally, ViewParity is OFF if the number of mirrors in the optic path is even – i.e. a refractor or an SCT. But OAG introduces an additional mirror, so ViewParity should be ON with OAG.
The other thing that happens on a meridian flip is the 180 degree rotation of the scene. The Meridian Flip button allows you to do a meridian flip and immediately resume guiding without recalibrating.
To make this concrete, here is how it works for me with my cge:
With OAG, ViewParity is ON because the number of mirrors is ODD.
When imaging in the East, NSReverse is FALSE.
With OAG, when I do a meridian flip, I must also rotate the OAG to recover the guidestar. Thus the view in the guide camera does not change at all – but NSReverse does turn ON in the west, so I must turn it on.
On the other hand, if I guided on my cge with a separate guidescope, I would have:
ViewParity is OFF because it is a simple refractor
Imaging in the east, NSReverse is FALSE
When I do a meridian flip, the view rotates 180 degrees, so I must press Meridian Flip
For my cge, in the west I must set NSReverse to TRUE
This may take some getting used to, but it greatly expedites and simplifies the meridian flip process – which often happens in the middle of the night when people need all the help they can get. If in doubt, do a QuickCal or FullCalibration.
The Shift button allows tracking a comet or asteroid that moves slowly relative to the stars. Enter the rates of motion in RA and Dec (“/hour), from a planetarium program or ephemeris, for the object and guide on a star as usual, but with Shift enabled. The guide star will be tracked as usual, but with a growing offset that tracks the comet. If the motion rates are set correctly, the comet can be exposed longer and reveal sharp detail that would otherwise be lost.
High end mounts can track at different rates directly, without the need for autoguiding, but this would not work for mid-range mounts because of periodic error and dec. drift, plus gearbox and bearing noise in both axes. The shift mode of guiding combines the precision of autoguiding with a reference star, plus the accurately know rate of motion of the comet or asteroid. Slow comets can be exposed much longer with no loss of detail, and fast comets remain sharp even though they might be noticeably blurred in a 30 second exposure without shift guiding.
on a fast comet requires some effort because the guidestar will
appear to move off the guide chip over time, while the comet remains
fixed on the imaging chip (as desired). This means you need to
leapfrog from one guidestar to the next as the comet moves along.
This requires pre-planning with a planetarium program. It’s
important to know the orientation of the tail in order to frame the
imaging camera and then choose good guide stars.
This image was guided with Shift enabled, using E/N motion rates read from TheSky. Shift allowed longer sub-exposures while still capturing the sharp edge of the tail. This is comet C/2006 VZ13, and this is a very sharp image of the tail that can be compared favorably with others on the web. This used 2m sub-exposures, which would have been greatly blurred without shift-guiding.
MetaGuide is aware of numerous messages from other applications on a PC that allow it to work in concert with applications by sending Windows remote messages as a broadcast.
MetaGuide supports the following messages – all of which except SHIFT ignore the values of wparam and lparam (so values of 0, 0 are recommended).
Please consult the file MGControl.py for all the latest commands and how to use them from Python.
Lock on current star
Disable lock, allowing lock on current brightest star
Set shift rates via wparam, lparam (details below)
Enable shift guiding
Disable shift guiding
Do meridian flip – rotate guide camera view 180 degrees
Reverse N/S directions, as for equatorial mount on merid flip
Disable N/S reversal
Mark the log with two integers, using wparam, lparam
Save an image, just like pressing the button
wp contains the rotator angle x 10 as an integer. Useful for maintaining calibration when using a rotator to set the OAG angle
Wp contains the dither radius in arc-seconds x 10 as an integer
Wp contains an integer corresponding to the DirectShow value for the exposure. Typically if n is sent, then the exposure will be 2^n – and n can be negative. If n = -2 the exposure will be 0.25s. The min and max values are shown in the Setup dialog
Wp contains an integer corresponding to the DirectShow value for the gain of the camera. The min and max values are shown in the Setup dialog
All commands can be sent with wparam = lparam = 0, except for MG_RemoteSetShift and MG_RemoteMarkLog. For that Shift, wparam and lparam map an unsigned integer to a signed, float value corresponding to the shift rates in “/hour. The mapping is:
float ewrate = (wparam-10000);
float nsrate = (lparam-10000);
The system broadcast command corresponds to the following code that applies to many programming environments:
UINT handle = RegisterWindowMessage("MG_RemoteSetShift");
WinPostMessage(HWND_BROADCAST, handle, wparam, lparam);
The high accuracy of the MetaGuide centroiding works best when the guidestar is imaged at long focal length, and there is little flexure in the optics. This is the case with off-axis guiding. Off-axis guiding has a bad reputation due to people years ago randomly searching for a guidestar. This is all different nowadays with planetarium programs that allow custom field-of-view indicators (FOVI’s) that let you pre-plan the best guidestar near your object. If the OAG is calibrated and has direct readout of angles, then the guidestar can be literally dialed in prior to imaging. Instead of hunting for the guidestar, you just hunt for the object itself and roughly frame it – which will then make the guidestar appear on the guide video view. You can then calibrate using that guidestar and commence guiding after optimally framing the target object. Note that the guidestar should be well focused and not too bright or faint – based on studying its intensity profile and fwhm. If the guidestar is too faint, try using Integrate mode.
Off-axis guiding setup showing simple degree indicators.
FOVI view in TheSky showing selected guidestar and corresponding angle.
The Logging button starts logging in two separate files. Each file is given a unique name that combines the user name, scope description, date, and an index. One log, MG_XXX_XXX.log, is a simple log of 3 columns containing time, NS, EW (pixels) if not guiding, or time, sumE, sumN if guiding. The former is for passive logging of the star, e.g. for periodic error measurement; the latter can also be used for PE measurement, but is based on the actual autoguider response. The other log file, MG_Full_XXX_XXX.log, is a very detailed listing of values. The file has a header line at the top describing the contents in each column.
It's important to note that the log lines are output every 0.5s even if the guide corrections are every 2s. That means you can see how the star is moving even while guiding.
Note that units of arc-seconds are associated with NS/EW measurements, and units of pixels are used for X/Y screen coordinates.
The PECPrep log is for the EQMOD PECPrep utility, which allows diagnostics of the periodic error and more.
For python users, the full log can be loaded and plotted easily with the following code:
# load the log file as a pandas dataframe
f = pd.read_csv(logname, sep=",", comment="#")
# plot e/w position vs. time
plt.ylabel("raw e/w (arc-sec)")
_ = plt.title("Raw view of e/w position of star in entire log")
For application developers and general hackers, MetaGuide is constantly broadcasting information over the local network as UDP strings, if you have enabled UDP broadcast in the Setup/Extra dialog. The strings are easy to parse and you just need to be able to catch the broadcast messages as they arrive.
For details see the MGListener.py file, which includes an MGListener.py class that documents the messages and allows you to capture them with your own code.
MetaGuide can measure small flexure and flop between two mounted telescopes by looking at drift and shift at the sub-pixel level. To do this, start two copies of MG on the same computer and associate each with a different camera. Under the covers the two copies will be communicating with each other. Then select the copy that goes with the longer focal length telescope, connect it to the mount, and calibrate it. The other MG need not be associated with the mount or calibrated.
Then press the Flexure button and a screen will appear as below
Now press Calibrate in the flexure dialog and follow the directions to move the telescope a small amount N/S and E/W. When the flexure dialog is calibrated, begin guiding using the long focal length MG that has already been calibrated and wait a bit as the flexure rate is measured. The radial dial indicates the instantaneous and averaged flexure rates.
To measure flop, place a star in the center of the field and zero the flexure measurement. Then either move the telescope and return to the star, or move to a different star and begin guiding again with the star in the center. The delta value shown, in arc-seconds, corresponds to the flop that has occurred.
Flexure measurement in realtime is particularly challenging since the slow motion on the sub-pixel scale is difficult to measure without noise. MetaGuide displays both the instantaneous values of displacement, along with a linear fit to the past several minutes of motion. Pressing reset will restart this calculation, and over time the flexure rate should stabilize.
It is very important to have the star autoguided and centered during the measurement, due to the precision required in measuring these small drift rates.
MetaGuide is entirely written in C++ with STL. The GUI components rely on MFC, but there is no use of .NET since MetaGuide is multi-threaded and delivered with most components statically linked – which I believe is disallowed by .NET. ASCOM is incorporated using the early binding mechanisms in C++. One detail is that ASCOM is encapsulated by a separate thread using a GIT, or Global Interface Table. This allows the emulation of an “infinite, interruptable pulseguide” by looping 1 second pulses in a separate thread. This gives greater control over the calibration process, and allows all calibration and guiding motions to be done using only PulseGuide calls. The only ASCOM command that MetaGuide uses to move the mount is PulseGuide.
The video component of MetaGuide relies on the DirectShow environment. MetaGuide uses an in-place transform filter, GuideFilter.ax, to handle the core image processing routines in its own thread.
MetaGuide users are encouraged to join the new forum at https://www.smallstarspot.com. Please review the content there and feel free to join and post your question.
I am a professional scientist with formal training in optics and physics. I have a range of publications and patents in a variety of disciplines and the list continues. I have always been interested in astronomy and optics, starting with a 3” f/5 refractor in the late 1960’s (Jupiter’s moons), to an Edmund 3” f/10 reflector (Saturn’s rings), to a home ground 4.25” f/5 reflector in the early 1970’s. I became interested in astrophotography and built a darkroom with used equipment, and shot hand guided (with guidescope) images of the sky on Tri-X and 103a-f hand-rolled film cartridges. In those days, Jack Newton was promoting Tri-X astrophotography and, with effort, getting faint renditions of objects like the Pelican Nebula. Somewhere I still have his astrophotography pamphlet from back then.
Later I used a range of telescopes including a 16” f/15 refractor on pendulum clock driven mount. I stopped amateur astronomy for some time, then got back into it with a Meade 7” Maksutov LX50, followed by a CGE1100, with which I did my first autoguided imaging using Guide Dog and Canon EOS with film. I felt that video with novel centroiding would help in many ways to aid collimation and document optical performance - and those same centroiding innovations would also improve AutoGuiding. That’s how MetaGuide came about.
I have imaged all over the earth, from the midwest, southwest, and northeast USA, to Ireland and now Melbourne, Australia. I enjoy persuing novel methods that can get maximum performance from my equipment – and I have no problem departing from conventional practice, which is largely driven by anecdote. I think that mid-range equipment can get much better results with the right software and technique, and much of what is currently considered best-practice can be improved. The images an amateur can capture with mid-range equipment today are far beyond what I would have imagined possible in the 1970’s.
Also at the main MetaGuide site: https://www.smallstarspot.com/metaguide/images
Many thanks to Andre Paquette for extensive beta testing. assistance in packaging the application, and the invitation to use the AstroGeeks site to host MetaGuide.
Many thanks to Andrew Johansen who did tons of testing and suggested valuable contributions.
Thanks to Dave Rowe for valuable suggestions on the phase detection and coma calculation.
Thanks to everyone in the AstroGeeks group for helpful feedback and testing, and for pointing out the need for support on small laptops
Thanks to Mark C. Malburg for his Oscilliscope control on The Code Project, which was the basis for the scrolling charts.
Thanks to Chris Rowland for improving the Celestron ASCOM driver
Thanks to Chris Shillito of EQMod for direct import of MetaGuide logs into PECPrep