Tips & Tricks / FocusMax
A Brief FocusMax Tutorial
Here is a brief description on how to get FocusMax working for you nicely:
  1. Slew to a 4th magnitude star...this is actually an important step.  Too bright or too dim a star will not give good V-curve results.
  2. After you connect to your focuser (I use MaxIM for camera/focuser control), find a way to settle it more or less in the middle of its range.  My PDF has 7000 steps, so I set it for 3500 to start, right in the middle.  Then, use your manual focus to focus things as well as possible.  (This will allow the focuser and FocusMax to have equal latitude in and outside of focus).  This will be a one-time step.
  3. In FocusMax (FM), on the setup tab (at the bottom), tell the system if you will be focusing in or focusing out to come to focus.  For me, 'in' is better so that the focuser is pulling up against the weight of the camera.
  4. Now click on V-Curve at the top, and the V-curve window will pop up.
  5. You will see two columns or methods of defining the V-curve parameters.  In the first column (called, "end points"), you can specify the start and end points of the V-curve run, along with the step increment.  With my PDF, it has 7000 steps.  If I were successful at having the initial manual focus point set more or less at 3500, then I might set the start at 4500, and the end at 2500, with maybe a step increment of 50.  This will give a V-curve of 40 steps.  This first run will be done by specifying the start and end points.  (The second method is the "half width" column, and we'll get to that later.)
  6. Click on the 'Run' button, and you will start your first V-curve run.
  7. FM will start at 4500, measure the half flux diameter (HFD) of the star, subtract 50 position points, do it again, until it gets to 2500...the end of the run.  Visually, you will see the "V-curve" on the screen, as it starts out measuring the fuzzy stars, gets sharper as it nears 3500, then gets fuzzy again as it hits 2500.
  8. At the end of the run, FM will take a moment to calculate the "slopes".  The right side of the curve has a slope, and intercepts the '0' on the X axis at the bottom at one value - like 3562.  The left side of the slope similarly has a slope, and intercepts the '0' axis at the bottom with a slightly different value, e.g., 3470.  This first V-curve run will be sloppy.  The difference of where the left and right slopes intersect the X axis is the Position Intercept Difference, or 'PID'.  (My FSQ eventually gets to 6-decimal agreement between the left and right slopes, or about a value of 3 for the position intercepts differential or "PID".)  It is mere microns away from perfect focus.
  9. Go back to the main window and click on the Focus button on the Focus tab.  FM will now do a quick run to do an initial focus.  (We are not done!  This is just for the next step).
  10. FM does NOT sniff around the bottom of the curve/tip of the V to "snoop" for the best focus position, as one might logically think.  It instead uses one side of the V-curve slope...figures out where it is on that slope, and calculates the best focus position.  According to the authors, this is far more accurate than the first method.
  11. FM will do an initial focus run...and populate this value into the V-curve window in the "Half Width" column, in the "Center" value.  Remember method # 2 two from above?  Time to use it.  For this, and all subsequent V-curve runs, we will allow FM to populate the "Center" value with the last focus value, and tell it the step value to use.  So, if the focus point (Center) were to be 3500, and the half width is 1000 (amount on each side of focus), with a step size of 50, then again you would have 40 measurement points on your next V-curve.
  12. Run the next V-curve.  After you have done one or two V-curve runs, increase both the width (e.g., 1500 instead of 1000), and decrease the step size (from 50 to 20).  This will increase the range over which the focus is being measured, as well as the number of points on the curve.  This will lead to more accurate results.  You will need a minimum of 12 good V-curve runs to get a good V-curve model!
  13. You will notice that the wider your selection, and the smaller step size you use, the more accurate will be your V-curve, and the more closely will the left and right sides of the V's intercept '0' at the same point.
  14. After you have done 12-14 V-curves, you can delete the original, unaccurate, ones.  Keep the good, accurate, ones.
  15. This whole process is one-time.  Once you have a set of V-Curves, FM will generate superlative, repeatable, focus results.
  16. If you ever change your system, like adding in a reducer, you need to go under "file" to create a new system profile, so that the data from the two different setups does not get mixed up with one another.  You can load the appropriate V-Curve file for each setup that you have.
  17. For those of us with temperature-finicky scopes like an FSQ or TOA, frequent refocusing now becomes easy.  I slew to my target, calibrate the scope for autoguiding.  I then slew to a nearby 4th magnitude star (there are plenty, after all), and do a focus run.  (This is not a V-curve run, we did all of that already.)  A focus run takes about a minute.  Since I use TheSky6, I just click on the "Slew Prior" button, and the scope goes right back to the target object.  Whenever I want to refocus, I just click on that "Slew Prior" button to switch between my target object, and the 4th magnitude star.  You can use CCDAutopilot or ACP to fully automate this entire process.