Advanced Setup Options

(Auto Test Variables)

Hello Again J

Below I will attempt to describe all the options under the Platform Settings / Auto Test Variables Dialog Screen.

Before we start messing around with any of these options, it Might be best if you read the Theory of Operation Page! I say this as a lot of what’s mentioned below will be rather difficult if you haven’t an understanding of the method the code uses for testing J

So Here we go:

1. Click Platform / Auto Test Variables from the main screen. Which will give you the following dialog box.
2. 
1. This is a duplicate of the last tab on the Platform Settings / Platform and Foucault Dialog box
2. These adjustments have NOTHING to do with the intensity of the light source!!!

i. Your light source should be set to get a full 255 pixel intensity reading on the bar graph display

ii. The light source might not make it ALL the way to 255 depending on the mirror F/Ratio and camera selection

iii. So long at the light source makes it above 200, you should be ok

3. What these sliders do is set the “Shade of Gray” used during testing.

i. For instance, lets say we just started testing a mirror.

ii. We start looking at the zone pairs for Zone 1 (Closest to center of mirror)

iii. The first thing we need to do, is get a reference shade of gray established.

1. This is done by laterally (sideways) moving the knife edge into Or out of the return light cone

2. During an auto test, the code will do the above until it gets the “Shade of Gray” between the Min and Max intensity sliders selected above!

iv. Here’s the Rub, if the Min = 30, and Max = 40, the code may never find this “Sweet spot” and will get stuck in a loop clicking the knife in one notch, then out another notch!

v. Conversely, If the Min = 30 and Max = 100, the code can easily find the “sweet spot”, but this will degrade the overall test result accuracy (Not sure why, it just does!)

vi. Here’s something fun to do while testing a mirror:

1. Run a test with Min = 30, Max = 60. Record Results

2. Run a test with Min = 60, Max = 90, Record Results

3. Run a test with Min = 90, Max = 120, Record Results

4. Compare the results of the three tests. I’ll wager test one showed a Larger amount of Longitudal Travel (to-from) mirror than test 3 did. Why, I’m not sure! However real world star and Ronchi-Star tests show that test 1 will be closest to what the mirror really is!

3. Next Tab “Knife In/Out Stuff” (how the code controls lateral motion of the knife)
1. Here’s what the dialog screen looks like:
2. 
3. First things First, Frame averaging

i. As you know any video system has inherent noise

ii. This noise can result from:

1. Electrical noise in the camera or Capture system

2. Air Currents

3. Wiggly floors

4. Trucks driving down the road blocks away

5. Etc

iii. All of this adds up to one problem. No single frame of video can be considered pristine enough for the code to make decision with!

iv. So, to overcome this, the code does A LOT of noise cancellation by “Stacking Frames” and averaging them.

v. On this tab, the top slider controls the minimum number of frames to average before making ANY decisions on moving the platform in any direction!

vi. If you’re testing a slow mirror (F/8 – F/12+) you may need to crank this value up! Especially if the code is “Hunting Around” for the correct lateral position of the knife-edge.

4. Next on this tab is the Knife In/Out variables, broken down into 3 categories

i. Delta Difference (perhaps poor choice of words J)

1. If the current “Shade of Gray” is greater or Less than our target range (30-60?) +/- the “Delta Difference” The code will move the knife laterally in larger steps.

2. This is handy for a few reasons, Lets say the mirror is fully illuminated + your step size is ridiculously small. If the code just moved the knife laterally one step at a time, it might take an eternity to get the “Shade of gray” correct!

3. BTW, there is built in logic to prevent continually overshooting the “Shade of Gray” value. Once the Overshooting flag is set, the code will only use the Small Motion step size for the remaining time the code tests at this longitudinal location

ii. Large Motion

1. This Is calculated by your entry in Lateral step size previously entered

2. Make sure this value is HIGHER than the small step size! (something I will fix in later releases!)

iii. Small Motion

1. Again, this is calculated by your lateral step size entry

2. Preferably this will be as small as possible

3. Caveats:

a. Too Small and the test will take longer to run

b. Too Big, and the code may Hunt back and forth for the correct Lateral position that results in the “Shade of Gray” in the correct value.

5. Finally, at the bottom of this dialog box is the resulting “Psuedo-code display”

i. Everything in my code is done in Nested Loops!

ii. When an Auto-test is started, the loop begins.

iii. Inside this loop are a ridicules amount of If-Then statements J

iv. So, using the above variables, the entire loop for getting the Knife edge in the correct position to create the “Shade of gray” looks something like this:

1. If Frame Average > 4 Then

2. If Zone Intensity > 60 + 10 {Max + Delta} Then:

3. Cut Knife Into Return Beam 0.00075 {Or 6 Half-Steps}

4. Reset Current Frame Average to 1 and Start All Over!

5. End If

6.

7. If Zone Intensity < 30 - 10 {Min - Delta} Then:

8. Cut Knife Out of Return Beam 0.00075 {Or 6 Half-Steps}

9. Reset Current Frame Average to 1 and Start All Over!

10. End If

11.

12. If Zone Intensity > 60 {Max} Then:

13. Cut Knife Into Return Beam 0.000125 {Or 1 Half-Steps}

14. Reset Current Frame Average to 1 and Start All Over!

15. End If

16.

17. If Zone Intensity < 30 {Min} Then:

18. Cut Knife Out of Return Beam 0.000125 {Or 1 Half-Steps}

19. Reset Current Frame Average to 1 and Start All Over!

20. End If

21.

22. Else

23. Get another frame of Video data,

24. Average that data

25. Start All Over at the top!

26.

27. End If ** If the code gets this far,

28. ** Then its time to Rough Adjust Distance from Mirror!

6. So far, we’ve only discussed options for LATERAL control of the knife edge, for more info on EXACTLY how its done, see the Theory of Operation page on the left.
3. NEXT TAB J Finding current zone null (step #1)
1. This page of variables tells the code how to ROUGHLY get close to the zone null, and looks like this:
2. 
3. At the top of this page is another frame average slider!

i. The purpose is again noise cancellation

ii. Also, during the process of frame average counting up, the Previous code that controls Lateral position of the knife is still in play!

4. Now for some interesting stuff. Forward Delta, and Forward Travel.

i. This is how the code tests for Oblate mirrors!

ii. In addition, this part of the code helps to eliminate back lash in the test system by Always approaching a zone null by moving away from the mirror.

iii. For instance:

1. Lets say we have a perfect paraboilic mirror and zone 1 is perfectly nulled

2. Also lets say the ‘Shade of Gray’ is in the appropriate range

3. The problem is the code has no way of knowing the amount of backlash in the system!

4. Consequently the first thing the code is going to do is ensure that the zone opposite the knife edge side is brighter than the zone on the same side of the knife edge

5. The amount of difference is the Fwd Delta Variable.

6. So in this case, zone left and right side are equal, consequently, the code will move the platform forward until there is a difference of 15 pixel counts!

7. The amount of distance to move forward is set by the Fwd Travel variable.

8. This will continue until the difference in pixel intensity (left vs right) exceeds the Fwd Delta variable

9. At the end of this loop, the code KNOWS FOR CERTAIN that the zone being tested is Further away from the mirror than the platform currently is. Thus each zone tested will always be tested with a motion that started closer to the mirror, and moves away from the mirror.

10. The goal for this is to eliminate backlash AND enable the testing of oblate mirrors!

5. Next fun stuff is the Away Delta, and Away Travel

i. This portion will only come into play once the Forward Delta and Forward Travel section has been satisfied

ii. When this section of code starts, WE KNOW that the zone being tested is further away from the mirror than the knife-edge is currently.

iii. Consequently the zone side opposite the knife edge will be in the Min/Max “Shade of gray” Range (30/60 by default)

iv. In addition, odds are that the zone side ON the same side of the knife probably has a pixel intensity of near zero.

v. This results in the platform moving the distance set in the “Away Distance” variable, ONLY if the Difference between the two zones is greater than the “Away Delta” variable

6. Sorry, I know this is strange stuff, its kind of hard to make it simple J Just know that for the most part you can leave this section alone. The only realy purpose for tweaking these variables would be to increase testing speed. Personally I favor a slower more accurate test, than a fast and possible sloppy test J
7. Here the resulting pseudo code for this bit of the loop (Keep in mind, the previous psuedo code for Lateral motion is always running as well!)

1. ***Assuming Knife On Right side of light source!***

2. If Frame Average > 15 Then

3. If Found_Oblate = False Then:

4. If Left Zone Value > Right Zone Value + 15 Then:

5. Move Knife Towards Mirror 0.005 {or 40 Half-Steps}

6. Reset Frame Average to 1 and Start Over!

7. Else

8. Found_Oblate = True

9. End If

10. End If

11.

12. If Found_Oblate = True Then:

13. If Left Zone Value > Right Zone Value + 20 Then:

14. Move Knife Away From Mirror 0.002 {or 16 Half-Steps

15. Reset Frame Average to 1 and Start Over!

16. End If

17. End IF

18. Else

19. Get another video Frame, integrate it, then and start over at the top!

20. End If

21. **** If we made it this far, then its time to Move onto Final Null Code!

5. Last Tab for this dialog box, the Final steps to finding a zone null, which looks like this:
1. 
2. At the top again, is the last and final frame average slider. The code will not evaluate the final null until the frame average has reached this level.
3. Next down is the step size to used to “Creep” up to the zone null.

i. Default value is 0.001” (if English used, else its (0.0254mm))

ii. You can make this value smaller, which might improve the test accuaracy, but at the cost of a longer test run!

4. Last is the button to reset all these values to what I consider an optimal level. (I might be wrong, but at least it will help you recover from a little bit of over creativity on variable changes!) J
5. OK, if you made it this far, perhaps you might be interested in seeing the ACTUAL VB code that does the automated testing
1. It’s a little long, and I’m a internet taught programmer
2. IN addition, this is the subroutine that does testing when the Knife-edge is on the RIGHT. There is another subroutine that is similar and is used when the knife edge is on the left!
3. Here it is, in all its ugliness J

Private Sub Magic_Right()

If ZoneOn = False Then

MsgBox "Please setup zones first ;)", vbOKOnly, "Do the Zones!"

Exit Sub

End If

abort = 0

Manual = False

Timer1.Enabled = False

DoEvents

Dim series As Long

Dim CurrZone As Long

Dim CurrFrame As Long

Dim Dummy As Long

Dim Delay As Long

Dim TextBuff As String

Dim TxtZone As Long

Dim TxtSeries As Long

Dim StartTime As Currency

Dim EndTime As Currency

Dim LastMoveIsLeft As Boolean

Dim LastMoveIsRight As Boolean

Dim FoundOblate As Boolean

Dim WW As Long

Dim Filehandle%

Dim data As String

Dim file As String

QueryPerformanceCounter StartTime

Move_Mot

Text1.Text = ""

For Dummy = 0 To 16 'Clear any old Knife edge readings!

KEZ(Dummy, 0) = 0

KEZ(Dummy, 1) = 0

KEZ(Dummy, 2) = 0

KEZ(Dummy, 3) = 0

Next Dummy

MirRad.Enabled = False 'keeps user from accidentally chaning mirror radius slider

Command3.Enabled = False

Command1.Enabled = False

Command5.Enabled = False

Picture3.Visible = False 'Hide the Manual speed indicator

Form_Resize

'This will be the new and improved main loop (I hope!)

For series = 0 To Reps

For CurrZone = 1 To NZ

APlay App.Path & "\Sound\Z" & CurrZone & ".wav"

Text1.Text = "Now working on Zone " & CurrZone & vbCrLf & Text1.Text

WorkZone = "Testing Zone " & CurrZone

For CurrFrame = 1 To TFFinalFrames

capGrabFrameNoStop lwndC

FrameAverage = CurrFrame

Call Plot_Zones

If CurrFrame > LRMinFrames Then 'Only do this if we have averaged enough frames!

If LZV(CurrZone) > Max Then 'Only do this if Left zone intenesity is > Max and we need to move the knife Left

If LZV(CurrZone) > Max + LRDelta And LastMoveIsRight = False Then

For Dummy = 1 To LRLgMove

x = x - 1

If x < 0 Then x = 7

Xpos = Xpos - 1

Out &H378, (coil(x, Abs(mnuinvert.Checked), Abs(mnuLRswap.Checked), Abs(mnuSwapMotors.Checked)) + coil(Y, Abs(mnuinvert.Checked), Abs(mnuTFswap.Checked), Abs(Not (mnuSwapMotors.Checked))))

Delay = 0

Do While Delay < Mindelay.Value * 2

Delay = Delay + 1

Loop

Next Dummy

LastMoveIsLeft = True

CurrFrame = 1

Else 'Since we got here with a need to move knife left, then if either zone is > max and < Max + LR Delta *OR* Last move was right (Since we don't want to overshoot!) then move in Small increments

For Dummy = 1 To LRSmMove

x = x - 1

If x < 0 Then x = 7

Xpos = Xpos - 1

Out &H378, (coil(x, Abs(mnuinvert.Checked), Abs(mnuLRswap.Checked), Abs(mnuSwapMotors.Checked)) + coil(Y, Abs(mnuinvert.Checked), Abs(mnuTFswap.Checked), Abs(Not (mnuSwapMotors.Checked))))

Delay = 0

Do While Delay < Mindelay.Value * 2

Delay = Delay + 1

Loop

Next Dummy

LastMoveIsLeft = True

CurrFrame = 1

End If

End If 'This will take care of Reducing zone intensity

If LZV(CurrZone) < Min Then

If LZV(CurrZone) < Min - LRDelta And LastMoveIsLeft = False Then

For Dummy = 1 To LRLgMove

x = x + 1

If x > 7 Then x = 0

Xpos = Xpos + 1

Out &H378, (coil(x, Abs(mnuinvert.Checked), Abs(mnuLRswap.Checked), Abs(mnuSwapMotors.Checked)) + coil(Y, Abs(mnuinvert.Checked), Abs(mnuTFswap.Checked), Abs(Not (mnuSwapMotors.Checked))))

Delay = 0

Do While Delay < Mindelay.Value * 2

Delay = Delay + 1

Loop

Next Dummy

LastMoveIsRight = True

CurrFrame = 1

Else

For Dummy = 1 To LRSmMove

x = x + 1

If x > 7 Then x = 0

Xpos = Xpos + 1

Out &H378, (coil(x, Abs(mnuinvert.Checked), Abs(mnuLRswap.Checked), Abs(mnuSwapMotors.Checked)) + coil(Y, Abs(mnuinvert.Checked), Abs(mnuTFswap.Checked), Abs(Not (mnuSwapMotors.Checked))))

Delay = 0

Do While Delay < Mindelay.Value * 2

Delay = Delay + 1

Loop

Next Dummy

LastMoveIsRight = True

CurrFrame = 1

End If

End If 'this should resolve all Knife IN/Out motions, and if we moved knife, then currframe is reset to one, so lets goto init To/From checks.

End If

If CurrFrame > TFRoughFrames And FoundOblate = False Then

If LZV(CurrZone) - TFOblatedelta < RZV(CurrZone) Then

Text1.Text = "Zone " & CurrZone & " Is Closer to mirror, moving forward 0.01""" & " LZV =" & LZV(CurrZone) & " RZV = " & RZV(CurrZone) & vbCrLf & Text1.Text 'Move tester forward .01"

For Dummy = 1 To TFOblateMove

Y = Y + 1

If Y > 7 Then Y = 0

Ypos = Ypos - 1

Out &H378, (coil(x, Abs(mnuinvert.Checked), Abs(mnuLRswap.Checked), Abs(mnuSwapMotors.Checked)) + coil(Y, Abs(mnuinvert.Checked), Abs(mnuTFswap.Checked), Abs(Not (mnuSwapMotors.Checked))))

Delay = 0

Do While Delay < Mindelay.Value * 3

Delay = Delay + 1

Loop

Next Dummy

CurrFrame = 1

LastMoveIsRight = False

LastMoveIsLeft = False

Else

FoundOblate = True

End If

End If

If CurrFrame > TFRoughFrames And FoundOblate = True Then

If LZV(CurrZone) - TFRoughDelta > RZV(CurrZone) Then

For Dummy = 1 To TFRoughMove

Y = Y - 1

If Y < 0 Then Y = 7

Ypos = Ypos + 1

Out &H378, (coil(x, Abs(mnuinvert.Checked), Abs(mnuLRswap.Checked), Abs(mnuSwapMotors.Checked)) + coil(Y, Abs(mnuinvert.Checked), Abs(mnuTFswap.Checked), Abs(Not (mnuSwapMotors.Checked))))

Delay = 0

Do While Delay < Mindelay.Value * 3

Delay = Delay + 1

Loop

Next Dummy

CurrFrame = 1

LastMoveIsRight = False

LastMoveIsLeft = False

End If

End If

If CurrFrame = TFFinalFrames And FoundOblate = True Then

If LZV(CurrZone) > RZV(CurrZone) Then

For Dummy = 1 To TFFinalMoves

Y = Y - 1

If Y < 0 Then Y = 7

Ypos = Ypos + 1

Out &H378, (coil(x, Abs(mnuinvert.Checked), Abs(mnuLRswap.Checked), Abs(mnuSwapMotors.Checked)) + coil(Y, Abs(mnuinvert.Checked), Abs(mnuTFswap.Checked), Abs(Not (mnuSwapMotors.Checked))))

Delay = 0

Do While Delay < Mindelay.Value * 3

Delay = Delay + 1

Loop

Next Dummy

CurrFrame = 1

LastMoveIsRight = False

LastMoveIsLeft = False

Else

If CurrZone = 1 And series = 0 Then

Ypos = 0

Xpos = 0

TextBuff = Text2.Text

capGrabFrame lwndC

Call Plot_Zones

End If

If Metric = True Then

KEZ(CurrZone, series) = Round(Ypos * StepSize * 25.4, 4)

Else

KEZ(CurrZone, series) = Round(Ypos * StepSize, 4)

End If

Text2.Text = ""

For TxtZone = 1 To NZ

Text2.Text = Text2.Text & "Zone " & TxtZone

For TxtSeries = 0 To Reps

Text2.Text = Text2.Text & vbTab & KEZ(TxtZone, TxtSeries)

Next TxtSeries

Text2.Text = Text2.Text & vbCrLf

Next TxtZone

Text2.Text = Text2.Text & vbCrLf & TextBuff

If Save_picture.Value = 1 Then

On Error Resume Next

MkDir textpath

MkDir textpath & "\" & WorkFile

MkDir textpath & "\" & WorkFile & "\Images"

MkDir textpath & "\" & WorkFile & "\Images" & "\" & "Test" & WorkNum

SavePicture Picture1.Image, textpath & "\" & WorkFile & "\Images\" & "Test" & WorkNum & "\" & "Zone" & CurrZone & "rep" & series & ".bmp"

SavePicture Picture1.Picture, textpath & "\" & WorkFile & "\Images\" & "Test" & WorkNum & "\" & "Raw_Zone" & CurrZone & "rep" & series & ".bmp"

'MkDir textpath & "\Images"

'MkDir App.Path & "\Images\" & WorkFile

'MkDir App.Path & "\Images\" & WorkFile & "\" & "Test" & WorkNum

'SavePicture Picture1.Image, App.Path & "\Images\" & WorkFile & "\" & "Test" & WorkNum & "\" & "Zone" & CurrZone & "rep" & series & ".bmp"

'SavePicture Picture1.Picture, App.Path & "\Images\" & WorkFile & "\" & "Test" & WorkNum & "\" & "Raw_Zone" & CurrZone & "rep" & series & ".bmp"

End If

Text1.Text = "Found Zone " & CurrZone & " @ " & Time & " LZV =" & LZV(CurrZone) & " RZV = " & RZV(CurrZone) & vbCrLf & Text1.Text

End If

End If

If abort = 1 Or UloadMe = True Then GoTo skip

Next CurrFrame

LastMoveIsRight = False

LastMoveIsLeft = False

FoundOblate = False

Next CurrZone

'Rewind our Platform to Where ever it ended up back to 0

If Ypos > 1 Then WW = -1 Else WW = 1

Text1.Text = "Rewinding " & Ypos & " Steps to " & WW & "steps" & vbCrLf & Text1.Text

DoEvents

For Dummy = Ypos To -WW Step WW

Y = Y + -WW

If Y > 7 Then Y = 0

If Y < 0 Then Y = 7

Ypos = Ypos + WW

Out &H378, (coil(x, Abs(mnuinvert.Checked), Abs(mnuLRswap.Checked), Abs(mnuSwapMotors.Checked)) + coil(Y, Abs(mnuinvert.Checked), Abs(mnuTFswap.Checked), Abs(Not (mnuSwapMotors.Checked))))

Delay = 0

Do While Delay < Mindelay.Value * 2

Delay = Delay + 1

Loop

Next Dummy

If Xpos > 1 Then WW = -1 Else WW = 1

For Dummy = Xpos To -WW Step WW

x = x + WW

If x > 7 Then x = 0

If x < 0 Then x = 7

Xpos = Xpos + WW

Out &H378, (coil(x, Abs(mnuinvert.Checked), Abs(mnuLRswap.Checked), Abs(mnuSwapMotors.Checked)) + coil(Y, Abs(mnuinvert.Checked), Abs(mnuTFswap.Checked), Abs(Not (mnuSwapMotors.Checked))))

Delay = 0

Do While Delay < Mindelay.Value * 2

Delay = Delay + 1

Loop

Next Dummy

If DebugWin = False Then 'Save the debug text so as to save Memory!

Text1.Text = ""

Else

data = Text1.Text

file = App.Path + "\Debug.txt"

Filehandle% = FreeFile

Open file For Append As #Filehandle%

Print #Filehandle%, data

Close #Filehandle%

Text1.Text = ""

End If

Next series

skip:

If abort = 1 Then

WorkZone = "Test Aborted!"

Else

WorkZone = "Test Complete! :)"

End If

capGrabFrame lwndC

Call Plot_Zones

MirRad.Enabled = True

If mnuinvert.Checked = False Then

Out &H378, 0

Else

Out &H378, MyMod

End If

If UloadMe = True Then Unload Me

Manual = True

Command3.Enabled = True

Command1.Enabled = True

Command5.Enabled = True

Timer1.Enabled = True

QueryPerformanceCounter EndTime

QueryPerformanceCounter TimeStart

Text2.Text = "Elapsed Test Time, " & ((EndTime - StartTime) / TimeFreq) \ 60 & " Minutes " & ((EndTime - StartTime) / TimeFreq) Mod 60 & " seconds" & vbCrLf & Text2.Text

Text1.Text = "Elapsed Test Time, " & ((EndTime - StartTime) / TimeFreq) \ 60 & " Minutes " & ((EndTime - StartTime) / TimeFreq) Mod 60 & " seconds" & vbCrLf & Text1.Text & Text2.Text

Chk_trak.Value = 0

If abort = 0 Then Play App.Path & "\Sound\transpec.wav"

If abort = 0 And UloadMe = False And ZoneOn = True Then

If Reps > 0 Then

Dim StdDevTmp As Double

Dim a As Double

Dim b As Double

Dim tmp1 As Double

StdDev = 0

For a = 1 To NZ

StdDevTmp = 0

For b = 0 To Reps

StdDevTmp = StdDevTmp + KEZ(a, b)

Next b

tmp1 = StdDevTmp / (Reps + 1)

StdDevTmp = 0

For b = 0 To Reps

StdDevTmp = StdDevTmp + ((tmp1 - KEZ(a, b)) ^ 2)

Next b

StdDev = StdDev + ((StdDevTmp / (Reps)) ^ 0.5)

Next a

StdDev = StdDev / NZ

If Metric = False Then

Text1.Text = "Standard Deviation for this test was: " & Round(StdDev, 5) & units & vbCrLf & Text1.Text

Text2.Text = "Standard Deviation for this test was: " & Round(StdDev, 5) & units & vbCrLf & Text2.Text

StdDev = StdDev * 25.4

Else

Text1.Text = "Standard Deviation for this test was: " & Round(StdDev, 5) & units & vbCrLf & Text1.Text

Text2.Text = "Standard Deviation for this test was: " & Round(StdDev, 5) & units & vbCrLf & Text2.Text

End If

Else

StdDev = 0.01

End If

Call Command5_Click

End If

If DebugWin = False Then 'Save the debug text so as to save Memory!

Text1.Text = ""

Else

data = Text1.Text

file = App.Path + "\Debug.txt"

Filehandle% = FreeFile

Open file For Append As #Filehandle%

Print #Filehandle%, data

Close #Filehandle%

Text1.Text = ""

End If

QueryPerformanceCounter TimeEnd

Picture3.Visible = True

Form_Resize

End Sub