Filter effect on back focus distance

[andyboy1970]

I am setting up my filterwheel with Astronomik RGB, CLS-CCD & HA filters installed and was wondering how to factor these into the back focus calculation.

I will be using this setup with my Starlight Xpress mono camera on a Williams optics Megrez 72FD with FF ver1 and on a Skywatcher Newt. with 0.9x Coma corrector.

CC & FF require 55mm from camera.

Camera mount to chip = 17.5mm

Filterwheel thickness = 21.5mm

Therefore I need to use 16mm of spacers but what effect does the thickness of the filters have on this calculation?

[wxsatuser]

Add 1/3 the filter thickness.

[Davey-T]

There's a long confusing thread on this subject, when they say " add " a third of the filter I think they mean include it in your 16mm so actually subtract it and add 16mm minus the filter third.

Told you it was confusing 

Dave

[RayD]

Simply put if you need a 16 mm spacer without factoring filters, and you have Baader filters @2 mm (just an example as I don't know how thick the Astronomik ones are), then you will need a 16.66 (or 16.7) mm spacer.  As noted by @wxsatuser you need to add 1/3 of your filter thickness.

Personally I would go with your 16 mm and make the difference up with delrin rings as there is no guarantee that your sensor is spot on 17.5 mm.  There may be a bit of trial and error to get it spot on.

Edit:

Here is a simple diagram which I believe explains what I see as the theory of the filter changing the light path and pushing back the focal point, which is why you need to add a bit (this is not my image).

Edited October 26, 2017 by RayD

[andyboy1970]

Thanks guys, I will go with 16mm and see if it needs adjusting.

[ollypenrice]

2 hours ago, RayD said:

Simply put if you need a 16 mm spacer without factoring filters, and you have Baader filters @2 mm (just an example as I don't know how thick the Astronomik ones are), then you will need a 16.66 (or 16.7) mm spacer.  As noted by @wxsatuser you need to add 1/3 of your filter thickness.

Personally I would go with your 16 mm and make the difference up with delrin rings as there is no guarantee that your sensor is spot on 17.5 mm.  There may be a bit of trial and error to get it spot on.

Edit:

Here is a simple diagram which I believe explains what I see as the theory of the filter changing the light path and pushing back the focal point, which is why you need to add a bit (this is not my image).

I always struggle with this so I took Ray's diagram and added the hardware around it. The image below is how it came out. It shows why Davey T is right. The lengthening of the light path means you need a shorter spacer. Even as I write this it feels wrong, but see what you think of the diagram. For me, at least, it has brought a bit of clarity since, in the past, I though adding length to the light path meant adding it to the spacer. It seems not... but maybe I'm up a gum tree again!

Olly

Edited October 26, 2017 by ollypenrice

[Filroden]

4 minutes ago, ollypenrice said:

I always struggle with this so I took Ray's diagram and added the hardware around it. The image below is how it came out. It shows why Davey T is right. The lengthening of the light path means you need a shorter spacer. Even as I write this it feels wrong, but see what you think of the diagram. For me, at least, it has brought a bit of clarity since, in the past, I though adding length to the light path meant adding it to the spacer. Mais non!

Olly

It always felt counter-intuitive to me, but your diagram shows you need to add length to the spacer for the light path to hit the chip and be in focus. If you shorten the spacer, it would push the focus point even further behind the chip. And that is how it worked for me in practice. I need 66mm without a filter and with the Astrodon (3mm) filters I need 67mm of space.

[RayD]

Yes I think your diagram shows it as needing to lengthen Olly. That's what I've found too. 

If the focal point is pushed back the sensor needs to go back to meet it. 

[ollypenrice]

Nightmare! Yes, I now think you're right! I've always said 'add' myself but now I can't see where the 'subtract' argument comes from.

Olly

Edited October 26, 2017 by ollypenrice

[RayD]

5 minutes ago, ollypenrice said:

Nightmare! Yes, I now think you're right! I've always said 'add' myself but now I can't see where the 'subtract' argument comes from.

Olly

Me neither. I think it is confusion where people work out the space, then subtract the 1mm,  then need to add that to the spacer. 

I took ages going through this as I really struggled with spacing initially. 

[ollypenrice]

13 minutes ago, RayD said:

Me neither. I think it is confusion where people work out the space, then subtract the 1mm,  then need to add that to the spacer. 

I took ages going through this as I really struggled with spacing initially. 

But if we take as gospel the sentence 'angle of convergence remains the same' then we get this:

????

Olly

[RayD]

I looked at it in a fairly simple way. If I need to place my sensor at 55mm behind my FF then the light cone gets lengthened by 1mm I now need to place it at 56mm.

[andyboy1970]

Of course! Got it, Because the distance from the filter to the chip is constant therefore shortening the FF to filter distance gives a wider light cone and the final focal point at the Ccd is shortened!

Right i'm going for a lie-down....

[Filroden]

41 minutes ago, ollypenrice said:

But if we take as gospel the sentence 'angle of convergence remains the same' then we get this:

????

Olly

You’re trying to move the light to reach the chip by changing the spacing but the light path cannot change (not without adding lenses). You have to build more space to move the chip to where the light now reaches focus. 

[ollypenrice]

51 minutes ago, Filroden said:

You’re trying to move the light to reach the chip by changing the spacing but the light path cannot change (not without adding lenses). You have to build more space to move the chip to where the light now reaches focus. 

OK but at least I can now understand the existence of the 'subtract' argument which I think I first heard of from QSI and which was adopted by Optcorp when I asked them for the correct spacer for my TEC140 flattener to full frame chip. (This works perfectly, for what it's worth.)

So what's the error in my graph? In the second version the overlay at the bottom brings the focal point to that of the chip distance. The angle of convergence had been respected (because it was a cut and paste job from the original light path with filter.) On the other hand my overlay shows the filter intersecting the light path further up the light path (within the pink 'spacer' section) so that's not geometrically correct and could be the sorce of the error.

What bugs me in this whole question is that the 'subtract' camp are not idiots but I've known myself for long enough to know that I might well be an idiot!

Olly

Edited October 26, 2017 by ollypenrice

[Filroden]

12 minutes ago, ollypenrice said:

So what's the error in my graph? In the second version the overlay at the bottom brings the focal point to that of the chip distance. The angle of convergence had been respected (because it was a cut and paste job from the original light path with filter.) On the other hand my overlay shows the filter intersecting the light path further up the light path (within the pink 'spacer' section) so that's not geometrically correct and could be the sorce of the error.

By copying and pasting you've "chopped" off some of the light path inside the spacer. In reality, the only way to shorten the light path is to pass it through a lens. You cannot move the light path without adding a lens. So in the lower example, the addition of the filter does alter the light path by extending its point of focus. That point is now fixed in space in relation to the rear of the reducer. You could have placed the filter anywhere between the reducer and the chip and its effect would be the same (though the filter's diameter would obviously need to be bigger the further it was from the chip).

So with a fixed point of focus there are only two ways to bring the system to focus:

1) move the chip so it is now positioned at that point (by adding more space between it and the reducer)

2) move the point of focus to where the chip is by placing an additional lens in the system that would reduce the light path by an equal and opposite amount as the filter. 

Option 1 is much more prefereable as it does not introduce any further glass/air interfaces.

[Edit: I'm in Spain atm so can't draw what I mean but if you move your diagram upwards so it is inline with the original you will see your light path and the original light path no longer align. This is the missing piece of the light path that cannot be accounted for simply through a spacer.]

Edited October 26, 2017 by Filroden

[RayD]

I think some flatteners are more sensitive than others Olly.  With my Esprit and the SW FF the tiniest of changes makes quite a bit of difference, but with my Hotech one on the TS80 1mm either way really isn't that noticeable.

I'd say to understand it take it to extremes, and imagine a really thick filter which straightens the light by 30 mm.  Your CCD sensor would need to be the 55 mm needed for the FF plus the 30 mm which has been added by the 100 mm filter straightening it for 30 mm for it to achieve focus.  

If you take the light cone and slice it in the middle laterally, move the bottom down 30 mm and then draw straight lines between the cut ends (similar to the drawing I posted), you will see the cone angle doesn't change, it is simply lengthened by the glass of the filter, all else remaining the same, we now need to push the CCD sensor back the 30mm to meet the focal point, which we do with additional spacers.  If I now do nothing but swap the filter for a thinner one, I will need to remove spacers as the focal point will move closer to the FF as the straightened section of light is now shorter.

That's my understanding and it has definitely worked for me when setting up spacing.

[RayD]

I suspect the technical optical guys when taking about reducing are talking about something entirely different, but actually meaning the same thing, but for them they are using technical speak rather than us asking do I need longer or shorter spacers.

[steppenwolf]

I have considered this question on numerous occasions and I always come to the same conclusion - if you add a filter to the light path then for the focal reducer/field flattener to continue to work OPTICALLY as required, you need to ADD spacing between the focal reducer/field flattener and camera sensor to compensate.

[Filroden]

Found this in an earlier post on the same topic:

Quote

 

The following is taken from Astrodon's website where Don refers to backfocus and the use of a 3mm Astrodon filter:

"There may be some confusion as camera manufacturers measure backfocus from the focal plane of the CCD to the outer surface of the camera. When they account for the thickness of the filters, they SUBTRACT 1mm, which is correct as measured from the CCD. However, most people measure backfocus from the back of their scope or from a corrector, and then add/subtract spacers to arrive at the correct backfocus. In this case, as measured from the scope, the 1mm must be ADDED. A subtle point, but does get people in trouble from time to time."

Now on the QSI website they state that the backfocus of the QSI690 is 50.2mm as measured from the focal plane of the CCD to the outer surface of the camera. They then go on to say that when used with a 3mm filter you subtract 1mm from this backfocus distance i.e. 49.2mm.

So, this indicates that Astrodon and QSI agree, does it not?

 

So QSI are saying the effect of a filter means the distance inside the camera reduces by 1/3 the thickness of the filter. So to maintain the original spacing you have to add that space outside the camera. We would never really talk about the optical length of light inside the camera itself so we simplify this to just what we have control over, how much space we have between the reducer and front face of the camera. We add space to compensate for the effect of the filter.

I think it's sangria time...

[RayD]

1 minute ago, Filroden said:

Found this in an earlier post on the same topic:

So QSI are saying the effect of a filter means the distance inside the camera reduces by 1/3 the thickness of the filter. So to maintain the original spacing you have to add that space outside the camera. We would never really talk about the optical length of light inside the camera itself so we simplify this to just what we have control over, how much space we have between the reducer and front face of the camera. We add space to compensate for the effect of the filter.

I think it's sangria time...

That's the technical speak I was referring to 

[ollypenrice]

Doh. I'll just go back to my original interpretation which was the same as Steve's! However, Optcorp did subtract 1mm when calculating my adapter. I didn't argue because if it didn't work it would be their fault, but it did work (and it's quite sensitive for a full frame chip with the TEC, I gather.)

I'll let someone else draw the right set of graphics!! I'm on toothache antbiotics so no sangria for me... Well, maybe just one!

Olly

[steppenwolf]

1 minute ago, ollypenrice said:

I'm on toothache antbiotics so no sangria for me... Well, maybe just one!

I have found that Sangria, as made by Janie, cures most of life's maladies and when you wake up on the patio the next morning, there's no hangover either, amazing.

[wxsatuser]

1 hour ago, Filroden said:

Found this in an earlier post on the same topic:

So QSI are saying the effect of a filter means the distance inside the camera reduces by 1/3 the thickness of the filter. So to maintain the original spacing you have to add that space outside the camera. We would never really talk about the optical length of light inside the camera itself so we simplify this to just what we have control over, how much space we have between the reducer and front face of the camera. We add space to compensate for the effect of the filter.

I think it's sangria time...

Exactly, subtract refers to camera back focus and add refers to optical path.

 

[knobby]

Old thread dig up alert ! So if my filter is 6mm thick, the glass being 3 mm ... Would I add 1 mm to the target distance and include the 6 mm in the distance ?

So the total would be the same as adding a 5 mm spacer ?