Do filters add to back focus?

[strutsinaction]

Agreed, Andrew. You subtract only when you need to maintain a fixed distance between a reducer/corrector and CCD.

For example, consider my case - Tak x0.73 + QSI690.

When using this reducer, the back focus distance from the reducer to the CCD imaging plane should be 72.2mm. The CA-35 (SKY90) adapter is 16.0mm and the QSI-690 back focus distance is 50.2mm when fitted with the standard T mount adapter plate. Therefore, the customer Wide T Mount (supplied by IKI) needs to add 6mm to the back focus.

However, when fitted with filters, you need to subtract a third of the thickness of the filter from the backfocus distance of the QSI-690. The Baader LRGB 1.25" filters have a width of 2.1mm which removes 0.7mm from the backfocus of the QSI-690 i.e. 71.5mm. The Astrodon NB 1.25" filters have a width of 3.0mm which removes 1mm from the backfocus of the QSI-690 i.e. 71.2mm. Therefore, the Wide T Mount adapter needs to add 5.3mm (72.2-16.0-50.2-0.7) when using a Badder filter and 5.0mm (72.2-16.0-50.2-1.0) when using an Astrodon filter.

I think 

Regards

John

[nightster]

I spent some time working with Vincent at Highpoint Scientific (my favorite stateside astro dealer.)  He helped me with the backfocus calculation for my scope and the new CCD today.  We found the error in the Astronomik thickness, so I was correct on that line.  He also contacted the maker of my OTA, Flattener and comfirmed the setback distance and the allowable margin of error.  The dealer stressed the importance of a tight tolerance to the prescribed 55mm distance.  I thought I'd share the worksheet he sent me as I thought this a great way both understand and calculate this space for any novice such as myself.    

Item Part # Backfocus remaining   Stellarvue flattener SFF6-25 55 55 Male T-threads (M48 spacers can be used) Stellarvue Spacers M48 Spacers 10 45 10mm thick M48 to M48 spacer Stellarvue Spacers M48 Spacers 3 42 3mm thick M48 to M48 spacer     0 42   Starlight Xpress SXUFW-1T 29 13 Telescope side (female T-threads), Camera side (male T-threads) Astronomik NB+LRGB -0.33 13.33 Assuming 1mm thick filters (1/3rd filter thickness) Atik 460EX 13.5 -0.17 Ends in female T-threads

[nightster]

Well that didn't work, and since I cannot edit my posts yet that garble will have to stay, not quite sure how to post a table to the thread.

[ollypenrice]

Merlin66's argument is compelling (see what I mean about great arguments out there!!). This would seem to confirm the QSI and Optcorp stance but if the light path has now been modified, how relevant is the original 50mm PHYSICAL spacing in the example above?

Fascinating stuff.......

Sent from my iPhone from somewhere dark .....

In reality there will usually be no physical increase in the distance brought about by inserting a filter because the 'increase' is contained with the thickness allowed for the filterwheel. Yes? No?? Arrrgh.

Olly

[steppenwolf]

In reality there will usually be no physical increase in the distance brought about by inserting a filter because the 'increase' is contained with the thickness allowed for the filterwheel. Yes? No?? Arrrgh.

I agree, Olly, the difference is well within the acceptable tolerance but, one them is right and one of them is wrong and the difference between the two is TWICE the change in light path length!

I see both arguments which is why I find this such an intriguing discussion!

[ollypenrice]

I agree, Olly, the difference is well within the acceptable tolerance but, one them is right and one of them is wrong and the difference between the two is TWICE the change in light path length!

I see both arguments which is why I find this such an intriguing discussion!

I was meaning no purely physical increase but I have now returned to my original belief that there must be an optical increase.

Indeed, most intriguing. I'm glad I don't need an LP filter as well!

Olly

[ollypenrice]

Although I don't use his filters, here's a quote from Don Goldman (Astrodon) who I believe knows a thing or two about filters but maybe hasn't got the price right yet! :-

You are quite mistaken Mr Wolf. My husband has the prices absolutely right.

Mrs Goldman.

[steppenwolf]

I was meaning no purely physical increase but I have now returned to my original belief that there must be an optical increase.

Aha, I misread but either way this is a very interesting academic argument and I would love to have a definitive answer but with big guns like Don V QSI not agreeing, that answer may not be forthcoming any time soon!

[strutsinaction]

I've changed my mind again and am now in the 'add' rather than 'subtract' camp   Andrew's ray diagram is too compelling, So, my Tak x0.73 reducer + QSI690 calculation is a follows:

When using the reducer, the back focus distance from the reducer to the CCD imaging plane should be 72.2mm. The CA-35 (SKY90) is 16.0mm and the QSI-690 back focus distance is 50.2mm when fitted with the standard T mount adapter plate. Therefore, a custom Wide T Mount needs to add 6mm to the back focus. When fitted with filters, the back focus distance increases by a third of the thickness of the filter. The Baader LRGB 1.25" filters have a width of 2.1mm which adds 0.7mm to the back focus i.e. 72.9mm. The Astrodon NB 1.25" filters have a width of 3.0mm which adds 1mm to the back focus i.e. 73.2mm. Therefore, assuming the custom Wide T Mount adapter contributes 6mm to the back focus an additional 0.7mm needs to be added using a spacer when using a Badder filter and 1.0mm when using an Astrodon filter.

I'm not sure the 0.7mm or 1mm would make any difference anyway but I'd still like to know if my calculation is correct nonetheless!

Regards

John

[andrew s]

Hi John - I think the confusion here is because there are two different case and different definitions of back focus.

Case one what does adding a filter do to the focal position if you add a filter without changing the geometric distances- in this case as my diagram show it moves the focal point further away from the objective i.e you add ~ 1/3 thickness to get the focal point.

Case two what to do if you place a filter between a corrector and ccd. In this case you what to maintain the optical depth between corrector and ccd and you need to subtract ~ 1/3 thickness (actually t(n-1) where t is the thickness and n is the refractive index of the filter) from the spacing between corrector and ccd. You will also need to refocus.

Regards Andrew 

[Merlin66]

Guys,

Stop.

Think about it slowly and clearly....

The two "issues" are - mechanical "fixed" distance and "optical" path distance.

The difference is +/- the filter "correction" depending on your frame of reference.

I'm thinking with a corrector ( as distinct from a reducer) - the optics of the corrector are designed such that when placed in the optical path (somewhere) that the design distance from the corrector to the focus, and hence the CCD is a fixed distance.

If we then place a filter element (say 3mm) between the corrector (say requiring 55mm spacing) and the CCD; looking at the optical path the design distance for the corrector remains unchanged but the "optical" path increases by 1mm. To me this says that with a filter the 55mm spacing would actually be 56mm - the corrector still functioning at it's 55mm plus the, added 1mm.

If we maintained the "physical" spacing of 55mm, the corrector would have to be placed (55-1=54mm) from the CCD, which is not it's design position.

[andrew s]

Guys,

Stop.

Think about it slowly and clearly....

The two "issues" are - mechanical "fixed" distance and "optical" path distance.

The difference is +/- the filter "correction" depending on your frame of reference.

I'm thinking with a corrector ( as distinct from a reducer) - the optics of the corrector are designed such that when placed in the optical path (somewhere) that the design distance from the corrector to the focus, and hence the CCD is a fixed distance.

If we then place a filter element (say 3mm) between the corrector (say requiring 55mm spacing) and the CCD; looking at the optical path the design distance for the corrector remains unchanged but the "optical" path increases by 1mm. To me this says that with a filter the 55mm spacing would actually be 56mm - the corrector still functioning at it's 55mm plus the, added 1mm.

If we maintained the "physical" spacing of 55mm, the corrector would have to be placed (55-1=54mm) from the CCD, which is not it's design position.

Ken - I am not sure I fully understand your post but I would say that the design distance is most probably the optical depth (geometric distance * refractive index) so if you place a 3 mm filter in the optical path it adds to the optical depth so you need to reduce the geometric distance to compensate. I agree with you earlier post on this.

The two issues I discussed above are not mechanical v optical but about two different questions you can ask both of which can have mechanical v optical distance issues.

Regards Andrew

[Merlin66]

Andrew,

The reason I didn't use the "optical" path for the corrector was that the corrector "doesn't" see the filter, and, as far as it's concerned it's still functioning as if it were at a 55mm spacing. The filter just takes the focus from the corrector (at 55mm) and moves it by 1mm to reach the CCD.

Think through the set-up if the filter were 18mm thick!

The 6mm change in the focal position would not change the design of the corrector.....it would be (55+6=61mm) from the CCD.

Yes No??

[andrew s]

Andrew,

The reason I didn't use the "optical" path for the corrector was that the corrector "doesn't" see the filter, and, as far as it's concerned it's still functioning as if it were at a 55mm spacing. The filter just takes the focus from the corrector (at 55mm) and moves it by 1mm to reach the CCD.

Think through the set-up if the filter were 18mm thick!

The 6mm change in the focal position would not change the design of the corrector.....it would be (55+6=61mm) from the CCD.

Yes No??

I think I understand you now but don't agree with your conclusion. A corrector works by adding  a specific optical path difference into the light path to bring the light rays to, say, a flat focal plane at a specific optical depth (distance) from the corrector. As air has a refractive index of ~1 the optical and geometric distances will be the same e.g. your 55mm. If you place another optical element after the corrector e.g. our 3mm filter and as its refractive index will be ~1.4 you no longer have the matching of optical and geometric distance so to get back to the correct optical distance of 55mm from the corrector to ccd you need the geometric distance to be ~ 54mm ( 51 mm of air and 3mm glass as 51*1 + 3*1.4 ~ 55).

This effect was use in the past to design very fast F1 Schmidt cameras which were solid glass with a cutout for the film!

Regards Andrew

[Merlin66]

Andrew,

You are 100% correct!

I apologise for muddying the water...

I'll take my pills now and lie down.....

[ollypenrice]

OK, it's the class Dunce again! Why wouldn't Andrew's ray diagram apply to a reducer? I don't understand why a reducer produces a beam different from the one produced by the objective.

Olly

[Merlin66]

Probably does.....

Any element corrector/ reducer etc. has a design spacing to the CCD chip to give optimum performance.

When the filter is introduced between the element and the CCD then the change to the optical length  needs to be considered.

Andrew's answer above gives the solution, taking into account the effect of the filter.

[strutsinaction]

I'm with Ollie on this.

I can understand that there's an optimum design spacing between the reducer and CCD chip. I would have thought that this was simply the optical distance to the new focal plane and assumes that the optical train is in air with a refractive index of 1 and no filters.

If you then add a filter between the reducer and CCD then the optical path increases so that the new focal plane is further away than the original focal plane and so the optimum design spacing is now increased. If you keep the optical path the same then surely the light rays from the filter will no longer converge at the CCD?

I'm sure a light ray diagram would clear this up!

Regards

John

[andrew s]

OK, it's the class Dunce again! Why wouldn't Andrew's ray diagram apply to a reducer? I don't understand why a reducer produces a beam different from the one produced by the objective.

Olly

As Merlin 66 says it does. With the case of an just an objective you don't have any option but to refocus to account for the focal shift due to the filter. With a reducer/corrector you have two degrees of freedom. You can adjust the reducer/corrector distance and refocus. This allows you to bring the reducer/corrector to ccd distance back to the design distance and then refocus to account for the focal shift.

In the case of the objective only it also cause a shift from the design optical distance and that is why a filter in a converging beam introduces aberrations.

As light relief you might like to ponder the following. 101 optics says that for a simple converging lens rays parallel to the optic axis (and close enough to it) go through the focal point. Now why is this so? Well one way to explain it is that all the rays have the same optical path length. The ray through the center of the lens goes through the thickest part of the lens but takes the shortest distance through the air along the optic axis while a ray through the edge of the lens has the least glass to go through but the longest route through the air from the edge of the lens to the focal point. In fact you find t*n + r*1 = constant ( where t is the thickness of glass gone through with n it's refractive index and r the air distance with refractive index~1). If you think about it it also explains why a simple lens has a curved focal surface!

This is an example of the Principle of least Action applied to optics, It led Feynman to his formulation of quantum mechanics QED the subject of another post!

Regards Andrew

PS I am sure you will be all pleased when I get some clear skies and so post here less! Got any free slots at your place Olly?

PPS John posted while I was typing but I think I covered the point. I am now worried my diagram will cause you all to do the wrong thing. 

Edited February 15, 2015 by andrew s

[steppenwolf]

I tried to discuss this subject with Jane last night - we decided that we are 'incompatible' when it comes to topics like this     

[strutsinaction]

It's all gone quiet on this thread, so let me open it up again having done some more research 

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?

In my case, to maintain the 72.2mm optical backfocus when using the x0.73 reducer with a 3mm Astrodon filter, I'd have to add a 1mm spacer to my custom T-Mount.

Regards

John

[NickMorris]

Hi John,

I think you are right in that Astrodon and QSI are saying the same thing, just worded differently.....  This means that behind the CA-35 Tak adaptor, it is expecting 55mm to CCD plane therefore the spacer needs to be 55-(50.2-1.0) = 5.8mm ?  Do you concur?

[andrew s]

I hesitate to post again on this but... ...I agree Astrodon and QSI are in agreement. What they are saying is consistent with the diagram I posted. That is adding a filter pushes the focus out by ~ 1/3 thckness i.e. 1mm for a 3mm filter. This means that camera plus filter consumes 1mm less of geometric backfocus than the camera alone. 

What to do if you have a corrector/reducer depends on if you want to maintain the geometric distance or the optical distance. 

If you want to maintain the geometric distance add 1mm as John proposes if you want to maintain the optical distance subtract 1mm and refocus. 

Given that CCD will have a window and cover plate then the optical distance is probably ill defined anyway!

I now retire from this thread.

Regards Andrew

[ollypenrice]

I hesitate to post again on this but... ...I agree Astrodon and QSI are in agreement. What they are saying is consistent with the diagram I posted. That is adding a filter pushes the focus out by ~ 1/3 thckness i.e. 1mm for a 3mm filter. This means that camera plus filter consumes 1mm less of geometric backfocus than the camera alone. 

What to do if you have a corrector/reducer depends on if you want to maintain the geometric distance or the optical distance. 

If you want to maintain the geometric distance add 1mm as John proposes if you want to maintain the optical distance subtract 1mm and refocus. 

Given that CCD will have a window and cover plate then the optical distance is probably ill defined anyway!

I now retire from this thread.

Regards Andrew

Retire at will, Andrew, as is your perfect right, but please don't do so before accepting my heartfelt thanks for the most edifying of posts! I must admit to functioning at about 15% of an already limited mental capacity following an accident which leaves me unable to sit in comfort for more than a few minutes, so my mind is not really on the task. I'll go through all this in a month or so when I'm back to what passes for normal!

Olly

[strutsinaction]

Hi John,

I think you are right in that Astrodon and QSI are saying the same thing, just worded differently.....  This means that behind the CA-35 Tak adaptor, it is expecting 55mm to CCD plane therefore the spacer needs to be 55-(50.2-1.0) = 5.8mm ?  Do you concur?

Hi Nick

The optimum back focus for the Tak x0.73 reducer is quoted as 72.2mm. The CA-35 (SKY90) adapter is 16mm and the back focus of the QSI690 is 50.2mm. So, without adding a filter into the mix the custom T-mount adapter needs to contribute 72.2-16-50.2=6mm. I purchased this customer T-mount adapter from Ian King specifically for this case.

If I'm using a 3mm then I have 2 options:

- If I want to maintain the geometric back focus distance I'd need to add a 1mm spacer. This I can do. I also think I wouldn't need to refocus.

- If I want to maintain the optical back focus distance then I think I'm stuck because I can't remove 1mm from the custom T-mount adapter (unless I buy a lathe ). I guess the T-mount should have been manufactured 1mm shorter so it could be used with and without 3mm filter (for the latter I would simply need to adda a 1mm spacer). With this option I'd also have to refocus.

For me, this is all hypothetical anyway, and don't think the 1mm would make much difference. Who knows. As Andrew correctly points out, there are other optical elements that would also need to be accounted for too, like the CCD window and coverplate, which Astrodon also mentions on his website.

I would also like to thank Andrew for his valuable insight. I'm not sure what his background is, but it sounds like he knows what he's talking about. I'm still learning 

Regards

John