Focusing screen calibration

I have a question for the mathematicians/theorists amongst you good gentlemen........

I have just changed to using my 1DMkIII as my macro body now that I'm happy with the 1DMkIV for larger wildlife species. Previously I've used 40/50D (occasionally 5DMkII) for macro and installed split prism focussing screens to them for accurate manual focus.

I have obtained several split prism screens (OEM versions) for the 1D/1Ds bodies and find that they are giving about 1mm back focus (clearly identifiable when comparing to Liveview and the final image). Canon supported the 40D/50D and 5DMkII with shims (washers in Canon parlance) for calibration but they don't have the same offerings for the 1Dsrs bodies.

Don't worry, we're there now.......
Is it possible to calculate the required shim (or shaving!) required based on the amount of back/front focus being experienced?

I know that I could find the result by simple (albeit fiddly) experimentation but I'd be much happier if I could achieve it by understanding the calculation.

Many thanks in advance folks.

Bob

Hi, Bob,

Bob Latham said:

I have a question for the mathematicians/theorists amongst you good gentlemen........

I have just changed to using my 1DMkIII as my macro body now that I'm happy with the 1DMkIV for larger wildlife species. Previously I've used 40/50D (occasionally 5DMkII) for macro and installed split prism focussing screens to them for accurate manual focus.

I have obtained several split prism screens (OEM versions) for the 1D/1Ds bodies and find that they are giving about 1mm back focus (clearly identifiable when comparing to Liveview and the final image). Canon supported the 40D/50D and 5DMkII with shims (washers in Canon parlance) for calibration but they don't have the same offerings for the 1Dsrs bodies.

Don't worry, we're there now.......
Is it possible to calculate the required shim (or shaving!) required based on the amount of back/front focus being experienced?

Click to expand...

First, let me be certain that you are speaking of a discrepancy between focus as indicated in the finder and the focus as experienced on the image or in Live View (that is, autofocus is not involved here).

Assuming that this is so (and assuming it is due to a discrepancy in the dimension of the focusing screen from its focal plane to the plane that locates it in the finder housing), then we can calculate the change in shim thickness needed to correct for that discrepancy.

The data that would be needed is the amount (and direction) that the plane of object focus for "perfect focus" on the screen differs from the plane of object focus for "perfect focus" on the image for some specific distance to the object plane (measured from some defined point, for example the camera focal plane).

It will take me a few minutes to reconstruct the calculation (the math is actually covered in one of my technical articles) but I'll do it as soon as I can (I just got up!).

An example of the data would be this:

For "perfect" focus on the focusing screen on a focus target nominally 10 m from the camera focal plane, the actual object plane of sharpest focus (on the image) is 3 mm closer to the camera than the target.

It is of course necessary to determine this in a scrupulous way. A common technique is with a slanted target. Unlike work with autofocus, it is not necessary to use an apparatus having a main target that is perpendicular to the line of sight.

Let me know if you feel you will be able to provide the needed input data, and in the meantime I will gather together the necessary math.

Best regards,

Doug

Hi, Bob,

Well, I've had breakfast, so here's the math.

f is the focal length of the lens used for the test.

S is the distance from the lens to the focus test target (ideally measured from the first principal point of the lens, but in practical cases that need not be precisely determined).

s is the distance from the plane of the focusing target to the object plane that will be precisely focused. s is positive if that plane is beyond the focus target plane (the situation we often describe as "back-focusing").

t is the needed change in the thickness of the focusing screen "shim" to correct the focusing error. t is positive if we need a greater shim thickness.

Then:

t =s*((f/S)^2)

So, for a numerical example: assume f= 100 mm, S=2000 mm (about 6.6 feet), and s = -5 mm (a "front-focusing" situation: the object plane of perfect focus is 5mm nearer to the camera than the focus target.)

Then t=-0.0125 mm; that is, we need to decrease the thickness of the focusing screen shim by 0.0125 mm.

The derivation of the basic relationship (it is the "longitudinal magnification" relationship) can be found here:

http://dougkerr.net/Pumpkin/index.htm#ShifOfFocus

(The trick of course is getting the signs right for this application!)

I hope this is helpful.

Best regards,

Doug

Doug,

Many thanks for ruminating this over your breakfast.....I'll confess that my early morning mental challenge is usually simply just trying to remember if I've already sugared my coffee or not.

I am indeed using manual focus and comparing the focus screen's opinion to the Liveview or captured image version....ie, the difference in distance between the registering surface of the screen and the sensor plane.....incidentally, using the MP-E65 at the higher magnification end of its range is where the error becomes obvious.

I'll take away your offering in the second post and see what I can decifer from it once I have figured out the focal length of my MP-E at a suitable magnification and then done a similar exercise with one or more of my other macro lenses (EF100L, EF180L, EF50 compact + Life size converter).....I suspect it'll be slightly longer timewise than it takes me to chew through two croissants though.

Again, many thanks.

Bob

Hi, Bob,

I note with interest that in the EOS 20D, the shims ("washers" indeed) are available in thickness increments of 0.05 mm. (I had been quite involved in the use of a third-party split image focusing screen on the 20D, and made the acquaintance of the "washers" during that adventure.)

By the way, I have assumed that their choice of the term "washer" is an extension of the provisions for adjusting the axial location and orientation of the sensor assembly, which of course does involve what we would recognize as "washers" of graded thickness (in the 20D they come in thickness increments of 0.02 mm).

Best regards,

Doug

Doug,

I have previously used Katz-Eyes offerings for the 30D, 40D and 50D but they don't have offerings for the 1D/1Ds series of bodies. I sold my 1.6x crop bodies before discovering the error in position of the Canon EC-B, EC-I and EC-L screens on the 1D3 and a quick check of one of the screens on the 1D4 showed a similar, but marginally different, offset. The 1Ds3 may be similarly out of register but I don't intending using it for such critically focussed subjects.
Canon sourced "shims" for the 5D have part numbers CB3-2850-000-003 to CB3-2850-000-030 rising in increments of 20 or 30um(?)...the last three digits being the thickness.

Why they don't have a similar series offered for the 1D series bodies is a mystery to me.

Bob

Hi, Bob,

Bob Latham said:

Why they don't have a similar series offered for the 1D series bodies is a mystery to me.

Click to expand...

I only have, in the 1-series, the parts catalog for the EOS-1D, but I note in it that the entire finder assembly is mounted atop the mirror box with two graded washers on one side ("finder focusing washer") and a graded two-hole rectangular shim on the other ("finder focusing shim").

The thicknesses for both run 0.03, 0.05, 0.10, 0.15, etc in mm. The corresponding part number suffixes are 003, 005, 010, 015, etc. (That is the same scheme used for the "washers" in the 20D.)

There are also rectangular shims used for lateral positioning of the finder assembly in each axis ("parallax adjusting shims"). They run in the same series.

This may differ in the "Marks" - I haven't taken the time to track down parts catalogs for them.

If you would like that document for the 1D, send me an e-mail and I'll e-mail it to you:

doug.kerr@att.net

Best regards,

Doug

Bob Latham said:

I am indeed using manual focus and comparing the focus screen's opinion to the Liveview or captured image version....ie, the difference in distance between the registering surface of the screen and the sensor plane.....incidentally, using the MP-E65 at the higher magnification end of its range is where the error becomes obvious.

Click to expand...

Hi Bob,

Just a gentle warning, the focal length in these macro lenses you are apparently using is not necessarily the one advertized, because they are internal "focusing" lenses with floating elements to also adjust magnification. Maybe it's better to use a regular lens at some distance, the focusing screen offset will be the same. Optical design becomes important for such determinations, and theorethical thin lens equations become less accurate at close focus distances. So use Doug's formula as a guidance, to figure out the order of magnitude, but you'll still have to do some incremental testing I'm afraid.

Cheers,
Bart

Hi, Bart,

Bart_van_der_Wolf said:

Hi Bob,

Just a gentle warning, the focal length in these macro lenses you are apparently using is not necessarily the one advertized, because they are internal "focusing" lenses with floating elements to also adjust magnification.

Click to expand...

I believe Bob said he planned to try to ascertain the actual focal length of the lens of interest at some focus setting.

Maybe it's better to use a regular lens at some distance, the focusing screen offset will be the same.

Click to expand...

It has seemed to me that almost all lenses of any interest today have some type of internal focusing (and thus the prospect of a change in focal length at different focus settings.

Are there lenses we are likely to encounter that do not (I assume that these would be those with "all group" focusing; that is, in effect, the entire lens moves back and forth - albeit inside a stationary external housing).

I know my ED 50 mm macro has all-group focusing!

Best regards,

Doug

Well, I see that the EF 50 mm f/1.8, 50mm f/1.4, 35 mm f/2, and 28mm f/2 are all-group focusing.

Best regards,

Doug

Doug Kerr said:

It has seemed to me that almost all lenses of any interest today have some type of internal focusing (and thus the prospect of a change in focal length at different focus settings.

Are there lenses we are likely to encounter that do not (I assume that these would be those with "all group" focusing; that is, in effect, the entire lens moves back and forth - albeit inside a stationary external housing).

Click to expand...

Hi Doug,

Of the lenses that Bob specifically mentioned, I only have the MP-E and the EF100L (new 100mm macro). The MP-E most likely(!) moves the whole group of elements/groups, but the 100mm one I'm not so sure of (I've seen some 'strange' differences in the exit pupil diameter at various magnification settings).

Cheers,
Bart

Hi, Bart,

Bart_van_der_Wolf said:

Of the lenses that Bob specifically mentioned, I only have the MP-E and the EF100L (new 100mm macro). The MP-E most likely(!) moves the whole group of elements/groups . . .

Click to expand...

"The optics uses a 3-group floating system, which moves three lens groups independently for focusing." (per Canon museum.)

. . .but the 100mm one I'm not so sure of.

Click to expand...

"Inner" focusing system (per Canon specs).

Best regards,

Doug

Doug Kerr said:

"The optics uses a 3-group floating system, which moves three lens groups independently for focusing." (per Canon museum.)

Click to expand...

Hi Doug,

Thanks for checking. An amazing piece of engineering this 'reversed' focusing lens is. Not only does it allow to increase the 'tube'length' by an amazing and stable amount, it also changes the lens group distances as it goes. And most importantly, the image quality for 1:1 to 5:1 magnification (no infinity focus possible, 1:1 is the smallest magnification factor) is very good (although it requires a lot of light and a good tripod (or a very steady hand and/or flash) for the extreme magnification settings). If only physics could be deceived into more Depth of Field (other than with focus stacking).

As for the EF100mm L:

"Inner" focusing system (per Canon specs).

Click to expand...

Doesn't surprise me a bit.

Cheers,
Bart

Bart_van_der_Wolf said:

Hi Bob,
Just a gentle warning, the focal length in these macro lenses you are apparently using is not necessarily the one advertized, because they are internal "focusing" lenses with floating elements

Click to expand...

Bob Latham said:

I'll take away your offering in the second post and see what I can decifer from it once I have figured out the focal length of my MP-E at a suitable magnification and then done a similar exercise with one or more of my other macro lenses (EF100L, EF180L, EF50 compact + Life size converterBob

Click to expand...

Bart,

As from my earlier post (quoted above) I realised that I would need to find out the focal length at a given magnification. I performed this exercise a while back with my older EF100/2.8 (non-IS L) version and found it to be 78mm at MFD.
The MP-E changing from 1x to 5x doesn't extend enough physically to be a simple internal linear extension so the moving rear lens group comes into play. My best guess at the moment would give it a focal length of 40mm at 5x (assuming it is a true 65mm at 1x)
The temperature over here has shot up to 33 degrees and I'm alternately steaming up the eyepiece whilst depositing sweat from my nose on the rear Liveview screen.....I'll get back to this tomorrow unless I can clear enough working space in the refridgerator.

Bob

Bob Latham said:

The temperature over here has shot up to 33 degrees and I'm alternately steaming up the eyepiece whilst depositing sweat from my nose on the rear Liveview screen.....I'll get back to this tomorrow unless I can clear enough working space in the refridgerator.

Click to expand...

Hi Bob,

I understand only too well. It's between 27 (at the seashore) and 35 degrees C (inland) here in the Netherlands (normal temp would be 22 C). It's 34+ C in my home office, and there is definitely not enough room in my fridge (I checked).

Take care,
Bart

Hi, Bob,

Bob Latham said:

As from my earlier post (quoted above) I realised that I would need to find out the focal length at a given magnification. I performed this exercise a while back with my older EF100/2.8 (non-IS L) version and found it to be 78mm at MFD.

Click to expand...

I'd be interested to know how you determined that.

Thanks.

Best regards,

Doug

Doug Kerr said:

Hi, Bob,

I'd be interested to know how you determined that.
Thanks.
Best regards,
Doug

Click to expand...

Maybe a method using flawed reasoning?

I set the lens to minimum focus and shot an accurate rule. I then added two EF25 extensions tubes (they're actually 27mm of extension each) and reshot the rule at minimum focus. I then calculated what the focal length would need to be to give the magnification increase provided by the 54mm's of additional extension.
Go on then Doug, shoot me down....I've been married for 29 years so I'm very well used to being told I'm wrong.

Bob

Hi, Bob,

Bob Latham said:

I set the lens to minimum focus and shot an accurate rule. I then added two EF25 extensions tubes (they're actually 27mm of extension each) and reshot the rule at minimum focus. I then calculated what the focal length would need to be to give the magnification increase provided by the 54mm's of additional extension.
Go on then Doug, shoot me down....I've been married for 29 years so I'm very well used to being told I'm wrong.

Click to expand...

No, that's a very practical method.

I believe the formula for the focal length, f, is:

f=L/(m'-m)

where L is the length of the extension tube, m is the magnification without the extension tube, and m' is the magnification with the extension tube (the lens proper being "set" the same in each case, the focal length being for that "setting").

Best regards,

Doug