A - the
optical register: what does it means ?
possibility of using a lens (with say, register X) on a DSLR (with
register Y) is essentially linked to the concept of optical register.
optical register is the distance (in millimetres) between the
focal plane (CMOS sensor plane) and the flange's top plane (see
the figure A). If the lens register is greater than the DSLR register
(i.e. X>Y), the adaptation is easy to make and we can simply
use an adapter ring on the DSLR's original flange .
If the lens register is shorter than the DSLR register, which
is the case with Minolta MC/MD lens and the Canon 5D, it is necessary
to modify the flange  (see
the figure B).
effect of reducing the register from 44mm to 43.5mm makes it possible
to achieve a "real infinity" focus with a Minolta MC/MD
lens at a fully open aperture, e.g. f1.4, without the need for
any internal lens. Therefore the Minolta MC/MD Rokkor can be used
without compromising either the optical design or the original
optical quality in any way. Note that "real infinity"
is quite different from the "virtual infinity" obtained,
for example, by stopping down a 24mm lens at f16 and focussing
on a subject 10 metres away.
a subject is only in focus on one plane, i.e. the main plane of
focus, at a fixed distance from the object focal plane (the film
or sensor) in accordance with the distance setting on the focus
ring. The depth of field (and the circle of confusion) gives us
the impression that the image is in focus a bit less than and
a bit further than the distance of the main plane of focus from
the object focal plane (the sensor). So, "real infinity"
means that the infinity focus is not only achieved by stopping
down the lens but is also possible with the aperture fully open
when lens is set at infinity or just before infinity distances.
from the point of view of optical performance, we can say that
when mounted on the EOS-MD flange type 3, a Minolta MC/MD lens's
behaviour is the same as it would be on a Minolta SLR film body
such as the X-700.
are the main aspect about
type 3 flange ?
the design phase of the EOS-MD type 3 flange some significant
points were considered:
The MC/MD Rokkor lenses must not suffer any alteration to the
mounting bayonet. This choice was made because Rokkor lenses lose
their monetary value when changes are made to the bayonet and
at the same time they can increase their value over time, especially
for rare and higher quality versions. By contrast DSLRs, both
APS-C and Full Frame, lose their value over time so ultimately
it is better to make changes to the camera whilst preserving the
MC/MD lenses in their original state.
The internal geometry of a Canon DSLR 
is the only one compatible with the shape of the Minolta MC/MD
It must not behave like conventional adapter rings which, when
mounted on the EOS body, restrict use to only short distances
(macrophotography or just a few metres) or indeed the type of
adapter ring with internal optical elements that allow you to
reach "real" infinity focus but at the cost of poor
optical quality of the final digital images. Also, they must avoid
causing any loss of light due to the thickness of the adapter
ring or add any multiplier factor to the focal length. On the
contrary, performance must be the same as when using the MC/MD
Rokkor lens on a Minolta SLR body in terms of optical behavior
and infinity focussing.
It must not involve any modification to the electronic circuits
inside the CANON DSLR and the installation steps must be fully
or partially reversible (see below for details).
It permits the following lenses only:
Minolta Manual Focus lenses (SR mount): Auto Rokkor, MC/MD
Rokkor lenses including the MD type.
universal lenses with Minolta MD mount (Vivitar, Tokina, Tamron
Adaptall 2, ect)
M42 (42x1) lenses (Zeiss Jena, Pentax Takumar, Russian lenses,
Pentacon, etc.) using the M42-Minolta MD adapter "on" the
Canon 5D (MK I and MK II) uses a larger mirror than the APS-C
DSLR. It comprises two different parts: a plastic support which
holds the mirror, (in dark gray in fig. C) and a thin glass "true"
mirror, (in light gray in fig. C). In order to use Minolta MC/MD
lenses on the flange type 3, both parts need to be trimmed by
about 2.5mm (see fig. D for details) otherwise the mirror would
collide with the base of the Minolta MC/MD lens as it flips up
during exposure (see fig. D).
we should consider: is this mirror trimming a reversible process?
It is possible to give three different answers to this question.
It is totally reversible if we replace both parts: the plastic
support and the thin glass. Obviously this is the most expensive
choice. The mirror will be as good as before the change.
6.2) It is partially reversible, by replacing only the thin
glass and retaining the trimmed plastic holder. This option
is less expensive and again, the mirror would still be as
good as before even if the underlying plastic support remains
It is not reversible if we leave both parts altered. This
option has no further costs but may be noticeable when we
replace the original Canon flange and lens (see fig. F).
that the trimming process does not affect the final image or the
image that we would see through the viewfinder. In fact, the missing
part of the mirror (about 2.5mm along the front side, see fig.
E, F) would still be masked by the thin metal frame of the focus
When using the EOS-MD type 3 flange, focussing the Minolta MC/MD
is MANUAL ONLY and aperture values can only be set by stopping
The EOS-MD type 3 flange has a mechanical lock that sets the
stopping point for the Minolta MC/MD lens but it does not have
the traditional "lock pin". The Minolta MC/MD lens is
held firmly (to prevent movement during focussing and stopping
down) to the flange by a combination of friction and pressure
effected by a sturdy steel ring. Is there a risk of the lens dropping
out of the camera during focussing operations? No, because the
lens is very tight on the flange when it is already near the point
of stop and the pressure is at maximum when the lens reaches the
point of stop.
5D operates correctly only in the Manual mode (M) or Aperture
Priority (Av) and light metering can be set to different modes
(matrix, center, semi-spot, spot). For example: setting the 5D
to Av mode and using a Rokkor wide-angle lens at hyperfocal distance
can be useful for a dynamic subject that moves within a certain
distance range. This is an example of true automatic mode operation
using a manual focus lens !
what about focus operations ?
it possible to achieve precise focussing using this flange
with a manual focus lens like the MC/MD Rokkor? The answer
the original 5D focus screens are not accurate enough to use
with a manual focus lens it is advisable to replace the original
focus screen (i.e. the Ee-A) with a split screen (easily available
new on www.ebay.com by searching with "eos split screen")
or with the original Canon Ee-s. The split screens are, in
general, accurate and precise but sometimes I have found that
they can introduce a slight "backfocus". Obviously this backfocus
(and also frontfocus) is an issue not related to the optical
register modification introduced by the EOS-MD flange but
to the thickness of the split screen which may be different
from the original one. Further, the split screen on the 5D
(and also on other DSLRs) introduces some exposure error in
low light of about 2 f-stops, especially using the "spot"
mode, because the spot area used by the DSLR's lightmeter
and the split/microprism area can overlap and give a false
reading. So, what is the best lightmeter mode? I have found
that the "matrix" setting on the 5D (and also on other DSLRs)
lightmeter gives the best results overall.
can understand from the fig. G what happens inside a DSLR.
G - inside a DSLR
order to get perfect focus the path P1 must be equal to path
P2, so P1 = P2.
P1 > P2 then we get backfocus. This is a situation that
can occur using a non-original focus screen (i.e. a split
P1 < P2 we have frontfocus.
it is possible to correct the backfocus by making an appropriate
adjustment to P1 by adding a shim (see fig. H) between the
split screen and the pentaprism (in order to obtain P1 = P2).
Did you know that this is the procedure used by Canon service
themselves by using a proprietary shim (see the Canon
5D service manual on page 11, CB3-2850-000)? So, we can
make our own shim by taping the outer perimeter (only !) of
the split screen, on the side facing the pentaprism itself,
with a thin strip 1 mm (max 1.5 mm) wide of adhesive tape
(my advice is to use the Scotch Magic 3 with thickness = 0.05mm).
Calibration consists of finding the correct thickness (usually
just 1 or 2 layers of tape would be necessary, with thickness
equal to about 0.1mm) to get the focus correct by using the
scale at 45 ° in the following
article (see page 18). Note that this thin
around the split screen will not be visible in the viewfinder
because it is covered by the thin metal frame that holds the
split screen below the pentaprism.
you are thinking: can I do all the above changes
to my Canon 5D by myself? How difficult are
they to do? Again, there are two responses.
they are quite easy to perform if you
have good skills in dismounting cameras
and have good tools for trimming the
NO, there is some risk of damaging the
camera (the mirror in particular) if
you do not have those skills and tools.
So what do you do? It would be much
better to take your own (expensive!)
Canon 5D to a camera repair laboratory.
we have shown that a CANON 5D camera can be
customized in order to mount all MC/MD Rokkor
lenses by using a new EOS-MD flange, the type
3, allowing us to enjoy "all" the
optical features of an original Rokkor lens
on a digital sensor with 24x36mm format.
you think it is a dream? Are there still doubts
about that ?