Tri-Elmars and Bayonet Flanges

[marknorton]

Advertisement (gone after registration)

Regular readers will know that the frame lines in the viewfinder are selected by different length bayonet flanges, three of them.

 

The existing Tri-Elmar has a clever mechanism so that as you change the focal length from 28 to 35 to 50mm, the effective flange length is changed and the correct frame is displayed.

 

The new Tri-Elmar takes us into a different world because none of the viewfinder frames available match the focal lengths selected. True, a Leica 21mm prime will already select the 28/90 frame set, but there is no such definition for the 16/18mm focal lengths.

 

Any thoughts as to whether changing the focal length on the new lens will change the frames selected? Does it matter? Do we care since an aux finder is always required?

[ho_co]

Gee--I hadn't thought of that!

 

The lens is relatively inexpensive (I mean compared to a house, for example), and the elegance of the earlier Tri-Elmar's linkage isn't needed, as you said--not to mention the increased need of upkeep that linkage could require.

 

In addition, keying different finder frames could be taken to imply a future model for super-wide lenses only.

 

So my guess is: Just use a single 21mm flange (but with its own zebra stripes of course?).

 

And I don't know whether they ever solved the question of how many angels could dance on the head of a pin; we might want to concern ourselves with that.

 

--HC

[marknorton]

We've always assumed - to be confirmed - that when the existing Tri-Elmar is mounted, the lens code will say "Tri-Elmar" but that the electronics will sense the position of the flange lever to determine the actual focal length in use. Only by doing that can the actual focal length be put into the EXIF file header and focal-length related processing applied.

 

With the new Tri-Elmar, there will certainly be its own unique code but without the flange mechanism, the camera will not know the selected focal length and will only be able to apply correction which is lens related, not focal length related.

 

OTOH, the new Tri-Elmar is a brand new design, so they may have been able to optimise the design with the M8 in mind.

 

The new Tri-Elmar has a modest "zoom" ratio - just 1.31 - compared to the existing one of 1.79 and that probably accounts for the lens being very expensive instead of outrageously so.

[wattsy]

We've always assumed - to be confirmed - that when the existing Tri-Elmar is mounted, the lens code will say "Tri-Elmar" but that the electronics will sense the position of the flange lever to determine the actual focal length in use. Only by doing that can the actual focal length be put into the EXIF file header and focal-length related processing applied.

 

That' s interesting - I hadn't thought about that problem. As you say, presumably Leica have had to introduce a connection somewhere between the camera electronics and the flange lever in the lens mount. Personally, I'd have preferred it if they had simply left the Tri-Elmar out in the cold when it comes to the lens recognition part of the M8 than build in an additional mechanical/electronical linkage that is only relevant to one lens. (I guess my experiences with the 'simple' DX coding mechanism adopted by Leica in the M7 makes me wary of Leica and 'Heath Robinson' solutions.)

 

The new Tri-Elmar has a modest "zoom" ratio - just 1.31 - compared to the existing one of 1.79 and that probably accounts for the lens being very expensive instead of outrageously so.

 

I guess compared to the German made Zeiss 15mm/F2.8, the Leica lens is reasonably competitive - after all, the latter provides three wide or ultra-wide focal lengths for the price (albeit at F4). Having said this, given the choice I'd imagine many buyers would prefer to pay half the amount and get a Leica single focal length ultra-wide - perhaps a 16mm/F4 or an 18mm/F4.

[Goldie]

If there are significant variations in the systemic vignetting between the three focal length settings of the new Tri-Elmar, Leica might well have put 3 codings in the flange which move into position as required. Its an idea that could either be dead simple to accomplish or fiendishly difficult to do.

[marknorton]

I already have my (current) Tri-Elmar coded and there's a single code, irrespective of the selected focal length. The only way the camera can figure out which focal length I have selected is to look at the setting of the frame lever. It moves, maybe 3 - 5 mm according to what lens is mounted so that is easy to sense electrically or optically.

[Goldie]

Advertisement (gone after registration)

Mark you have missed my point.

I am suggesting that if it is needed optically, then it is feasible to change the 6-bit code as you change the focal length. This could be accomplished by having a sliding plate under the bayonet with 3 separate codes which move into view/position as you twist the focal length ring. It would be a very simple mechanism; space would be the only issue.

However given the small focal length changes I would be surprised if there are any significant changes in optical characteristics. Indeed with the original Tri-Elmar I seem to recall Erwin Puts describing very similiar performance at all three settings, and that with much larger movements in focal length.

[marknorton]

I agree that it is feasible to change the lens code as you change the focal length. I'm not sure it's likely because they would then use up 3 of the available 63 codes with just a single lens. As I say, the existing Tri-Elmar has a single code so either they make do with that single code or else sense the position of the frame lever to identify the focal length.

 

I agree with your point about the lens performance being similar across a narrow range of focal lengths but on the other hand, it's in the focal length range where the correction is most required and which is the reason for the coding in the first place.

[Guest hammertone]

Having said this, given the choice I'd imagine many buyers would prefer to pay half the amount and get a Leica single focal length ultra-wide - perhaps a 16mm/F4 or an 18mm/F4.

 

Seems Zeiss has heard your wishes. They are introducing a ZM 18mm f4 (and a 21mm f4.5) at Photokina.

[Goldie]

On further reflection another way of resolving this issue has occured to me. The frame selector dial on the auxillary viewfinder could be electronically connected to the camera via the hotshoe and thus transmit focal length data that way. After all flash can't be used when the viewfinder is in place, so a contact could be re-assigned another function. Indeed does anybody know what that line of serrations does at the back of the hotshoe stirrup, it may be for this very function.

Mark, Leica have said so far that the coding is to deal mostly with vignetting. Now how tightly the compensation has been programmed is an unknown, it could be that variable within the Tri-Elmars is within the compensation envelope.

What I would like to know is how the camera knows the lens aperture actually in use, since the vignetting variation can be as much as 2 stops, between full stop and say f5.6.

[leicapfile]

It may be possible on the coded bayonet mount of the 28-35-50, that at each of the focal length settings, a different set of two of the codes are bridged mechanically when changing focal lengths. Perhaps that is why six spaces were chosen. (2-2-2, each different) The same solution for the new wide TRI.

 

As was suggested, it may be a function resolved by the aux finder driven by the frame selected and then transmitted to the body. I would tend to doubt that however and would vote for the 'secret' (if there is one), being in the coding. That makes more sense, and I'm sure they thought about. Remember they knew they had two different Tri-Elmars they had to contend with.

 

It would be interesting to compare the code patterns on the different TRIs for a clue.

 

Best,

 

Jerry

[marknorton]

Here's a scan of the mount of my first generation Tri-Elmar - a single code and the pictures of the new lens suggest it will be a single code too.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[cbretteville]

The coding reads 21 decimal - with white as ones if read from the left and with blacks as ones if read from the right - cool. If read from the right with whites as ones it reads 50 decimal.

 

Ok, geek humor - I'll crawl back under my laptop.

[ho_co]

The coding reads 21 decimal - with white as ones if read from the left and with blacks as ones if read from the right - cool. If read from the right with whites as ones it reads 50 decimal.

 

Ok, geek humor - I'll crawl back under my laptop.

========================================================

There are only 10 kinds of people in the world: those who

understand binary, and those who don't.

========================================================

 

--HC