[pathdoc]

Think about what happens when we put a K or M lens (or third-party equivalent) on our Pentax DSLRs. We put the camera into manual mode then push the green button, which stops the lens down and takes a meter reading, and then we push the shutter button, which stops the lens down and takes the shot.

All of this is handled through the camera's electronics. That means all of it is handled in firmware.

So here's an idea - have an option in the menu for "K mount mode". Pushing the shutter button to the halfway point would give focus confirmation (as is the case now with manual lenses on our DSLRs) and would also stop down and meter via the firmware. If you didn't follow through and take the shot within (say) a second, it would go back to full-open again until you pushed it all the way down.

We could go one further and shift to a mode in which the aperture ring is allowed to stop down the lens constantly and the lens behaves like a Takumar in manual-clutch mode.

I see no reason why this could not be applied retrospectively to existing bodies. Even if it didn't formally uncripple the mount, it would at least make the process of using K and M lenses more seamless.

The other thing which comes to my attention is that of course K and M mount lenses do not get their aperture settings written into the firmware. Now, wide-open metering works because the aperture arm on the lens (tied to the ring) pushes on its opposite number in the camera to tell the meter how many stops down from maximum it will be when the shutter goes off. In other words, metering in uncrippled cameras is relative to the full-open state. Now turn that on its head - we can manually enter the focal length (although I wish my K5 would write this into the EXIF regardless of whether SR was active), so why not the aperture? The camera can tell how many stops of light it loses between full open and when the green button is pressed, and write the relevant value accordingly.

Third - for an A mount lens and subsequent, IIRC the camera controls the aperture in firmware by the degree to which the stop-down lever is moved. I know the relationship between stop-down arc swing and degree of aperture closure is different for the K and M mounts, but it is still predictable. It should therefore not be too difficult to write a firmware routine which would allow you to mount a K or M lens (at least a prime), tell the camera the focal length and maximum aperture, and then (again in "K/M mount mode") have the computer swing the lever by the appropriate amount.

If I'm missing something here, or I've got it wrong, by all means tell me - but it seems there should be no reason why the K mount cannot effectively be uncrippled in firmware, at least for older primes and constant-aperture zooms. For variable aperture zooms, there is of course always the status quo. And if that's possible, it should be possible to uncripple it retrospectively across the board for all Pentax DSLRs.

At least until such time as a proper K mount is restored.

---

Last edited by pathdoc; 08-12-2015 at 08:11 AM. Reason: correction of typos

[boriscleto]

The way I see it is you're lucky you got backwards compatibility at all. When the *ist D came out in 2003 it didn't have stop down metering, non A lenses didn't function at all, just like the crippled MZ/ZX series cameras. Stop down metering was only added later. Nikon didn't offer any backwards compatibility till 2006.

The goal of any camera manufacturer is to sell you new lenses. Any backwards compatibility is a bonus.

[pathdoc]

Granted, but I'm looking solely at what ought to be possible.

---------- Post added 12-08-15 at 13:11 ----------

Addit: I'm suggesting all this because I know it's probably a pipe dream to expect Pentax to fully restore the mechanical contacts that handle pre-A mount lenses. But if a firmware "get-around" could be worked, it would involve no physical design changes and we might stand a better chance of getting it implemented.

[Not a Number]

So here's an idea - have an option in the menu for "K mount mode". Pushing the shutter button to the halfway point would give focus confirmation (as is the case now with manual lenses on our DSLRs) and would also stop down and meter via the firmware. If you didn't follow through and take the shot within (say) a second, it would go back to full-open again until you pushed it all the way down.

With the iris stopped down there may not be enough light for focus detection to work. The half-press would have to turn on focus detection - pause until focus is confirmed then stop-down and meter. Or stop-down and meter first and open for focus detection,. You may as well use the green button.

It would be simpler to input in the min and max apertures as Nikon does and then treat it as an "A" MF lens. Even though there is a non-linear relationship on the stop-down level. Your exposures will likely be all over the place because of this.

[pathdoc]

Even though there is a non-linear relationship

I seem to recall that although it is non-linear, it is also well known and predictable - the issue would be applying the right amount of displacement, but surely the current generation of microprocessors and controllers should have no problem dealing with this?

[wombat2go]

which would allow you to mount a K or M lens (at least a prime), tell the camera the focal length and maximum aperture, and then (again in "K/M mount mode") have the computer swing the lever by the appropriate amount.

As I understand it, the lever on the right side of the lens ( looking down at top of camera) is not for metering or positioning, it is simply 2 position: "hold wide open" or "momentarily allow lens to stop down to its setting".
In the second position the camera lever merely swings out of the way at high speed, to a stop.
The metering lever on the left side has a wider arc and in the old cameras, used to position a potentiometer ( variable resistor) in the metering circuit.

My thought would be that attempts to re-introduce the left side lever for -M lens metering in a dslr would cause user unhappiness because it would be too inaccurate. That lever evolved to the in-tandem 3 (?) bit pattern for the -A cameras followed by the digital comms in the DA etc.
There was obviously a reason to keep improving it.

[pathdoc]

As I understand it, the lever on the right side of the lens ( looking down at top of camera) is not for metering or positioning, it is simply 2 position: "hold wide open" or "momentarily allow lens to stop down to its setting".

In that case, how is the stopping down on K-A mounts controlled when they're in auto mode?

The metering lever on the left side has a wider arc and in the old cameras, used to position a potentiometer ( variable resistor) in the metering circuit.

This much was clear actually looking at an uncrippled (P30T) K mount and finding the camera-side counterpart to the aperture ring position arm. I've always been lousy at circuits, but I understand enough to imagine in broad terms how it might be done even with 1960's technology.

My thought would be that attempts to re-introduce the left side lever for -M lens metering in a dslr would cause user unhappiness because it would be too inaccurate.

If they could make a potentiometer circuit work then, why not now?

[glasbak]

I also did think about this problem, already a decade ago or so (when I bought the *istD)

Your first idea is a very old one, in screw mount times it was already implemented by the chinon 2000 iirc.
Of course this is possible, but pentax want to sell lenses, so it wil not make life with old ones easier.

Your second idea is simply too inconvenient, you would not only have to enter the working aperture, but also the full aperture.

Your third idea goes down by the assumption it is nonlinear, but it is probably much more linear than the absolute nonlinear A lever to aperture movement.
Only the A lever movement is known (the glass area halves for the same linear displacement of the aperture pin), and calibrated.
The pre A lenses lack the calibration, the 'gain' is not the same across lenses, and the 'offset' is not the same, even for lenses of the same type.
This means the moment the aperture starts to close is not guaranteed at the same position of the aperture pin, and the amount of diameter change is different for each different type of lens.

I have come to the conclusion that, to regain full functionality of the pre-A lenses on a crippled camera, we must 'upgrade' the bayonet.

If Pentax should offer a upgrade service, where they replaced the bayonet (which is remarkably similar between almost all K and M lenses) with a 'chipped' one.

This chipped one could tell the camera everything it needs to know :

1. focal length (zooms will be a problem though)
2. Max aperture, and aperture range. (optional add a sensor)
3. The aperture lever motion to aperture relation (calibration has to be done for each individual lens when the bayonet is replaced, or the camera has to do it similar to how now the AF focus offset is calibrated and stored in the camera)

The A position is missing on the old lenses, so best is to sacrifice the smallest manual aperture setting for this.

All of this will need the camera to be aware of this type of lenses, so this upgrade service has to be supported by pentax who have to add something to the firmware.

[stevebrot]

I seem to recall that although it is non-linear, it is also well known and predictable

ROFL

It all depends on the lens mounted and the chosen aperture. Truly! Some lenses require full actuator travel while others trip fully with little movement. Some third-party offerings may provide linear actuation due to multi-platform support, but that would be incidental and not predictably present.

I know the idea of having some sort of automation with the non-A contact lenses would be a nice thing, but the stop-down option provided actually works pretty well with current Pentax bodies.


Steve

[stevebrot]

My thought would be that attempts to re-introduce the left side lever for -M lens metering in a dslr would cause user unhappiness because it would be too inaccurate.

Accuracy in center-weighted mode would be similar to that of the bodies the lenses were originally made to fit. I have several late-70s to mid-80s K-mount SLRs and all provide excellent meter linearity and accuracy. Those with SPD cells also have sensitivity equivalent to current cameras having similar meters. A variable (strip) resistor may not be the best choice today, but digital coding based on lever position would be an obvious alternative and may even be possible without a mechanical coupling.


Steve

[Lowell Goudge]

I seem to recall that although it is non-linear, it is also well known and predictable - the issue would be applying the right amount of displacement, but surely the current generation of microprocessors and controllers should have no problem dealing with this?

it is often reported as linearly changing the radius of the aperture with movement as opposed to the area

As I understand it, the lever on the right side of the lens ( looking down at top of camera) is not for metering or positioning, it is simply 2 position: "hold wide open" or "momentarily allow lens to stop down to its setting".
In the second position the camera lever merely swings out of the way at high speed, to a stop.
The metering lever on the left side has a wider arc and in the old cameras, used to position a potentiometer ( variable resistor) in the metering circuit.

My thought would be that attempts to re-introduce the left side lever for -M lens metering in a dslr would cause user unhappiness because it would be too inaccurate. That lever evolved to the in-tandem 3 (?) bit pattern for the -A cameras followed by the digital comms in the DA etc.
There was obviously a reason to keep improving it.

Based upon movement, the smaller the selected aperture the larger the error, but likely much better than present metering ussues with the green button and manual lenses

The left side lever gave the number of stops the lens was stopped down, which permitted open aperture metering, using the right side lever as an "on/off" switch.

It worked just fine,mand then all you would need to input to the camera is the maximum aperture so that the non linear metering of the focusing screen could be adapted for (an issue well costume noted when the K10D first came out)

[npc]

I also did think about this problem, already a decade ago or so (when I bought the *istD)

Your first idea is a very old one, in screw mount times it was already implemented by the chinon 2000 iirc.
Of course this is possible, but pentax want to sell lenses, so it wil not make life with old ones easier.

Your second idea is simply too inconvenient, you would not only have to enter the working aperture, but also the full aperture.

Your third idea goes down by the assumption it is nonlinear, but it is probably much more linear than the absolute nonlinear A lever to aperture movement.
Only the A lever movement is known (the glass area halves for the same linear displacement of the aperture pin), and calibrated.
The pre A lenses lack the calibration, the 'gain' is not the same across lenses, and the 'offset' is not the same, even for lenses of the same type.
This means the moment the aperture starts to close is not guaranteed at the same position of the aperture pin, and the amount of diameter change is different for each different type of lens.

I have come to the conclusion that, to regain full functionality of the pre-A lenses on a crippled camera, we must 'upgrade' the bayonet.

If Pentax should offer a upgrade service, where they replaced the bayonet (which is remarkably similar between almost all K and M lenses) with a 'chipped' one.

This chipped one could tell the camera everything it needs to know :

1. focal length (zooms will be a problem though)
2. Max aperture, and aperture range. (optional add a sensor)
3. The aperture lever motion to aperture relation (calibration has to be done for each individual lens when the bayonet is replaced, or the camera has to do it similar to how now the AF focus offset is calibrated and stored in the camera)

The A position is missing on the old lenses, so best is to sacrifice the smallest manual aperture setting for this.

All of this will need the camera to be aware of this type of lenses, so this upgrade service has to be supported by pentax who have to add something to the firmware.

I don't think it needs to be that complicated at all. It all can be done purely in software/firwmare. It is a pity that the "traditional" photo compaines still fail to embrace the fact that today the cameras are just computers with a lens mount and do not take full advantage of this.

Where right now all Pentax cameras ask you to select the focal length it can actually be a menu where you can select a complete, user-created lens profile say M50_1.4

The profile can store besides focal lenght also max aperture, distortion corrections, etc that the camera can apply automatically.

FUJI, Sony (via downloadable app) and most imortantly RICOH GXR M-mount support this kind of functionality. Also since the number of different lenses is somewhat finite - the Pentax community could quickly create a repository for these profiles so you can just upload those to the camera.

Then you can dial the f-number on the body like you would do it for a modern lens (but of course also setting it in the aperture ring) - so while camera metters wide-open - it can calculate the correct exposure since it knows the actual f-stop the lens is set to. Sure, one extra step - but if you don't what to use this feature - do not use a profile or leave the f-stop at the default 0 to get the current behaviour.

As a side efect you can also get correct exif information recoreded.

Would work best only for primes but who cares about old zooms anyway

[jonby]

Assuming it's not possible to automate aperture control with any of the methods suggested, I have a few other suggestions for improvements in the handling of K/M lenses which wouldn't require any major changes/technical challenges.

Regarding the problem of the aperture value not being recorded in EXIF with K/M lenses, my suggestion would be to deal with it in the following way: when a K/M lens is mounted, the camera allows you to use the e-dial as normal to set the recorded aperture, which is shown in the usual places in the VF, top LCD and back LCD. It is then up to the user to simply match this to the aperture set on the lens ring. Whatever aperture is set on the camera when the exposure is made is recorded in the EXIF. I guess this could be seen as confusing behaviour to some users so perhaps it would need to be enabled in the setup options. It would probably be a good idea to have a small 'warning' mark next to the aperture value to remind you that the displayed value is not necessarily the actual value in use. But I think this would be a really simple way to do it. If my understanding is correct, then if there were also a way to input the maximum aperture for the lens in use, then it would be possible for the camera to display a manual exposure meter reading based on the displayed aperture, as per normal manual mode behaviour. This would mean that you could half-press the shutter button to begin metering, set the shutter speed and aperture while watching the meter as normal, and then just set the corresponding aperture on the lens before taking a shot. This could even give you a form of aperture priority - as long as the aperture set on the camera matches that on the lens then the camera could set the shutter speed.

It would also be useful to be able to set and record focal length in EXIF when not using SR, as the OP suggests. As a regular tripod and K/M lens user I find it frustrating that FL is not recorded. Not quite sure of the best way to do this. A setup option to force the SR setting to be recorded whenever a K/M lens is in use, even if SR is disabled, would be a start, but there should also be an easier way to change it without going into the menus, as this is too slow when changing FL on the fly with a zoom lens. One way would be to have it as one of the 'quick menu' options.

---------- Post added 08-12-15 at 05:45 PM ----------

Then you can dial the f-number on the body like you would do it for a modern lens (but of course also setting it in the aperture ring) - so while camera metters wide-open - it can calculate the correct exposure since it knows the actual f-stop the lens is set to. Sure, one extra step - but if you don't what to use this feature - do not use a profile or leave the f-stop at the default 0 to get the current behaviour.

As a side efect you can also get correct exif information recoreded.

I think we're suggesting more or less the same thing here. I was writing mine as you were writing yours.

[wombat2go]

Accuracy in center-weighted mode would be similar to that of the bodies the lenses were originally made to fit. I have several late-70s to mid-80s K-mount SLRs and all provide excellent meter linearity and accuracy.

I looked at some of my lenses ( -M, -A, & off brand) to see the aperture metering tang slop, while giving them a wriggle, under the venerable B&L,....
It would be completely inappropriate to report any conclusions from such a rudimentary test.

One lens click stops reminded me of the method the automobile builders of the time ( 1970's and on) would determine the thickness of steel in the body panels:
"Well it ain't 18 gage, and it ain't 16 gage, well it must be about 17 gage."
No joke, that was in some standards of the time..

[JohntheThird]

There is one mayor comment to this. Because closing the aperture increases DOF, for best focusing results you want the lens to be wide open. Stopping down for metering will oppose this. So the half-press first needs focus confirmation, - then stop down metering to store the results, compute shutter speed (or aperture), then you can fullpress. This will make it slower.