SMC 28/3.5 aperture oddity

<p>Hi All.</p>

<p>I recently purchased a Pentax K-series 28mm f/3.5 lens. I'm generally happy with it and will probably keep it, but I've noticed something odd about the apertures. The ring is marked "3.5, 5.6, 8, 11, 16, 22". In testing on my K20d, using reciprocity as a standard, I found that 5.6, 8, 11, and 16 all gave results that were consistent with each other. But 3.5 and 22 both gave results that were about 1/3-1/2 stop underexposed, compared to what reciprocity would have predicted.</p>

<p>For instance, f/8 and 1/60 will give the same exposure as f/5.6 and 1/125, which is what you would expect. But then, for the same subject with the same lighting, f/3.5 and 1/250 will also give the same exposure, which implies that what's marked as f/3.5 is really more like f/4. The effect is similar at the narrow end. F/11 and f/16 will give the results you would expect based on reciprocity, but then f/22 and half the f/16 shutter speed will be under by about 1/3-1/2 stop.</p>

<p>If somebody has an interesting point to make about it, I'll describe my informal testing procedure. It gave consistent results over three trials. I am looking at the histogram, not just trying to judge exposure on the LCD.</p>

<p>Two questions come to mind:</p>

<p>(1) Is there a known issue with this lens? E.g. is it really an f/4 lens and Pentax cheated a little in the description? I have done some research and haven't come across any description of such an issue.</p>

<p>(2) It's striking that the two extreme stops are roughly consistent with each other, while the intermediate stops are also consistent with each other, but the extreme stops are inconsistent with the intermediate ones. Is it possible that the aperture ring could have gotten out of alignment, so to speak? The largest stop is set by the pressure of the camera's aperture tab against the lens's lever. The smallest stop presumably is built into the lens. The intermediate stops are set by the aperture ring. So if it's possible for the aperture ring to get out of alignment with the lens, that could explain the results I'm seeing.</p>

<p>The lens is in very good cosmetic shape for such an old lens and doesn't show any obvious signs of having been worked on by an amateur. When I take it off the camera and work the aperture lever with my finger, it seems snappy and free-moving. I can't observe any stickiness. If I look through the unmounted lens while moving the aperture ring, at f/22 the pentagon formed by the blades is distinctly irregular, with one obviously longer side. As I open the aperture the irregularity becomes less and less noticeable. At f/3.5 the blades appear to be fully retracted, as one would expect.</p>

<p>Probably the most sensible answer is "Quit worrying about it and go take pictures", and that's what I intend to do. I am puzzled by this, though, and would be interested in hearing from anyone who knows more about it.</p>

<p>Your conclusion to quit worrying about it is probably the right one, there's probably nothing wrong with your lens. I have sometimes wondered if things like this can be explained by falloff--like if only the center of the frame is considered the f/3.5 rating might seem more justified but if you look at the whole frame, when shot wide open it will appear some fraction darker.</p>

<p>I have wondered why Pentax doesn't include 1/2 stops at the aperture range extremes for most of their lenses. In cases like the lens you mention, not only do they not provide the half-step for f/4.8, but the difference between the extreme and first click-stop is 1 1/3 EV (as rated, ignoring the exposure issue you mention). I don't necessarily think that from an exposure point of view that extra click-stop is all that critical but from an ergonomic point of view, consistent intervals makes sense.</p>

<p>Reciprocity is the wrong term, but I presume that you changed the shutter speeds as you changed the aperture, for the same exposure value.<br>

Light falloff (at the corners) will be more at the wider apertures. This may be part of what you're seeing. Compare exposure in the middle, not the corners.<br>

You may also be seeing slight quirks in exposure time accuracy.<br>

It's also possible that there's slight misadjustment of the aperture mechanism, or that the blades are shutting a trifle slowly. I had to clean the whole aperture mechanism on my Pentax-M 28/3.5.</p>

<p>I have noticed the same thing with some older MF lenses and cannot explain it.</p>

<p>Thanks for the replies. John, what's wrong with my use of "reciprocity" and what would be the correct term for what I describe in my second paragraph?</p>

<p>Andrew, re your second point, on this particular lens (K 28/3.5) there is a click stop between f/3.5 and f/5.6. I'm not sure what f-stop it corresponds to. I was trying to figure that out when I started to notice the issue I asked about. Once I noticed that issue, I stopped paying attention to the intermediate click stops. (I believe you about the general point. Thus far, I have only one other K mount lens with an aperture ring.)</p>

<p>Check your formula for exposure. Pentax and Minolta use a different brightness constant than CanNikon and Sekonic.<br /> http://en.wikipedia.org/wiki/APEX_system<br /> Pentax uses a measuring constant of 14. The others use 12.X. Wiki suggests that this leads to a metering differential of about 1/6 EV. So, if this is applied to the problem of evaluating predicted exposures, every so many cycles of the problem, an error would accrue. <br /> A^2/T = (BS)/K</p>

<ul>

<li>A = aperture</li>

<li>T = time (shutter duration)</li>

<li>B = brightness</li>

<li>S = sensitivity</li>

<li>K = 14 for Pentax.</li>

</ul>

<p>For Brightness, I estimate by approximate values of sunny 16 against some numbers I found in my spotmeter manual. My spotmeter is Minolta. My cameras are Pentax. I don't push for exactness with exposure.</p>

<ul>

<li>EV 10, twilight, 140</li>

<li>EV 11, full shadow, 280</li>

<li>EV 12, open shade, 560</li>

<li>EV 13, cloudy bright, 1100</li>

<li>EV 14, weak sun, 2300</li>

<li>EV 15, bright sun, 5600</li>

</ul>

<p>Using those values for brightness, and the APEX formula above, we can solve for an exposure well enough to get the picture done without a light meter. I suspect a measurement constant value difference may be a contributing factor to the prediction problems. I think it's unlikely to be a mechanical problem with the equipment.</p>

<p>Maybe instead it's two different approaches, by metering constants, to solving the same problem. Plug in two different constants, it'd stand to reason that we'd get two different answers. Yet, if we don't recognize that were using two different constants, then maybe we'd suspect there was an error where we can't otherwise observe one.</p>

<p>I suspect your equipment works, but the trouble is with the math. How are you evaluating the idea that the far-end apertures are incorrectly exposed?</p>

<p>When you say you are looking at the histogram, it's likely that the bar graph is showing the math, directly. So, if the graph looks a little short to us, maybe it looks dead-on to the engineers. Since, they'd probably evaluate the graph by the actual data. Sometimes we have to be careful to deduce and not induct results.</p>

<p>If you are using the little bar graph histogram provided by the camera, notice that those axises are not labeled. There's also no table report showing actual data collected. The graph itself is probably designed to be an approximation.</p>

<p>So, if that graph looks bent a little one way, when you were expecting the other, maybe it goes back to that metering constant.</p>

<p>When Minolta introduced the 58/1.2 MC Rokkor it had to contend with the considerable amount of light fall-off at the corners wide open. To prevent people from underexposing their film Minolta moved the aperture number of the f/1.2 setting slightly. This would make the corners less underexposed and the center a little overexposed. If you used stop down metering you would get a sligltly different reading. When full aperture metering was introduced some photographers worried that while it was more convenient it might introduce metering errors because of mechanical differences between lenses set to the same aperture. Some later electronic cameras solved this problem a different way. If I use a Minolta X-700 in Aperture Priority mode and the f/8 I have set is off by a little, the camera will adjust the shutter speed to compensate for it. Things got more complicated when variable aperture zoom lenses came into use. Some older cameras are still better used in stop down mode with variable aperture lenses. I have two 50-250/4-5.6 Tokina AT-X lenses in Konica mount and I find I need to make a manual adjustment to keep from getting underexposure at the long end. The 28/3.5 SMC K model is a cult lens. If you don't want it, plenty of other people do. </p>

<p>Reciprocity is the failure of the same "amount of light" to produce the same negative density when the exposure time is longer, and is a failure of film.<br>

In general, reciprocity is not a failure of digital sensors.<br>

It's possible that the f-stops are not quite accurately as marked on the lens. For the cameras that it was built for (before "crippled" mounts), all that mattered were that the aperture follower correctly reported the relative T-stop relative to wide open. What that peg says is how much less light will really come through this lens during exposure than is coming through right now with the aperture wide open. That's relative to the point that Jeff is making, on some lenses the light falloff at widest aperture may result in a T-stop at that aperture that is far less than one stop larger than for one stop down. That is, you stop down the lens one stop, but the amount of light goes down by less than one-half, because the falloff is much less.</p>

<p>John O'K-O, thanks for those links. Of course I've heard people talk about EV but I didn't know about the APEX standard. If I have the gist of it, though, you're giving me more credit than I deserve. All I was doing was using the camera histogram to check that the aperture settings on the lens were consistent with each other. I was making no attempt to check whether the exposures were objectively correct by any standard. So I don't think there can be a question of different constants.</p>

<p>Aron, your post prompted me to check my own copy of the K 28mm f3.5 lens using one of my MX bodies. The MX has a half-stop LED meter indication (and, hence, a +/- 1/4-stop resolution) in the viewfinder. The results were consistent with yours.</p>

<p>With the aperture set to f5.6 and the -1/2-stop LED lit, opening the aperture to f3.5 produced a maximum increase of 1 stop in the meter indication, rather than the 1.5 stop indication increase that I expected. (The difference from f5.6 to f3.5 is actually 1&1/3 stops, just 1/6-stop short of 1.5 stops, and hence should be shown as 1.5 stops given the MX's +/- 1/4-stop resolution). The slight shortfall between 1&1/3 and 1.5 stops leaves some room for doubt, so one ought not be dogmatic, but in this respect, the lens appears to be behaving as if it has a maximum aperture of f4. I agree that light fall-off is a a good, though untested, explanation.</p>

<p>On the narrow end, closing the aperture from f16 to f22 only ever produces a 1/2-stop decrease in the meter indication, rather than the 1 stop decrease expected. This really is a bit of a surprise and does differ to the metering behaviour of other lenses I've used. My copy of the lens has a symmetrical pentagon, so I suspect that tolerances in the blades or linkages is not the reason.</p>

<p>So I don't think the aperture ring on your lens is out of alignment as you suspect. An interesting oddity, as you term it.<br /><em></em><br /><em>Andrew, re your second point, on this particular lens (K 28/3.5) there is a click stop between f/3.5 and f/5.6. I'm not sure what f-stop it corresponds to.</em><br />Pentax states in the 1977 version of the (orange) lens instruction booklet that the click stop between f5.6 and f3.5 on all lenses having a maximum aperture of f3.5 is f4.8.</p>

<p>Hope this helps.</p>

<p> </p>

<p>Rod, interesting, thanks. Anyhow, the return period for my lens has expired and I'm happily taking pictures with it, as everyone recommended.</p>