Odd metering with Ricoh XR-1s

I have two SLR bodies, Ricoh XR-1s and Carena CX-300, and I've noticed Ricoh's strange behavior. While Carena measures more or less OK and simply loses the plot in low light (1EV error at most under normal conditions, completely unreliable below EV5 @ 50 asa, something to be expected), Ricoh does something odd: with Pentax-M 50/1.7 it constantly underexposes by 1 EV, and with other lenses it's accurate. This behavior is repeatable - every time my 50/1.7 goes on, the meter would underexpose an image by~1EV.

I've tested it as well, as possible: flat surfaces, evenly lit. I've compared it with an exposure meter that always worked well - not that it does matter... I've changed ISO and lighting conditions, the lenses was changed many times per body during these tests. Same lens works OK with the other body.

Any thoughts? Or is it me being wrong somehow?

 

I've decided to go with the two instruments offering 1/3 fstop measurement accuracy: the suspected body and my light meter - the meter is not important, it's just a mean of helping myself or memorize stuff and relate to and it just happens to work correctly. What is odd, is lens X giving different reading, than lens Y by as much as 1 fstop. I use evenly back lit surfaces, like lamp shade or LCD screen, so focal length is not a factor to consider.

Lenses: Pentax-M 50/1.7 and Auto Revuenon zoom 80-200 1:4.5 were somewhere around -2/3 fstop off wide open (meaning an underexposure), Miranda 28/2.8 and Pentax F was within 1/3 fstop from what the light meter said.
Rotating iso dial works as expected from 14 din up, changing aperture also - all changes are accurately shown by the meter. Iso dial set at 13 works as if I'd set it to 12 though.

One more thought crossed my mind. I've changed the batteries quite recently... But maybe it's the batteries, that make the problem here? Other button cells I have are half-depleted, these two are the best freshiest newest couple, two months old or so. But, assuming that all the couplers/mechanics and sensors are OK, and the internal meter is calibrated properly, maybe 1.7 is simply bright enough to make the batteries act as if they were depleted, and the cheap Revuenon is imprecise to a similar degree, even if it feels heavy and solidly built?

It's not the camera being off by a minor degree that gives me the headache, but not knowing the reason.

 

Yups there is a battery check, and it would be my first camera actually requiring silver oxide and refusing alkaline, but I know that's possible, batteries are the first factor I'll try to rule out in a couple of days, most shops supply only alkaline, maybe watchmakers will have silver ones.

 

So, after visiting a shop nearby... Xmas, this is not a battery issue.

Back to square one, I'm gazing at aperture tabs in my lenses, there's something odd with how they're placed. I assume wide open position of a "leading edge" of a tab should be consistent, but it's not. Lenses giving consistent results have this tab set perfectly in the right spot. Pentax M has the tab set 0,6mm more "away", Revuenon - 2mm, hitting the spot closed to first full aperture (so from 4.5 to 5.6). The movement of the tabs seems to be far greater than possible error resulting from 0.6 offset, it would result in ~0.25 fstop error only.
Edit: this still ay be an issue with how the rest of the bayonet was machined...

 

I'm opening aperture ring entirely, and I'm measuring the distance in between the follower and the bayonet tab. I'm pretty sure that's the edge of the follower that's coupled with the body, but I'm not sure about if all bayonet tabs in all the lenses have to be machined perfectly. What I'm getting with good lenses is 8.8mm.

For today, my guess is as follows: The lenses (apart from old Revuenon, that's not even the biggest problem with this lens...) are machined tolerably well. The body itself measures things tolerably well. Hence the confusion. Added up, these small errors grow to considerable size, visible and measurable with Ricoh's needle, but not with Carena's LEDs with 2-3x bigger tolerance. Hence it looks OK with the followers off by over 1/2 stop, not mentioning 0.6mm (1/4 fstop or even less) in case of Pentax-M. It's still odd. But at least it's a kind of repeatable error one can live with.

K1000 is the only model I want to try some day in whole K-mount world, you don't need to convince me here. As soon, as something breaks in what I have currently, I'll think of buying it. Cheap, reliable in dust, cold, hot and humid, well documented in terms of repairs (only with Leicas and Pentax LX I've been able to find repairs documented as well, as here) and otherwise featureless, so there's less breaking points - this is what I need. Even if it's a 1kg brick, I'll carry it if it's worth the effort.

Also every model with Wheatstone bridge (not match needle meter) will be pretty ignorant when it comes to battery voltage and easy to adjust.

 

Yaay, thanks!
At least I know it's lens fault partially.

I've had a fair glance at K1000's repair manual, including schematics. It was a part of my recent research. It's a pity there are so few complete service books available. Optical off switch is really clever feature there, at least for what I want to do - not much with sport od wildlife work, when the cap is off for most of the time. Ricoh is way more frustrating, turning on the meter with the wind lever often works better, than expected.

Specs, including weight, are important for me, but more so reliability. Carena weights 450g without feeling flimsy, there's quite a long way to 620g of K1000, enough for a spare (or the only) lens (pancake is ~110g IIRC). OM1 is not small, but it's really low. Weight-wise, there are cameras way lighter, like *ist, 335g, pity it works only with auto-aperture lenses, not as hardy, as M series.

I take extreme measures when it comes to reducing weight - from not taking a cup on a week long trip in the woods to cutting my toothbrush in half. Not that I can't carry more, but being less tired simply makes more sense (and more opportunities for a good photo). But I won't save weight at the cost of reliable photography gear, as I won't on a comfy backpack or a spacious hammock.

 

Xmas, I have to confess I'm an alcohol stove addict... Thanks for the tips!

Nikon Collector, in XR-1s I have mirror lockup with self-timer, Carena CX-300 (same as Cosina CT-1ex) is feature-wise K1000, shrunk down, with LEDs rather than a needle. At least I know where I'm going. I know I'll be missing some features, but it's not about feature-full camera in my case. I doubt there's an analog one in K-mount world offering fair weather-sealing with matrix metering, and that's the mandatory stuff for some people in similar circumstances. I need simplicity, and if this means taking a cable release with me - I'll take it, no problem.

 

Ralph, my Ricoh XR-1s measures the light wide open.

 

1 stop off would be enough to ruin slide film, even with proper bracketing in +1/2 -1/2 manner.
KR5 Super 2 looks like a great camera, similar to my Carena in many regards Thumbs up!

I've used flat, featureless surfaces: wall, cloth, table, lcd, ground, with daylight, fluorescent and tungsten light where possible. Metering pattern has little to do in such cases, but I always remember to set focus to infinity, I think focusing from 0,25m/10in would result in measurable brightness drop, this is how close my 28mm lens would go. Regarding pattern, there are some cameras that are compensating for backlit subjects, and there are cameras compensating for light coming through the viewfinder, so I was trying not to let too much light in. Focal length - I'd say there may be a tiny difference in how bright image is formed with advanced viewfinders (fresnel lenses etc.), but it's only a speculation, and the difference can't be that big.

Initially I've noticed strange behavior of both bodies: one was acting strange in low light, while other was more lens-dependent. One gave greater variance, than the other. First issue is simple to explain - Carena doesn't work in low light, I have to live with it. Second issue was a bit harder, related to the measurement itself - Carena gives green light for way more, than 1/2 fstop range, while OK for me in Ricoh was within 1/3 fstop range. I'm quite happy to pin-point such issues and their outcomes... So I've reduced down amount of tested factors, getting rid of Carena body from the equation. I've eliminated battery issues with Ricoh and I've been left only with three lenses, rather than four, as the aperture tongue in my Revuenon, being off by 2.2mm, makes any testing with this lens pointless.

My only conclusion is 0.6mm of deviation within aperture transfer mechanism is enough to introduce considerable error. Smaller, than what Carena allows, but enough to confuse me with what Ricoh shows. so I'll have to trust my external light meter at least now and then, and I'll try to service Pentax-M lens myself, I hope it's doable with grinding stone in worst case, or screwdriver alone if things are designed to do it this way. For now, I'm unable to do anything more.