Mamiya 645 PD Finder Meter - Does not work in low light

I acquired a Mamiya 645 1000s recently and have only shot two rolls of film on it. My last roll was spent shooting a cityscape at golden hour and blue hour.

The metering seemed to work fine during golden hour (I checked it against an iPhone light meter app), but as soon as the light started fading after the sunset, the meter shows under exposed no matter how open the aperture was and how slow I set the shutter speed (I ended up shooting manually using my iPhone as an impromptu meter).

I also tested it indoors. The same issue arises where no matter what, the meter shows underexposed unless I point it directly at a light source (in which case, it would show overexposed). The camera is properly set up (lens set to Auto; full light going through the viewfinder; fresh LR44 batteries; shutter speed on camera set to red dot, which defers to the PD prism).

Is this normal? I thought the PD prism was really good in low light. And why would it only work in lit situations and then crap out when it's dark?

The meter in the PD prism finder is not particularly sensitive in lower lighting conditions: it has a range of EV 2.85-17, vs the typical range of EV 1-17 in most 35mm SLR meters. The lowest light it can read before becoming unstable is roughly f/2.8 at 1 second with ISO 100 film (1/4 sec at ISO 400).

Depending on the scene you're trying to shoot, it could simply be that the light falls below what the meter can accurately measure, so it just defaults to "underexposed" indication. Many early medium-format TTL meters from the 1970s were similar: not well-suited for twilight landscapes or indoor use.

You can verify this by comparing the prism readout at its lowest level with a separate handheld meter, or phone app. Aim the camera at a scene just above its measuring cutoff at ISO 100, so you get a recommendation of f/2.8 and 1 sec. Then measure the same scene with the handheld meter or phone app: if these display an EV of approx 3.0, with the same suggested settings, your Mamiya prism is accurate at its lowest limit. If the separate meters are grossly different, the Mamiya meter may need repair. Repeat the comparison while aiming directly at a bright lamp indoors, and a bright scene outdoors, again checking for a match.

There are at least two different types of metering finder for the metal Mamiya 645 system. The first version was designed way back in the 1970s. So the fact that it's still working at all, nearly 50 years on, should be seen as a bonus.

BTW, you need to remember to turn the camera-powered auto-exposure metering prism off after use, otherwise it'll quickly flatten the camera battery. The older metering-only version with its own battery isn't quite so power hungry, but as I suspect you've found, it's not very sensitive either.

I thought the PD prism was really good in low light. And why would it only work in lit situations and then crap out when it's dark?

You have to remember that 400 ISO was about as fast as it got in the 1970s, and TTL metering was a relatively novel thing in a rollfilm camera. In those days, if you wanted to meter candlelight, or whatever, you reached for your trusty Gossen Lunarsix. The TTL prism was designed for quick 'snapshooting' in good light conditions.

If you want all mod cons and cameras that can focus and meter flawlessly when it's too dark to even see the subject, then stick with digital.

There are at least two different types of metering finder for the metal Mamiya 645 system. The first version was designed way back in the 1970s. So the fact that it's still working at all, nearly 50 years on, should be seen as a bonus.

That's what I think every time I acquire another vintage camera with built-in metering: if it meters at all, I'm pleasantly surprised, if its accurate, I feel so ahead of the game I want to party.

I just double-checked the Mamiya brochures/manuals, and count four documented meter prisms for the first wave of M645, M645J, and M645 1000S cameras: CdS prism, PD Prism, PD-S prism, and AE prism (there was likely also an AE-S prism or factory mod to match the PD-S for compatibility with the 1000S model, for a possible total of 5 meter prisms). To my surprise, all of the prisms have the same range specification (EV 2.85-17). One would have thought the PD and AE versions, with silicon blue cells, would read a bit lower light than the CdS. Apparently Mamiya felt consistent range across all the prism variations would be more useful to the professionals most likely to buy them?

Nikon had the opposite philosophy: their "PD" prisms could read several stops lower light than their CdS prisms (-2 EV vs +1 EV). One reason the F2 S, SB and AS prisms are prized over the standard CdS versions.

Thanks for all the insight. It's likely that the metering range is just not that great. I basically can't get anything to read other than underexposed at f2.8 (no matter how slow the shutter up to 8s) indoors under 60W lights. I still think my particular prism is especially limited. I might just get a waistlevel finder (I hate holding the camera to eye-level) and meter with my Sekonic or iPhone (in a pinch).

Also, the PD prism I have has a metering button that turns it on for about 10 seconds before turning off automatically. I think it's a more recent version (has the 1/1000s top shutter speed).

Also, the PD prism I have has a metering button that turns it on for about 10 seconds before turning off automatically. I think it's a more recent version (has the 1/1000s top shutter speed).

Yes, that would be the PD Prism Finder S. Same as the first PD meter prism, just updated to accommodate the faster 1/1000 top shutter speed of the 1000S model. Yours is probably working within spec: vintage TTL film camera meters often have hard cutoffs at their limits. Mamiya designed the PD-S lower measurement limit as f/2.8 at 1 sec at a given film speed, so it will display underexposure at 2, 4, and 8 secs. Even if the exposure would technically be correct at the slower shutter speed.