This is unexpectedly long and probably quite boring, but I’ve typed it now so it might as well get an airing.
Hacking P-TTL for manual lenses
Normally I prefer manual flash, in a situation where lighting is relatively static and the camera isn’t being moved around, complete control over lighting gets you hopefully the shot you had in your head.
For live insect photography outside though, manual flash can be hard work. The slightest movement of the clouds or turning a few degrees and the lighting changes - the exposure settings you worked out are now wrong. It’s less of an issue if you illuminate the whole scene with flash, but black backgrounds are rarely as nice as a blend of ambient and flash.
P-TTL (Pre-flash Through The Lens) is the automatic metering system which Pentax uses on its modern cameras. It creates an initial very quick (usually too quick for the eye to register) ‘pre-flash’ and the metering system uses the feedback from this together with information about the current camera and lens settings to calculate how much light the flash needs to put out for the exposure. The camera then takes the picture with these settings and, hopefully, you come out with a nice image.
P-TTL isn’t perfect and it gets confused by high contrast or shiny objects, but you can use exposure compensation in these situations. I thought it might be useful for insect photography, so I could concentrate more on composition and less on trying to chimp my way to the right settings. While a lot of people swear by manual flash even for insects, I want to see how P-TTL will work out…it might be useless…or it might not..I might end up using a mixture of both.
Both my macro lenses are manual models from the 80s, they offer great image quality but have no electronics in them at all. P-TTL relies on feedback from the lens such as its aperture and focal length in order to make exposure calculations. Officially, P-TTL only works with automatic lenses, it can’t be made to work with manual lenses. Except it can. Sort of.
I carried out a couple of reversible modifications to try this out. None of these are new ideas at all, just new to me.
Telling the camera a manual lens is an automatic lens
The camera mounting ring has a slightly recessed connection. Automatic lenses have a protruding pin which shorts this connection and tells the camera that they are automatic. If the connection isn’t shorted, i.e. when a manual lens is mounted, the camera reverts to more basic settings – which don’t include P-TTL and multi-segment metering because it doesn’t know anything about the lens that’s connected, so can’t make the necessary exposure calculations. What you can do though is unscrew the camera mounting ring, place a tiny amount of foil over the recessed connection to short it and then screw the ring back on (first image).
Now the camera will grudgingly go along with the lie that there’s an F1.2 – F22 auto lens connected when you connect a manual lens. If you then set the aperture on the lens ring to its smallest diameter (F22 for my lenses) you can then control the aperture of the manual lens using your camera dial. Sort of.
Aperture arm mismatch :(
From what I understand, the issue is the movement of the aperture arm on a modern Pentax dslr camera is equivalent to the surface area of the lens opening – so the arm moves different distances between each f stop. The aperture arm on the older Pentax cameras which manual lenses were designed for moved the same distance between each stop and the conversion to the correct surface area of the lens opening was done by the lens. As a result, if you put an old manual lens on a new Pentax dslr camera and try to control the aperture from the body, the f stop you set on the camera is rarely the same as what the lens actually gets set to. It is the same wide open and fully closed and roughly similar on my lenses between f8 and f13, however around f4 to f5.6 there can be up to two stops of light difference between what the camera thinks and what the lens is actually set to.
P-TTL performs its measuring pre-flash with the aperture on the lens wide open and calculates how many stops to tone the flash down by for the actual shot according to what aperture you have defined. Unfortunately because of the ‘arm mismatch issue’ the real stopped down aperture is not quite the same as what the camera thinks it is and the wonky calculations result in in under or overexposed shots. If we need to keep using different exposure compensations for various apertures to get P-TTL to expose properly then this is no easier than just sticking to manual settings.
As above, the metering works fine if the camera is shot with the aperture on the body set to wide open. This is because the camera doesn’t apply any misinformed stop-down calculations – the aperture it pre-flashes with is the same aperture it’s going to take the picture with.
So, what if I disengage the lens from the camera’s aperture arm and leave the aperture setting on the camera at its widest?
I didn’t want to perform any physical changes to my lenses and as they’re close focusing anyway, losing infinity focus wasn’t an issue, so I’m using a short extension tube to separate the lens from the aperture lever. I didn’t need to worry about this previously as the steel base of the lens connected the pins automatically, but as the extension tube is anodised aluminium, I had to add a little strip of foil to the bottom of it to short the data pin on the camera body to ground (this pin must also be shorted in order for the automatic lens pin mod to work) (second image).
Now the lens is functioning like an even older manual stop down m42 lens. It can be set to any aperture but as the camera think it’s always wide open it should meter correctly. Nice!
There is a small downside – modern cameras not in the process of actually taking a picture hold the aperture of the lens wide open because it provides the maximum amount of light for composing and focusing the picture. With a manually stopped down lens, what you see through the viewfinder is the same amount of ambient light that will enter the camera when a picture is taken. In some ways it’s quite useful for macro because you can see exactly what your depth of field is - however at higher f stops, the viewfinder will get increasingly dim…fine in daylight, but at dawn or dusk, possibly so dark it’s unusable. I’ll have to see how much of an issue this is, but I’m thinking I can probably live with it.
So, did it actually work? From a few tests (Images 3 & 4), it seems to work just fine :)Looking forward to trying it outside….
- reallysmall posted this