Ghost image with Contaflex

[bernard_L]

Made the first test of my second Contaflex (Super) body, bought last June. Some images show a ghost image, offset from the main image by a small amount. Most noticeable on pictures taken with the 85mm pro-tessar. Some pics with the same 85mm (and same body) do not show that problem. Still unsure whether the problem might be present with pics taken with other lenses (50mm Tessar, 35mm pro-tessar). See attached pics: full frame, and two details @2700dpi (yes, I know; but how can I communicate my problem??)


The secondary image appears to be offset along a direction between 2 o'clock and 3 o'clock; the twigs along that direction appear as single. Nothing is rattling inside the Pro-Tessar 85mm (Zeiss construction). The only (tentative, incomplete) explanation I can come up with is some mechanical jolt occurring in the middle of the exposure. The problem might be made more visible by, but not caused by, the longer FL of teh 85mm. Mirror slap? Is that a Known Issue (as they used to call bugs at Microsoft)? Thank you in advance for insights.

 

[OlyMan]

Surely motion blur, caused by mirror slap or anything else, would be just that: a blur. The ghosting appears as a faint overlapped and offset image, as distinct and sharp as the main image, so surely it must be caused by some kind of reflection or refraction. Are you using a screw-on filter? Screw-on filters won't inherently produce these kind of artefacts; I'm clutching at straws really.

 

[E. von Hoegh]

Could the film be moving within the camera ?
It's an odd one, that's for sure.

 

[Nodda Duma]

Open the back, put the lens in a bulb setting, and view the image plane with a magnifier (directly or with scotch tape at the image plane for impromptu ground glass). That will quickly tell you if it's in the lens or not without having to shoot a roll of film.

-Jason

 

[bernard_L]

Thanks to responders so far.
@ OlyMan. I'm not so sure; if a mass inside the camera, at some time during the exposure, moves (in, say, 1 millisecond) from position A (at rest) to position B (dead stop), the rest of the body, from conservation of momentum, will experience an opposite motion, in inverse proportion to the respective masses (or moments of inertia if we are talking about rotation motion). Lots of thingies move swiftly inside a Contaflex during one shutter cycle. To me the jury is still out. In the shot I show, I had no filter, and, as you point out, one does not expect a filter to produce this effect. Thank you for submitting your thoughts.
@ E. von Hoegh. That crossed my mind, but I suppose the pressure plate friction would not allow such motion.
@ Nodda Duma. Just now I took the camera in my hands. In contrast with other cameras, like Oly35RC or Minolta Himatic 7SII, there is absolutely no play between the lens and the body. Kudos to Zeiss, or rather to Deckel, who supplied the whole lens assembly (except the glass) as OEM to Zeiss. Will do what you suggest, but, the focal plane sees light only for a few milliseconds; my eye and brain are not fast enough.
Next step, I'll take a series of exposures, spanning a range of shutter speeds. First loosely supported, then repeat on a reasonably strong tripod.

 

[ransel]

Perplexing - but I have seen similar effect before. Brings up more questions than answers:
1 - was the camera on a tripod - if yes, is it a very sturdy tripod and was a cable release used?
2 - what was the shutter speed - I seem to remember that 1/30-1/15 are rather tricky due to shutter and mirror movement, even on a tripod.

 

[bernard_L]

@ ransel. Camera was hand held. Shutter speed was most probably 1/125, certainly no slower than 1/60.

I noticed there is inside the body of the Contaflex, close to the rear of the lens, a gear in the shape of an arc of circle, that engages a smaller gear below the lens. It seems to move by approx 60° when the film is advanced and the shutter cocked; presumably it moves back much faster during the shutter sequence. On the other hand, any camera has parts that move during exposure, without noticeable adverse effects...

 

[Tim Sewell]

This may be a pointless comment, but have you confirmed that the issue is on the negatives and not just the prints/scans?

 

[shutterfinger]

If you have confirmed the ghost image is on the film then waste a roll of film and verify that it is staying taught at the film plane during exposure. A very slight film movement during exposure caused by the shutter may be the cause. Worn sprockets on the body may let the film slip.

 

[John Koehrer]

I'm curious about the lighting and the stability of camera or table. I've seen similar ghosting with flash and a slow shutter speed
if the camera/table were to move slightly. Hey it's just a guess.

 

[MattKing]

Were these shot with flash?

 

[bernard_L]

Thanks again to those who submitted suggestions.
@ Tim Sewell. Scanned first with V700; re-scanned with LS2000 for confirmation; it is the latter that I posted.
@ shutterfinger. Ghost is on film (see above). "waste a roll of film and verify that it is staying taught at the film plane during exposure". How can I verify that? I do have rolls to waste (found film that I don't care to develop) suitable for such mechanical tests. I see two difficulties: (a) the amount of motion, deduced from the digitized image, is of order 40-50 micrometers; the diameter of a human hair; (b) to observe the film, the back must be open, which means that the pressure plate --which plays a crucial role-- is absent.
@ John Koehrer and MattKing. No flash. Sunlit from behind my right shoulder, against a background in the shade (among my favorite cliché's for dry weeds). In 50+ years of film photography, this is the first time I see this, and it is with the first roll using a new-to-me camera; so I feel justified in suspecting the camera.

To answer a question that might come: the example I showed is not unique. Another example, also shot with 85mm FL, which helps to put in evidence small camera motion:

See the ghosting on the bright upper edge of the branches sloping from upper left to lower right. Because the ghost image is weaker than the main, it tends to be washed out for dark features on a brighter background, and to show better for bright features on a back backdrop.
And, BTW, the fact that ghosting is more apparent on images shot with the longer FL argues for camera motion as a whole, rather than film motion inside the camera, which would produce identical amounts of separation independent of the lens FL.

So today I'll make a systematic series of exposures at various shutter speeds, as I already stated yesterday.

To all who commented. Even though I may sound negative, because I dismiss most of the proposed explanations, this exercise is actually very helpful and stimulating for me.

 

[shutterfinger]

How can I verify that?

I see two difficulties: (a) the amount of motion, deduced from the digitized image, is of order 40-50 micrometers; the diameter of a human hair;

Being I'm not familiar with the camera I googled it then looked at an instruction manual on the go to site.
When changing lens does only the front element come off or is there a direct path to the film? Remove front and rear lens elements if needed so that the film/pressure plate can be accessed through the shutter with it open on B and aperture at maximum.
With film loaded Open the shutter on B (use locking cable release if possible) and physically check for film movement at each frame along the roll.
On 35mm cameras always turn the rewind knob/lever in the rewind direction until the film tightens for best advancing. Loose film in the cassette can cause stripped sprocket holes, advance jams, or spacing problems.
If the film cannot be accessed through the shutter then remove the back at the end of film advance and check the looseness or advance the film with the back off holding the supply cassette in place as it will be with the back on. Lockouts sometimes have to be deactivated as they will be when the back is on to be able to operate the camera. Careful inspection will be necessary to find the cause if it is film tightness.

 

[Ian Grant]

I wonder if the shutter blades are not closing fully immediately after exposure so you're getting a second ghosting extra exposure. It's definitely a double exposure effect.

Ian

 

[Brett Rogers]

I wonder if the shutter blades are not closing fully immediately after exposure so you're getting a second ghosting extra exposure. It's definitely a double exposure effect.

Ian

Yes. This was the possibility which occurred to me. Wondering if the lens shutter blades are sticking. They might be slow to close when the release is depressed, permitting a partial exposure after the mirror and capping plate has retracted, before the actual exposure occurs. Alternatively they may be slow to close fully after exposure. I have witnessed this with some Contaflexes if the shutter has not been cleaned, though I have not personally had the issue on a film as I've not shot with a camera thus affected. But I've seen blades leave a residual pinhole opening after firing because they are bogged down in lubricant residue. And when this happens they can take some seconds to finally creep fully closed depending on how badly they are gummed up. Before shooting more film I would wind and fire the camera off at various speeds looking at the shutter blades from the front to observe their action directly, but also observing though the lens whilst it is pointed at a light source. If the shutter is not operating correctly you should be able to spot it visually. It does not appear to be a capping plate issue as you'd get uneven fogging along one edge as a result of that if the timing was out a little or it was not seating correctly. Check it anyway to be thorough as occasionally the timing needs a little fine tuning. But it's not the issue here.
Cheers
Brett

 

[Ian Grant]

Brett, I actually have a problem like this with a shutter at the moment, although it's not a Compur, and I've seen it with Luc style shutters as well before I stripped and re-built them.

Ian

 

[Brett Rogers]

Brett, I actually have a problem like this with a shutter at the moment, although it's not a Compur, and I've seen it with Luc style shutters as well before I stripped and re-built them.

Ian

OK, that's interesting, I have not had experience with those shutters, Ian, but it's likely the same sort of problem. With the Contaflex it is probably the shutter blades themselves that are sticking together and running slow. Although as it needs a rapid stop down system to actuate when the release is depressed, it is also important to ensure these are clean, too. You won't get a double exposure from those but, if they're not moving quickly enough, that Eg f/11 aperture you selected may end up being f/2.8 instead, or effectively something on between the two, depending on how badly they need cleaning.

The complex stuff inside the body of the Contaflexes rarely gives much trouble. Once a shutter is cleaned they will generally run again OK. Given that to my eye it definitely looks like a double exposure I believe you are correct and the shutter is not moving with as much alacrity as it should.
Cheers
Brett

 

[bernard_L]

And when this happens they can take some seconds to finally creep fully closed depending on how badly they are gummed up.

If I understand correctly what you are writing, one should observe a "dragged" image, like with ambient plus flash on first curtain. While one is observing a distinct second image. So, could we be seeing the effect of shutter bounce? But I have a feeling we are getting close. If only I had a fast photo-detector ready... I do have the digital scope. This would allow to actually observe the shutter bounce.
I might have to take a dive into a synchro-compur in the near future. With some help from: http://www.suaudeau.eu/memo/rep/Compur-shutter-repair-manual.html

Before shooting more film...

The test film is exposed but not yet developed nor s*****d. Experimental data cannot hurt.

 

[Brett Rogers]

If I understand correctly what you are writing, one should observe a "dragged" image, like with ambient plus flash on first curtain. While one is observing a distinct second image. So, could we be seeing the effect of shutter bounce? But I have a feeling we are getting close. If only I had a fast photo-detector ready... I do have the digital scope. This would allow to actually observe the shutter bounce.
I might have to take a dive into a synchro-compur in the near future. With some help from: http://www.suaudeau.eu/memo/rep/Compur-shutter-repair-manual.html

The test film is exposed but not yet developed nor s*****d. Experimental data cannot hurt.

If the shutter was slow to close after exposure you'd get a dragged image. You'd have to. It's a single exposure. I've never heard of bounce in a Compur shutter (which doesn't in itself mean it can't occur, of course) but it's an issue normally associated with a focal plane shutter curtain re-bounding into the gate after it's been released.

On the other hand, the sequence of operation that occurs when a Contaflex is fired begins with the lens shutter already open and the capping plate shielding the film from light. When the release is depressed the shutter must close, first, in order that the mirror and capping plate can retract out of the way of the film gate before the shutter opens and closes for the exposure. If the shutter is sticking badly enough, I wonder if the blades could still be in the process of closing after the mirror and capping plate have lifted? This would produce two separate exposures: the first, incorrect one after the capping plate has lifted; and the second, (normal) one, when the shutter opens and closes for the set exposure period (or some incorrect variation of that if it is not running freely).

The spring that controls the blades before they shut for the exposure is not as powerful as the shutter main spring, so, whilst I am hypothesising here, if the shutter was set to a high(ish) speed without the escapement slowing it down much, I can envisage that it *might* be possible for the blades to cap the film gate too slowly when the capping plate has lifted, but still open and close more or less correctly to make the exposure.

When the shutters of these cameras stick, they can be so bad that even though the body is wound and released the blades refuse to cycle fully because they're so badly glued together. On the other hand, they can also run OK at most speeds, but play up in the pallet range, or they can run slow across all the speeds. It depends on whether the mechanism is running freely and the blades are a bit sticky, the mechanism is not running freely at all speeds and the blades are sticky, or anything in between, according to how badly it all needs cleaning. The solution is cleaning the shutter, because the body gearing, mirror, capping plate etc are rarely troublesome, and, once the shutter is clean and lubed, the plot will usually come together and run well as designed.

You could try isolating the functions of the shutter closing and mirror and capping plate lifting from the exposure process, by setting the self timer, so that the exposure process is delayed. This will enable you to assess the initial phase of the mechanism by itself and, with the back off the camera, should any light penetrate through the len before the timer has run off and released the shutter, it may be visible. If so, it would be confirmation that I may be onto something above
Cheers,
Brett

 

[Ian Grant]

If I understand correctly what you are writing, one should observe a "dragged" image, like with ambient plus flash on first curtain. While one is observing a distinct second image. So, could we be seeing the effect of shutter bounce? But I have a feeling we are getting close. If only I had a fast photo-detector ready... I do have the digital scope. This would allow to actually observe the shutter bounce.
I might have to take a dive into a synchro-compur in the near future. With some help from: http://www.suaudeau.eu/memo/rep/Compur-shutter-repair-manual.html

The test film is exposed but not yet developed nor s*****d. Experimental data cannot hurt.

It takes very little dirt or grease to prevent the shutter blades functioning exactly as they should, and it's very hard to spot. It can't be an optical problem as there's that definite camera movement and double exposure, which is quite a lot weaker than the first.

About 3 or 4 years ago I bought a Kodak 170mm f7.7 Anastigmat (Dialyte) in a Wollensak Velosto (Optimo in the US) shutter, the shutter had some internal corrosion. It's a self cocking shutter and initially although I got it working wouldn't quite close consistently. It took a lot of cleaning before it ran smoothly. Just a tiny drag on the blades can prevent a shutter closing as it should.

Ian

 

[bernard_L]

Test shots -- As promised
Yesterday evening I developed the test shots, and this morning s*****d them. See attached pics. The subject was a 15 watt 220V bulb, powered with 120V. Located approx 3m from the camera. Exposure guesstimated to have room in the dark and filament above middle grey (anybody has a spotmeter that can measure an incandescent filament??); in that respect the results meet my expectations. Each series runs from 1/500 f:4 to 1/15 f:22, except 85mm hand held where I stopped at 1/30. Images shown are 240x240 crops @ 2700dpi.

Hand held, 50mm Tessar

Just like in the "natural" images that prompted my post, the ghost image seems to be to the left and slightly below the main image. The ghost becomes relatively weaker a longer exposure times; and conversely at 1/250, the ghost is brighter than the main image (or the offset direction has reversed). At 1/500 no ghost is visible

Hand held, 85mm Pro-Tessar

More or less the same observations as for the previous series. Again at 1/500, no ghost is apparent.
I'm surprised and gratified that, ghosts apart, the main image shows no motion blur at 1/30 hand held (85mm); didn't know i was that good

Tripod mounted, 85mm Pro-Tessar

Ghosts are still there (!) and with an offset globally of the same order of magnitude as in the previous series. The direction of the offset, almost horizontal, is also the same. And now we also see a ghost at 1/500. I didn't know my tripod was that bad

Bulb "flash" with open shutter

A test proposed by IanC in a so-called conversation. Original proposal: shutter in B. Open shutter, trigger flash, close shutter. Which I did, except instead of firing flash at reflective target, I turned on/off the tungsten bulb, same as previous tests. Approx 3s @f:22, more exposure than other test shots, but the dynamic range of the film copes with that. Possibly some optical ghosting of the lamp, but the filament shows none of the duplicate ghost seen previously (that would have been quite a puzzle, which is not a reason for not doing the test).

What to make of this experimental material?
The data might fit with what is proposed by Brett Rogers (thank you for your detailed write-up). Brett, do you agree?
A further confirmation would be to record the light in the focal plane during one cycle, and to confirm that an extra, spurious, pulse of light occurs before the main exposure. This would confirm Brett's explanation, and definitely kill the idea of shutter bounce. Setting up that test my take a few days, so anyone interested in the outcome should make sure he/she has thread watching turned on.

Again thanks to all for their attention and suggestions.

 

[bernard_L]

You could try isolating the functions of the shutter closing and mirror and capping plate lifting from the exposure process, by setting the self timer, so that the exposure process is delayed. This will enable you to assess the initial phase of the mechanism by itself and, with the back off the camera, should any light penetrate through the lens before the timer has run off and released the shutter, it may be visible. If so, it would be confirmation that I may be onto something above

I had overlooked this (sorry about that). I did the test. 1/125 f:2.8, self-timer, camera back removed, LED flashlight into lens, me looking from the back. Normal exposure (after self-timer delay) produces a very distinct flash of light, nothing at all is seen when shutter button is pressed and capping plate lifts, uncovering the rear of the lens. Repeated 5 times.
<<scratches his head>>

 

[Ian Grant]

Seems to indicate what I was saying, you'd expect the ghost exposure to be greater as the lens is opened up.

Ian

 

[ransel]

Bernard, I suggest changing your point of view...not physically...but mentally, and consider this fluke a benefit to your image making. Then, in the future when someone points out the ghost imaging you can respond that you like it that way (in other words, I am out of ideas).

 

[albada]

An earlier poster suggested that the film might be moving slightly. The rear cover-flap moves away from the film rapidly, pulling a slight vacuum on the film, so one can imagine that such tugging could move the film. A way to check this would be to maintain some clockwise force on the rewind knob with your fingers when firing (on a tripod). This will keep the film taut, which will probably prevent horizontal slippage.

Mark Overton