Horizon 202 - banding problems

[Keith Davison <keith.davison@xxxxxxxxx>]

I have just obtained an Horizon 202 and am experiencing 
problems with horizontal banding.

Firstly, let me emphasize that these horizontal bands run
parallel the to the long axis of the frame, they are NOT the vertical
flare bands that appear across the short axis of the the frame
when bright light sources such as the sun are in the field.

I mainly use Royal Gold 25 neg film, but had occasion to put two
rolls of 200ASA colour neg through the Horizon, both on bright,
sunny days.  Both of these films showed banding in smooth toned
areas such as sky, on some frames.  More importantly, the pattern
of the banding striations was identical on the two films, despite
their being different brands and being processed at different labs.
Banding seemed to be virtually absent on my RG25 negs.

I decided to investigate, and ran a series of tests with 200ASA 
colour neg film, I shot light overcast sky in the evening from a 
shadowed location, allowing the gradual dimming of nightfall to 
allow me to shoot a range of f-stops on the 1/250, 1/125 and 1/60 
speeds.  I then contact printed the negs on b&w paper and scored 
them for banding before checking the frame numbers against my 
taking-notes (to eliminate observer bias), the results were very 
consistent.

Banding is most severe with the 1/250 speed, but for any given 
speed it increases with f-number. The worst case in the test was
1/250 at f11, but banding is still faintly visible with 1/250 at f2.8.  
Banding is almost invisible below f8 at 1/60, (which explains
why I had not noticed it with my RG25 negs!).
I then examined the bands using colour scans of the worst frames.  
There is no chromatic shift associated with the banding, only a 
change in density.
Converting the scans to greyscale, I used the Optimas image 
analysis package to extract an averaged luminance profile across 
the bands. 
The bands show both density increase and decrease relative to 
local unbanded areas, and are of non-regular widths and 
separations, but these are constant within the pattern from film to 
film.

Examining the taking slit at the back of the drum, I notice that the
non-adjustable side of the slit seems to effectively be the edge of
the outer drum, in my case, anodised brown.  The edge of this drum
is very poorly machined, being quite rough.  My guess is that the
striations are caused by this rough edge, this would explain why the
banding is worse at 1/250.  With a narrower slit, the variation in
width caused by the "hills and valleys" of the rough edge will be
proportionately greater, and hence cause more pronounced 
banding.  Also as the lens is stopped down, the image forming light 
cones become narrower and will project a sharper impression of 
the edge back to the film, also intensifying the banding.

Has anyone else experienced this problem?  Bear in mind it will
only be really obtrusive at f8 and above with 1/250 or 1/125, and
that the banding (although still present), is masked to the human
eye by image content, being very apparent only on large smooth
toned areas, such as sky.

I will be returning my 202, and will test the replacement when I 
receive it, but I am worried that this may be a common problem.
Given that the only control over close focus with the 202 is
depth of field, I do not wish to have this further restricted
when hand holding with fast film.

Does anyone know if it is feasible to strip the drum out of the
camera?  If this were possible, the rough edge could then be ground
perfectly flat to eliminate this problem, I would rather live with
the resulting slight overall exposure increase, than with the f-stop
limitations imposed by the banding.

As a last thought, there have been several reports of 202s with an
extra 1\500 speed, this is presumably achieved by an even narrower
slit than usual, it would follow that these cameras would be even
more susceptible to the banding problem at the 1\500 speed. 
Perhaps these 202s have improved machining on the slit edge to 
allow the extra speed?  Anyone willing to check?


Cheers,
Keith Davison