Ever since getting the Arca Swiss last winter I’ve wanted to make a bail back for my shop built 5×7. Unfortunately the design of that camera back doesn’t lend itself very well to adding a bail. The torsion spring housing is on the tall side, so the pivot for the bail would need to be pretty high to bridge that and would increase the depth of the camera more than I’d like. Also I think a leaf spring arrangement would work much better for a bail, which is what the Arca and most other cameras use. Rather that modify the original back and risk screwing everything up, I should probably make a new back. This camera is starting to feel like the ship of Theseus.
The Arca Swiss bail back is unusual because the spring load on the ground glass frame is at the bail’s pivot at the center of the ground glass frame. The leaf springs themselves attach to mandrels installed on rails on the back itself, not to the ground glass frame. The resulting clamping force seems barely adequate for the Arca 5×7, so I think I’ll attach the springs to the corners of the frame itself, and fix the leaf spring to the back in the center of the rails. This should apply more clamping force on the film holder and also reduce the stress on a wooden frame when the spring is under load.
Peruvian walnut, same as the original build. This was a piece I’d avoided using on the original build because it was bowed, so had to spend a lot of time with a hand plane to true up a surface before sending it through the thickness planer. It’s an odd wood. Freshly milled it has a pretty sickly appearance, almost looks spalted, but it eventually oxidizes to a more even brown. Lacquer warms it up quite a bit as well. But it’s light weight, stable, and fairly durable.
For the back platform I usually build up 1/8″ thick strips of wood in layers, sort of like a log-cabin. It’s a simple matter to overlap the corners to form strong joints just by cutting the pieces to staggered lengths. It also spares me from puzzling over how to cut the corner joinery around the film holder recess rabbets- you need only reduce the width of the top strips of wood to provide the recess. There’s another rabbet cut around the perimeter on the bottom that nests inside the rear standard, but that doesn’t need to be cut until the back is fitted to the rear standard.
This time around I built it with solid pieces of wood 7/16″ thick, with rabbets and lap joints milled into the parts. I forgot to take a picture of the parts before glue up, but it’s just a half lap joint, with the rabbets cut away on the styles where they intersect with the rabbet on the rails. The front rail, where the rib-lock is located, is a little different since it’s open for inserting the film holder. It’s the thickness of the half lap joint itself, and once everything else was assembled I glued in the parts for the rib-lock and light trap. It was pretty fussy fitting the thin leading strip of the riblock slot. It would probably be fine to just omit this thin strip, but once it was milled down to the correct size it was easy enough to glue it on with some spring binder clips. In hindsight I should have made the front rail piece the thickness of the lap joint plus the film holder rabbet, and just cut the rib-lock slot shoulder to shoulder with 1/8″ router bit before gluing everything together.
Overall though I like this method better. Laminations are tricky to glue up because the parts tend to slide around quite a lot during glue up unless you use a lot of clamping jigs. Also, due to the amount of glue used in the laminating process, the back can be prone to warping unless you leave it clamped up and dead flat for days until the wood recovers it’s original moisture level. This assembly was just clamped for a few hours, then left under weight on a flat slab overnight.
The ground glass frame was made from 7/16″ stock. The T-dimension rabbet for the ground glass is cut before the joinery on the corners. I checked the T-dimension on several Fidelity “Deluxe” holders, but everywhere I measured I got a slightly different reading which surprised me a little, so just took the dimension from the old back and checked that against the Arca Swiss before finally cutting the ground glass rabbet.
The corner joinery is mortise and tenon with asymmetrical shoulders. The tenon isn’t centered in the stock due to the depth of the T-dimension rabbet, so I cut the mortises on the mill using a 1/8″ downcut spiral router bit. I had to offset the bit 1/16″ away from the rabbet to get enough depth in the tenon without damaging the rabbet- 1/8″ bits can only cut about 1/2″ deep due to the flare of the bit where it meets the shank. Due to this offset I had to cut a stepped shoulder on one side of the tenon. The X/Y table on the mill made it a little easier to move the fence to re-locate the router bit after flipping the work pieces. The tenons were cut on the table saw with a crosscut sled and it was a little fussy setting all the depths and widths.
All of this would have been much easier if just I’d planned the thickness of the stock with the T-dimension in mind so the joints would be dead center in the stock. I always tell myself to sketch these parts up in CAD before milling stock but never do. Anyhow after the frame was glued up and trimmed to final size, I cut the rabbet around three sides on the bottom of the frame to nest into the back recess.
I’m using stainless steel feeler gauge stock for the leaf springs, based on a recommendation on a forum. Not sure if this material has a true spring temper but it rebounds nicely, certainly a lot better than some NOS ‘Deardorff’ 5×7 camera springs that actually creased under load. I think those were nickle plated cold rolled brass. Ebay.
The ends of the springs need to be formed for attachment to the ground glass frame. It’s not an easy bend to make in tempered stainless steel so I made a simple jig. Two 1/8″x 2.5″ steel roll pins are inserted side by side into the base- an scrap of old counter top- the Formica laminate helps to resist compression when the pins are under load. The roll pins are spaced the thickness of the spring. For reinforcement the pins are gusseted on top by a scrap of aluminum. The pivot pin is left long and fitted topside with a piece of 3/16 steel tubing which fits over the roll pin and into a flanged bronze bushing. The roll pin next to the pivot pin is cut flush with the top of the aluminum gusset, which is then attached to the base through a spacer (plywood scrap) to prevent the pins from twisting under load. For the bending arm, a short piece of steel bar is drilled for the pivot bushing, and is attached to a plywood fence beneath. Admittedly the jig materials and design are pretty underwhelming, but it was good enough for just 4 bends.
Worth noting that with this type of bending jig you can’t fully loop the stock without having to disassemble the jig to remove it, so the fence only needed to swing about 200°. I thought about making the jig so the pins were open on the top for easy stock removal after fully closing the loop, but the forces of the bend are probably too much for 1/8″ roll pins only fixed at the bottom, no matter how deeply the pins are set into the base of the jig. Anyway, each spring only needs to be looped on one end. The front of the springs are only bent 180° in a longer U shape, so the spring are really only attached at the back. In use the mandrels up front need to slide back and forth as the springs bend and the distance between the loops shortens. But it’s good to have the rear loops formed as tight to the mandrel as possible so the ground glass frame doesn’t wander and loose registration to the back when the bail is used. I’ll close off those loops using a vise and some mandrel stock as a form.
I decided to install a shoulder around the perimeter of the ground glass frame. This isn’t really needed for light blocking, or to help guide the frame into the back’s recess, but they may help to stiffen the back platform along the load axis of the springs. Also the back looked sort of plain without it. The springs will be installed on the outside of the shoulder. The strips are only 3/16″ to keep the distance the spring mandrels have to span relatively short. I inlaid the longer side strips into a dado cut in the back for strength and to make gluing them in easier. The rear shoulder strip has a much wider glue surface and is flush with the back edge, so that didn’t need a dado.
Next I need to make the mandrels and mounting hardware. After that’s done I’ll start on the bail itself.
Not much to show for the weekend. Sprayed some lacquer. Also made the bail pivot, spring mounting hardware, and a few mistakes as well.
I decided to use 1/4″ stainless steel stock for these parts. For the spring connections I’m going to use roll pins that are press fit into the spring blocks (photo above), and aluminum would fatigue pretty fast under load. The center pivot blocks (photo below) will have a pivot pin perpendicular to the block, so wanted that as strong as possible as well. My first attempts at making identical parts. Nice to work with something other than aluminum for a change, but these tiny parts are surprisingly heavy.
I routed out the ground glass frame to accept the hardware, but it’s starting to look like I located the pin blocks too close together. I wanted the spring tension directly on the rib lock in front, and just flipped the frame to make mirror mortises in back with the same setup on the router table. I really should have moved the rearward pin blocks farther back so they’re symmetrical to the back itself, not the ground glass frame. Oops. I guess I can always saw away the recesses and glue some new wood strips on the side. I could use a continuous piece of metal on each side for the hardware, but that would be a lot of additional weight.
Anyway, nice to see some finish on the walnut, it looks a fair bit better already. I got an odd batch of semi gloss Deft lacquer, the stuff wrinkled beyond belief. I thought I hadn’t let it cure enough before re-coating, but no matter what I tried the second coats always always wrinkled right away. After scraping to bare wood for the fourth time, I went to town and got a can of the Deft gloss, which went on much better. I don’t care for high gloss finishes so buffed it down to even luster. Everything surface inside the shoulder strip on the back will be painted matte black, including the underside of the ground glass frame and the underside of the back itself. I haven’t decided yet if I will flock the film holder seat.
Finished the loop on the leaf springs. Using a set up block to keep the spring square to the vise, I inserted the roll pin into the loop and snugged it down in the vise, using the V groove in the jaw to locate the pin and close the loop over it. This is plain steel roll pin in the photo, have some stainless steel pins on order from McMaster-Carr that will be used on the back.
The mere act of finishing the wood parts in this project is getting a bit conspicuous. The project has all the elements of a full camera build- woodwork, metalwork, and finishing with mistakes scattered throughout – plus I forgot that I have to match the color of the existing camera.
I re-located the spring blocks and blended the wood repairs nicely and was feeling pretty good about that until I set the freshly lacquered new back on the camera this morning and saw the glaring difference in wood tone. I finally remembered that I used a tinted grain filler in pores of the walnut after the sanding sealer for a smoother final finish. As a result, the original camera is quite a bit more mahogany-toned than the new back.
Rather than strip to bare wood again and fill the pores, I got some ‘brushing’ lacquer in town. This sprays pretty well if thinned a bit, I’ve used it before for small projects. I have a Fuji HVLP rig, but it seemed like overkill for the very small parts in this project. If I’d actually remembered how nice this spray rig is to use I never would have bothered with the rattle cans. I still have the same dye that I used on the wood filler. The transparent reddish-brown dye (TransTint) is alcohol based, so it works great added directly to lacquer, shellac, and even water based finishes. This is the best way I’ve found to tweak the tone of wood without obscuring the grain. If the wood is sealed with a plain sanding sealer, then the tint can be added to the following coats and so it goes on very evenly without blotching. It’s an especially nice method for toning down the contrast between sapwood and heartwood as you can selectively spray lighter areas to blend them in.
Before and after tinting. The overspray on the paper masking represents how little dye is needed to dramatically change the wood tone.
Already I want to refinish the entire camera, so if I ever do: for this project I sprayed a few coats of plain lacquer, wet-sanded it level with 400 grit and paint thinner to smooth out the open grain, then hit it with some fine synthetic steel wool. For the first two tint coats I added 6 drops of tint to 10 oz of finish, roughly cut 5:1 lacquer to thinner. For the topcoats I used the same cut but without the tint, which will seal in the tint so a final polish (800 grit, wet-sanded with paint thinner) wont cut through the color. On the gun I used a 1.5mm tip and dialed back the air and spray pattern until I got a small even pattern with minimal overspray. The cure time was about two hours between the coats, which went on fairly heavy. Based on past experiences, light finishes turn a little gritty when cured.
At last, on to the bail itself. Actually there’s not much left to do since the design and parts have been decided. The bail is made of two pivot arms of 1/2 x 1/8″ aluminum joined together by handle of 3/16″ stainless steel tubing. The ID of the tubing makes it easy to thread 6/32 screws in to attach it to the aluminum. On the other ends of the bail arms, about 1 1/8″ away from the pivot, is a bearing on each arm that leverages the bail away from the back. It rolls atop the leaf spring, keeping pressure on the spring at the same time as the bail is lifting the springs, which keeps the spring load from stressing it’s screw connections to the back. A spring mounting screw stops the bearing travel a little past perpendicular in a hold-open position. The aluminum parts were ganged together and the holes for the pivot, the bearings, and the handle are drilled at once.
The problem with the bearing though is finding a place for the front spring mounting screw. I ended up countersinking that screw head on each side so the bearings could roll past. Putting these screws too far forward would have increased the spring tension considerably. The distance between bearing and pivot gives about 1 1/4″ of room when the back is open.
Finished. The bearings are press fit onto a 3/16″ sleeve that’s threaded to mount to the bail arm with a 6/32 screw. To attach the spring I used machine screws and nuts. The nuts are recessed into the other side of the back.
Open position. I thread-locked the pivot handle connections to keep the bail action from unscrewing the parts. A bearing stop keeps bail handle from slamming down on the opposite side. The stop is a 6/32 threaded spacer that fits over the leaf spring connection screw. The springs are a little heavy for a 5×7, but should be fine. There’s quite a lot of load on the corners of the ground glass frame, so I replaced the 1/8″ roll pins with 2/0 taper pins for the spring mandrel connections, which lock in place quite a bit more securely. I’m quite pleased with the open position, the opening is wider at front. I biased the spring attachment screws towards the rear mandrel blocks in the hope that this would happen.
So far I really like the action of the bail, but I’m dismayed by the weight of the new back. The old back was 19 oz. This one is about 27 oz, a half pound heavier. The wood construction is essentially the same, but there is a lot more metal on the new back. I may try to come up with a design for lightweight shackle blocks to replace the stainless steel, maybe something that mounts on top of the ground glass frame instead of the sides. The springs and bail handle are pretty light, not much weight to be saved there- even if I switched to a wire handle style or cut the springs in half by width I’d probably only save an ounce or two.
I suppose it’s not really a fair comparison, the torsion spring design of the old back is remarkably efficient, only a few small torsion spring coils and minimal aluminum stock were needed for that design. Maybe I’ll try to adapt a bail to the old back after all.
Testing for light leaks in the darkroom. Light fixture fits onto the lens board stage, and inserting a film holder should seal off all light. I keep the tripod fairly low to more easily look around the camera. Haven’t checked the bellows or lensboard stage in a while.
Usually I just toss 2 or 3 LED headlamps in the bellows and close off the ends of the cameras when testing for light leaks, but thought a dedicate fixture that radiates light from the center of the bellows would be much more appropriate. It’s just a cheap lamp socket kit attached to a blank lens board. I used a 1100 lumen bulb, LED but it still emits heat so kept the test to 10 minutes. After 10 minutes in the dark I’ll know if there’s a problem. The room must be totally dark- even faint reflections off the outside of the camera will look like light leaks, so there can’t be any light in the room at all. I saw some very faint bleed along the sides of the inserted film holder that was hard to distinguish from a retinal afterimage. Finally I just compared the new back to the old, and saw the same faint bleed. I never noticed light fogging from the old back, but I will route out a small groove for foam strips in the film holder seat just to be on the safe side. Unlike the old back, the seat on the new back has no velvet flocking, the interior surfaces are simply painted matte black.
I set up the mill to route a 3/16″ dado for the foam strip inlay. The neoprene foam I used is 1/8″ thick, so I set the depth for 3/32″, and clamped a block on the Z-column to serve as a depth stop to keep the depth exact around the perimeter. There’s a fence clamped to the x-y table, with a hold-down that clamps to that and keeps the back tight to the table surface. After the back is clamped to the table as well, the dado is cut using the X-axis crank. A lot of setup, but wanted to make sure the depth of the inlay was consistent.
Dado for foam strip inlay. After the picture I went ahead and cut a dado along the top as well on the other side of the riblock, just to keep light out when composing. The installed foam raises everything up about 1/32″. This doesn’t affect the T-dimension, since it raises the film holder and ground glass frame equally.
Jig for cutting 3/8″ neoprene strip down to 3/16″. Wood fence has 3/16″ rabbet milled into it to hold strip at the correct width. The aluminum bar keeps the strip tight to the fence. Neoprene foam (https://www.mcmaster.com/#8694k118/=18xiiep) isn’t idea for a friction surface, but having the bail means holders aren’t slid in and out, they are dropped in place. Still, I got the most wear-resistant foam I could find.
Installing the foam was surprisingly fussy. If you stretch it at all the dimensions will change quite a bit. Even simply removing the adhesive backing will make the length of the strips shrink. The adhesive is very aggressive as well. It took me a few tries to get it evenly installed around the perimeter with the same reveal. I did the light leak test again after installing the foam, and it sealed up nicely.
Fucking hell what a drawn out project, but finally ready to run some film through it now.
After all this work I’m already using the old back again. The new back does work, strickly speaking- it’s light proof, the T-dimension is correct, the bail itself functions as expected. But I made a lot of design errors. The new back adds awkwardness, uncertainty, and even bulk to very routine steps- exactly the opposite of what I set out to achieve. If I’d paid more attention to the details of the Arca-Swiss bail and kept the design simpler I could have avoided most of these errors. Well, except maybe weight- the Arca back is pretty heavy, about a quarter of the weight of that entire camera.
- The ground glass frame and shoulder that it nests into should have only been as wide as an actual film holder. Even with all the room provided by the bail, the film holder is tricky to seat correctly in the shallow back recess- it’s much too easy to skew it so it doesn’t seat properly. In testing it out I always felt like I should remove the back to make sure the film holder was seated properly, which totally defeats the point of having a bail. I made myself not check though, just to see how error-prone it was. While all 12 test shots turned out fine, the anxiety is still there whenever I think about using it again.
- The bail handle itself interferes with the dark cloth. I use a cloth with an elastic band that fits around the back, and it’s fussy to put on and remove the cloth without it getting snagged on both ends of the bail handle and stressing the rear standard.
- I’m not confident that the feeler gauge leaf springs will hold up over time. Not sure if they have a true spring temper, they seem to be loosing some rebound already. The torsion springs on the old back still feel great even after 4 years of heavy use.
- Too heavy. The old back is 8 ounces lighter that the new back.
I may try to modify the back to fix some of these issues, but it would probably be better to come up with a better design first and then rebuild it from scratch. The main issue is trimming down the ground glass frame to the size of a film holder and still having enough wood to securely mount the spring hardware. The GG frame on the old back is oversized too, but with that design you are only opening the front enough to insert the holder- the remaining spring tension keeps the holder tight in the shallow recess and guides it along until there’s a tactile snap when it’s fully seated. When that tension is fully released by the bail, it’s just too difficult to tell when the holder is in the recess. If the shoulder that locates the GG frame was only as wide as the holder itself, that problem would vanish. I could make the recess for the film holder much deeper for more positive registration, but that would weaken the back unless I really increased the thickness of the back in a rebuild, which would also increase the weight.
The bail handle is a pretty easy fix. I would put a dogleg in the pivot arms so the bearing is always touching the spring. This angle would also allow the handle to be flush with the back in the closed position. Without these gaps under the bearings and the handle the dark cloth would not snag on them, at least not nearly as much. Alternately, I could make the bearing a fixed point on the rails, a secondary pivot instead of a moving part, but that would shift the gg frame forward in operation quite a lot, causing more stress on the springs.
The feeler gauge springs can always be reinforced with another layer of shorter spring steel shim stock, so that’s not a big deal.
As for weight, I think aluminum would probably be fine for the pivot and pin connections. A .25″ thick bar running the length of each side of the ground glass frame would still only weigh half as much as the stainless steel parts. The feeler gauge springs aren’t heavy, neither is the aluminum bail handle- most of the extra weight is in those tiny stainless blocks.