(Editor's Note: the following article was published in the April edition of the Darkroom Underground.)
To make a short story long (after all that's what I do):
Several years ago, my oldest daughter expressed curiosity about old fashioned film. I dug out my Olympus OM-1 and we shot a few rolls. She got a new smart phone and that was the end of her interest in film. (There's still hope for my youngest daughter, she's the artist in the family.)
However, my old passion was rekindled and I was soon playing with all the cool gear that I couldn't afford back when I was shooting for my high school year book.
Before I realized it, I had bought and sold a Century Graphic and moved up to a 4x5 Crown Graphic. That's when I discovered that the only equipment available to process sheet film was, shall I say, non-optimal. Most of the tanks had been designed in the 1930s/40s; the newer products were either out of production or required copious amounts of chemistry and three hands to load.
A cursory marketing survey suggested that I wasn't the only photographer who was looking for a better method for processing 4x5 film. With a background in Product Development, it was inevitable that I would attempt to remedy the situation.
Let's introduce the first axiom of product development: You don't know what you don't know.
While this statement can be interpreted several different ways, in the product design/development world, it is a reminder that you probably don't even know what questions need to asked, much less what the answers are. These yet-to-be-discovered surprises will be randomly divided to technical issues, marketing challenges and feature optimization. The only way to unravel these mysteries? Just get started, you'll figure out soon enough just how much you don't really know.
So we started experimenting. We tried "taco" style adapters, tubes, light-proof trays and several inversion based designs before finally settling on an upright tank with a "plunger" for agitation. It was a simple design and eliminated the need for a water-tight seal.
The plunger design met all of our preliminary design goals:
- It required less than 500 ml of solution.
- The NRE cost estimates were reasonable (Non-Recurring Engineering: design effort, mold costs etc.)
- Easily made light-tight. (Note that there's a big difference between being water-tight and being light-tight.)
- It filled and drained very quickly.
- Minimum of four sheets of film.
- The film holders were easy to load.
- Estimated production cycle times and costs where within budget.
Test results looked good. (Bear in mind, almost all early testing was done using E6 color chemistry, since that's what I was shooting. E6 development times and agitation cycles are pretty much pre-determined.)
With a working prototype on the bench, we launched a Kickstarter. Now most people assume that the main function of a crowd funding campaign is, well, funding. The truth is a bit more subtle. An equally important side effect, (maybe even more important), is market validation. After all, there are lots of reasons people may act enthused when you tell them about your latest brainstorm. Maybe they just don't want to hurt your feelings or really weren't listening and want you to leave them alone. Who knows? However, ask them for their credit card number and you'll learn just how enthused they really are!
Back to our project, we were surprised, thrilled but still surprised, when the project funded in only three days! By the time the campaign closed, we had almost a thousand people waiting for film tanks.
Now that the project was real, we ramped up our testing. After evaluating numerous combinations of black and white film/developer, we discovered that the "plunger" agitation worked best with specific patterns of plunger speed and cycle time. In fact, "pumping" it too fast would sometimes create "swirls" in the highlights on the film.
Meanwhile, people were posting comments in the various forums. Most comments fell into one of the following categories:
- Love it! Anything new for large format is awesome! Sign me up.
- Interesting but I want inversion agitation.
- If trays were good enough for Ansel, they're good enough for me...
I don't have firm statistics (other than that #3 was about 2%), but it became obvious quickly that the market really wanted an inversion based system. Now there are only two reasons not to give your customers what they want, (assuming it's technologically possible): a) you're too arrogant; b) you're just plain stupid.
Remember the axiom: you don't know what you don't know? Now we knew. It was time to head back to the proverbial drawing board, or more literally, the 3D CAD program, and revisit the inversion agitation concept.
Bear in mind that product development is generally an iterative process: refine what works; redesign what doesn't. We liked the film holders and the baffle system of the original design. We just needed to add a water-tight lid. How hard could it be to add an O-ring?
Actually, the first surprise we encountered with the inversion prototype was the fill/drain times. The unit "gurgled" badly, filled slowly and "burped" when draining. (The plunger design lid didn't need to be water-tight and thus allowed air to vent as the tank filled.)
Should be simple enough, just add a vent. Of course that needs to be light-tight as well. The obvious answer was to add duplicate baffles on the opposite end of the tank. This would offer the advantage of filling/draining the tank from either end at the cost of the additional parts.
Frankly, the cost of another set of baffles was minimal, the real reason we abandoned the idea was the volume of liquid required. Simulation showed that the symmetrical design would require over 700 ml of solution, well past our design goal of 500 ml.
Time to introduce our second product development axiom: Every design is an optimization of compromises.
We had to balance the fill/drain time against the surface area of the baffles. The larger the baffles, the faster we could fill/drain the tank, but the more liquid required. Of course, the increased volume of liquid would take longer to drain.
Another compromise needing resolution: testing also showed that increasing the distance between the baffle plates would increase liquid flow. Unfortunately, that required smaller holes in the baffles to keep everything light-tight, which in turn slowed the transfer down.
Further optimization is always possible, and designers have to know when to call it "good enough". After all, does it really matter if the tank fills in ten seconds instead of eight? After a great deal of empirical testing, we had reached a point of diminishing return and we froze the design.
Back to the vent. Obviously, the vent didn't need to be the same size as the fill/drain port. It just has to let air in and out. So we could use a smaller, and therefore cheaper, cap for the vent. So why did we make both outlets the same size?
Here's our third, and frankly most important, axiom: Every engineering decision comes down to economics (or politics..)
We made both ports the same size. It was actually more cost effective to procure and inventory just one cap rather than two different sizes.
Another feature that fell victim to the third axiom: a bottom drain. We seriously considered adding a drain at one end toward the bottom of the tank. But this greatly complicated the mold design which increased both the NRE and the production costs. So we dropped the idea. (After our most recent wash tests, we're glad we did: FAQ 1.)
Not all surprises are bad. The "squeeze play" technique not only helps keep the tank water-tight, it also exerts enough pressure (3.5 - 4.0 lbs) on the lid that we didn't need to incorporate a latch. This reduced cost and improved usability.
I won't bore you with all the other trials and tribulations of the project but would like to list some of the lessons learned:
You can test too much. Kickstarter backers might remember an update entitled: Someday I'll laugh about all this... We spent several weeks chasing light leaks in the final prototype. (Turned out the clear epoxy used to hold the prototype together was leaking light during the prolonged test.) We're still not laughing.
Curse those *%($&#@ caps. We had a minor miscommunication with our warehouse in Hong Kong. The shipping envelopes used for the Kickstarter rewards weren't as padded as we expected. It wouldn't have been a big deal except for the caps. About three dozen caps cracked during shipment. We still don't really know why, we've bounced them off of concrete and not broken any. We suspect that they may have been screwed on too tight and then cracked when bumped.
Film holders, Rev 3. We never expected the film holders to be such a challenge. Enough said, read our blog posts for details.
That's a relatively short history of the SP-445 and a general introduction to the joys of product development. We're still working on other ideas and hopefully will have another product development story to share soon.