I'm continuing on from the basis of my prior post. This is gonna be about my views on capacity and minimum volumes of developer.
In color processing the manufacturers set spec limits for "process control charts" (see my prior post). A commercial processor will generally keep their "process" more or less centered within these limits. Now, I'm not gonna talk about WHY the limits are what they are, but rather what it takes to stay within these limits.
People who have worked with process control in color systems know, from experience, that when it is necessary to change a replenishment rate then 10% increments are generally good enough. If your control charts are getting near the lower limit then a 10% boost in developer replenishment rate will generally nudge them back near the center. If you're more critical, say with a C-41 film process, you may go to 5% increments, but there is not much need to go smaller than this. (Anyone with such experience, feel free to confirm or deny, as you wish. )
With something like a C-41 developer, its "activity," from a chemical concentration viewpoint only, relies largely on two things. First, the concentration of the developing agent, CD-4. And second, the concentration of the "restrainers," mainly bromide ion released as a byproduct of development. These two things are gonna track closely, but in opposite directions, with the "amount" of development that occurs. For each molecule of silver halide "developed," some developing agent becomes destroyed, and the halide (mainly bromide) is released. My point is that IT DOESN'T MATTER how many ROLLS are processed; it matters how many molecules of silver are developed. (If you "develop" unexposed film, it will NOT exhaust your developer.) Now, things are generally spec'd in terms of the number of rolls - this is based on an assumption of "average" exposure. But if the "average exposure" is off, then the developer "capacity" is likewise gonna be off.
Above, I specifically referred to C-41 because I know that the developer is sensitive to development byproducts. In contrast to this is the RA-4 paper developer. I also worked with its predecessor, commonly called EP-2 (or 3). EP-2/3 WAS very sensitive to the byproducts, and process control was very finicky. So the introduction of RA-4 was almost like a minor miracle; by comparison it was largely insensitive to replenishment rate. If someone forgot to start up their replenishment it almost didn't make much difference. Why? Well, RA-4 is a silver-chloride based paper, and chloride ion barely restrains that developing agent, CD-3. (The prior system used silver- bromide, and bromide ion strongly restrains CD-3 developing agent.) So when looking at a developer capacity one should consider how strongly the specific developing agent is affected by the development byproducts. As an example, consider the replenished XTOL systems, which are "self-replenished" and relatively immune to replenishment errors. I don't know much about the Xtol makeup, but this performance tells me that the developing agent is relatively immune to byproduct buildup. I could be wrong, it's just a guess. In the opposite direction is another example. I used to wonder why replenishment D76 had fallen so out of favor that Kodak stopped making the replenisher. (I never used it commercially, just observed that replenisher was no longer made.) But when I look up the data sheet I see that Kodak had special instructions for replenishment of D-76 when using the T-grain films. It looks to me like D-76 may be especially sensitive to T-grain development byproducts, which essentially affects the necessary replenishment (or capacity, I presume).
Anyway, back to the C-41 developer capacity. From what I said previously, about using 10% increments when changing replenishment rates, and that a 10% change would likely stay, more or less, within process specs, one can get an idea about the capacity. If you had a C-41 process running near center of a control chart, you could probably reduce the replenishment rate by about 10% and still run, a bit raggedly, near the lower control limit. Doing this would roughly increase the byproduct concentration by about 10%, as well as reduce the developing agent concentration by about 10%. So an in-spec process could roughly run under those conditions. And someone looking at the so-called capacity of the developer could make some estimates based on those ideas. If you can keep those two things within about 10% of aim specs, your development can likely stay within the spec range on a process control chart. If you go further off, say 20%, most likely you will be solidly out of spec.
Regarding single use of C-41 developer, a number of people have observed that Kodak gives a capacity of only about 4 "rolls" per liter. Which might be seen as 250 ml per roll. I made some rough calculations, based on mostly public information, about this. (Disclaimer: I did NOT carefully check these nor confirm with tests, so could be way off.) If one were to develop a single, average roll, in 250 ml of developer, the developer would start out chemically "correct," and end up about 10% off on both CD4 and bromide. Giving an average error, more or less, of about 5%, which ought to be comfortably inside of the process control spec limits. But... consider what happens with an overexposed roll, say between one and two f-stops. This seems to roughly double the development "demand," starting with CD4 and bromide on spec, and ending up about 20% out of spec, for an average of 10% off. If the average is representative of real processing, it's probably on the lower edge of in-spec process control. So it looks like Kodak's capacity is more or less ok for commercial processing. Probably, I think.
A lot of people wanna stretch the use, say to 8 rolls per liter. If they do this in 8 individual batches, 125 ml each, here's what I figure. Your CD4 and bromide start out on spec, ending up about 20% off. So the average error is about 10%, likely running on the bottom edge of a process control chart. But if the film is overexposed most likely it will fall out of process spec limits.
Varying off the topic a bit more, some people are probably still wondering about that idea of replenished systems, and why would anyone wanna go through those headaches of using a densitometer, etc? Well, if you start out with a little developer, say a liter or two to process with, and you do your development directly in that tank, here's what happens. If you use Kodak LORR replenisher you add about 26 ml for each roll, and the control chart stays on the money. In other words each liter of LORR replenisher is good for about 38 rolls. That's about 10 times more capacity than the Kodak 4 roll per liter spec. And... every roll is on spec. Got a roll 3 stops overexposed? No problem, the tank size of a liter (or more) will easily swamp out the effect of a single roll. A bunch of overexposed rolls? You'll have to increase the replenishment rate. Anyway, this is why commercial labs work that way. You just have to process enough film to be worth it.
Anyway, that's my take on things. Different developers, different situations.