Some surprising (and good) results:

1. The Trial-130119 developer (described 3 postings back) will also accept Dimezone S without crystallizing. At least it's now been a week with no evidence of crystallization, and I'll truly believe it when it's gone one month. D316 quickly crystallizes with Dimezone S. I used 0.1 g of Dimezone S in place of 0.05 g of Phenidone. Based on resulting dev-times, I should have used 0.096 g of Dimezone S to keep dev-times the same as Phenidone. With 0.1 g, dev-times must be reduced by 4%.

2. The Dimezone S version produces finer grain and lower fog. I only tested Delta 400 because I've found it's unusually sensitive to developer composition. Here are scans of strips developed in XTOL, D316, Trial-130119/Phenidone and Trial-130119/Dimezone. The contrast of these scans was boosted digitally to exaggerate differences in grain:

XTOL:130201-D400-F1320-Xtol-Ctr13R.jpg D316:130201-D400-F1332-D316-Ctr13R.jpg T119/Phen:130201-D400-F1306-T119Phen-Ctr13R.jpg T119/Dime:130201-D400-F1318-T123Dime-Ctr13R.jpg

To me, the T119/Dime is noticeably better than the others. Clue: Once you've viewed the images, you can quickly switch between them, and the instant image-changes make grain-comparison easy.
The curves of this brew versus XTOL show a good match with the usual higher shoulder:

130201-D400-T119Dimez-Xtol.jpg

I was surprised that it did not crystallize with Dimezone S. I suspect the lower metaborate/ascorbic ratio is the reason. It obviously is not related to concentration-ratio because Trial-130119 is more concentrated than D316.

Fog was another surprise. Simply switching from Phenidone to Dimezone S dropped fog by 0.1 log-units (1/3 stop). That's a lot. Delta 400 has higher fog than most films, and metaborate/glycol/Phenidone worsen it. But merely switching to Dimezone S dropped the fog down to XTOL's level (0.38). I like Dimezone S. PE says it stores better, and I find it also gives better image-quality.

I think my next step will be to mix new concentrate using 0.096 g of Dimezone S and verify that image-quality is unchanged with the longer dev-times.

Mark Overton