A few weeks ago, I had a nice concentrate that matched XTOL in all respects -- density-curves, grain and even dev-times. But that work was halted when I discovered that the concentrate crystallizes and separates. This occurred with DimezoneS but not Phenidone, so I was back to using Phenidone. In that formula, using Phenidone was slightly grainier than DimezoneS.

After more experimenting, I've come up with the formula below. Its grain is slightly better than XTOL, sharpness is the same, and dev-times are longer. The 1-litre formula for mixing directly into water:

Sodium sulfite ................. 45 g
Sodium metaborate ........ 2 g
Ascorbic acid .................. 4.5 g
Phenidone ...................... 0.05 g
Target pH = 8.05 to 8.10.
13:40 minutes for TMY2.

For a concentrate, mix everything except sulfite into 16 ml of propylene glycol.
Here's the density-curve compared with XTOL, which is a perfect match:


Grain looks a tad better than XTOL in my 22x loupes. Here are full-resolution crops of neg-scans of XTOL (left) and concentrate (right):

10-00-XTOL1-678crop.jpg - 10-2A-678crop.jpg

An objective way to gauge grain is by the sizes of JPEG files, keeping dimensions and contrast identical. Finer grain (textures) produce smaller files. In the crops above, the concentrate's files are 2.25% smaller.

The sharpness of the high contrast horizontal line looks the same to me, but the low contrast vertical lines look a bit softer for the concentrate. Shutterspeed was 1/8 second (tripod), but there still might have been a bit of horizontal vibration. I'll have to watch future test-strips.

I discovered this formula back in February and posted it to apug on Feb 5, 2012 here. But I was investigating the effect of sulfite, and didn't pursue this discovery.
There's still much experimenting I want to do, varying ingredients by small amounts and gauging the effects. But I wanted to post this because it's a good milestone and will be my baseline for further work.

Mark Overton