Yutzy et al patented a few different implementations of acid coagulable gelatins and method of making emulsions using the same. His patents have some omission in the area of how the acid should be added. There is an art in the gelatin-to-silver ratio, temperature, stirring, and other secondary coagulants used in conjunction with this method. Depending on how you add these things, the emulsion coagulum may come out like big chunks or tiny granules. Of course, the latter is better for complete washing, but then the granules take longer time to settle after the first wash cycle. As you see, the desalting stage is a big chunk of proprietary know-how and not very fully described in patents, and this topic is also rarely discussed in academic literature either. But these things, if you have sufficient knowledge and experience in photographic chemistry, you can try them and fill in the missing parts by experimentation.
For Yutzy's acid coagulaion methods, see 2597915, 2614928, 2614929, 2718662 and compare.
In my experience, the coagulation method disclosed in Roth, US Patent 3118766 is superior. It coagulates more easily, enhances adhesion of the emulsion, and the yield of acylation reaction is more complete within the desirable pH range.
One big difference in terms of ease of using these techniques is that phthalic anhydride must be first dissolved in a suitable water miscible solvent like acetone, while trimellitic anhydride can be added to the water phase directly, as the latter is water soluble. Another big difference is that the pH range required for the acylation reaction. Trimellitic anhydride can react faster, even at a lower pH. During the acylation reaction, the pH slowly drops, and you can't use common buffering agents (despite the agents listed in the patents) because these are nucleophilic reactions and the buffer will likely interfere. So you will have to have the pH maintained in the usable window by addition of NaOH or KOH while monitoring the pH and temperature. A wider reaction pH window of trimellitic anhydride makes this process quite a bit easier.
On the problem side, these acylatin-in-emulsion methods are incompatible with ammonia ripened emulsions, because these acid anhydrides will react with the ammonia faster than gelatin. I personally do not use ammonia as the ripener (it smells very bad unless your emulsion making setup is in a fume hood---my lab is on 250cfm ventilation but I don't like the smell) and ammonia also increases fog level very much. One way to make them compatible is to use acylated (i.e., phthalated or trimellitiated) gelatin from the beginning. This technique is actually very useful for making tabular grain emulsions, but not conventional cubic or octahedral emulsions. Fog level, contrast and grain shapes are affected.
On the other hand, if you use phthalated/trimellitiated gelatin from the beginning, your nucleation pH is inevitably higher than pH of 5.5, to stay out of the range where gelatin starts to coagulate. This is problematic in making low fog emulsion of good contrast. In order to make good emulsion, I often have to adjust the pH and temperature after nucleation but before growth stage, and pH of 3 is definitely preferred for nucleation for most emulsions. Phthalated gelatin in presence of salt coagulates at pH 3 and is unusable.
Another problem, a real practical problem, is that these acylated gelatins do not harden with common organic hardeners AT ALL. If you use these hardeners, you must keep the proportion of the acylated gelatin low, and add a non-acylated gelatin after washing stage. This makes gelatin-to-silver ratio too high. Of course, there is work of Yudelson, US Pat 3017280, and more, but these agents are very nasty to make, and many are nasty to use.
I have tried all these methods and came to a conclusion: keep trimellitiated and phthalated gelatin for tabular grain purposes, and for subbing purposes.
For conventional emulsion work, I use more convenient polymer coagulation method. See De Pauw US Pat 3884701. This method is very much easier to use, and superior in many ways. The problem is that the polymer described in this patent is not commercially readily available, although it is not very difficult to make from styrene and styrene sulfonic acid, with AIBN or other starter, in a suitable organic solvent. Of course, this area is again a lot of proprietary know-how, since you have to adjust the molecular weight range by adjusting the temperature, concentration, and radical scavenger. Making the polymer in the form of a block copolymer is also useful, but not described in this patent (but any polymer scientist can think of this). Some of the polymer methods don't even require a pH adjustment. (That is, if your kettle pH is 5.5 or 6.5 or something in that range at the conclusion of growth stage, which is very typical, then you can just add the coagulating agent without even monitoring or adjusting the pH. This is really convenient.)