Our Recent Silver Bromide Gelatin Emulsion Workshop

[rwhb12]

Photo of all the finished glass plates on GEH course



Russ

 

[lunarcaustic]

The computer code that provided the most complete information of how the process worked is titled “COMPSELECT”. From this reference, the RIT papers, and other patents, I was able to cobble together the parameters for controlling the Crystal Habit from CubicCubo-octahedralOctahedralTabular. The targets were given for the initial potassium bromide concentration, temperature, along with the minimum and maximum bromide concentration breakpoints for each crystal habit zone. The formulas for the Nernst and Debye Huckle equations were embedded as well. The code also contained two-point calibration checks for the silver billet electrode. From all of this I was able to extract enough information to gain an understanding for controlling the silver ion concentration during the precipitation by modulating the salt and silver jet flow rates under PID control and ramp segments. The evolution of this process from the art of making silver halide emulsions to the science of making them is interesting.

George

 

[Photo Engineer]

George, much of the details and references are also in my book, as you know.

COMPSELECT was written in QNX and so would have to have a bit of translation. It was replaced by MERLIN. Also, there were programs for the Taylor, Siemens and WDPF computers.

There are patents on the Debye Huckle and Nernst equations. Lots of detail.

BTW, click on the title of a workshop in the GEH list and you get an expanded description of the course.

PE

 

[Photo Engineer]

George, I can find no on-line reference to COMPSELECT.

PE

 

[lunarcaustic]

COMPSELECT

Ron,

This program is written in BASIC. It contains routines for selecting the Halide type: Br, I, and Cl and allows the user to enter a make with up to 15 segments in the flow-profile with holds.

It has ability to pre-heat the z-batch and utilizes both shallow and deep temperature sensors. It also has routines for storage and retrieval of the various “makes” should you want to repeat a particular flow profile. There are also subroutines for VAg calibration whereby the silver billet, triple junction reference and temperature probes are immersed in standard solutions and checked for proper operation.

There are detailed error routines for detecting any drift along with suggested possible causes such as noisy signals caused by faulty connections, plugged junctions, leakage rate of the salt bridge solution and poorly platted silver electrodes. It offers corrective actions to replate the silver billet electrode, and replace the lower bridge junction. This combined with the RIT papers and your many posts on VAg have given me a clear understanding of the topic. All-in-all, it lays out the details necessary to construct a working system migrated to todays programming languages and single-board microcontrollers.

On a related thread you mention the need for higher purity silver for the electrodes and importance of the uniformity of plating. I purchased two lengths of high purity silver rod, diam. 7.0 mm, 99.99%, Sigma-Aldrich item number 26,562-4 (10 grams = 25 mm or 40 grams = 100 mm) for fabricating the silver Billet electrodes. I am planning to round-off and polish the end of the rod to achieve a "bullet" geometry and mount it in the end of a Teflon rod prior to plating. I hoping this smooth-surface geometry will improve the uniformity of plating over wire-type electrodes.

George

 

[Photo Engineer]

The program you have is not COMPSELECT then. It sounds like a blend of portions of the COMPSELECT ported to BASIC and blended with KEDS (Kodak Emulsion Design System). KEDS was (is?) far more comprehensive in capabilit as it produces the code to run the process control equipment directly along with the model and error messages.

All of this was to be replaces by ESP (Emulsion Scaling Program).

I build my electrodes from Silver "wire" as shown in the book, and plate them as described. But using them in critical situations requires a bridge of neutral salts to prevent Cl contamination of the make. The double junction electrode "leaks" Cl enough to disturb measurements. The book shows examples of a bridge.

There is so much more to this that you are missing which I left out of the book as I did not want to get overly complex. You see, even what you have is a dumbded down version of what we used which was capable (in the end) of handling up to 60 segments, but generally handled 25. That was an average make, not 15.

PE

 

[lunarcaustic]

Here are the two references I found online while searching the topic of VAg. These are the programs I am referring to. Although they are old as you indicate, they are a step above pushing a syringe plunger with ones thumb.

The first is a patent document that contains the COMPSELECT program in appendix A.

The other is the RIT paper which also has a Program listing in the Appendix.

http://www.google.com/patents/EP0356342B1

https://ritdml.rit.edu/bitstream/handle/1850/11354/NNatanThesis09-1987.pdf?sequence=1

 

[Photo Engineer]

Thanks. The problem is that no interpreted BASIC program is fast enough for this type of work, and a compiled BASIC program is just on the edge of being fast enough. What was needed was a C++ DLL that could be run in the background to do the work.

See my PM.

PE

 

[kb3lms]

The Natan thesis is very interesting and useful. I've built the setup shown on page 13 and it works quite well. One of my plans this winter was to actually try to follow the data for a controlled make. Haven't gotten that far yet, though.

Thank you for the additional patent reference. I don't think this one is in my library.

For all practical purposes, BASIC is no longer interpreted - although microsoft has taken a step back in that direction. Knowing the effort that Eastman Kodak dumped into the development of the C++ language (along with other companies) I am glad to see they put it to practical use. QNX was quite popular in those days. In many ways the mid-90's were the end of the golden age of computer science.

-- Jason

 

[Photo Engineer]

There are several more patents referenced in my book on vAg measurements. Most notably are those by Judd and Lin.

I was charged with developing a method of process control for the PC and had ordered the Burr Brown board for the core controller. At that time, a new hire came in named Yun Chang. I was asked to turn all of my hardware and notes over to him and in the end, he and several others did a wonderful job of turning it into reality. I doubt if I could have done as good a job. You see their names on the patents and some appear on the article by Natan.

However, it should be noted that this work is old. When it appears, it is nearly 10 years out of date. The "current" status of COMPSELECT, KEDS, ESP and other internal programs is far beyond what is shown.

PE