View Full Version : stepper motor / PC control of emulsion making
04-02-2007, 12:28 PM
I am a new member to this forum, and have been catching up on the technical information presented here as fast as I can.
First, I wanted to thank 'photo engineer' for posting as much information as he has so far to date. The importance of these contributions cannot be overstated.
From what I am reading on this forum, it seems like the process of emulsion making on a small scale would benefit tremendously from using small stepper motors, servos, solenoids, etc (machine control) to keep certain steps in this process more exact and repeatable. There are plenty of off the shelf components, subassemblies, and free software to take care of most steps (more precise injection of ingredients, timed sequences, mixing, heating, cooling, coater transport, paper or film movement, etc). And you wouldn't need a machine shop to put this all together. And, you would use a PC to control and monitor the process. If you were to perfect this, you could maybe modify the same small scale coating apparatus to coat more than one emulsion at the same time (ok, maybe not right away).
I am aware that the basic idea of making emulsion and coating is to do this by hand, and I know some (or most) of the idea of machine control is a bit much for some folks. I am just suggesting that the idea of using electromechanical control in very small scale emulsion making may not be difficult at all. There are companies that specialize in supplying smaller motion control components for the hobbyist (or small scale operator). There is also e**y as a source of parts. Once the system is created, it can be duplicated by who ever wishes to. If you have more precise control over certain steps in the process, you could tailor the results any way you like.
That is, of course, if the emulsion gods are kind to you that day, and assuming every other variable (ingredient quality, etc) is in your favor.
04-02-2007, 01:36 PM
Stepper motors with geared pumps (gravimetric) or peristaltic pumps (volumetric) can be used in making emulsions. Other types of slow speed motors can also be used. All you need is a control method. Burr Brown makes an excellent control board for the PC that does just this. All of these have certain advantages and disadvantages.
However, few people want to get into this complex of an operation as it also would imply feedback control of flows to control things precisely as it can fly out of control with even small errors. There are other problems inherent in this as well such as pump chatter to name one.
I could give a whole workshop on this subject alone, as it was one of my specialties for about 15 years.
My approach was KISS and I have adhered to it in all of my formulas. The total investment for making my emulsions would be well under $1000, but I suspect that your approach could run up as high as $5000 or more and it would take a lot of time. A single peristaltic pump (new) might run $250 alone.
I assure you that a single run emulsion such as I make can be done in as little as 1 hour including full prep time and cleanup, but an emulsion as you describe might take as much as a full day for setup (including programming the equipment), running and teardown - cleaning. BTDT.
Just debugging problems associated with line fill and line purge would create an interesting bit of research for you up front. After all, if you are adding 100 ml of chemical, you can't leave 10 ml in the line and some systems will not deliver without a 'head' of fluid behind them.
But, I don't want to discourage you from trying, and if you are interested in pursuing it, I am here to help as much as I can.
04-02-2007, 02:08 PM
I am very interested in investigating this. My plan is to attend your workshop, and learn all I can how the process is done manually, and what's involved. I know I will have many questions for you in the meantime.
04-02-2007, 05:45 PM
If a vacuum-driven approach is used, then fluid pumps are not required, right? All that is needed is flow measurement and regulation.
(Yes, I know this is absolutely not needed for home stuff, but the idea of cottage industry is one I enjoy...)
A small controller utilizing dynamic fuzzy logic should do a good job of regulating flows. Designing a system that continuously coats plates is a little more challenging. However, Eastman did it in the 1880s. Therefore, a system must be feasible. Whether it fits into my bathroom is a different question.
04-02-2007, 07:03 PM
Brian, I'm not quite sure what you mean by a vacuum driven approach for pumping fluids into a vessel for making emulsions.
Could you elaborate?
A continuous plate or film coater has its size mandated by rate of coating and speed of drying. In some ways it is easier to design a plate coater than a film coater and easier to design either than design an automated emulsion making setup. OTOH, coaters are often much larger than making apparatus due to the dryer, even the fully automated variety of making station.
04-02-2007, 07:16 PM
I suspect that the primary benefit would not be to be able to recreate the hand-process of your existing method, PE, but to gain enough control to do more sophisticated emulsions, fully pan films, etc.
I also suspect that the greatest benefit would be simple-homebrew-but-not-full-production sorts of ideas. e.g. a stepper motor into a leadscrew into a syringe instead of a proper peristaltic pump. (Assuming, of course, you can properly linearize the process). Sure you need to refill the syringe after each batch, but we're not talking about trying to compete with Ilford or Kodak here..
I suspect that it would be easy to assemble sufficent amounts of hardware to do the job and control it via computer...... but the software to do it properly... that's hard.
04-02-2007, 08:15 PM
The leadscrew into a syring is an excellent method of automation and probably better than any of the others, provided the proper syringe is used.
Having a pan film however, has nothing to do with automation, sorry to say. The pan sensitization takes place long after the making of the emulsion and is merely a dump and stir. It is coating the pan emulsion by hand that is he problem and requires an IR lab for fog prevention.
We used a syringe system at Kodak for small makes or small additions to large makes. That is essentially what I am duplicating but with manual operation which works well enough at up to 600 ml. Beyond that, syringes would still work, but then the problem becomes one of mixing and the type of delivery, and that is one of the essences of my workshop. Method of delivery is critical as is stirring. This is also scale dependant and is something I have not discussed here on APUG at all to this point.
With some emulsions, I have seen up to 4 or even more delivery systems running at one time with a very particular method of delivery to prevent 'hot spots' in chemistry. No matter how accurate pumped delivery is, if a hot spot develops, the emusion is scrap. Sometimes this even includes the use of geometrically placed inlets and baffles.
Delivery systems may be running Silver, Salt 1, Salt 2 and dopant, and in addition the flow rates of 1 or more may be ramping up or down (at different rates) and may start and shut down for specific periods of time. I know of formulas that include upward and downward ramps in concentration as well, and I have never mentioned this before on any forum.
In addition to those 4 flows mentioned above, base and acid may be run for specified times for kick starting ammonia digestion, temperature may ramp up and down in a specific fashion, and a UF ultrafiltrator may kick in and out periodically to maintain a given ionic strength.
So, in the final analysis, with the methodology described in this thread, you can move from the 40s and 50s to the 60s and 70s, but not much beyond the 80s, and certainly not into 90s technology.
And, one of the big features of control is the control circuitry with a measurment device and feedback loops (for both pH and pAg). You need an algorigthm and the right PID setting for each factor for the proper flows to prevent oscillation of the 1 - 3 salts being delivered to prevent undue banding in the emulsion and any acid or base.
Gee, I'll bet the Director of Research at Ilford is drooling right now.... :D
I encourage you interested parties to try, and I'll try to answer what questions I can when problems arise.
I think that one of the threads here already has a block diagram of a simple emulsion process. Just multiply delivery systems by about 3 and control systems by 2 and you are approaching modern 90s making.
04-02-2007, 09:02 PM
Okay, I think I forgot about the whole coating-pan-film part... :) And tuning PID is mostly what I was referring to with the software taking forever... :D
The problem, of course, is that if you can spend a few weeks hunched over the soldering iron to build the electronics and run a GPL'd software package and get an improvement in home-emulsion-screwing-around, that's one thing.... but auto-tuning-PID software is going to shoot the price up awful high.
I know one can source ready-made pH sensors with digital output... but what about pAg sensors? and are the ready-made pH sensors adequate for the purposes they'd be used for?
I guess the big question is... with levels of automation acquired through dedicated screwing around, could one potentially control the pAg levels well enough to do T-grained films or at least consistent cubic grains with no K-grain?
04-02-2007, 09:56 PM
No, there are no really good pAg sensors around but there are good pH sensors.
Yes, you can build - program a good PID loop and tuning software. It isn't that hard.
No, you don't need one if you really want to do this. There is a short cut.
Yes, you can make a pretty good t-grain emulsion but it may be a bit "Klunky" and you can probably make it pretty mondisperse and repeatable.
Yes, I can get you help on converting a unit that reads millivolts to a meter for pAg control, but it will take a lot of solder, and a lot of silver and electrode know how. Yes, I know how to contstruct a decent model from the Debye Huckle equations for this job.
But, in the long run you want to use vAg control for precision and accuracy.
I think what you really want is the ability to make something mid 60s to mid 70s which will be a decent t-grain and moderate process control.
This will take 2 pumps (peristaltic) and a batch of solenoid controlled valves, 4 addition systems and vAg detection with manual control. We can talk about addition methods later.
What do you want, a free book on-line? :D
04-02-2007, 11:20 PM
Hi, PE! I was thinking that instead of using a fluid pump to move chemicals, why not let them flow into a vacuum? Then there wouldn't be any pump agitation. Since the vacuum is pulling the chemicals into the mixing container, it would also remove air bubbles and such, right?
(BTW, if I can pay for this information with home-roasted coffee, please let me know. I am currently roasting Ethiopia Harar, but I also have Tanzania Songea, Zimbabwe AAA from Dandoni Estate, and Uganda Bugisu.)
04-03-2007, 04:50 AM
I think I'll chip in here, as I have spoken to Ron via Private messages in the past with regard to automation.
There are a few very low cost DAC/ADC boards that are connected via USB.
I don't know whether these would have sufficient ADC range for this application though.
One thought I had recently that may help with getting rid of pumps and standing lines of fluid after the make:
Have a reservoir which is filled with the correct quantity of solution for the make. This has a sealing lid, and an air feed into the top. Thus it may be pressurised with a small low pressure compressor.
From the bottom of that, take a feed into a Solenoid valve array.
The array would be comprised of 8 solenoid valves, each of which has a different sized constriction on the output (possibly a needle valve). The constrictions would run as follows: 1/256, 1/128, 1/64, 1/32, 1/16, 1/8, 1/4 and 1/2.
By switching the valves on and off in an appropriate fashion, one could obtain a variable flow rate
You would obviously need one of these systems for each solution to be added.
The first problem here is that a system with several banks of 8 valves quickly becomes rather heavy on I/O pins. This could be worked around by the addition of a PIC chip to drive each set of valves.
These (PICs) would all be connected to one serial line, and given individual ID's.
You could then start a given solution at a given rate as follows:
From the control PC, send 1, 255. This would tell the PIC with ID 1 to listen, while all the others would ignore the next command. The 255 would start that solution running at full speed (by turning on all the valves). To obtain half speed, you would send 1, 8. This would activate only the valve with the 1/2 constriction.
Any thoughts? Is this idea complete piffle?
With regard to the boards I mentioned earlier, I am intending to purchase at least one to experiment with in the near future.
Can anyone tell me how I would go about measuring pAg?
What kind of sensor do you need to use to measure pAg - can they be built relatively easily?
What kind of measurement would you make - voltage or resistance?
04-03-2007, 09:26 AM
Sorry, but I cannot envision how you would do this without very complex plumbing. But, maybe you can design something. Draw it up, scan it and post it.
My taste in coffee is mainly instant. I'm really a tea person when it comes to discriminating choices. Thanks though.
04-03-2007, 03:50 PM
Its like Ben-s', but with the pressure regulated the other way around.
Ben-s' chemicals are sealed and pressurized, and feed into the mixing container.
In mine, the mixing container is sealed and a vacuum is applied. Then the chemicals are fed in as per Ben-s' design.
04-03-2007, 04:08 PM
Ben-s, you just need a driver board to operate your solenoid design. The simple ones are just a basic diode-transistor arrangement to energize the solenoid.
The only thing I would recommend for it is that the software should ramp the flow up and down, and not let it jump between values.
What is your acronym for PIC? The application looks like it needs a 16-bit digital I/O, like this Kadtronix unit (http://www.kadtronix.com/digio.htm). Ganging things like you have described is quite common.
04-03-2007, 04:09 PM
I suspect that using a stepper into a leadscrew into a syringe would accomplish roughly the same thing as you propose, ben-s, but with less plumbing.
Dono, my situation is that I have the ability to fab & program bits, but not necessarily the ability to get a complete set up at the moment. OTOH, I'm perfectly willing to help others with the software and electronics bits necessary to do it if somebody's willing to fund the raw materials.
Solder Flux is the smell of happieness in my geekroom. :D
I'm confused about vAg vs. pAg. pAg I can interpolate from chemestry, but what is vAg?
04-03-2007, 04:35 PM
To start with, all of the methods here seem feasible to me except possibly the vacuum method. In fact, all are or have been used at EK.
As for pAg, it is the negative log of the silver ion concentration similar to pH which is the negative log of the hydrogen ion concentration.
OTOH, vAg is the actual voltage developed for this measurment. This entire procedure requires a special electrode setup different than that used for pH and different electronics as well.
The advantage to vAg is that you can measure -136.15 mv easily but pAg is measured as 5.54 or some such (those fake values were not equivalent, just an example) and so you can get 5 digits from vAg and only 3 from pAg. That is a simplified view (see the real stuff below - I decided to do the calcs).
So, in the end, vAg has more precision and accuracy for us than pAg. And, contrary to posts here by another APUG member, pX or vX (halide) is NEVER used, because it will be nearly the same for Cl, Br and I at similar concentrations, but vAg may vary by hundreds of millivolts and this difference is crictical in some cases.
Here is an example (please appreciate the lengthy calculations needed for this :D )
Calculations are all at 40 deg C for 0.1 molar solutions
KBr pAg = 10.49, vAg = -51.3 pX = 1.115
KI pAg = 14.07, vAg = -272, pX = 1.112
KCl pAg = 8.09, vAg = 98, pX = 1.117
You see how much the vAg swings, and how similar the pX values are. The pAg are logarithmic terms, but are harder to interpret and meters do not have the capability of rendering logarithmic factors properly.
So, vAg values are used for critical work at Kodak.
The calculations are approximations based on the Debye Huckle and Nernst equations as noted above. (I should have mentioned Nernst above)
04-03-2007, 04:42 PM
Okay, so if we want to measure pAg / vAg, what does the electrode look like?
04-03-2007, 05:06 PM
Well, for starters, one electrode cannot be a salt electrode that 'leaks' or is porous to salt and the other should be a silver billet plated with the halide ion of interest (Cl, Br or I) or should be a bare electrode if one is doing mixed halides. The system is then calibrated using calculations as I did above and the entire electrical system is similar to the old Wheatstone bridge conductivity meters that we used when I went to undergraduate school to make pH measurements.
If the electrode leaks salt or is porous to salt, it will affect the measurment it is trying to make.
A conductivity meter would be a good starting point or a Wheatstone bridge.
That is all I feel free to discuss at this time
04-04-2007, 04:12 AM
I did say I would keep out of this forum until I had at least made my first klunker emulsion...
...but stepper motors and such are very much my thing, as is computer control.
A word of warning, though. A few years ago (maybe 10) I aquired a beautiful Ross wide angle aerial photography mapping lens. A huge, heavy beast.
It covers 8 by 10 easily and works out something equivalent to a 25mm lens on 35 mmm.
I temporarily mounted it on an 8 X 10 camera and took a few amazing pictures. It was very hard to use, though, as it had no iris or shutter.
Shortly after this I managed to find a huge leaf shutter that just fitted over the rear elements, although it was just the blade assembly without a mechanism.
I am no precision engineer and whilst I can do DIY with the best of them, designing and building a mechanical shutter mechanism was beyond me...
So, thinks I, just knock up a quick, stepper motor operated, electronically controlled one.... simple!
Well, after a few hours of labour (like.... 200?!) I still have the prototype shutter which occupies about 4 square feet, consisting of circuit boards, spaghetti wiring, LED bariers, stepper motors and a huge power supply to operate it all.
It sort of works, but not reliably enough at the speeds I need.
About another 50 hours of work might get the bugs out,
Another 50 might get the design tidied up enough to mount in a camera,
... but might not ;)
But it would still need some further development!
And I haven't even thought about the camera to mount it all on, yet.
Will I ever finish it? I doubt it. It was fun in its own way, but it sure stopped me from doing a lot of photograhy!
I'm sure some people will be better disciplined (and smarter) than me when it comes to these projects, but my personal experience of moving from the idea to the practicalities suggests that even if you keep it very, very simple and anticipate it will take 10 times as long as you think.. you will probably still grossly underestimate the time needed :-)
04-04-2007, 05:03 AM
Ron; Many thanks for what you have provided. I'm sure we all appreciate that you are limited in what you can say due to trade secrets.
Wirehead; Yes, I agree with you. A syringe based system would work well, quite likely better than my idea.
Brian; Yes, I would have to use a pack of transistors and diodes to do the heavy current switching and EMF suppression. I didn't want to muddy the waters of an already complex explaination by explaining specifics of drive circuitry in the original post.
The PIC is a series of very nice microcontrolers made by Microchip.
They are reprogrammable, low cost, and can be clocked from virtually DC up to 20MHz
I'm still learning them, but they are very nice.
What I was suggesting was to make all the individual systems "inteligent" so there is less load on the controling PC. The PC simply tells each chip what to do, and they get on with the job until the next instruction.
I would ramp flow rates in steps of 1/256th of the total flow. Although there would be much clunking from the valves!
Steve; I know the feeling!!! I still have a half built PIC controlled rotating panoramic camera.
I think that the syringe idea is best for accurate metered dosing. It's a lot cheaper than a peristaltic pump, although capacity limited.