PDA

View Full Version : Automated Emulsion Making - more modern formulas



Pages : 1 [2]

rmazzullo
04-20-2007, 08:04 AM
Hello all,

I have an idea for a solution pump that needs some input from the list members who are more familiar with the programming or the mechanics that might be required. This is only addressing a portion of a possible delivery and feedback system. Measuring vAg and mixing is not discussed in this message.

If you tie a stepper motor to a small peristaltic pump unit, shouldn't it be possible to measure out exactly what you need to add to the emulsion based on the amount of steps the motor has to complete? Also, if you tie an optical encoder (a cheap one is a few bucks, or you can take them from an old computer mouse) to the output of the pump shaft, couldn't this be used as a type of "flow meter" to get feedback into the motor control circuit?

The idea is that if the fluid amount in a given length of a particular diameter of tubing is known, the peristaltic pump can be designed to move that same unit of fluid between cam lobes (or 2 one-half cam lobes [not two-and-a-half] if you want to reduce pulsations in the line) over 'x' amount of time.

A peristaltic pump is a loop of flexible plastic tubing, held in a semi circular housing, and an rotor with 'x' number of sides [4,5,6 etc] with rounded corners rotates within the same housing as the tube, squeezing the tubing as the rounded corners of the rotor travel past it, pushing fluid along the length of the tubing. These are physically simple devices, and you can make the pump housing out of plastic using a forstner bit, and use small blocks of nylon, teflon, etc, for the rotor. The caveat here is that you must make the pump rotor as geometrically correct as you can (square, pentagon, hexagon, etc must be precise, not lopsided, and round the corners slightly). The rotor has to travel within the housing, but far enough away from the housing edge to allow for the tubing to be mounted along the edge, and squeezed as the rotor travels past. Too close to the tubing and the rotor will bind and stop.

What I am trying to figure out is if a small piece of slightly larger diameter tubing inserted somewhere in the middle of the original diameter tubing, after the pump output, would act like an 'expansion chamber' smoothing pulsations in the line. Or, would building a pump cam with say, 8 lobes instead of 4, would smooth out the pulsations to a manageable level.

Again, the thinking here is that accurate delivery of ingredients would require feedback of the actual amount of solution delivered to the mixing vessel (if you add the recycling feature mentioned in an earlier post by Photo Engineer, the motor will be running at the right speed when the valve is switched to deliver the solution to the mixing vessel at the right rate and duration).

I know this is adding up to more complexity, but while it is more involved, this may not be expensive to build at all. Another forum member mentioned using free CNC software to control the motors. I haven't looked yet, but this might fit the bill, if we can control more than 3 axis. Where the problem lies is in accurately measuring vAg, and how to integrate that feedback signal to adjust solution flow as desired.

Your comments are appreciated.

Thanks,

Bob Mazzullo

Photo Engineer
04-20-2007, 09:19 AM
Bob;

1. The solution motion is a series of pulses that have to be turned into a steady flow at the delivery end of the tube.

2. The tube changes diameter (relaxation) during use and changes volume with every use and so must be calibrated.

3. etc... BTDT. It works. It just has several years worth of problems to work through and program. We had a whole team of engineers and $M or more. :D

PE

rmazzullo
04-20-2007, 11:26 AM
PE,

Do you mean that over time, the tubing wall itself "weakens" and the diameter increases? Was this effect seen only at the portion of the tubing directly impacted by the pump rotor?

Thanks,

Bob M.

Photo Engineer
04-20-2007, 12:30 PM
Bob;

Yes indeedy do. The pump rotor gradually weakens all plastic tubing and the diameter changes gradually by a fair amount.

In addition, any replacement tubing is different in diameter when averaged over even one foot of length as it is not made to very exacting tolerances. Therefore, before each run, the tubing must be carefully calibrated for delivery and of course a correction must be put in for the calibration process itself which further weakens the tubing. You get this as a dv/dr (change in volume per run) in which time of run is part of that polynomial. That term is dv/dt so you might say it is one or two integrals depending on viewpoint. So, how is your calculus?

And then of course, the fluctuations change (pulsations) due to the rotors depending on speed and tubing wall thickness. This also has to be solved. We did this quite elegantly at EK IMHO. I had nothing to do with it. It is unpublished work by that friend who did the scuba diving.

PE

rmazzullo
04-20-2007, 01:28 PM
PE,

My calculus is extraordinarily rusty. Something else I am going to have to brush up on....

Thank you for your help,

Bob M.

David Grenet
04-20-2007, 08:01 PM
Bob,
If you're using a stepper motor, why would you need to use the optical encoder? Surely any feedback you get would just be a scaled version of the number of steps of the motor (which you have complete control over anyway).

PE - my calculus is fine but as an EE student dv/dt means something else and to smooth it one would just add a capacitor... something tells me this wouldn't work here :P
Would flexible (rather than stiff) tubing with a smallish help to give a similar effect?
David.

rmazzullo
04-20-2007, 08:15 PM
Hello David,

The idea of the optical encoder was to provide a second source of feedback in the event the pump rotor came loose from the motor shaft, as well as a type of "flow meter" to verify that what the stepper was sending was being "delivered' by the pump. If the two signals didn't coincide ('x' number of steps = arc separated by two adjacent lobes, for example) you can use the event (signal?) to abort the procedure. That was the original intent, anyway.

Thanks,

Bob

Photo Engineer
04-20-2007, 10:23 PM
Bob,
If you're using a stepper motor, why would you need to use the optical encoder? Surely any feedback you get would just be a scaled version of the number of steps of the motor (which you have complete control over anyway).

PE - my calculus is fine but as an EE student dv/dt means something else and to smooth it one would just add a capacitor... something tells me this wouldn't work here :P
Would flexible (rather than stiff) tubing with a smallish help to give a similar effect?
David.

I recognize that David.

I meant that the delivered volume in 5' is different today than it was 1 week ago due to tubing changes. All tubing we tested changed. Some was incompatible with the chemicals.

One way is to use a geared pump with stainless tubing that goes on gravimetric flow. There are other ways.

PE

PeterB
04-21-2007, 08:18 AM
....I have an idea for a solution pump that needs some input from the list members who are more familiar with the programming or the mechanics that might be required. This is only addressing a portion of a possible delivery and feedback system..... Bob Mazzullo

Bob,
I've just re-read all the relevant posts on this topic and it has become overwhelmingly clear to me that peristaltic pumping is not the way to go.

Its fraught with problems [2] that can only be solved by a team of dedicated experts doing it for a living - and that ain't us ! The syringe idea has the most merits and I suggest that we should accept that and move on.
The benefits of the syringe idea are as follows:

It is proven technology which is available now.
It is the cheapest automated delivery method.
You do not need to model a complex system in order to work out how much solution you have just delivered - the delivery volume is solely dependent on the distance the plunger has moved (assuming no air in the line [1]- because at the pressures being used, the compressibility of the fluid is insignificant)
Additionally, I would recommend servo motor control of the syringe's lead screw rather than stepper motor control. (Stepper is the only method discussed so far.) A servo motor will attempt to get to its destination because its position is controlled using a closed feedback loop, whereas stepper motors generally don't use encoders or pots to track their position as they operate open loop (without feedback).

If we used a stepper motor and the syringe exerted a force of greater strength than the motor, then it will interfere with the motion and you won't end up delivering the volume you requested. With a servo motor, you won't miss any steps unless the syringe jams up.

regards
Peter

[1] Even if there is a head of air in the syringe, then it's compressibility can be estimated using the eq'n P1V1/T1=P2V2/T2. I say estimated because the system is not closed and therefore the compression isn't adiabatic.

[2] PE has recently said that "We were always concerned about developing pulsations when using a peristaltic pump at low flow rates." and that
"It works. It just has several years worth of problems to work through and program. We had a whole team of engineers and $M or more."

rmazzullo
04-21-2007, 09:02 AM
Hello Peter,

The replies I have received from both PE and you have been very helpful. The syringes are inexpensive, and are available from www.smallparts.com, as well as others, so I will look at using that method of delivery. The largest syringe I can find so far is 60cc. I was trying to come up with an idea that would have more capacity than 60 cc, but still be considered small scale.

Any further thoughts are most welcome.

Thanks,

Bob

Photo Engineer
04-21-2007, 09:10 AM
Peter has put this in perspective. Thanks much Peter.

The syringe will work best at our scale. If you go over 1 liter, then you will have to use pump delivery. We used peristaltic pumps at 10 L volumes but also syringes. Above 10 L, say 100L and 1000L we used more traditional centrifugal pumps with geared motors that delivered by gravimetric flow.

In all cases except syringes, some sort of pulsation had to be contended with, and the solution differed with pump. At small scale we used tygon tubing and recalibrated for every run, and at large scale we used stainless tubing. That route gets expensive fast.

Our solutions to these problems were also quite expensive.

PE

Photo Engineer
04-21-2007, 01:42 PM
Here are additional facts on automated emulsion making.

Emulsions with mixed halides generally end up with banding unless the addition of both halides is carefully controlled. You have missed this point totally, I think.

If you add silver, and then a bromoioide, it bands if the halide is controlling. Kodak had many solutions to this.

If you put the iodide into the kettle, it ends up all in the center and must be 'churned' to the surface by using ammonia or other silver halide solvents (Kodak had 2 unique solutions to this problem)

If you make a bromoiodide that is not banded or core shell or 'converted' you must have a smooth addition of silver and iodide, with bromide doing the controlling. This is a key to this type of making. Agfa used methods very old compared to those used by Kodak. I know nothing about the Ilford methods.

Pulsations introduced by pumps must either be minimized, evened out or eliminated by some means. Elimination is possible as noted by Peter in his exellent post, by several means. We used some of these, but there are many more of them that can and must be used, especially at smaller scale.

This is why I have said it can be done, but is expensive and time consuming. Ryuji has posted elsewhere that he is doing automated precipitation but when I asked him to post examples, he appears to have refused and may now be ignoring my posts completely.

I have offered publicly and privately to work with him, but he has refused my offers in the strongest terms in all instances.

I believe that both of us together may be able to offer a good, workable automated method of making monodisperse emulsions or t-grains for about $5000 in equipment. It is possible. I am also willing to work with anyone else on a 1:1 basis.

In the mean time, I continue to stress that older emulsions from the 20s to the 40s or 50s are more in keeping with what the average hobbyist might attempt, but advanced experimenters can move beyond this. I continue to be willing to help, but caution you that use of automation is a difficult road. I am willing to help anyone, but if they ignore my help, suggestions, or consider me unable to 'produce' I will be unable to help you all in the long run. I cannot do all of this myself.

Pursuant to this, I have mentioned before, I am willing to develop a workshop covering these advanced topics, but due to the limited interest in even the fundamental workshop, I am not willing to go ahead with the 'advanced' version. This is all very expensive when one considers the cost of just the silver involved.

Someone jokingly said to me that if one more major B&W producer quits making film or paper, people will beat a path to my door. I don't expect that to be true as by that time, I will be long gone - hoping for a long lifetime for analog photography.

But, in the mean-time, I will do what I can.

Now, for a specific hint. If you want to automate, and if pulsations are a problem, a very narrow inlet orifice with back pressure minimizes the pulsations. HINT:HINT:HINT..... You go from there.

PE

Photo Engineer
04-21-2007, 01:44 PM
After I posted this, I realized I forgot one important point.

I want these ideas to go to the public, not to the private companies. I don't want to pass on this information, only to have them become secrets of a big company.

PE

PeterB
04-22-2007, 07:25 AM
Here are additional facts on automated emulsion making.........Now, for a specific hint. If you want to automate, and if pulsations are a problem, a very narrow inlet orifice with back pressure minimizes the pulsations. HINT:HINT:HINT..... You go from there.

PE

Hi PE, thanks again for adding more value to this thread. One day somebody must gather all the piecemeal bits of info into one requirements spec. document.; There is an obvious need for this when more than one person is trying to understand the proposed system(s).
Perhaps I'll cobble one together when I have time, although others are welcome to do this 'cause I'm not about to come across much spare time in the near future.

Your HINT a the end of the post reminds me of this piece of equipment (http://www.dhinstruments.com/prod1/pdfs/brombloc_s.pdf) that is used for calibrating or measuring mass flow rates of gases. It relies on the principle of "Critical Flow (http://www.flowmeterdirectory.com/flowmeter_artc/flowmeter_artc_03120401.html)". Although I thought critical flow only applied to gases. Well perhaps I'll look into that another day.

regards
Peter

Graham06
04-29-2007, 01:11 PM
Hi PE, thanks again for adding more value to this thread. One day somebody must gather all the piecemeal bits of info into one requirements spec. document.; There is an obvious need for this when more than one person is trying to understand the proposed system(s).
Perhaps I'll cobble one together when I have time, although others are welcome to do this 'cause I'm not about to come across much spare time in the near future.

...

regards
Peter

Why don't you guys start a Wikibook? (http://en.wikibooks.org/wiki/Main_Page) . These forums are fine for discussing specific problems, but it's the worst format for pressenting the information you are aiming to gather.

Anyone who contributes to these forums could transcribe information into the wikibook in a logical place. Minor copyright problems are solvable

Emulsion
05-16-2007, 06:58 AM
This US patent 5442562 seems to be related to the topic you are discussing...the example at the end of the patent is particularly relevant.

The patent can be view and download as a PDF at:
http://www.google.com/patents

PeterB
05-16-2007, 11:10 PM
This US patent 5442562 seems to be related to the topic you are discussing...the example at the end of the patent is particularly relevant.

The patent can be view and download as a PDF at:
http://www.google.com/patents

Wow, that is great to be able to view the specific patent (http://tinyurl.com/2d78ff) with so much detail !! I knew I could do this, but the Google GUI is SO much more user friendly that the USPTO one.

Did you know that Eastman Kodak Company has only filed 33 patents (http://tinyurl.com/39z4wq) in the last 2 years, but a total of over 600 patents (http://tinyurl.com/2qt4vw) from June 2003 until May 2007 !! (Google Patent search won't return more than 600 hits for one date range)

One could have a field day trawling through the patents (while suitably refining one's search parameters) to get ideas for designing automated emulsion making systems.

If only I had more time on my hands.... :(

regards
Peter

rmazzullo
05-16-2007, 11:21 PM
Wait until you discover what is lurking in the other Kodak patents, some before 1960.

Bob M.