Skimming Principles

[BeanAnimal]

reply directed towards goby ..... read first word of post

[goby1]

My question was why it is better, not if. In my 100gph+100gph scenario, 200gph enter in the top in both cases, with that amount going from top to bottom. In one case, 200gph leaves the skimmer, and the input is 200gph sump water (non-recirculating). In the other case (recirculating), 100gph leaves the skimmer and the input is 100gph sump water and 100gph from-the-bottom-of-the-skimmer water. I've read escobal's book and many of these threads, and don't see why you'd use the other pump for recirculation. I'm not saying you're wrong, just that I don't get it yet.

In the non-recirculating scenario, you're processing twice as much water from the tank. The situation regarding the water flow is 200gph in both cases, but in the non-recirculating setup, the skimmer is fed less filtered water. And it outputs twice the flow back into the tank. Still seems better to me.

G1

[goby1]

Ok tiny just about to read your post... just popped up while writing the last one.... G1

[tinygiants]

Originally posted by goby1
Ok tiny just about to read your post... just popped up while writing the last one.... G1

It is all about dwell time. I wanted flow and dwell.

[jnarowe]

bean, I get it now. DOH!

[goby1]

So with 4gpm from the sump, you'd have 4x turnover per day. Which, utilizing escobal's other equations, gives a better purity #.

I posted some time ago to what used to be Calfo's forum about surface skimming. I presented an argument that says that once you utilize even a small amount of suface skimming, then the surface is only marginally more concentrated than the rest of the water. Sure, if you have an overflow from under the surface of the tank water, crud will build up. But once you skim even a small amount, the surface is for all intents and purposes the same as the rest.

When you talk about proteins needing more time to be attached to the bubble, it's an issue of diffusion, not surface science. A molecule will either preferentially attach to the interface or it won't. It's not attracted to the surface in the sense that it is pulled there from other places in the water. It gets there by diffusion and convection. Once there it might stick. If it does, in this sense it is attracted to the interface by VERY short range forces. More 'trys' at sticking won't make it any more likely to stick. Hope this makes sense.


G1

[tinygiants]

Originally posted by goby1
So with 4gpm from the sump, you'd have 4x turnover per day. Which, utilizing escobal's other equations, gives a better purity #.

I posted some time ago to what used to be Calfo's forum about surface skimming. I presented an argument that says that once you utilize even a small amount of suface skimming, then the surface is only marginally more concentrated than the rest of the water. Sure, if you have an overflow from under the surface of the tank water, crud will build up. But once you skim even a small amount, the surface is for all intents and purposes the same as the rest.

When you talk about proteins needing more time to be attached to the bubble, it's an issue of diffusion, not surface science. A molecule will either preferentially attach to the interface or it won't. It's not attracted to the surface in the sense that it is pulled there from other places in the water. It gets there by diffusion and convection. Once there it might stick. If it does, in this sense it is attracted to the interface by VERY short range forces. More 'trys' at sticking won't make it any more likely to stick. Hope this makes sense.


G1

Yes 4 gpm turns my tank over every 6.5 hours. That is using the purity formula. My goal is to turn it over every 12 hours.

Interesting point about the surface not having any more protiens. It is however the first place that you will see evidence of skimmable material when you do not have a surface fed overflow.

I understand your point about protiens not being attracted like a magnet. However the protiens at the ingterface need time to fully attach to the bubble exterrior. When they do attach the bubble becomes much more stable. In escobals book it talked about bombardment rate. By increasing the throughput of the skimmer you are driving down the bombardment rate. Again a practical example of this is the NW and beckets. Many people started off with a big pump feeding the becket injector as the only source of water for the skimmer. This put them with a very high feed rate. When they went to a recirc set up and a more reasonable feedrate they reported better skimmate production. The proof is in the skimmate.

As for my recirc pump, I am not sure that it is worth it. My bubbles are still to fast to be really captured by reasonable water flow. My recirc only adds 1 - 2 seconds to the dwell time. I need smaller bubbles in order to justify the recirc.

Dale

[jnarowe]

about the surface issue: if there is already turbulence within the tank, then there will be some protein already attached to bubbles on the surface. IMO it is very important to bring these through the skimmer, otherwise they would just accumulate on the surface.

Many skimmer manufacturers state in their literature that the best way to feed their skimmer is by gravity feed from an overflow rather than from a sump. They sell lots of "in-sump" designs for practical purposes but it seems the gravity feed is more important. That is I think the main reason why overflow length is such an important issue for today's reef keeper.

My skimmer will be gravity fed and in-sump. I have over 6 linear feet of overflow and I am positive that will be beneficial for skimming.

[BeanAnimal]

Goby, I think your basing your thought process here on some flawed (unproven or disproven) logic.

It is not just the "surface skimming" it is the surface area. The surface of the water is where the protiens are attracted. Lets look at it from a number standpoint.

say the tank is 20 X 50 inches. That is 1000sq inches of surface area. Now lets say the overflow is pumping 1000 gallons an hour and the "overflow" is 25 inches wide with a water depth of 1/2".

Given time T, about 1/2 of the water going over the falls is NOT from the surface, and has had NO air water contact. Because it has not had air contact, the water is not as protien rich, nor have the protiens in that water been partialy dislodged.

Given the same elapsed time T with a 50 inch wide overflow and the same turnover rate, there is nearly 50% more water that has made the air/water contact before going over the falls. This water is thus much more protien concentrated, and therefore more likely to have protiens ready to be dislodge by the skimmer.

This has been proven many times. The easiest way to do this is not scientific, just feed your skimmer overflow water from a narrow surface skimmer and then from a wide surface skimmer. the results will speak for themeselves.

Regarding your second point, it does not matter if it is a close proximity issue or not. Simple math a disollution or dilution will show you that a givin protien has a much better chance of being removed in a smaller cleaning loop (skimmer recirc), than it does in a larger skimming loop (sump recirc). This goes back to the efficiency of cleaning by dilution.

Bean

[goby1]

'Time to attach,' if one were to quantify one, would be on the order of nanoseconds. Diffusion times would be much, much larger, and in our case is the time of interest for 'attachment.'

Bombardment... this is the same in both cases as there would be 4gpm going down the skimmer.

G1

[BeanAnimal]

Firstly "bombardment" as it is calculated in the spreadsheet seems like a useless number. It does not appear to be valid if you observe how it changes with regards to skimmer size and flow. Actually it is counter intuative in many cases.

Secondly, you seem to be missing something with regards to the whole process. You assumption is that the water is the same no matter where it comes from. This is not true. I think we have all tried to show this and I am not sure why you don't see it.

With regards to "time to attach" are you talking about the actual jump the molecule makes accorss the interface, the time it takes for the molecule to become glued enough to make the jump (reported to be as long as 2 minutes) or the time it takes the molecule to be ready to jump?

The TIME T I was talking about was simply elapsed time (for purposes of an average our actual count) of protiens going over the overflow into the skimmer.

I would contend that if the surface water is more protien rich and the protiens are more easily stripped from it (proven many times by many different studies) that water already in the skimmer that has been made to contact the air would also be more protien rich. Running this water in a recirc fashion is therefore then more efficient than introducing more water into the skimmer and "starting from scratch".

If you follow your premise to it's logical extreme, then faster flow through the skimmer means better performance, no matter what size the skimmer is, or how much air it uses. Your premise also again ignores dilution as a factor in regards to an actual protien being pulled away.

Bean

[goby1]

No, it hasn't been proven many times. I'm basing my arguments on proven logic. Fick's laws of diffusion are true. A water molecule is a couple of angstroms in size. The forces that attract the protein to the surface are also on the order of angstroms.

Sure 50% more for the larger overflow, but not 50% of what? With 20 by 50 inches, half an inch thick, my calculations tell me that there are 2.2E26 molecules in that water, with 5.8E18 on the surface. That mean that about 2 millionths of a percent are on the surface.

G1

[goby1]

Bean, I'm talking about the same flow through the skimmer. In my example, it was 200gph. In Tiny's it was 4gpm. I agree it matters where it comes from. But with 200gph or 4gpm, why not take more in for processing?

G1

[BeanAnimal]

You are assuming that the 1/2 thich sheet of water is stagnant. With the turbulance of the water, many of these small molecules make surface contact with the air. This is the point where the contention is that they become bound to the surface tension of the water, making the surface more concentrated.

This makes the idea of removing as much "surface skim" as possible valid. Removing water from below the surface is removing water that has not made this initial air/water contact and has not therefore released it's protiens.

Something else to think about. Knowing that air contact provides an interface for the protiens to stick or move, we want to use as much water as we can that has recent air contact. As the surface of the water is agitated amd protiens stick to it, the water "going back down" is now protien deficient as compared to the rest of the tank water below it. So moving this "thick" sheet over water over the overflow is not gathering as many protiens as would moving the same volume of water in a thinner sheet.

I hope this makes sense.

Bean

[goby1]

The protein molecule doesn't get 'ready' to jump. The important time is the diffusion time for a protein to cross the diffusion layer near the interface, coupled with the state of convection that exists between the air bubbles. The convection determines the thickness of the diffusion layer, and the time to cross the layer is determined by the diffusivity of the protein in water, which is determined by the identity of the protein and the temperature. Fick's law is the relevant equation for diffusion. I don't know how to work out the convection part.

G1

[goby1]

Not to get two things mixed up here, as were talking about the surface skimming issue as well as skimmer operation. In regards to the surface skimming, the rate at which proteins concentrate at the interface is miniscule compared to the rate at which water flows over the overflow.

G1

[jnarowe]

Geez guys...can we go back to pictures?

[tinygiants]

Goby,

You passed me with your math on the number of molecules. I have seen with my own eyes an increased skimmate production when you slow down the feed rate. That is the bottom line. In my mind you need to decide how many times a day you want your water skimmed and create a skimmer sized to suit that flow.

As I was starting this project, I thought that I needed a 8" body. I then stumbled onto the idea of entraining the air from post in this thread. With that thought I started to work on flow characteristics. What I came up with is the skimmer I have now. Unfortunantly my recirc path does little to the average bubble. It will hold a microbubble in suspension for over a minute. So my skimming is based on water dwell time. I try to get as much out as I can in the first pass. I want the water as clean as I can get it without sacrificing my turnover rate. Faster turnovers yields less skimmate even though the water comes back around more times.

Dale

[BeanAnimal]

goby,

I am not quite sure that the "miniscule" time that a molecule spends at the surface is irrelavant. Many protiens are immedaitly bonded. Of course many take up to two (according to the research) 2 minutes.

In any case, the math and molar density of the water are all rellavant but maybe out of the scope of this conversation, in regards to what we are able to prove or disprove. The only thing we have is experience or trial and error to show what works and what does not.

I would suggest testing your theory. The best way I know is to use natural seawater that is collected for this purpose. There are several ways a side by side test can be run. Both a limited proties test with closed systems both of the same quantity if NSW and test with an open ended system that uses the ocean as the "tank" (endless supply of homogeneous protiens if they are sucked in side by side to the test skimmers).


Bean

Bean

[goby1]

Sure, immediately bonded. But they have to be transported to the interface first though. The only two ways this can happen are via convection and diffusion. Diffusion is very easy to describe when you know the diffusivity (diffusion constant) of the molecule of interest. I don't know this value, as we don't know what we are trying to skim. Most of the nutrient export doesn't even involve proteins to begin with, I'd guess. Probably phytoplankton. Convection is more difficult, and the relevant result of solving the convection equations determines the diffusion thickness, or the thickness of water where diffusion is the dominant transport mechanism. This layer is the one close to the interface. In any case, back of the envelope calculations can illustrate that the rate at which 'proteins' are transported to the interface is very small compared to the rate at which water overflows the tank. This means that the water would have to flow very slowly over the overflow, in order to have a noticeable increase in the concentration of 'protein' in the overflow water.

It would be good to test issues related to skimmer design. I think the most important one would determine what we are trying to remove. Only then can you have an idea about rates to use in the calculations. You would also have to know the identity of the 'pollutant' so you know what to test the skimmate for. Testing an idealized protein would be useful to compare geometries though.

I'd think that testing for the removal rates of a homogeneous (did you mean monodisperse?) protein would be best conducted in NaCl in distilled+DI water, buffered to 8.4. Testing (for rates of removal) wouldn't be too far out of line for us to do, assuming the pollutant we dosed with was easy to test for. I would really like to do this sometime. Let me know what you think about the testing idea. How would you know what to look for by skimming NSW? We could save a lot of time,effort and money by doing a scaled down experiment. Maybe 2'' PVC with cheap airpumps/diffusers? It might be better to test the concentration of the skimmed water rather than the skimmate.

I do know of at least one person who has done a lot of research on foam fractionation, but it was to do with mineral flotation. I really should run this by them to get an opinion.


Let's keep this topic going. I think it is interesting, and at the same time very relevant to our hobby. Good chunks of cash get spent and made on skimmers, that for many of use are the principle means of water treatment.

G1

[BeanAnimal]

Test could be approached in two basic ways that everybody understand quantitative and qualitative...

You are looking at the issue from a numbers standpoint (not a bad approach either). I am looking at it from a "wow that is stinky skimmate" standpoint.

Before any real numbers testing is done, it would be nice to use several setups to see wich pulls out the most noticeable skimmate product. I suggested NSW because it is readily available in large quantities and can be used to feed any number of blind comparisson skimmers. (directly from the ocean if needed). In this way we don't have to worry about the limited size of a tank and it's finite source of gunk.

With regards to a test to prove efficiency in a closed system, NSW again provides an endless source of of a "standard" test medium.

Once these test were carried out, I am sure a clear line would develope between what works, and what works well. At this point quantitative test could be undetaken to show what is being removed and possibly why.

It is pretty much common sense that the better skimmate is dark and smelly. Just turn your skimmer off for a few days and fire it back up, the skimmate will be darker and nastier, showing that there was more "stuff" to be removed.

Thoughts?

[H20ENG]

I wonder how much skimmate you would get from NSW in a good clean location ie Catalina or offshore.
Though I do agree about it being a good constant for testing.

[BeanAnimal]

I am not sure how much you would get and never looked at it that way. But now that you mention it, maybe NSW would not skim well at all due to lack of nutrients. I guess that is what we are striving to produce is NSW in our captive systems.

So the easy fix is to skim NSW from the jersey shore

BTW Chris.. I just got my JANDY actuator. Have not tested it yet and still need to either buy the jandy valve or find a way to hook it to a standard ball valve.

Bean

[H20ENG]

Sweet Bean!
Usually some small box tubing and a grinder will make a good valve coupling

Maybe get some water change water from a huge fish growout system. There a good test.
You can also just run them side by side on a large system and see which one shuts down first or gets the best skimmate.

[tinygiants]

I have been thinking about this debate over surface skimed water and such.

There is another way for the protiens to get to the surface. Current. Now maybe that still does not make for a greater saturation, maybe it does. In either case, most tanks I know of use a surface skimmer to keep the water surface clean.

That being said most tanks drain lines are surface fed. And I am reasonably certain that the concentrations of protiens is greater in my display than in my sump. For 1 thing my sump is filtered by the skimmer for nearly 20% of the water. The rest goes through a UV then refugium or the sock. So I would think the tank fed skimmer would be more optimum than getting water from the sump.

Dale