Mix up some lamp black and run it thru the filter. If the oil goes in black and comes out clear on the first pass it is as good as TP.
Ralph
Ralph
DIY bypass filter
- Thread starter Fitzhugh
- Start date
- Status
- Joined
- Sep 28, 2002
- Messages
- 39,798
Fitzhugh, you give a sound presentation there. I would sorta compare it to someone entering a foreign society that was sustained on corn flakes (which were also alien to you)..and had generations of healthy residents ..and questioning the nutritional value of cornflakes ..but I do not find any flaws in your disposition. You require beyond "apparent" results or inferential evidence. The thing is ..there are no filter manufacturers that offer TP as a stand alone filter. There's no money in it. Filter companies are in the filter business and not in the tp holder business. There's no recurring sales involved with tp filtration. Some use it in combination with traditional paper media ..but none, that I'm aware of, use it exclusively. Hence the data that you deem essential ..is not out there to be referenced.
Motor Guard runs into this problem. Most buy the fixture (as Ralph does) and merely "cheat" and figure a way to use tp for air filtration. Now I'm sure that MG has the type of data that you're looking for ..but it would involve a slightly different filter ..and would be dealing with air ..hence there would other "alterations" to the data to invalidate it for you.
I would like to add, however, that I don't think too many bypass filter manufacturers have one standard test with which they use to determine their Beta numbers. If you ask Mel (Filter Guy), he'll inform you that the data, that you're accepting as canon in this tale of two filters, is done under very broad latitudes under their respective SAE/ISO standards. That is, manipulation of the data is wide spread making even published Beta numbers variable in their "truth", so to speak. Aside from the tendancy for more expensive, premier, filters having better filtration, we too are left with only anecdotal evidence as to their worth (PC with a PureOne vs. PC with a WallMart ST Ecore for example).
One offering that you may accept as valid ..is a somewhat unconventional installation that uses DOE (double open end) wound filters. These are sold in McMaster Carr, Grainger, and many others. They are rated by micron and flow limitations. They are used for air, water, oil, solvents ..lots of stuff and are distinguished by their composition. This has been accepted practice for decades of producers and users in industry ...so I think you can give at least a few of those who utilize these filters credit of doing the homework on them. These are offered by premier filter manufacturers (the aforementioned PALL, for one) ..all the way down to smaller filter enterprises that operate on a much smaller scale and offer custom sizing.
Even these filters will lack the level of data that you require for an informed choice since the data, much like the case with tp, is somewhat universally accepted.
Good luck in your research.
Motor Guard runs into this problem. Most buy the fixture (as Ralph does) and merely "cheat" and figure a way to use tp for air filtration. Now I'm sure that MG has the type of data that you're looking for ..but it would involve a slightly different filter ..and would be dealing with air ..hence there would other "alterations" to the data to invalidate it for you.
I would like to add, however, that I don't think too many bypass filter manufacturers have one standard test with which they use to determine their Beta numbers. If you ask Mel (Filter Guy), he'll inform you that the data, that you're accepting as canon in this tale of two filters, is done under very broad latitudes under their respective SAE/ISO standards. That is, manipulation of the data is wide spread making even published Beta numbers variable in their "truth", so to speak. Aside from the tendancy for more expensive, premier, filters having better filtration, we too are left with only anecdotal evidence as to their worth (PC with a PureOne vs. PC with a WallMart ST Ecore for example).
One offering that you may accept as valid ..is a somewhat unconventional installation that uses DOE (double open end) wound filters. These are sold in McMaster Carr, Grainger, and many others. They are rated by micron and flow limitations. They are used for air, water, oil, solvents ..lots of stuff and are distinguished by their composition. This has been accepted practice for decades of producers and users in industry ...so I think you can give at least a few of those who utilize these filters credit of doing the homework on them. These are offered by premier filter manufacturers (the aforementioned PALL, for one) ..all the way down to smaller filter enterprises that operate on a much smaller scale and offer custom sizing.
Even these filters will lack the level of data that you require for an informed choice since the data, much like the case with tp, is somewhat universally accepted.
Good luck in your research.
I'm perhaps excruciatingly aware that I don't have the history in this strange land, and that I don't know all the ins and outs. That's one thing that makes me interested in finding some objective data.
I do apologize if I sounded irritated, and I tried to pre-apologize because I knew it might come off sounding that way. I know that this is purely my own emotional needs for data. I have a long-term investment it being able to measure things to predict future performance. It's a quirk that I tried to explain.
And I do appreciate the economics of the situation - it is hard to justify the expense of some testing for $1 elements if you don't think you'll make your money back. And at the bottom of all this is the obvious test - do engines live longer with TP filters than some other kind of filters as bypass elements. That testing is hugely too expensive and time consuming to get to, although it's what we all are really after. We just can't get at it directly.
I do think that some enlightened soul could make a nice business out of what I so broadly hinted at - grab a source for consistent TP rolls, maybe just private-label Scott 1000's if that works, and then go cook some filter tests. If they turn out vastly better than other available filters, sell them at a premium - like $2.00 each, which would be making a killing on TP. Publish the numbers and start making TP filters something that you don't have to explain to people - a lot - to get them to buy. I actually think that's a win-win for the TP filtering and bypass markets. I considered doing it myself while I was typing my response to Ralph, and voted no based on the other fish I'm frying. But it's an open market, doesn't require much investment (the testing is trivial compared to the advertising bills) and would actually do the world a service.
Especially where, as you note, the test results may be ... um... interpreted in the best light - wink, wink, nudge, nudge. Of course that goes on. It's a game. But the rules on testing and advertising at least establish bounds.
As I said, I don't mean to be disrespectful, and I very much appreciate the good pointers that you, Ralph, and others have given me here. You are, almost all of you, more experienced here than I am.
But I have seen what I think is a representative cross section of anecdotal evidence for TP filters, acknowledge that they may well be as good as represented, but still can't make the step personally to using them. That's based on perceived inconsistency rather than ideas that they're ineffective. The inconsistency is not anyone's fault - it just is because no one's trying to make them consistent necessarily, or prove that they are if they are.
I will go look at the filter options you mention and see what's there. And thank you, Ralph, and slalom for messing with my questions.
I do apologize if I sounded irritated, and I tried to pre-apologize because I knew it might come off sounding that way. I know that this is purely my own emotional needs for data. I have a long-term investment it being able to measure things to predict future performance. It's a quirk that I tried to explain.
And I do appreciate the economics of the situation - it is hard to justify the expense of some testing for $1 elements if you don't think you'll make your money back. And at the bottom of all this is the obvious test - do engines live longer with TP filters than some other kind of filters as bypass elements. That testing is hugely too expensive and time consuming to get to, although it's what we all are really after. We just can't get at it directly.
I do think that some enlightened soul could make a nice business out of what I so broadly hinted at - grab a source for consistent TP rolls, maybe just private-label Scott 1000's if that works, and then go cook some filter tests. If they turn out vastly better than other available filters, sell them at a premium - like $2.00 each, which would be making a killing on TP. Publish the numbers and start making TP filters something that you don't have to explain to people - a lot - to get them to buy. I actually think that's a win-win for the TP filtering and bypass markets. I considered doing it myself while I was typing my response to Ralph, and voted no based on the other fish I'm frying. But it's an open market, doesn't require much investment (the testing is trivial compared to the advertising bills) and would actually do the world a service.
Especially where, as you note, the test results may be ... um... interpreted in the best light - wink, wink, nudge, nudge. Of course that goes on. It's a game. But the rules on testing and advertising at least establish bounds.
As I said, I don't mean to be disrespectful, and I very much appreciate the good pointers that you, Ralph, and others have given me here. You are, almost all of you, more experienced here than I am.
But I have seen what I think is a representative cross section of anecdotal evidence for TP filters, acknowledge that they may well be as good as represented, but still can't make the step personally to using them. That's based on perceived inconsistency rather than ideas that they're ineffective. The inconsistency is not anyone's fault - it just is because no one's trying to make them consistent necessarily, or prove that they are if they are.
I will go look at the filter options you mention and see what's there. And thank you, Ralph, and slalom for messing with my questions.
- Joined
- Sep 28, 2002
- Messages
- 39,798
No no no, pal!!
No reason to think that my response was intended to sound at all harsh (note the
) or punitive in any way. Pardon me if it did. Rereading it ..it does appear somewhat "deadpan". My bad
You gave very good reasons for your
By all means ..when travelling in uncharted waters ..it's best to be cautious and thorough ..even if they're only knee deep (you don't know).
I was merely pointing out that even those that appear savvy don't necessarily have the type of data you seek ..they have to rely on reputation and "accepted practices".
May I suggest page 335 (about halfway down) at McMaster-Carr
then go to page 337 for the filters that fit in them . There's a chart that shows what filter is appropriate for the intended liquid to be filtered.
I was merely pointing out that even those that appear savvy don't necessarily have the type of data you seek ..they have to rely on reputation and "accepted practices".
May I suggest page 335 (about halfway down) at McMaster-Carr
then go to page 337 for the filters that fit in them . There's a chart that shows what filter is appropriate for the intended liquid to be filtered.
Ralph, I do respect your experience, and your dedication to what has worked for you. And I mean what I'm about to say with all of that respect. I do not mean to be offensive; this is just a difference of opinion.
1. I believe that toilet paper is a good, perhaps even a great filter medium.
2. I do not believe that I can depend on every roll of toilet paper I find to do an equally fine job. This comes from my professional training and experience.
3. I prefer predictability to uncertain greatness.
I would happily reverse item number 2 if the data was available to show me that toilet paper reliably meets filter effectiveness tests on an ongoing basis; that is, the predictable performance of toilet paper could be demonstrated.
I make a distinction between things I can predict as a result of measurements and things I can't. Things which I cannot predict on the basis of measurements I classify under the headings of random chance or athletic contests.
Can you tell me who's going to win the Super Bowl?
You may have an opinion, but the honest answer is no, you cannot. There is no way to calculate beforehand who will win. Even if your opinion proves correct, that does not mean you can predict the winner of the next one. These are athletic contests. While both teams may be great, play professionally and excellently, there is no way to predict on a numerical basis who will win.
My opinion of toilet paper is something like that. At my local grocery, there are many brands of toilet paper. Some are softer than others. Some have different sized rolls; some have single sheets, some double. Some have moisturizers and scents added, for pity's sake. Is one a better oil filter than the other? Clearly, yes. Would some of them be terrible filters? Probably.
But let's focus on good, tightly wound rolls from a single maker. To evaluate how good a filter it is, we need to know things like what size particles actually get through; what the residual acidity/alkalinity of the paper is; what the consistency of the tightness is between rolls; what metal and grit is left in the paper pulp when the paper is made. The only good way to find that stuff out before we put it into an engine is to test the TP roll.
The idea of testing is pretty simple. Making predictions used to be done by sacrificing a goat or a chicken and reading the entrails. Some people used tea leaves, some threw yarrow stalks, and some people used Tarot cards. All of these methods shared the results that occasionally they would be right. However, people found out that if they actually subjected something to repeatable testing, they could make predictions about the things that they tested that were right much more than sometimes. An engineer that's right no better than sometimes get fired - and fast, before he gets someone killed. Testing is what the modern scientific world has developed for making predictions, and it works remarkably well where it's done correctly.
So to predict how well a filter will work, we ought to test it, that being the best predictive tool we have available. There happen to be predefined, standardized tests for filter effectiveness as a function of particle size.
Given sufficient money and time, rolls of TP could be tested for particulate efficiency, chemical composition, and winding tightness, and for other things that matter to filters. And we would know what performance that one roll of TP that we tested would have delivered in the car before we ran the tests and ruined it for car service.
With even more money and time, we could run a statistically valid sample of TP rolls from that maker and come up with a good set of results that statistics say are usable as predictors of the performance of that run of TP, including the likely average and statistical spread of performance. That's starting to be predictable.
With even more money and time we could run statistically valid samples of many lots of TP and make valid predictions of the performance for that TP maker (at least within that paricular brand.) If this was kept up in some way, it would be a useful predictor, and if I had that kind of information, especially if the someone, preferably the company that makes the stuff, was legally liable for telling me the truth about what the TP does, I'd flip right over to using TP given only that the numbers were good enough. It would then not be an athletic contest - I would know, within bounds, what performance the next roll of TP would give me.
So far, I have not found even one test-data report of the filtering efficiency of even one roll of toilet paper. The closest I have found is that some folks determine whether their oil is clean by how clear it looks or by whether it wipes off their fingers without leaving a stain. This is not a statistically valid test, needless to say.
Just to illustrate, you point out that a TP filter will remove lamp black from oil in one pass. Is that because the filter nets out the carbon granules, or because the carbon granules are chemically or electrically attracted to the TP fibers? Or did some chemical reaction inside the filter eat the carbon? I've seen and done chemistry experiments showing "magical" clearing of solutions from opaque to clear by adding one little bottle of reagent. Who's to say that if you contaminated oil with 1 micron particles of steel, silicon, or rust the particles would not pass through?
One could *test* for whether the carbon, steel, rust, silicon, whatever is going through the filter. That would be the beginning of knowledge.
For me, until I get that kind of information on an ongoing basis, the filtering efficiency of TP must remain an interesting natural phenomenon. One roll is great. Does that mean the next roll is great? In my mind no. With my luck, I'll get a run of uniformly bad rolls. I'd rather have predictably good than unpredictably great where my engine's lubrication is concerned.
So additional anecdotes about TP filtering efficiency are not very useful to me. I believe some of them are great. I believe some are awful. And I have no rational basis for selecting the good from the awful. Anecdotal evidence is NOT a rational basis for decisions. TP may be the most effective oil filter on the planet, and I'm going to miss out on it because I cannot find predictive testing data to support that. So be it.
So Ralph, thank you very much for your help, your experience, and your and willingness to share it with us. I appreciate your insights. But I cannot compare TP in an apples-to-apples fashion to industrial filters at this time. And until that is possible, I cannot change my opinion of the use of TP for filtering the oil for my car's engine.
So please - tell me what the beta profile of a TP filter is; better yet, tell me a statistically valid set of data on a sampled set of TP. But you don't need to tell me more anecdotes - I've already read a large number of them to come to my current opinion. More of them will not change my opinion.
1. I believe that toilet paper is a good, perhaps even a great filter medium.
2. I do not believe that I can depend on every roll of toilet paper I find to do an equally fine job. This comes from my professional training and experience.
3. I prefer predictability to uncertain greatness.
I would happily reverse item number 2 if the data was available to show me that toilet paper reliably meets filter effectiveness tests on an ongoing basis; that is, the predictable performance of toilet paper could be demonstrated.
I make a distinction between things I can predict as a result of measurements and things I can't. Things which I cannot predict on the basis of measurements I classify under the headings of random chance or athletic contests.
Can you tell me who's going to win the Super Bowl?
You may have an opinion, but the honest answer is no, you cannot. There is no way to calculate beforehand who will win. Even if your opinion proves correct, that does not mean you can predict the winner of the next one. These are athletic contests. While both teams may be great, play professionally and excellently, there is no way to predict on a numerical basis who will win.
My opinion of toilet paper is something like that. At my local grocery, there are many brands of toilet paper. Some are softer than others. Some have different sized rolls; some have single sheets, some double. Some have moisturizers and scents added, for pity's sake. Is one a better oil filter than the other? Clearly, yes. Would some of them be terrible filters? Probably.
But let's focus on good, tightly wound rolls from a single maker. To evaluate how good a filter it is, we need to know things like what size particles actually get through; what the residual acidity/alkalinity of the paper is; what the consistency of the tightness is between rolls; what metal and grit is left in the paper pulp when the paper is made. The only good way to find that stuff out before we put it into an engine is to test the TP roll.
The idea of testing is pretty simple. Making predictions used to be done by sacrificing a goat or a chicken and reading the entrails. Some people used tea leaves, some threw yarrow stalks, and some people used Tarot cards. All of these methods shared the results that occasionally they would be right. However, people found out that if they actually subjected something to repeatable testing, they could make predictions about the things that they tested that were right much more than sometimes. An engineer that's right no better than sometimes get fired - and fast, before he gets someone killed. Testing is what the modern scientific world has developed for making predictions, and it works remarkably well where it's done correctly.
So to predict how well a filter will work, we ought to test it, that being the best predictive tool we have available. There happen to be predefined, standardized tests for filter effectiveness as a function of particle size.
Given sufficient money and time, rolls of TP could be tested for particulate efficiency, chemical composition, and winding tightness, and for other things that matter to filters. And we would know what performance that one roll of TP that we tested would have delivered in the car before we ran the tests and ruined it for car service.
With even more money and time, we could run a statistically valid sample of TP rolls from that maker and come up with a good set of results that statistics say are usable as predictors of the performance of that run of TP, including the likely average and statistical spread of performance. That's starting to be predictable.
With even more money and time we could run statistically valid samples of many lots of TP and make valid predictions of the performance for that TP maker (at least within that paricular brand.) If this was kept up in some way, it would be a useful predictor, and if I had that kind of information, especially if the someone, preferably the company that makes the stuff, was legally liable for telling me the truth about what the TP does, I'd flip right over to using TP given only that the numbers were good enough. It would then not be an athletic contest - I would know, within bounds, what performance the next roll of TP would give me.
So far, I have not found even one test-data report of the filtering efficiency of even one roll of toilet paper. The closest I have found is that some folks determine whether their oil is clean by how clear it looks or by whether it wipes off their fingers without leaving a stain. This is not a statistically valid test, needless to say.
Just to illustrate, you point out that a TP filter will remove lamp black from oil in one pass. Is that because the filter nets out the carbon granules, or because the carbon granules are chemically or electrically attracted to the TP fibers? Or did some chemical reaction inside the filter eat the carbon? I've seen and done chemistry experiments showing "magical" clearing of solutions from opaque to clear by adding one little bottle of reagent. Who's to say that if you contaminated oil with 1 micron particles of steel, silicon, or rust the particles would not pass through?
One could *test* for whether the carbon, steel, rust, silicon, whatever is going through the filter. That would be the beginning of knowledge.
For me, until I get that kind of information on an ongoing basis, the filtering efficiency of TP must remain an interesting natural phenomenon. One roll is great. Does that mean the next roll is great? In my mind no. With my luck, I'll get a run of uniformly bad rolls. I'd rather have predictably good than unpredictably great where my engine's lubrication is concerned.
So additional anecdotes about TP filtering efficiency are not very useful to me. I believe some of them are great. I believe some are awful. And I have no rational basis for selecting the good from the awful. Anecdotal evidence is NOT a rational basis for decisions. TP may be the most effective oil filter on the planet, and I'm going to miss out on it because I cannot find predictive testing data to support that. So be it.
So Ralph, thank you very much for your help, your experience, and your and willingness to share it with us. I appreciate your insights. But I cannot compare TP in an apples-to-apples fashion to industrial filters at this time. And until that is possible, I cannot change my opinion of the use of TP for filtering the oil for my car's engine.
So please - tell me what the beta profile of a TP filter is; better yet, tell me a statistically valid set of data on a sampled set of TP. But you don't need to tell me more anecdotes - I've already read a large number of them to come to my current opinion. More of them will not change my opinion.
- Status
