Tweet
I've taken the liberty of reprinting this here due to the wealth of information included. It was written by a filter systems professional and posted on another forum. It references diesel but I think the info applies to us gassers as well.
Reprinted, copied with permission © 2010 - R. Hampel
IN-LINE recreational marine diesel fuel filtration.
Standard industrial and high tech maintenance change-out is one year. Fuel filters are made from 'resinated' cellulose fibers; cellulose easily degrades/digests after time in contact with water - water being a component of MOST fuel oil in a macroscopic or emulsion form, not necessarily 'visible' water. If the tankage is vented to the atmosphere, there *will be* water in the fuel. Once such filter media (eventually) absorbs water it rapidly loses its mechanical strength, due to water absorption, then there is no 'surety' of the retention rating.
A filter rated at 1µM (one 'micrometer' in modern terminology) is almost meaningless, without an efficiency rating that accompanies the rating. The 'best' a resinated cellulose fiber filter media (most common for fuel filtration usage) can retain is 3µM at 97%(wt.) efficiency, polypropylene micro fibers @ 1.2-1.5µM @ 97% .... this would approximately 'translate' to 25µM and 10µM respectively approaching a 100% (absolute) rating basis.
Such a filter rated at 1µM is most probably 'nominally' rated by the manufacturer and I would guess/assume (with the lack of support data) that its removal efficiency would be in the range of 50-80%, meaning in 'filter-speak that it could pass a basketball sized particle and still be 'rated' at 1µM - called a 'nominal' rating. Filter ratings without accompanying 'efficiency of rating' are meaningless and leaves the 'consumer' very vulnerable to 'shysterism'. Typically fuel/oil filters 'should' be correlated to ASTM/SAE methods or at least to the Oklahome State University (OSU) "F2" test .... a measure of influent particles vs. effluent particles vs. particle size distribution ..... otherwise a manufacturer can 'tout' meaningless ratings, etc. A lot of cheap and poor performing 'asian' filters are dumped on the market this way. Resinated Cellulose filter media for fuel filtration is now essentially a 'monopoly' from essentially ONE world source, so the above statements should be considered to be quite valid.
Most marine diesel engine manufacturers seem have chosen 15µM-17% (97%) as their 'choice' for the small engine mounted 'guard filter'. The typical Racor filter-train proceeding this would be 30µM (95-97%) ---> 10µM (95-97%) ----> engine 'guard' 15-17µM (95-97%) .... so it seems that they have defined that the 'most damaging' particle for such equipment is ~20µM (100%) - a pretty common engine industry accepted standard.
It takes 'work' to do filtration. A 1µM (97%) will take 10 times the 'work' to do the same job as a 10µM (97%) filter AND will have 10 TIMES the differential pressure drop across it at the SAME flow rate; or it will have 1/10th the flow rate at the SAME operating pressure differential .... plus it will 'plug' at approximately ~30 times (or more) faster than a 10µM.
So, if someone erroneously thinks that a smaller (1 or 3µM) is somehow 'better', then they MUST increase the surface area to keep the differential pressure across the filter at the same level as the 10µM - 10 times the surface area for the 1µM and 3.5 times the surface area for the 3µM ---- if they dont do this then, in typical/poor vacuum motive service on marine diesels, they risk the premature breakage their lift pumps.
Recirc-polishing filtration for recreational marine diesel service.
Recirculation polishing filtration is a 'mathematical trick' use to DILUTE the number of resident particles in a fuel tank. It involves constantly recirculating a LARGE amount of fuel through relatively 'open' (low flow resistance) filters whose ratings are (should be) approximately 3-10X the µM size of the particle*µM level desired 'in the tank'. The operative scientific concept follows "Newton's Law of Exponential Decay" ... in this specific case 'exponential dilution' of particles IN THE TANK.
Filter considerations for 'recirculation' filtration.
Filters used for fuel oil filtration are typically made from micro-fibers --- so one, by analogy, is essentially filtering out 'golfballs with sticks' .... and a lot a golfballs do get through on the 'first pass' through the filter. Such filter media contains a 'broad distribution/retention' of µM (statistical 'pores') ratings and most 'so-called pores' are larger than the target particle µM rating!!!! ... and a quite a few 'capture sites' that are smaller than the 'target particle' or µM 'rating'.
The typical efficiency of a resinated cellulose filter - the most common for fuel filtration- would be: Filter manufacturer rating = 15µM. Removal efficiency (approximate/guesstimate) = 30µM-100%, 20µM-99%, 15µM-95%, 10µM-80%, 5µM 50%, 1µM-15%, 0,5µM-5%. In nature, the smaller the particle the exponentially MORE the amount/number of particles.
As one can see the above 15µM rated filter example does have 'some' removal capacity of 0.5µM; but at quite low %. So if one filtered with the above 15µM filter the total tank contents 20 times (20X). 5% + 5% + 5% .... 5% + 5% , one would ultimately get 4E15 'reduction' of particles AT 0,5µM (and greater µM).... but that is by filtering from one tank to another, filtering the entire contents of the one tank, then 'reversing' the filtration each time; in recirculation filtration you are simply 'diluting' the particles in the tank ... meaning that there will ALWAYS be 'some' (but low numbers) particles in the tank!!!!!
In recirculation filtration "Tank Turnover" (how fast the entire tank volume can be filtered over and and over again) is THE factor for the efficiency of 'recirculation filtration'. A filter located on a 'return line' of ~50 cc per minute (0.013 gallons per minute - ~3/4 Gallon per HOUR) ISNT going to turnover any tank ---- as such would take 67 hours to result in ONE 'turnover' of a 50 gallon tank ... and @ 5% (removal) of .75 GPH only 'cleans up' ~140 cc worth of Oil and there is still 50 - 3/4 = 49.25 gallons of 'unfiltered oil' still in the tank' .... the DILUTION RATE of particles is toooooo damn slow if one erroneously decides to the put a filter on the 'return line'. With 'return line' filtration it would take ~1350 HOURS to get down 'reasonable' particle distribution (0,5µM+) !!!!!!!!!!!!!!! Alternatively (and I recommend) If one would pump at 1.5 gallons per MINUTE (90 GPH per 50 gallon tank) on a totally separate circuit one would 'turnover' the entire tank contents every 40 minutes. 40min X 20 times (a 5% reduction 'per pass') = 13 hours to result in virtually NO particles greater than 0,5µM in the tank .... and as the filter become 'dirtier' the retention becomes more efficient (using dirt to filter dirt') ..... and using a (cheap) 15µM rated filter to do the job!!!!!! A 1µM Filter has 1/15th the flow of a 15µM filter so using a 1µM filter in the above example would take approximately 13 hours X 15 = 195 hours (for ONE single 'turnover').... because the pump is now 'straining' against the differential pressure of the 1µM filter (15 times the differential pressure to operate the 1µM in comparison to the 15µM).
What causes most tank contamination? Answer: Fungus !!!!! Typically a 'resin forming' fungus which uses the OIL as its nutrient source - Cladosporium Resinae, etc. The more the water content (doesnt have to be 'visible' water) the faster the metabolism/growth of the fungus. Such fungus 'particles' are 'gelatenous' (cell structure and water) and are 'soft'/stickly particles. Such particles can easily 'slime over' the outer surface of a pleated filter, which reduces the effective surface area of the filter ..... potential sudden flow stoppage. Such 'soft' particles are easily 'extrudable' ... can deform and easily pushed through a filter under increasing differential pressure across a filter - and (some) can 'coalesce' back to original size downstream of the filter. Such particles dont 'burn well' and (I claim) are the probable chief cause of 'coke' in exhaust portion of the engine. The larger the surface area the 'better' when filtering 'deformables', keeps the 'velocity IN the filter down.
HighTurnover Recirculation polishing systems, quickly removes the SPORES of these fungals, quickly removes any 'free' fungals from the oil, keeps the 'particle distribution' at acceptable levels, .... and MOST IMPORTANT will QUICKLY restore the particle levels back to 'normal' if a large segment of the fungal deposition (living and dead cells + products of metabolism, etc.) on the WALLS breaks loose into the oil (Tank "recovery" after 'slugs of debris'). Such a system needs HIGH TURNOVER rate, an 'open flow' (low ∆P across the filter - so the pump has little resistance).
My personal system: Main: 30µM ---> 10µM ---> 3 gallon 'day tank' ---->15µM 'engine guard"; Recirculation Filtration (100 gallon tank): 15 µM 'depth type' 2.75"Ø X 10" Length @ 3.5 Gallons per MINUTE pump (pressure feed to filter !!!!!!). Filter housing and 'cartridge is 'standard industrial' - allows usage of 'common' filters from multi-MULTIPLE sources, cheaper. Pump (Walbro transfer pump) is wired to engine panel, so that Im recirculating oil anytime my engine is on. My oil is always 'visibly' CLEAR* indicating that the oil is at submicronic µM levels. I never ever 'challenge' my main in-line filters (changed yearly, and I use upwards of ~500-1000+ gal. per yr.).
Note*: Clear oil= oil sample in clear glass container which is held between my eyeball and STRONG white light (sun) .... if there is no 'haze', then I can assume that the particle distribution is ~ sub-micronic; if haze, then I run the recirc. system when in port until 'clear'.
I still need to thoroughly mechanically CLEAN-OUT the tank every ~3 years; although, Im starting to 'prefer' Starbrite Tank Cleaner™. Keeps the particle 'nucleation sites' and spores at LOW numbers, as such 'particles' agglomerate or 'grow' !!!!
Note/suggestion: avoid usage of 'toilet paper' or 'paper towels': the cellulose fibers are/is 'unbound'/unresinated and may/will break loose and enter the flow stream; and, the FLOW RATE though such is 'very small' (low 'tank turnover') with HIGH operating differential pressure (extrusion risk of 'deformables').
Reprinted, copied with permission © 2010 - R. Hampel
IN-LINE recreational marine diesel fuel filtration.
Standard industrial and high tech maintenance change-out is one year. Fuel filters are made from 'resinated' cellulose fibers; cellulose easily degrades/digests after time in contact with water - water being a component of MOST fuel oil in a macroscopic or emulsion form, not necessarily 'visible' water. If the tankage is vented to the atmosphere, there *will be* water in the fuel. Once such filter media (eventually) absorbs water it rapidly loses its mechanical strength, due to water absorption, then there is no 'surety' of the retention rating.
A filter rated at 1µM (one 'micrometer' in modern terminology) is almost meaningless, without an efficiency rating that accompanies the rating. The 'best' a resinated cellulose fiber filter media (most common for fuel filtration usage) can retain is 3µM at 97%(wt.) efficiency, polypropylene micro fibers @ 1.2-1.5µM @ 97% .... this would approximately 'translate' to 25µM and 10µM respectively approaching a 100% (absolute) rating basis.
Such a filter rated at 1µM is most probably 'nominally' rated by the manufacturer and I would guess/assume (with the lack of support data) that its removal efficiency would be in the range of 50-80%, meaning in 'filter-speak that it could pass a basketball sized particle and still be 'rated' at 1µM - called a 'nominal' rating. Filter ratings without accompanying 'efficiency of rating' are meaningless and leaves the 'consumer' very vulnerable to 'shysterism'. Typically fuel/oil filters 'should' be correlated to ASTM/SAE methods or at least to the Oklahome State University (OSU) "F2" test .... a measure of influent particles vs. effluent particles vs. particle size distribution ..... otherwise a manufacturer can 'tout' meaningless ratings, etc. A lot of cheap and poor performing 'asian' filters are dumped on the market this way. Resinated Cellulose filter media for fuel filtration is now essentially a 'monopoly' from essentially ONE world source, so the above statements should be considered to be quite valid.
Most marine diesel engine manufacturers seem have chosen 15µM-17% (97%) as their 'choice' for the small engine mounted 'guard filter'. The typical Racor filter-train proceeding this would be 30µM (95-97%) ---> 10µM (95-97%) ----> engine 'guard' 15-17µM (95-97%) .... so it seems that they have defined that the 'most damaging' particle for such equipment is ~20µM (100%) - a pretty common engine industry accepted standard.
It takes 'work' to do filtration. A 1µM (97%) will take 10 times the 'work' to do the same job as a 10µM (97%) filter AND will have 10 TIMES the differential pressure drop across it at the SAME flow rate; or it will have 1/10th the flow rate at the SAME operating pressure differential .... plus it will 'plug' at approximately ~30 times (or more) faster than a 10µM.
So, if someone erroneously thinks that a smaller (1 or 3µM) is somehow 'better', then they MUST increase the surface area to keep the differential pressure across the filter at the same level as the 10µM - 10 times the surface area for the 1µM and 3.5 times the surface area for the 3µM ---- if they dont do this then, in typical/poor vacuum motive service on marine diesels, they risk the premature breakage their lift pumps.
Recirc-polishing filtration for recreational marine diesel service.
Recirculation polishing filtration is a 'mathematical trick' use to DILUTE the number of resident particles in a fuel tank. It involves constantly recirculating a LARGE amount of fuel through relatively 'open' (low flow resistance) filters whose ratings are (should be) approximately 3-10X the µM size of the particle*µM level desired 'in the tank'. The operative scientific concept follows "Newton's Law of Exponential Decay" ... in this specific case 'exponential dilution' of particles IN THE TANK.
Filter considerations for 'recirculation' filtration.
Filters used for fuel oil filtration are typically made from micro-fibers --- so one, by analogy, is essentially filtering out 'golfballs with sticks' .... and a lot a golfballs do get through on the 'first pass' through the filter. Such filter media contains a 'broad distribution/retention' of µM (statistical 'pores') ratings and most 'so-called pores' are larger than the target particle µM rating!!!! ... and a quite a few 'capture sites' that are smaller than the 'target particle' or µM 'rating'.
The typical efficiency of a resinated cellulose filter - the most common for fuel filtration- would be: Filter manufacturer rating = 15µM. Removal efficiency (approximate/guesstimate) = 30µM-100%, 20µM-99%, 15µM-95%, 10µM-80%, 5µM 50%, 1µM-15%, 0,5µM-5%. In nature, the smaller the particle the exponentially MORE the amount/number of particles.
As one can see the above 15µM rated filter example does have 'some' removal capacity of 0.5µM; but at quite low %. So if one filtered with the above 15µM filter the total tank contents 20 times (20X). 5% + 5% + 5% .... 5% + 5% , one would ultimately get 4E15 'reduction' of particles AT 0,5µM (and greater µM).... but that is by filtering from one tank to another, filtering the entire contents of the one tank, then 'reversing' the filtration each time; in recirculation filtration you are simply 'diluting' the particles in the tank ... meaning that there will ALWAYS be 'some' (but low numbers) particles in the tank!!!!!
In recirculation filtration "Tank Turnover" (how fast the entire tank volume can be filtered over and and over again) is THE factor for the efficiency of 'recirculation filtration'. A filter located on a 'return line' of ~50 cc per minute (0.013 gallons per minute - ~3/4 Gallon per HOUR) ISNT going to turnover any tank ---- as such would take 67 hours to result in ONE 'turnover' of a 50 gallon tank ... and @ 5% (removal) of .75 GPH only 'cleans up' ~140 cc worth of Oil and there is still 50 - 3/4 = 49.25 gallons of 'unfiltered oil' still in the tank' .... the DILUTION RATE of particles is toooooo damn slow if one erroneously decides to the put a filter on the 'return line'. With 'return line' filtration it would take ~1350 HOURS to get down 'reasonable' particle distribution (0,5µM+) !!!!!!!!!!!!!!! Alternatively (and I recommend) If one would pump at 1.5 gallons per MINUTE (90 GPH per 50 gallon tank) on a totally separate circuit one would 'turnover' the entire tank contents every 40 minutes. 40min X 20 times (a 5% reduction 'per pass') = 13 hours to result in virtually NO particles greater than 0,5µM in the tank .... and as the filter become 'dirtier' the retention becomes more efficient (using dirt to filter dirt') ..... and using a (cheap) 15µM rated filter to do the job!!!!!! A 1µM Filter has 1/15th the flow of a 15µM filter so using a 1µM filter in the above example would take approximately 13 hours X 15 = 195 hours (for ONE single 'turnover').... because the pump is now 'straining' against the differential pressure of the 1µM filter (15 times the differential pressure to operate the 1µM in comparison to the 15µM).
What causes most tank contamination? Answer: Fungus !!!!! Typically a 'resin forming' fungus which uses the OIL as its nutrient source - Cladosporium Resinae, etc. The more the water content (doesnt have to be 'visible' water) the faster the metabolism/growth of the fungus. Such fungus 'particles' are 'gelatenous' (cell structure and water) and are 'soft'/stickly particles. Such particles can easily 'slime over' the outer surface of a pleated filter, which reduces the effective surface area of the filter ..... potential sudden flow stoppage. Such 'soft' particles are easily 'extrudable' ... can deform and easily pushed through a filter under increasing differential pressure across a filter - and (some) can 'coalesce' back to original size downstream of the filter. Such particles dont 'burn well' and (I claim) are the probable chief cause of 'coke' in exhaust portion of the engine. The larger the surface area the 'better' when filtering 'deformables', keeps the 'velocity IN the filter down.
HighTurnover Recirculation polishing systems, quickly removes the SPORES of these fungals, quickly removes any 'free' fungals from the oil, keeps the 'particle distribution' at acceptable levels, .... and MOST IMPORTANT will QUICKLY restore the particle levels back to 'normal' if a large segment of the fungal deposition (living and dead cells + products of metabolism, etc.) on the WALLS breaks loose into the oil (Tank "recovery" after 'slugs of debris'). Such a system needs HIGH TURNOVER rate, an 'open flow' (low ∆P across the filter - so the pump has little resistance).
My personal system: Main: 30µM ---> 10µM ---> 3 gallon 'day tank' ---->15µM 'engine guard"; Recirculation Filtration (100 gallon tank): 15 µM 'depth type' 2.75"Ø X 10" Length @ 3.5 Gallons per MINUTE pump (pressure feed to filter !!!!!!). Filter housing and 'cartridge is 'standard industrial' - allows usage of 'common' filters from multi-MULTIPLE sources, cheaper. Pump (Walbro transfer pump) is wired to engine panel, so that Im recirculating oil anytime my engine is on. My oil is always 'visibly' CLEAR* indicating that the oil is at submicronic µM levels. I never ever 'challenge' my main in-line filters (changed yearly, and I use upwards of ~500-1000+ gal. per yr.).
Note*: Clear oil= oil sample in clear glass container which is held between my eyeball and STRONG white light (sun) .... if there is no 'haze', then I can assume that the particle distribution is ~ sub-micronic; if haze, then I run the recirc. system when in port until 'clear'.
I still need to thoroughly mechanically CLEAN-OUT the tank every ~3 years; although, Im starting to 'prefer' Starbrite Tank Cleaner™. Keeps the particle 'nucleation sites' and spores at LOW numbers, as such 'particles' agglomerate or 'grow' !!!!
Note/suggestion: avoid usage of 'toilet paper' or 'paper towels': the cellulose fibers are/is 'unbound'/unresinated and may/will break loose and enter the flow stream; and, the FLOW RATE though such is 'very small' (low 'tank turnover') with HIGH operating differential pressure (extrusion risk of 'deformables').
Comment