Millbank Bag flow test:
For a while I have been curious about how one millbank bag seems to flow quickly and another does not seem to flow at all. It seems understandable that there would be a variation in the tightness of the weave and this could be due to an infinite variety of reasons including quality, condition or age of the cotton or even the particular 'loom' it was made on. The guideline '5 minute filter' for a litre that is printed on all these bags sometimes seems at best optimistic and as you will see from the test a whole range of factors come into play to produce a result both in terms of time and also filtering performance.
[SIZE=+0]What I am trying to understand is, if there is a practical average one should expect - where the flow is not so quick that it indicates ineffective filtration (over-large 'pore' size) and it being so slow (with a very tight weave) that it becomes impractical: dying of boredom or dehydration might become a real issue![/SIZE]
For the test I took 13 Millbank bags from various 'eras' - 1945 to 1996 - both issued and unissued at various grades/conditions.
I soaked all the bags for the same period and in this case I settled upon an overnight soak. This of course is not a realistic scenario. The reason for leaving it so long was for several reasons. I wanted to test at 'maximum swell'; due to the long nature of the test I needed to ensure that they had all soaked for a comparative length of time thereby ensuring fairly even performance. If I was to quickly immerse them in a bucket and then pull them out one by one (after a bit of squeezing as instructed) it would have meant that some would have had a relatively short immersion period of a few minutes whilst others a proportionally much longer soak of up to several hours. This would mean that the soaking times relative to each other would be of a wide proportional range. Leaving overnight would mean that the test could be undertaken consecutively and that they would have all been in for similarly 'long' time and an hour's difference would not have made a significant difference. On the downside the resulting long soak meant that the Millbanks bag fibres were significantly more swollen than would be the case when the bags was used from dry with a quick soak. This resulted in a very stiff material and consequently a much finer filtering/slower flow.
Soak times are an interesting point and important to make. The first observation and assumption from this test is that the longer you soak it, the more swollen the fibres, the tighter the weave, the more effective it will filter particles out but will be much slower in doing so. This is quite an important factor to bear in mind.
Another related factor that is important to note is that using the bag for the first time on a trip (let us suggest that it is a multi-day overland trip) the flow time will be quicker as the fibre will not be as swollen. If the bag is left damp/wet between uses the fibres may continue to absorb moisture and swell...making subsequent filterings longer but possibly also more effective. It also makes sense that the bags will also become more clogged with trapped organic matter - which might suggest increased micro-biological contamination or more effective filtering of organic matter!
So...to the test itself.....
The objective of the test: to understand the nature of the variance in performance between various Millbank Bags and ascertain which might be the best compromise in terms of flow/duration.
Main assumption: the greater the flow the courser the filtration, the slower the flow the finer the filtration.
Test Observations:
Bag Year
A 1992 B 1992 C 1945 D 1996 E 1993 F 1992 G 1983 H 1993 I 1974 J 1974 K 1993 L 1990 M 1996
All the Millbank bags were loaned by Kevin Endicott of Endicott's Army Surplus - www.endicotts.co.uk
For a while I have been curious about how one millbank bag seems to flow quickly and another does not seem to flow at all. It seems understandable that there would be a variation in the tightness of the weave and this could be due to an infinite variety of reasons including quality, condition or age of the cotton or even the particular 'loom' it was made on. The guideline '5 minute filter' for a litre that is printed on all these bags sometimes seems at best optimistic and as you will see from the test a whole range of factors come into play to produce a result both in terms of time and also filtering performance.
[SIZE=+0]What I am trying to understand is, if there is a practical average one should expect - where the flow is not so quick that it indicates ineffective filtration (over-large 'pore' size) and it being so slow (with a very tight weave) that it becomes impractical: dying of boredom or dehydration might become a real issue![/SIZE]
For the test I took 13 Millbank bags from various 'eras' - 1945 to 1996 - both issued and unissued at various grades/conditions.
I soaked all the bags for the same period and in this case I settled upon an overnight soak. This of course is not a realistic scenario. The reason for leaving it so long was for several reasons. I wanted to test at 'maximum swell'; due to the long nature of the test I needed to ensure that they had all soaked for a comparative length of time thereby ensuring fairly even performance. If I was to quickly immerse them in a bucket and then pull them out one by one (after a bit of squeezing as instructed) it would have meant that some would have had a relatively short immersion period of a few minutes whilst others a proportionally much longer soak of up to several hours. This would mean that the soaking times relative to each other would be of a wide proportional range. Leaving overnight would mean that the test could be undertaken consecutively and that they would have all been in for similarly 'long' time and an hour's difference would not have made a significant difference. On the downside the resulting long soak meant that the Millbanks bag fibres were significantly more swollen than would be the case when the bags was used from dry with a quick soak. This resulted in a very stiff material and consequently a much finer filtering/slower flow.
Soak times are an interesting point and important to make. The first observation and assumption from this test is that the longer you soak it, the more swollen the fibres, the tighter the weave, the more effective it will filter particles out but will be much slower in doing so. This is quite an important factor to bear in mind.
Another related factor that is important to note is that using the bag for the first time on a trip (let us suggest that it is a multi-day overland trip) the flow time will be quicker as the fibre will not be as swollen. If the bag is left damp/wet between uses the fibres may continue to absorb moisture and swell...making subsequent filterings longer but possibly also more effective. It also makes sense that the bags will also become more clogged with trapped organic matter - which might suggest increased micro-biological contamination or more effective filtering of organic matter!
So...to the test itself.....
The objective of the test: to understand the nature of the variance in performance between various Millbank Bags and ascertain which might be the best compromise in terms of flow/duration.
Main assumption: the greater the flow the courser the filtration, the slower the flow the finer the filtration.
Test Observations:
- The first litre (due to water column pressure) flows quicker than the second litre. The ratio of the flow is 2:1 to 3:1 or the first litre flows two-three times quicker on average than the second litre.
Times for the first litre ranged from 15 seconds to 20 minutes with the average being 6.8 minutes. If you took away the highest and lowest values and then averaged this then this would be 3.7 minutes. So we have a practical range of somewhere between 3.7 and 6.8 minutes (with 5.25 minutes being the mid-point) - Times for the second litre (nearer the black line on the bags) ranged from 2 minutes to 52 minutes with the average being 15.8 minutes. If you took away the highest and lowest values and averaged this like I did for the first litre then this would become 11.4 minutes. So we have a practical range of somewhere between 11.4 and 15.8 minutes (with 13.6 minutes being the mid-point)
- The total average of the 1st and 2nd litre combined (all values included) is 11.25 minutes per litre or if you remove the highest and lowest values then this is 7.5 minutes per litre. It makes sense then that the target zone for a Millbank bag is not 5 minutes but lies somewhere between 7.5 and 11.25 minutes for the filtration of a litre of water that is taken from a point somewhere between the 1st and 2nd litres. The reason why I say this is that it is recommended that you let the Millbank bag flow clear for a while (down to the black line) in order for it to 'rinse' the outside of the bag and to 'set' the fibres in the weave.
- All these values are for 'visually clear, running, stream water with a minimal amount of visual suspended organic matter'. This would roughly equate to a running mountain or moorland stream where the water was collected over a stable substrate (rock, gravel, sand) that has not 'muddied' the water. Time values would increase exponentially for water that has increased amounts of suspended organic matter in it. None of the bags have been tested for a point at which filtration of cloudy water becomes visually clear (or not) because of the infinite variety of particle size and quantity of suspended matter in various waters would render such an test undertaking to impractical proportions.
- This test is imperfect (due to the infinite variation in performance of any set of bags) and no one should rely on the efficacy of its findings. This is especially the case when preparing a safe/hygienic water supply. The water used in this test was visually clear and so no conclusions should be drawn on the size and quantity of organic matter filtered by these bags.
- It should be noted that the resultant water (regardless of its clarity) will NOT necessarily be to safe to drink and it should always be assumed that it is not. Please subsequently purify the filtered water in such a way as to guarantee safety for consumption. Millbank bags - regardless of how finely they filter - should always be classed as course filters and should not be compared to other methods of mechanical/membrane filtration. Some level of organic matter will always pass through these filters and may harbour/shield contaminants and microbiological nasties.
- An assumption (rightly or wrongly) has been made that if is takes longer to filter then it is filtering more finely. However there is a trade off between time and filtering performance.
- My personal conclusion is that (for the tested water) I will be aiming to use bags that filter towards the top end of the averages as a trade off between time and performance this means a time range of 3.7 - 6.8 minutes (5.25 mid point) for the first litre and 11.4-15.8 (13.6 minutes mid point). This will, when combining the two mid-point values for both litres, give a single 'generic' value (for each of litres) 9.5 minutes (per litre) - or twice the indicated average on the label of the Millbank bags.
- I would also consider taking along two bags on a trip (they don't take up much room or weigh very much) and taking one that filters quickly and one that filters slowly - this would keep it flexible to two of the most common scenarios - access to a course filter for 'slightly suspect' and very clear water (such as directly from an exposed spring in some rock) and needing to have some water quickly, to 'fill-and-leave' at at a camp-spot where the water is further 'downstream' and with more organic matter suspended within it but you have time on your hands and can leave it to drip.
- I was amazed that all the bags (bar one) shared the same 'weave' - despite some being 50 years apart. However one of them 'D' from 1996 had a noticeably different weave and had the shortest flow time. It will be interesting to test this one for 'filter power' in part 2 of this test.
Bag Year
A 1992 B 1992 C 1945 D 1996 E 1993 F 1992 G 1983 H 1993 I 1974 J 1974 K 1993 L 1990 M 1996
All the Millbank bags were loaned by Kevin Endicott of Endicott's Army Surplus - www.endicotts.co.uk
Last edited:
