"multi-pass" beta test for oil particle filter 3

[electricpete]

I'm a little bit confused by the terminology "multi-pass" beta test. I thought the beta ratio was ratio of input/output at a given particle size, which seems like it should inherently be a once-through measure.

Why do they call it a "multi-pass" test? If we use it on a once-through application such as filtering while transferring oil into a component, can we expect the same reduction as identified by the "multi-pass" beta rating?

=====================================
(2B)+(2B)' ?

 

[MikeHalloran]

The multipass beta rating exists because a lot of commercially successful filters are crap.

... or at least, not very successful at cleaning oil in a once-through process. The beta rating only has meaning in a 'polishing' system, where a pump recirculates the same oil from a tank, through the filter, and back to the tank. ... for tens of hours.

For once-through, as in a transfer pump system, you need a filter with an 'absolute' rating. Expect a fairly high pressure drop, or a physically large filter, to get a usable flow.





Mike Halloran
Pembroke Pines, FL, USA

 

[electricpete]

Are you sure about that?
Most places I read just suggest the “multi-pass method” is some kind of improved method for determining beta ratio. For example here:

http://www.filtrationtips.com/2005_09_08.htm

It would seem like a complete redefinition of beta to define it as something like an initial concentration over a final concentration after hours of filtering, rather than upstream/downstream concentration.


=====================================
(2B)+(2B)' ?

 

[tomwalz]

I believe Mike Halloran understated the situation in his initial comment.

We build filter systems and have found filter rating systems to be worse than useless. In fifteen years the only reliable method we have found is experimentation in an actual production environment. Even then, when we sell a filter system without first analyzing the customer’s fluid, we sell the filter system with the advisory that we may have to try more than one kind of filter to get optimal results.

We have not found a system that reliably describes filters to the point where we could switch suppliers safely based on nothing but the description.

However a similar situation applies with fluids. Two fluids with an identical designation may have significantly varying properties. Even if the properties are identical in the new fluid the properties can vary widely after a significant amount use.

Even in the lab it is difficult to get complete particle removal of a given size in a single pass.

When you consider the properties of the material comprising the filter, the manner in which these materials are used and the highly irregular nature of most particles it provides insight into the problems in single pass filtering.

Many filters involve the use of a flexible material and this flexible material can be distorted to allow particle passage. Whether a filter is felted, wound or assembled in another manner there will be some passages that will allow larger particles to pass through than is the case with other passages. In addition most rating systems specify that the filter will trap a certain percentage particles. This percentage may be 95%, 98% or something similar. In my opinion this is a highly suspicious figure.

The assumption is usually made that the particles are spherical however I do not believe I have ever seen a truly spherical particle unless it was introduced intentionally. These particles can assume an infinite variety of shapes and can present themselves in different aspects. Thus the long slender particle can pass through as though it were one micron in size the first time and be trapped because it presents itself as being many microns on the second pass.


Thomas J. Walz
Carbide Processors, Inc.

http://www.carbideprocessors.com

Good engineering starts with a Grainger Catalog.

 

[electricpete]

Thanks Tom. I understand there a tremendous number of variables and uncertainty. I am really just trying to get at the basic intent of the number beta (measured using the multi-pass standard).

Let's say I have a 5 micron, beta = 200 filter (determined using multi-pass method).

Do I expect to see roughly:
A - The number of particles >=5 micron on filter outlet is 1/200 of the number on the filter inlet
OR
B - A tank initially with concentration X will be reduced to concentration X/200 after some large unspecified (?) number of passes through the filter?

(which is closer to the intent?)

=====================================
(2B)+(2B)' ?

 

[MikeHalloran]

Pete, do not confuse filter math, which is mostly hyperbole, with the precise math associated with electronics.

In testing any fluid filter with size rating given in microns, and some beta ratio, you can expect to see, downstream of the filter, many particles larger than the filter's rating,
>including many particles that came from the filter itself<.

That last is called "media migration", and it plagues even, maybe especially, certain brands that advertise themselves as particularly effective/ clean/ borderline magical.



Mike Halloran
Pembroke Pines, FL, USA

 

[electricpete]

So I understand there are many caveats and precautions and variables to consider. Before discussing those, I would like to simply understand the basic idea of what beta is supposed to represent.

In some ideal or historical sense, beta was intended as an estimate or measure of ratio of contaminant on input to output, correct?

I think it's true, especially after I have read some more, but I was initially confused by the name "multi-pass" and Mike added to my confusion in his response by saying it represents a ratio achievable only after many many passes of the same fluid.

Here they discuss the multi-pass test.

http://www.machinerylubrication.com/Read/564/filter-beta-ratios

They also show Beta = Upstream/Downstream.

They provide a detailed discussion of how it is calculated. It seems geared to ratio upstream/downstream.

Why the name multi-pass? That is sort of my original question. I see here they discuss the fact that you circulate the fluid in a loop with CONTINUOUS INJECTION OF CONTAMINANT.

http://standards.sae.org/j1858_200206/

So, the fluid has to traverse the loop many times to do the test which may be why it's called multipass (right?), but that is not saying that if you didn't continuously inject contaminant you would have to circulate many times to achieve that ratio.

I think I understand it, but if you guys would like to acknowledge that I'd feel better.

=====================================
(2B)+(2B)' ?

 

[MikeHalloran]

The continuous injection of contaminant obscures the fact that the typical filter doesn't work very well on a single pass, and requires many passes to clean a fixed volume of fluid with no contaminant added. ... which would be a better test of filters in service, but would make the math more difficult.

Note also that the 'standard'test dust' is literally stuff that GM sweeps up off its Arizona proving ground. ... probably not the sort of thing you'd expect as a calibration standard.




Mike Halloran
Pembroke Pines, FL, USA

 

[electricpete]

So, the meaning of multi-pass is that we set up a steady state closed-loop system with contaminant injection to allow prolonged testing and evaluation of the once-through parameter beta.

If in the field we do once-through filtering, to the extent that our field conditions resemble the once-through conditions mimicked by the multipass test (including all the errors associated with fluid type, release from media etc), our best estimate of expected removal of the contaminant of size X (for example 5 micron) is beta (for example 200). True?

=====================================
(2B)+(2B)' ?

 

[MikeHalloran]

Something like that, yes, to the extent that you trust the filter manufacturer.

If you want the fluid really clean, you won't get there in one pass of a 'depth' filter, e.g. one that contains fibers like paper or cloth or toilet paper.
You have a chance with a 'membrane' filter comprising microscopic holes in an impervious sheet, but you need a lot of filter area or a lot of drive pressure or both to get a flow rate that you'd like for transfer duty.

E.g., I'd suggest transferring fluid with a simple gear pump or similar, equipped with a depth filter to trap whatever large pieces fall out of the pump.

If you want the transferred fluid to also be clean, follow up by polishing the fluid in the target tank by recirculating it repeatedly through a depth filter until you can measure a low particle count. If you don't have access to a particle counter, you have to either trust the filter manufacturer (not recommended) or use a membrane filter, once through, somewhere in the transfer process and play solitaire while the transfer proceeds.



Mike Halloran
Pembroke Pines, FL, USA