Non-destructive Ruffles® Bypass Valve testing

[Hatsuwr]

So you have no baseline control to compare to. Sloppy test methodology and won't give any real good comparison if you're just comparing bad ones to bad ones. The "best one" you test may still be way worse than a 100% "good one" baseline. Baselines exist in testing for a reason. Don't know why you can't seem to grasp that you need a known "good one" baseline to make any real sense in the testing.

I think you are forgetting the context of this post, which is non-destructive testing. Obviously you can do a lot more with destructive testing, but then that would be a different topic. I'm guessing that's why you are under the impression that I'm not understanding you.

You wouldn't want to choke it down too far which may do what you described, so maybe block half the inlet holes. I think with the vacuum level higher inside the filter, if the leak gap around the media is present then it would effect the level of vacuum more on the guage compared to if the vacuum level was at a very low level to start with. So essentially making your vacuum gauge more sensitive to a gap leak path. Try it and see if you're going to play around with this "science project".

What is being compared is the pressure differential across the filter media and bypass. The idea is that this restriction will be lessened (and therefore the differential lowered) if there are leaks. We can assess the difference in pressure differential at different flow rates, and the chart will look something like below. As you can see, at higher flow rates, the difference between the two pressure differentials is larger, which will be easier to detect. Adding more restriction elsewhere will reduce flow rates, and make the difference harder to detect.

Yes, the air is at ATM pressure before it flows into the base holes - doesn't matter after the air flows into the inlet holes. So that will be a constant (as well as the ATM air temperature) if you did comparative testing withing a short time frame. In other words, not one filter tested today and the other tested a week from now when the ATM pressure & temp may be slightly different due to weather conditions.

The air between the inlet holes and the filter media/bypass will be a bit below atmospheric pressure due to the restriction created by the inlet holes, which is what is relevant for testing the filter media and bypass. It doesn't matter much for comparative testing, but it's worth mentioning for accuracy.

 

[ZeeOSix]

I think you are forgetting the context of this post, which is non-destructive testing. Obviously you can do a lot more with destructive testing, but then that would be a different topic. I'm guessing that's why you are under the impression that I'm not understanding you.

You'd only have to destroy one filter to create a baseline known non-leaker filter to compare the other tested filters to. If you don't have a baseline to compare too, then the test is mostly useless. You don't know how a known good filter will behave, so without that you can't determine anything about the other tested filters except that one "might" be leaking more than the others tested.

What is being compared is the pressure differential across the filter media and bypass. The idea is that this restriction will be lessened (and therefore the differential lowered) if there are leaks. We can assess the difference in pressure differential at different flow rates, and the chart will look something like below. As you can see, at higher flow rates, the difference between the two pressure differentials is larger, which will be easier to detect. Adding more restriction elsewhere will reduce flow rates, and make the difference harder to detect.

Is this a positive displacement vacuum pump? Most cheap vacuum pumps typically are.

^^^ In the graph above, the green line would essentially be the curve if the vacuum level was raised higher by making the inlet holes a bit more restrictive by blocking some of them. The slope of the curve is higher compared to the red line, so any change in vacuum level due to a leaky leaf spring should show up better on the vacuum gauge - ie, make the vacuum gauge a bit more sensitive to restriction changes inside the filter.

If you could calculate/estimate the flow rate of the air through the vacuum pump (if positive displacement), it might be possible to get an estimate on the percentage of leakage at the leaf spring, but only if you have a known non-leaking baseline filter to compare to.

The air between the inlet holes and the filter media/bypass will be a bit below atmospheric pressure due to the restriction created by the inlet holes, which is what is relevant for testing the filter media and bypass. It doesn't matter much for comparative testing, but it's worth mentioning for accuracy.

Sure, but all you care about is what you see on the vacuum gauge. In every test, the inlet pressure on the inlet holes will be the ATM pressure acting on the inlet holes. As said before, if you test multiple filters in a short time period, the ATM pressure and temp will be a constant.

 

[Hatsuwr]

You'd only have to destroy one filter to create a baseline known non-leaker filter to compare the other tested filters to. If you don't have a baseline to compare too, then the test is mostly useless. You don't know how a known good filter will behave, so without that you can't determine anything about the other tested filters except that one "might" be leaking more than the others tested.

You'd have to destroy one filter every time. It's a completely different scenario. The title of the thread is non-destructive for a reason. If you are going to be opening filters, then there are many better ways to test things like this (like, just looking at the filter lol).

Is this a positive displacement vacuum pump? Most cheap vacuum pumps typically are.

^^^ In the graph above, the green line would essentially be the curve if the vacuum level was raised higher by making the inlet holes a bit more restrictive by blocking some of them. The slope of the curve is higher compared to the red line, so any change in vacuum level due to a leaky leaf spring should show up better on the vacuum gauge - ie, make the vacuum gauge a bit more sensitive to restriction changes inside the filter.

If you could calculate/estimate the flow rate of the air through the vacuum pump (if positive displacement), it might be possible to get an estimate on the percentage of leakage at the leaf spring, but only if you have a known non-leaking baseline filter to compare to.

I think it would be considered a multi stage centrifugal. But you are misunderstanding the graph. The lines are for the differential pressure across the media/bypass of two filters. One less leaky (green), and one more leaky (red). The difference between them increases in magnitude as airflow across the restriction increases. Adding inlet restriction that decreases the airflow rate would decrease the difference between the two measurements.

Sure, but all you care about is what you see on the vacuum gauge. In every test, the inlet pressure on the inlet holes will be the ATM pressure acting on the inlet holes. As said before, if you test multiple filters in a short time period, the ATM pressure and temp will be a constant.

I'm not really concerned about what atmospheric pressure is since, as I explained, it makes no sense to try and compare results separated by any significant amount of time. This type of test only makes sense for batch testing unless you start controlling many more variables.

 

[ZeeOSix]

You'd have to destroy one filter every time. It's a completely different scenario. The title of the thread is non-destructive for a reason. If you are going to be opening filters, then there are many better ways to test things like this (like, just looking at the filter lol).

No you wouldn't need to destroy a filter ever time. Once you have a known non-leaking baseline filter that becomes the control filter that others tested are compared to. The ones you test to compare to the baseline are the ones that are not destroyed. Without a known baseline as a control, then measurements are kind of useless expect to compare unknown configurations with other unknown configurations.

I think it would be considered a multi stage centrifugal. But you are misunderstanding the graph. The lines are for the differential pressure across the media/bypass of two filters. One less leaky (green), and one more leaky (red). The difference between them increases in magnitude as airflow across the restriction increases. Adding inlet restriction that decreases the airflow rate would decrease the difference between the two measurements..

Re: bold sentence. The less restriction there is across the filter, regardless if there's a leak gap involved, the lower the vacuum level will be, which would make any changes due to the flow paths going on inside the filter harder to detect on the gauge reading. If the restriction was near zero, the vacuum gauge would essentially be near zero, and you couldn't tell if anything is going on - not enough gauge resolution. All I'm trying to say is if the vacuum level is made higher across the filter that it may be easier to see a change on the gauge due to a leak path bypassing the media. Test it out to see if that's the case or not ... test to validate theory.

I'm not really concerned about what atmospheric pressure is since, as I explained, it makes no sense to try and compare results separated by any significant amount of time.

That's basically what I've said many times about the ATM side of the flow ... it's not a factor to be concerned about unless you are comparing tests ran on different days and the barometric pressure and air temperature is a lot different. But I don't think this test is sensitive enough to even see an impact from ATM conditions changing much.

This type of test only makes sense for batch testing unless you start controlling many more variables.

This test only makes sense if you have a known non-leaking baseline filter to compare all other tested filters to.

 

[Hatsuwr]

No you wouldn't need to destroy a filter ever time. Once you have a known non-leaking baseline filter that becomes the control filter that others tested are compared to. The ones you test to compare to the baseline are the ones that are not destroyed. Without a known baseline as a control, then measurements are kind of useless expect to compare unknown configurations with other unknown configurations.


This test only makes sense if you have a known non-leaking baseline filter to compare all other tested filters to.

You would have to destroy a filter every batch to get a new baseline. Unless you can keep all significant variables static, this test weeks from now has no meaningful comparison to this test today. The baseline is useless for non-destructive testing like this, and arguing against a necessarily non-destructive test by suggesting it could be better by being destructive is just silly.

Re: bold sentence. The less restriction there is across the filter, regardless if there's a leak gap involved, the lower the vacuum level will be, which would make any changes due to the flow paths going on inside the filter harder to detect on the gauge reading. If the restriction was near zero, the vacuum gauge would essentially be near zero, and you couldn't tell if anything is going on - not enough gauge resolution. All I'm trying to say is if the vacuum level is made higher across the filter that it may be easier to see a change on the gauge due to a leak path bypassing the media. Test it out to see if that's the case or not ... test to validate theory.


That's basically what I've said many times about the ATM side of the flow ... it's not a factor to be concerned about unless you are comparing tests ran on different days and the barometric pressure and air temperature is a lot different. But I don't think this test is sensitive enough to even see an impact from ATM conditions changing much.

Yes, but the restriction must be in the component being measured (the filter and bypass), not external to it (the inlet holes). Maximizing flow rate by minimizing external restrictions is how you get that maximized pressure differential across the filter/bypass.

Today's been a bit busier than expected, but I think I'll be able to do the tests this evening. I plan to do three runs of each filter to ensure repeatability is being preserved, and I'll compare a pair with only one inlet hole open too, just to be sure.

 

[ZeeOSix]

You would have to destroy a filter every batch to get a new baseline.

I was thinking you are just going to do this for one brand/model. Hey, being a tester sometimes isn't cheap., lol.

Unless you can keep all significant variables static, this test weeks from now has no meaningful comparison to this test today.

I've said that all along, so don't know why you keep saying what I've already said many times.

The baseline is useless for non-destructive testing like this, and arguing against a necessarily non-destructive test by suggesting it could be better by being destructive is just silly.

The baseline isn't useless. I really don't get why you think you can get any meaningful information when you don't have a known non-leaking filter as a baseline. Only the baseline filter would be sacrificial in order to create a baseline to compare all others of the same brand and model to.

Yes, but the restriction must be in the component being measured (the filter and bypass), not external to it (the inlet holes). Maximizing flow rate by minimizing external restrictions is how you get that maximized pressure differential across the filter/bypass.

In a vacuum situation like in your test setup, you get whatever flow the vacuum pump is going to flow based on the restriction it's trying to draw air through. If the flow restriction is increased, the vacuum level goes up. If the flow restriction is reduced, the vacuum level goes down as you elude to above ... but do you want the gauge in the mud, or higher?

If there was no flow restriction, the vacuum level would be near zero. If the gauge is in a higher vacuum range, it may be a bit more sensitive to the effect of a leak gap on the resulting vacuum level seen ... that's all I'm saying here. Maybe the change in vacuum level due to a leak would be the same delta on the gauge, regardess of where the vacuum level is sitting. Only experimentation would tell you for sure, and you would need a non-leaker in the same inlet hole block-off configurtion as a baseline to know.

Today's been a bit busier than expected, but I think I'll be able to do the tests this evening. I plan to do three runs of each filter to ensure repeatability is being preserved, and I'll compare a pair with only one inlet hole open too, just to be sure.

You could experiment by blocking 1 then 2, then 3 etc inlet holes to see how the vacuum gauge reacts. But you'd never know how the readings on a leaking filter would ever compare to one that's not a leaker. That's true no matter what you do without a baseline non-leaker filter to compare to.

 

[Hatsuwr]

Alright, first results are in. There will be more to come though.

Three FE3600s were measured for differential pressure across the filter media and bypass. The first is marked in red, the second in green, and the third in blue. Three measurements were taken of each, and then the first one was measured a fourth time to ensure repeatability.

I need to figure out an easier way of displaying the results, but this will work for now. I have marked the centerline of the needle for each measurement. I was unsure if I'd be able to achieve the resolution to see differences between filters on a needle gauge rather than with a transducer, but there is very clear grouping. I'm very happy with the consistency, especially considering my setup.

My interpretation of this is that filter 1 (red) is able to achieve the highest pressure differential, and therefore likely has the least leakage through tears, bad glue, bypass, etc. Again, this assumes that the media presents a relatively consistent restriction across the filters.

Filter 2 (green) performed significantly worse compared to the other two, and I suspect is has much greater combined leakage paths. If judging based only on this test, it seems clear that filter 1 should be the choice among the three to install on your vehicle.

As a side note, this test would work much better on a smaller filter. You would be able to achieve a higher pressure differential, and the same size leak on a smaller filter would be responsible for a larger portion of the pressure differential change.

 

[ZeeOSix]

^^^ But what would a known non-leaking filter register at? Even the one you think is the best of the three may be leaking like a sieve compared to one that doesn't leak at all.

If you do the same test with some inlet holes blocked on each filter, do you see the same vacuum delta amount between filters?

 

[Hatsuwr]

Oh, almost forgot dates!

Filter 1: 2023, 178th day
Filter 2: 2023, 022nd day
Filter 3: 2023, 178th day

Interesting that filters 1 and 3 were made on the same day and have nearly identical results, while the older filter 2 performed worse. There were no visible differences in the construction of the three.

 

[CharlieBauer]

the best of the three may be leaking like a sieve

Time for the oil test!

I think that a consistently measured amount of oil with a timer for how long it takes to disappear could then be mathematically converted into how much the gap would leak at operating temperature and pressure.

 

[CharlieBauer]

Oh, almost forgot dates!

Filter 1: 2023, 178th day
Filter 2: 2023, 022nd day
Filter 3: 2023, 178th day

Interesting that filters 1 and 3 were made on the same day and have nearly identical results, while the older filter 2 performed worse. There were no visible differences in the construction of the three.

The last digit on the Fram date code is the shift. First one is the plant code which is probably A on all of them.

 

[Hatsuwr]

The last digit on the Fram date code is the shift. First one is the plant code which is probably A on all of them.

The full codes for the FEs are 23178AF, 23022AF, and 23178AF. The XGs below all start with A though. Also, oil tests are next! Just want to complete all of the vacuum tests first.


Here are three XG3600s, with the same method as above. Filter 1 is red, filter 2 green, and filter 3 blue. Dates are:

Filter 1: A40462 - 2024, 046th day
Filter 2: A40462 - 2024, 046th day
Filter 3: A41413 - 2024, 141st day

Again there is no overlap in the results, the two filters manufactured on the same day have close results, and the newer filter performed the best. There was no visible difference in the construction of the filters.

With these filters, there was a resonance from the ADBVs. Filter 1 was nearly silent, while filter 3 was very loud. 2 was somewhere in between.

 

[ZeeOSix]

Interesting that filters 1 and 3 were made on the same day and have nearly identical results, while the older filter 2 performed worse. There were no visible differences in the construction of the three.

Can't tell what's going on inside by looking at the outside. If the leaf spring gap is a little different between them, then there will be a slight difference in vacuum level between them. And who knows what the difference would be on one that's known not to leak ... it could be substantial, or not ... only way to know is to test it.

 

[CharlieBauer]

The full codes for the FEs are 23178AF, 23022AF, and 23178AF. The XGs below all start with A though. Also, oil tests are next! Just want to complete all of the vacuum tests first.


Here are three XG3600s, with the same method as above. Filter 1 is red, filter 2 green, and filter 3 blue. Dates are:

Filter 1: A40462 - 2024, 046th day
Filter 2: A40462 - 2024, 046th day
Filter 3: A41413 - 2024, 141st day

Again there is no overlap in the results, the two filters manufactured on the same day have close results, and the newer filter performed the best. There was no visible difference in the construction of the filters.

With these filters, there was a resonance from the ADBVs. Filter 1 was nearly silent, while filter 3 was very loud. 2 was somewhere in between.

View attachment 242512

Ah yes, the Champ made Frams have a different date code format.

I can't tell from the close ups - which filter is more restrictive overall? FE or XG?

 

[Hatsuwr]

Ah yes, the Champ made Frams have a different date code format.

I can't tell from the close ups - which filter is more restrictive overall? FE or XG?

I should have included that, will do when I summarize everything. The FEs sit between about 10.5-11.5 kPa, while the XGs are 12-13 kPa. So the XGs are more restrictive to air, but I don't know if that ranking would definitely remain when using oil.

 

[CharlieBauer]

The FEs sit between about 10.5-11.5 kPa, while the XGs are 12-13 kPa.

Based on the inHg gauge lines, just before you posted this, I calculated that the difference between the highest and lowest in each batch is 0.1psi.

Your spread of 1kpa puts it at 0.145 psi.

When Ascent tested the Royal Purple (supposedly the same as the FE), unused it had a pressure differential of 9psi. So the differences you are seeing between highest and lowest are 1% to 1.6% of the PD of a similar filter that performed as expected.

 

[ZeeOSix]

I should have included that, will do when I summarize everything. The FEs sit between about 10.5-11.5 kPa, while the XGs are 12-13 kPa. So the XGs are more restrictive to air, but I don't know if that ranking would definitely remain when using oil.

Two things could account for that: 1) The media is not the same, and 2) The media area is not the same.

 

[ZeeOSix]

So the differences you are seeing between highest and lowest are 1% to 1.6% of the restriction of a similar filter that performed as expected.

But nobody knows how much each one is leaking at the leaf spring. That throws a whole big wrench into all this unless you know how a non-leaker performs.

 

[CharlieBauer]

But nobody knows how much each one is leaking at the leaf spring. That's throws a whole big wrench into all this unless you know how a non-leaker performs.

I agree which is why I recommend the oil test. But the other thing is, we're unlikely to find out how a non leaker performs because they don't make them anymore.

 

[ZeeOSix]

I agree which is why I recommend the oil test. But the other thing is, we're unlikely to find out how a non leaker performs because they don't make them anymore.

I've posted a dozen times in this thread how you could know a how non-leaker performs ... make one.