PureONE Flow vs. PSID Data from Purolator

Status
Not open for further replies.
Joined
Mar 2, 2009
Messages
2,371
Location
WA
This will be very interesting to the guys with C5 Vettes using the LS1 & LS6 engines, as well as anyone who believes that the PureONE is "restrictive to oil flow".

I own a 2002 Z06 Vette and often rev it to the limits. The LS1 & LS6 engines have the filter bypass valve built into the engine block near the filter mount. I believe the built in bypass valve is set for around 8~10 psi. Since the filter bypass valve is built into the engine, the filters for this car does not have a bypass valve. The PureONE for this engine is the PL14006.

I contacted the Purolator Engineering Dept to ask them if running the PureONE PL14006 at track events could possibly cause the filter bypass valve to open with hot oil at high RPMs when the flow rate is maximized. This is an experpt from my original email:
--------------------------------
"I'm running Mobil 1 5W-30 full synthetic oil, and the viscosity is specified by Mobil as 11.3 cSt @ 100º C. This would be the typical "nominal" hot viscosity. The oil pump in this Vette engine puts out ~10 GPM at 75 psi.

Can the PL14006 flow this maximum hot oil volume of 10 GPM at 11.3 cSt without creating more than 8 psi difference across the filter media? I'm worried that at high RPM the filter bypass (in the engine block) might operate and send unfiltered oil to the engine."

--------------------------------

The Engineer at Purolator actually tested the PL14006 on their flow bench and sent the "Flow vs. PSID" test data, along with some other inputs. This is their response:
--------------------------------
"The test stand was set at 203 degrees F. Note the higher temp will lower the viscosity of the oil even more. So at 212 degrees F the differential pressure of the filter will be even slightly lower. The test oil used is very close to the same viscosity of 11.3cST at the tested temp."
--------------------------------

He sent an Excel file with the test data (shown below), and a plot. Here is the data, so if you want to see the graph just plot it up in Excel.

Flow (GPM) --- Delta P
1 --- 0.4
2 --- 0.7
3 --- 1.1
4 --- 1.5
5 --- 1.9
6 --- 2.3
7 --- 2.7
8 --- 3.2
9 --- 3.6
10 --- 4.1
11 --- 4.7
12 --- 5.2
13 --- 5.8
14 --- 6.4
15 --- 7.0
16 --- 7.6
17 --- 8.3
18 --- 8.9

Here's the PL14006 ... it's not a huge filter (3" dia x 3.4" long).

http://www.purolatorautofilters.net/resources/Popup/Pages/PartDetailPopup.aspx?partnum=PL14006

This is very good information of flow vs PSID with hot oil, as this kind of data is hardly ever seen from the manufactures who have the required test equipment to measure this kind of data.

As you can see, the PureONE is NOT very restrictive to flow.
thumbsup2.gif
It's even less restrictive than I thought it would be ... and certainly much better than the internet rumors that make it sound as though they are "too restrictive" to flow.
 
Great info! Purolator seems to be a very forthcoming company with information, looks like all you have to do is ask!
 
Heck, it's quite clear from the data you presented here ..that from flow alone, no filter, or at least a PureOne ..can POSSIBLY go into bypass at all.
grin2.gif
 
Originally Posted By: Gary Allan
Heck, it's quite clear from the data you presented here ..that from flow alone, no filter, or at least a PureOne ..can POSSIBLY go into bypass at all.
grin2.gif



Based on the data, that seems to be very true. Of course, as the filter loads up it might be a different story. This test data was with a new filter. And as we all know, if the oil pump is modified to put out some insane amount of flow volume then all bets are off. But for this particular PureONE, it certainly looks like its good for lots of flow volume before the PSID gets significantly high.

If the same flow rates and viscosity were put through many different filter brands, it's possible there could be some significant flow vs. PSID performance difference, as shown in the relative flow graph below. I would suspect the Mobil 1 and FRAM Toughguard to also show worse flow performance under the same test hot conditions based on the data shown in this graph.

pi_filt_oil_gold_coldoil_thumb.jpg


Also, based on the size of the PureONE, there would probably be some difference in the hot flow curve, but I wouldn't expect a huge difference ... maybe a few PSID more throughout the upper flow range.

This could change the light on the 12 GPM Subaru flow issue given this data showing that the media in the PureONE flows pretty well. If the size of the filters used on the Subarus is a decent size and the bypass valve in the filter is in the 14 ~ 16 psi range, then it looks like a PureONE would be fine for a high flow rate of hot oil.

Someone with a high flow Subaru should ask Purolator to do the same flow vs. PSID measurement for the PureONE filter specified for the high flow engine.

Without actually seeing the measured data, it was anyone's guess or theory on what the actual PSID might be in some of these filters with high volume hot oil.
 
Nice info. Good on them for taking care of the customer like that!! It almost seems strange?? LOL...

I use this same filter on my LS-1 Z-28.

Slightly OT..

It does confuse me why Purolator specifically does not recommend the Pure-One filters for motorcycles given the good flow it seems to have?? I was thinking flow may have been a reason, guess I was wrong... again! Maybe I'll ask them.

I'll stick with the stock Suzuki filter on the bike (TL 1000s) and Pure 1 on everything else.
 
Originally Posted By: cven
Nice info. Good on them for taking care of the customer like that!! It almost seems strange?? LOL...

I use this same filter on my LS-1 Z-28.


Yes, I was quite surprised to get this data. From the email chain with Purolator, it sounded like they actually went and ran the bench test to gather the data. Maybe they were curious themselves, or didn't have any historical test data for that filter.

Originally Posted By: cven
Slightly OT..

It does confuse me why Purolator specifically does not recommend the Pure-One filters for motorcycles given the good flow it seems to have?? I was thinking flow may have been a reason, guess I was wrong... again! Maybe I'll ask them.

I'll stick with the stock Suzuki filter on the bike (TL 1000s) and Pure 1 on everything else.


Good question on the motorcycle filter applications. I can't remember their exact reasoning, but if it was due to flow or PSID reasons it seems unwarranted. I can't see a motorcycle oil pump putting out any more pressure or flow than a large V8 engine.

As I mentioned earlier above, if Purolator ran the same test on a smaller sized PureONE I'm sure the flow restriction would be slightly more, but I couldn't see it being more than a few PSI on the high end of the flow range.
 
Wow, I am really impressed to find a modern company which takes the time to respond intelligently to a customer's inquiry. So often these days you can't get anything more than a robo-answer from a clerk or an automated "response system".

Makes me glad I've been buying PureOnes as my filter of choice lately.
 
good to know-picked up 4 Pure-1's at AAP with 10 dollar coupon and 8 dollar rebate from Purolator. I've got the same vehicle--02 Z06---was planning on using it only on the Tahoe, but good to know it will work on the Vette as well


Steve
 
I wonder how the flow rate decreases at say 3K, 5K, and 7.5K miles? It would seem that more efficient filtration must mean the PureOnes catch more crud and therefore plug up quicker.
 
+1

Originally Posted By: wgtoys
Wow, I am really impressed to find a modern company which takes the time to respond intelligently to a customer's inquiry. So often these days you can't get anything more than a robo-answer from a clerk or an automated "response system".

Makes me glad I've been buying PureOnes as my filter of choice lately.



+1 I emailed my gas company asking what their policy was on grounding my gas line. The NEC says do if the gas company approves. They emailed back saying ask an electrician. I grounded it, but back under the house where it will be hard to see.
 
Originally Posted By: AuthorEditor
I wonder how the flow rate decreases at say 3K, 5K, and 7.5K miles? It would seem that more efficient filtration must mean the PureOnes catch more crud and therefore plug up quicker.


Per the info on Purolator's website, the PureONE is designed to hold 13 grams of crud. I'm sure Purolator has designed the filter to hold that much crud before the flow restriction becomes an issue.

http://www.purolatorautofilters.net/products/oil_filters/Pages/pureoneoilfilters.aspx

The info on Purolator's website also makes it look like the PureONE is a long OCI filter.

PurolatorOilChart.jpg


As with any high efficiency filter, you probably wouldn't want to run it that long on an engine that's sludged up or going through a cleaning process.
 
I use a Pure One myself on my Subaru Forester but the differential pressure across the filter data provided by Purolator is not what I would have been expecting. The problem I see is that data shows a pressure drop that is linearly proportional to flow rate when I would expect to see a pressure drop that is squared in proportion to flow rate increases. Google the “Darcy–Weisbach Equation” and select the Wikipedia link for the equation. The equation represents the head loss or pressure loss due to friction in a fluid system. This equation is commonly used for piping systems but it can be used to represent the pressure drop across an oil filter. The pressure drop across an oil filter should primarily be due to friction losses.

There are several terms in the equation.

L = Length of pipe
D= Diameter of pipe
g= gravity
f= Darcy friction factor
V= fluid velocity
H=Head Loss or pressure drop due to friction

For an oil filter the terms L, D, g, and f should be fixed or constants. This leaves fluid flow and pressure drop as the only variables. Notice that H increases as VxV or V squared. That means if you double the velocity, the pressure drop should be 4 times as great. The data provided by Purolator does not behave in this way.
The Purolator data show the pressure drop at 8 gpm to be 3.2 psid and at 16 gpm to be 7.6 psid. With a doubling of flow rate the pressure drop only increase by 2.375 times as much not 4 times as much. The predicted pressure drop should be 12.8 psid.
The Purolator data result might be explained if the terms, D or L where not fixed or constants. This would be the case if there was a bypass valve on the filter opening but this filter has no bypass valve. The other possibility might be the resistance to oil flow of the ant-drain back valve might change if it mechanically deforms or flexes as flow increase allowing a wider opening for oil to flow through. I would like to believe the good numbers provided by Purolator by my education in physics and fluid dynamics as well as experience operating nuclear power plants with lots of fluid systems makes me suspicious of the results. To believe them I would have to know more about the test set up or see the test myself. It may be possible I am missing something or making the wrong assumptions about the properties of the filter.

That famous graph showing the pressure drop across different brands of filters with the at different flow rates is also suspicious to me. The Fram ToughGurard pressure drop curve has the expected characteristics. Many of the others do not. The curves should look like a parabola turned on it side, not a straight line.


+1 To SuperBusa recommendation to ask Purolator to run a test on Purolator Subaru oil filter at Subaru engine oil flow rates.
 
I don't see where exponential results should be expected. The non linear results would be expected through orifices, but should, at some point anyway, move away from a linear appearance (2X the rate 2X the pressure) and more toward an exponential appearance (2X the rate 4X the pressure).

In line (as in an integrated filter and not a "swatch" test), the filter is a deceleration zone. You've got transitions that are velocity changes that are going to be evidenced by pressure differentials.

That is, you're going from a (+/-) .6^2" conduit to 100-150^2" of media that then transitions to some complex orifice equivalency. It then decelerates again ..only to accelerate upon exiting.

How much of a linear model appears when you transition from a 2' pipe to a 6' pipe with a 2' orifice mid span ..transitioning back to a 2' pipe ...that ultimately has a 2" outlet? While each section will have it's linear pressure changes, they're not on the same scale/index (I think I'm saying this right to trigger a light bulb in a smarter person).

Those velocity changes are going to be shown in pressure alterations ..but the "chamber" has to fall into something like an inverted "valve constant" in that simple views are confounded.
 
Originally Posted By: hofcat
I use a Pure One myself on my Subaru Forester but the differential pressure across the filter data provided by Purolator is not what I would have been expecting. The problem I see is that data shows a pressure drop that is linearly proportional to flow rate when I would expect to see a pressure drop that is squared in proportion to flow rate increases.


If you plot the data from Purolator, you can see it's not totally linear. It's someplace between linear and squared.

Originally Posted By: hofcat
The Purolator data result might be explained if the terms, D or L where not fixed or constants. This would be the case if there was a bypass valve on the filter opening but this filter has no bypass valve.


The PL14006 doesn't even have a bypass valve. That's on reason I asked Purolator if this filter could be used on the LS6 engine with a high oil flow rate that has the filter bypass valve in the engine block and set to 8 ~ 10 psi I believe (lower than some PureONE filters have inside the filter). So for this data, there is no way that a bypass valve is entering the equation because there isn't one.

Originally Posted By: hofcat
To believe them I would have to know more about the test set up or see the test myself. It may be possible I am missing something or making the wrong assumptions about the properties of the filter.


I fell that the data is pretty accurate. They obviously have a flow bench that is capable of heating oil, and with the proper instrumentation to measure the data (delta P, flow rate and oil temp). It’s far better data than any home grown garage test I’ve seen so far.

Originally Posted By: hofcat
+1 To SuperBusa recommendation to ask Purolator to run a test on Purolator Subaru oil filter at Subaru engine oil flow rates.


Yes, someone should see if Purolator would run the same test. I think the PureONE filter for the hi-flow Subaru's has something like a 14 ~ 16 PSI byapss valve setting ... so my felling now is that at 12 GPM the PSID will not go as high as 14 ~ 16 PSI.
 
Originally Posted By: Gary Allan
I don't see where exponential results should be expected. The non linear results would be expected through orifices, but should, at some point anyway, move away from a linear appearance (2X the rate 2X the pressure) and more toward an exponential appearance (2X the rate 4X the pressure).


The pressure drop across the oil filter is not expected to increase exponentially with flow, but with the square of the flow. That is if you 2X the flow, the pressure differential should increase 4x, if you 3x the flow, the pressure differential should increase 9x. The head loss equation does not have any exponent terms, it is a quadratic equation or a polynomial of the second degree. The exponential and quadratic curves have similar shapes to the eye when plotted but they are different.

Originally Posted By: Gary Allan

In line (as in an integrated filter and not a "swatch" test), the filter is a deceleration zone. You've got transitions that are velocity changes that are going to be evidenced by pressure differentials.

That is, you're going from a (+/-) .6^2" conduit to 100-150^2" of media that then transitions to some complex orifice equivalency. It then decelerates again ..only to accelerate upon exiting.


Sure but if these physical characteristics remain fixed with flow, that is the openings and passages the oil moves through when going through the filter do not change with flow, the resistance of flow through all components in the filter can be modeled with a single constant and the only pressure drop dependency would the squared dependency on oil flow.

Originally Posted By: Gary Allan

How much of a linear model appears when you transition from a 2' pipe to a 6' pipe with a 2' orifice mid span ..transitioning back to a 2' pipe ...that ultimately has a 2" outlet? While each section will have it's linear pressure changes, they're not on the same scale/index (I think I'm saying this right to trigger a light bulb in a smarter person).


But those pressure changes through the sections in piping system you describe are not linearly proportional to the fluid velocity in the section, they are proportional to the square of the fluid velocity. If a section widens to cause flow to be cut in half, the pressure will drop to ¼ the value.

If you take the system you described above, 2” pipe to a 6” pipe with a 2” orifice mid span, back to a 2” pipe, the resistance of the that system to flow from inlet to outlet can be modeled with a single constant as the physical properties are fixed.

If you 2X the flow through that system the pressure loss measured from the inlet to the outlet will increase by 4X. If you 1/2X the flow in half the pressure loss will be ¼X.


Originally Posted By: Gary Allan

Those velocity changes are going to be shown in pressure alterations


Yes and those pressure changes with fluid velocity are not linear, the fluid dynamics equations show they are proportional to the square of the fluid velocity.

Originally Posted By: Gary Allan

..but the "chamber" has to fall into something like an inverted "valve constant" in that simple views are confounded.

Not sure what you mean by this.
 
I'm not disciplined. Hence there is going to be some need for you to "reduce yourself to your lowest terms". You're not talking to a peer.

..but seeing as you've sent Soupy into scratching his head on how Baloney is turning in his grave (Bernoulli) ..and going back to read up so he can assure that what you're saying is what he meant all along ...

It's a good thing
lol.gif


Quote:
If you take the system you described above, 2” pipe to a 6” pipe with a 2” orifice mid span, back to a 2” pipe, the resistance of the that system to flow from inlet to outlet can be modeled with a single constant as the physical properties are fixed.



Sure. Now isolate the effects at the midspan orifice representing the filter ..or the velocity/pressure changes as seen at that point - (a highly debated topic)

Use small words.
grin2.gif


You don't happen to write technical manuals, do you? btw- people don't talk this way. They may teach/instruct/write/learn this way ..but no one talks this way.
 
The Purolator data is not exactly linear, but is close to being linear. It is certainly not proportional to the square of fluid velocity or flow. If it was exactly linear that would be a little suspicious. Real lab measurement results usually have some amount of instrument error.

To given an example of this flow to pressure relationship in fixed fluid systems, look at any of the charts and online calculators used in designing sprinkler systems for the home. Notice how when flow rate doubles on the charts the pressure goes up by 4X.

You exactly make my point about the bypass valve. This filter did not have one and it is not a factor in this case. That is what I was trying to say.

From what you describe they do seem to have the right test rig and even heated the oil to operating temperature. This is far beyond the capabilities of most if not all members of this forum unless they work in the automotive engineering industry. This is what confounds me the most. The data is not what would normally be expected based on basic engineering principals and assumptions. If it is accurate then there is something more complex going on in the filter that has never been explained on any forum or industry paper I have seen. A possible explanation may be the filter media resistance coefficient is not fixed like that of the inlet or exit holes to the filter but is a value that changes with fluid velocity due to some kind distortion of the media. That is the pores in the filter media may some how expand as the fluid velocity increases. This is just speculation on my part. Another effect not discussed yet is transition from laminar flow to turbulent flow. I assumed the flow is always turbulent. I do not see the characteristics of a laminar to turbulent flow transition in the data provided.
 
They do talk this way in engineering circles. It can be a challenge to describe fluid dynamics equations and principals to those who have not worked in this field, but I try.
 
Status
Not open for further replies.
Back
Top