People have discussed how oil pressure sensors on car engines are measuring viscosity too (at a given RPM). Does anybody know if any engine computers out there use pressure, temperature, and RPM sensor info to estimate viscosity? Sounds like a good idea.
Viscosity Measurement by Temperature-Pressure-RPM
- Thread starter fredfactory
- Start date
- Status
Originally Posted By: fredfactory
People have discussed how oil pressure sensors on car engines are measuring viscosity too (at a given RPM). Does anybody know if any engine computers out there use pressure, temperature, and RPM sensor info to estimate viscosity? Sounds like a good idea.
I'm not sure any sensors in an engine are measuring viscosity directly. An oil pressure sensor measures the resistance to flow in the oiling system, often reported in pounds per square inch (PSI) in the United States, or in kilopascals (kPa) in other parts of the world.
I don't think an estimation (or a direct measurement) of actual viscosity is possible. And I'm not sure that I can think of a situation where it'd be helpful, either. The viscosity grade is usually recommended or chosen to provide a certain operating oil pressure, and as long as that operating oil pressure is sufficient for the duty, then the viscosity grade selection is a good one.
People have discussed how oil pressure sensors on car engines are measuring viscosity too (at a given RPM). Does anybody know if any engine computers out there use pressure, temperature, and RPM sensor info to estimate viscosity? Sounds like a good idea.
I'm not sure any sensors in an engine are measuring viscosity directly. An oil pressure sensor measures the resistance to flow in the oiling system, often reported in pounds per square inch (PSI) in the United States, or in kilopascals (kPa) in other parts of the world.
I don't think an estimation (or a direct measurement) of actual viscosity is possible. And I'm not sure that I can think of a situation where it'd be helpful, either. The viscosity grade is usually recommended or chosen to provide a certain operating oil pressure, and as long as that operating oil pressure is sufficient for the duty, then the viscosity grade selection is a good one.
MolaKule
Staff member
Originally Posted By: fredfactory
People have discussed how oil pressure sensors on car engines are measuring viscosity too (at a given RPM). Does anybody know if any engine computers out there use pressure, temperature, and RPM sensor info to estimate viscosity? Sounds like a good idea.
No vehicle sensors i know of measure viscosity.
What pressure, rpm, and temperature relationships would you use to determine this?
People have discussed how oil pressure sensors on car engines are measuring viscosity too (at a given RPM). Does anybody know if any engine computers out there use pressure, temperature, and RPM sensor info to estimate viscosity? Sounds like a good idea.
No vehicle sensors i know of measure viscosity.
What pressure, rpm, and temperature relationships would you use to determine this?
Originally Posted By: MolaKule
No vehicle sensors i know of measure viscosity.
What pressure, rpm, and temperature relationships would you use to determine this?
At a given temperature, the Barus equation might work:
I'm wondering if engine computers out there determine viscosity from pressure-temperature-rpm information.
Possible Scenario: An engine computer gets oil pressure data, and then some kind of clues about oil temperature (either estimate temp from coolant temperature and running time with ambient air temperature, or measure it directly). It also gets RPM data. From all that, the engineers have calculated a lookup-table of oil pressure values expected, or, alternatively, after the oil is changed fresh, the engine computer gathers pressure info at various temperature (mostly hot) and RPMs to build a table over a week or so of driving. Then, if the pressures are too low, within some tolerance, it could be a sign the oil has been fuel-diluted and/or the VII broken up too much with age. If the pressure is too high, it might be a sign viscosity is too high. Note: Its probably important to calibrate itself with fresh oil at the start of an oil change interval to establish a comparative baseline.
No vehicle sensors i know of measure viscosity.
What pressure, rpm, and temperature relationships would you use to determine this?
At a given temperature, the Barus equation might work:
I'm wondering if engine computers out there determine viscosity from pressure-temperature-rpm information.
Possible Scenario: An engine computer gets oil pressure data, and then some kind of clues about oil temperature (either estimate temp from coolant temperature and running time with ambient air temperature, or measure it directly). It also gets RPM data. From all that, the engineers have calculated a lookup-table of oil pressure values expected, or, alternatively, after the oil is changed fresh, the engine computer gathers pressure info at various temperature (mostly hot) and RPMs to build a table over a week or so of driving. Then, if the pressures are too low, within some tolerance, it could be a sign the oil has been fuel-diluted and/or the VII broken up too much with age. If the pressure is too high, it might be a sign viscosity is too high. Note: Its probably important to calibrate itself with fresh oil at the start of an oil change interval to establish a comparative baseline.
I'm not even saying an engine computer even has to implement a Barus equation directly. The Barus just tells us oil has different pressures at different viscosities (at a given temperature). Just establish a baseline using fresh oil over a week or two of driving, then re-check oil pressure data once a month or so (at various temperatures and RPMs) to see if the oil has thinned or thickened too much.
BMW already uses a resistance-capacitance sensor in the oil to determine oil quality, although a quality loss is defined as some combination of contaminants and/or viscosity changes which the electrical sensors pick up, so close but not quite what I'm asking.
BMW already uses a resistance-capacitance sensor in the oil to determine oil quality, although a quality loss is defined as some combination of contaminants and/or viscosity changes which the electrical sensors pick up, so close but not quite what I'm asking.
Last edited:
In my (Subjective) experience, the relationship between oil pressure and viscosity is not always what we might predict.
Sometimes, in the same engine, lower viscosity oil yields higher gauge pressure. Other times, the opposite is true. Too many variables in different designs, of engines, suction screens, oil pumps etc.
Sometimes, in the same engine, lower viscosity oil yields higher gauge pressure. Other times, the opposite is true. Too many variables in different designs, of engines, suction screens, oil pumps etc.
As per my bearing tutorial (for want of a better word), bearing side leakage is affected by viscosity....demonstrated in this paper where a main bearing was removed from the engine oil supply and provided with oil of varying viscosity, coming up with this result.
For a given supply pressure, oil consumption through the bearing was related to the HTHS (not KV in non newtonian oils).
The non Newtonian bit needs further analysis. A straight Non VIIed oil has no difference between it's kinematic and HTHS behaviour...a multigrade with VII has two behaviours depending on shear rate.
The test has to be under high shear rates, 10^6/sec or thereabouts...engine in the above test as can be seen, depending on oil to be in the second Newtonian phase at around 3,000RPM.
Your viscometer needs to have a built in understanding of this.
CATERHAM's premise is the inverse of the above, in that he's using a pump of constant supply volume (he needs the relief closed, as per his assertions), and measuring the backpressure generated from the bearings. He is getting positive correlation between the HTHS and the oil pressure thus obtained, and I beleive that his test is quite accurate.
However, if you are using the viscosities recommended by the OEM, there's a pretty good chance that the oil relief is flowing, and depending on design, may have a lot of squirters, cam phasors and things that are sending volume through pressure/density, and KV low shear flows.
My question for your in service viscosity checking tool is what can you do with it...if you are monitoring for fuel dilution and permanent shear, I can see some merit.
CATERHAM asserts that as long as your oil pressure is higher than the OEM min oil pressure (for their recommeneded grade, in an engine considered "OK") that the viscosity is adequate...keep dropping viscosity until you've got minimum OEM oil pressure.
I maintain that the viscosity measurement obtained IS sound, but that tells you nothing about the behaviour of the loaded side of the bearings, pistons etc. as you don't have the manufacturer's understanding on where on the Stribeck curve your engine started...Minimum OEM oil pressure on their designed lubricants is a condemn limit, not a target.
For a given supply pressure, oil consumption through the bearing was related to the HTHS (not KV in non newtonian oils).
The non Newtonian bit needs further analysis. A straight Non VIIed oil has no difference between it's kinematic and HTHS behaviour...a multigrade with VII has two behaviours depending on shear rate.
The test has to be under high shear rates, 10^6/sec or thereabouts...engine in the above test as can be seen, depending on oil to be in the second Newtonian phase at around 3,000RPM.
Your viscometer needs to have a built in understanding of this.
CATERHAM's premise is the inverse of the above, in that he's using a pump of constant supply volume (he needs the relief closed, as per his assertions), and measuring the backpressure generated from the bearings. He is getting positive correlation between the HTHS and the oil pressure thus obtained, and I beleive that his test is quite accurate.
However, if you are using the viscosities recommended by the OEM, there's a pretty good chance that the oil relief is flowing, and depending on design, may have a lot of squirters, cam phasors and things that are sending volume through pressure/density, and KV low shear flows.
My question for your in service viscosity checking tool is what can you do with it...if you are monitoring for fuel dilution and permanent shear, I can see some merit.
CATERHAM asserts that as long as your oil pressure is higher than the OEM min oil pressure (for their recommeneded grade, in an engine considered "OK") that the viscosity is adequate...keep dropping viscosity until you've got minimum OEM oil pressure.
I maintain that the viscosity measurement obtained IS sound, but that tells you nothing about the behaviour of the loaded side of the bearings, pistons etc. as you don't have the manufacturer's understanding on where on the Stribeck curve your engine started...Minimum OEM oil pressure on their designed lubricants is a condemn limit, not a target.
You can make a chart (labor!), using the variables like at such rpms at such temps = such psi ... You gonna need the gauges and make a lotta notes, though.
Shannow, what do you thing of bearing holes enlargements? Would it make less film, and more flow, if any?
I mean like this
https://www.youtube.com/watch?v=1iPKr2vT59M
Wouldn't be needed a higher volume/flow pump to give same pressures?
I mean like this
https://www.youtube.com/watch?v=1iPKr2vT59M
Wouldn't be needed a higher volume/flow pump to give same pressures?
Last edited:
I've just never heard of an engine computer doing this. Seems easy. Has anybody heard of this built in to the OLM system?
In a given engine, measure oil pressure, get say 40 psi. Now get temperature and RPM and look up what the pressure "should" be insid the engine computer's memory, based on new oil (at the start of the last oil change). If it "should" be 50 psi, then its low by 10 psi, which the engineers have pre-determined is too much, so a warning light to change the oil comes on, alerting the driver of a fuel dilution problem.
I think a BMW resistance-capacitance oil sensor CAN do this (sense a viscosity change), not sure, so maybe they already have a "fuel-dilution alert" built-in to their engine computer already.
Since most cars don't have the BMW-style sensor, then maybe they can do it with oil pressure sensors, temperature sensors, RPM sensors alone.
In a given engine, measure oil pressure, get say 40 psi. Now get temperature and RPM and look up what the pressure "should" be insid the engine computer's memory, based on new oil (at the start of the last oil change). If it "should" be 50 psi, then its low by 10 psi, which the engineers have pre-determined is too much, so a warning light to change the oil comes on, alerting the driver of a fuel dilution problem.
I think a BMW resistance-capacitance oil sensor CAN do this (sense a viscosity change), not sure, so maybe they already have a "fuel-dilution alert" built-in to their engine computer already.
Since most cars don't have the BMW-style sensor, then maybe they can do it with oil pressure sensors, temperature sensors, RPM sensors alone.
If the engine had all electronic gages accurate, it could be a secondary calculation for the ECM. Possible, yes.
Originally Posted By: Pontual
You can make a chart (labor!), using the variables like at such rpms at such temps = such psi ... You gonna need the gauges and make a lotta notes, though.
Thats a great practical idea. That way a person can do what the engine computer might do if the software programmers at the car companies included that feature in the OLM algorithms.
You can make a chart (labor!), using the variables like at such rpms at such temps = such psi ... You gonna need the gauges and make a lotta notes, though.
Thats a great practical idea. That way a person can do what the engine computer might do if the software programmers at the car companies included that feature in the OLM algorithms.
"When no stress is applied, there is no viscosity loss.
When the stress is infinitely high, the viscosity instantly
drops to the base oil viscosity."
If I have that right, adding 5W50 to 5W40 may lower the
base oil viscosity and possibly the HTHS with it.
When we use oil pressure as evidence of a preconceived
condition that the bearings are experiencing, there
are many variables.
For fredfactory's idea to work, the engine designer and
software programmer would have to work with a lubrication
team and come up with a spec engine oil.
The paper from which the quote I posted brings up some interesting points.
Blending VII containing engine oils does not yield predictable results,
even when using the same VII of the same chemical composition.
If I have that right, adding a liter of 0W40 to our engine
that was programmed for 5W40, even of the same brand, would
throw the algorithm out beyond the mathematical viscosity
blend of the two would predict.
Engine oil pressure follows bulk oil temperature up and
down, because that is the viscosity the pump sees.
Oil temperature first and pressure second, at the
bearing furthest away from the pump and pressure gauge,
should be the location where our minimum operating
conditions are measured.
Applying the quote from the 1st paragraph, this is
what I extracted from it;
When blending two engine oils together to achieve an
imagined finished product, the result may be more
predictable by adding a mono grade (Newtonian) oil
to a multi-grade (Non-Newtonian) oil instead of blending
two multi-grade engine oils together.
And;
The application of low viscosity engine oils such as
5W16 that will likely be Newtonian, may have a higher
base oil viscosity than 0W20 or 5W20.
If the big-end bearing is HTHS and not KV100 dependent,
then the lighter SAE grades should impose a risk to
plain bearings in an engine.
When the stress is infinitely high, the viscosity instantly
drops to the base oil viscosity."
If I have that right, adding 5W50 to 5W40 may lower the
base oil viscosity and possibly the HTHS with it.
When we use oil pressure as evidence of a preconceived
condition that the bearings are experiencing, there
are many variables.
For fredfactory's idea to work, the engine designer and
software programmer would have to work with a lubrication
team and come up with a spec engine oil.
The paper from which the quote I posted brings up some interesting points.
Blending VII containing engine oils does not yield predictable results,
even when using the same VII of the same chemical composition.
If I have that right, adding a liter of 0W40 to our engine
that was programmed for 5W40, even of the same brand, would
throw the algorithm out beyond the mathematical viscosity
blend of the two would predict.
Engine oil pressure follows bulk oil temperature up and
down, because that is the viscosity the pump sees.
Oil temperature first and pressure second, at the
bearing furthest away from the pump and pressure gauge,
should be the location where our minimum operating
conditions are measured.
Applying the quote from the 1st paragraph, this is
what I extracted from it;
When blending two engine oils together to achieve an
imagined finished product, the result may be more
predictable by adding a mono grade (Newtonian) oil
to a multi-grade (Non-Newtonian) oil instead of blending
two multi-grade engine oils together.
And;
The application of low viscosity engine oils such as
5W16 that will likely be Newtonian, may have a higher
base oil viscosity than 0W20 or 5W20.
If the big-end bearing is HTHS and not KV100 dependent,
then the lighter SAE grades should impose a risk to
plain bearings in an engine.
Originally Posted By: fredfactory
I'm not even saying an engine computer even has to implement a Barus equation directly. The Barus just tells us oil has different pressures at different viscosities (at a given temperature). Just establish a baseline using fresh oil over a week or two of driving, then re-check oil pressure data once a month or so (at various temperatures and RPMs) to see if the oil has thinned or thickened too much.
fredfactory, I didn't get the Barus stuff, as I was typing, eating honey crumpets (honey is thixotropic, which is an interesting shear rate related phenomenon, opposite to what we are discussing), and having a morning cup of tea, all of which took about 20 minute...post overlap.
Barus is the change in viscosity with pressure, not a representation of oil pressure versus viscosity...and the difference between the KV of the oil in your sump and atmospheric is nearly nothing.
Take the mythical 10cst oil at 100C that 101 is premised on, and an 80 psi oil pressure (550,000 pa)
The pressure viscosity co-efficient is 16 or 17 ... x 10^-9.
Therefore going from 10cst in the sump at ambient, to 80psi in the gallery, the KV is increased by 0.094cst, a little less than a percent.
The effect that I was discussing was CATERHAM's "rotary viscometer", where pressure is an artifact of viscosity, bearing design and viscosity...although he does confuse pressure viscosity co-efficient occasionally.
I'm not even saying an engine computer even has to implement a Barus equation directly. The Barus just tells us oil has different pressures at different viscosities (at a given temperature). Just establish a baseline using fresh oil over a week or two of driving, then re-check oil pressure data once a month or so (at various temperatures and RPMs) to see if the oil has thinned or thickened too much.
fredfactory, I didn't get the Barus stuff, as I was typing, eating honey crumpets (honey is thixotropic, which is an interesting shear rate related phenomenon, opposite to what we are discussing), and having a morning cup of tea, all of which took about 20 minute...post overlap.
Barus is the change in viscosity with pressure, not a representation of oil pressure versus viscosity...and the difference between the KV of the oil in your sump and atmospheric is nearly nothing.
Take the mythical 10cst oil at 100C that 101 is premised on, and an 80 psi oil pressure (550,000 pa)
The pressure viscosity co-efficient is 16 or 17 ... x 10^-9.
Therefore going from 10cst in the sump at ambient, to 80psi in the gallery, the KV is increased by 0.094cst, a little less than a percent.
The effect that I was discussing was CATERHAM's "rotary viscometer", where pressure is an artifact of viscosity, bearing design and viscosity...although he does confuse pressure viscosity co-efficient occasionally.
Originally Posted By: Shannow
As per my bearing tutorial (for want of a better word), bearing side leakage is affected by viscosity....demonstrated in this paper where a main bearing was removed from the engine oil supply and provided with oil of varying viscosity, coming up with this result.
For a given supply pressure, oil consumption through the bearing was related to the HTHS (not KV in non newtonian oils).
The non Newtonian bit needs further analysis. A straight Non VIIed oil has no difference between it's kinematic and HTHS behaviour...a multigrade with VII has two behaviours depending on shear rate.
The test has to be under high shear rates, 10^6/sec or thereabouts...engine in the above test as can be seen, depending on oil to be in the second Newtonian phase at around 3,000RPM.
Your viscometer needs to have a built in understanding of this.
CATERHAM's premise is the inverse of the above, in that he's using a pump of constant supply volume (he needs the relief closed, as per his assertions), and measuring the backpressure generated from the bearings. He is getting positive correlation between the HTHS and the oil pressure thus obtained, and I beleive that his test is quite accurate.
However, if you are using the viscosities recommended by the OEM, there's a pretty good chance that the oil relief is flowing, and depending on design, may have a lot of squirters, cam phasors and things that are sending volume through pressure/density, and KV low shear flows.
My question for your in service viscosity checking tool is what can you do with it...if you are monitoring for fuel dilution and permanent shear, I can see some merit.
CATERHAM asserts that as long as your oil pressure is higher than the OEM min oil pressure (for their recommeneded grade, in an engine considered "OK") that the viscosity is adequate...keep dropping viscosity until you've got minimum OEM oil pressure.
I maintain that the viscosity measurement obtained IS sound, but that tells you nothing about the behaviour of the loaded side of the bearings, pistons etc. as you don't have the manufacturer's understanding on where on the Stribeck curve your engine started...Minimum OEM oil pressure on their designed lubricants is a condemn limit, not a target.
Shannow, this is the most even-keeled summary of your two positions I have observed. Thank you. I wonder if CATERHAM agrees...
As per my bearing tutorial (for want of a better word), bearing side leakage is affected by viscosity....demonstrated in this paper where a main bearing was removed from the engine oil supply and provided with oil of varying viscosity, coming up with this result.
For a given supply pressure, oil consumption through the bearing was related to the HTHS (not KV in non newtonian oils).
The non Newtonian bit needs further analysis. A straight Non VIIed oil has no difference between it's kinematic and HTHS behaviour...a multigrade with VII has two behaviours depending on shear rate.
The test has to be under high shear rates, 10^6/sec or thereabouts...engine in the above test as can be seen, depending on oil to be in the second Newtonian phase at around 3,000RPM.
Your viscometer needs to have a built in understanding of this.
CATERHAM's premise is the inverse of the above, in that he's using a pump of constant supply volume (he needs the relief closed, as per his assertions), and measuring the backpressure generated from the bearings. He is getting positive correlation between the HTHS and the oil pressure thus obtained, and I beleive that his test is quite accurate.
However, if you are using the viscosities recommended by the OEM, there's a pretty good chance that the oil relief is flowing, and depending on design, may have a lot of squirters, cam phasors and things that are sending volume through pressure/density, and KV low shear flows.
My question for your in service viscosity checking tool is what can you do with it...if you are monitoring for fuel dilution and permanent shear, I can see some merit.
CATERHAM asserts that as long as your oil pressure is higher than the OEM min oil pressure (for their recommeneded grade, in an engine considered "OK") that the viscosity is adequate...keep dropping viscosity until you've got minimum OEM oil pressure.
I maintain that the viscosity measurement obtained IS sound, but that tells you nothing about the behaviour of the loaded side of the bearings, pistons etc. as you don't have the manufacturer's understanding on where on the Stribeck curve your engine started...Minimum OEM oil pressure on their designed lubricants is a condemn limit, not a target.
Shannow, this is the most even-keeled summary of your two positions I have observed. Thank you. I wonder if CATERHAM agrees...
Which reminds me....
In at least two of the papers you posted, there
is a mention of the need to maintain oil pressure
sufficient to prevent gas bubbles from occurring
in the bearing space from the heated oil.
Because, as you have mentioned, the rotational
action of the journal can pull a couple of PSI,
which will lower the boiling point of the oil.
The difference in the crankcase pressure and
oil pressure, and not oil pressure/atmospheric
pressure, (which gauges are zeroed at) will
determine the boiling point.
In at least two of the papers you posted, there
is a mention of the need to maintain oil pressure
sufficient to prevent gas bubbles from occurring
in the bearing space from the heated oil.
Because, as you have mentioned, the rotational
action of the journal can pull a couple of PSI,
which will lower the boiling point of the oil.
The difference in the crankcase pressure and
oil pressure, and not oil pressure/atmospheric
pressure, (which gauges are zeroed at) will
determine the boiling point.
All the discussion above is good, yet we know people report differences in steady-state hot-runnin oil pressures when going from using a 40 weight to a 30 or 20 oil.
-------------------------------
For example, your sump is cool, 1/2 minute after cold start, rev the engine up to 1500 rpm in neutral and oil pressure is high.
-------------------------------
After the engine is up to operating temperature, rev to the same rpm (1500) and notice oil pressure is reading less.
-------------------------------
Therefore, we know the different oil pressure is due to viscosity differences, and the engine computer could use that.
Many readers here know OBDII engine computer fault detection algorithms are pretty sophisticated (EGR checks for example), so why not oil viscosity checks too, using pressure & temperature? I've never heard of an engine computer doing that, and wondered if any do, especially since fuel dilution in DI engines is a worldwide problem.
-------------------------------
For example, your sump is cool, 1/2 minute after cold start, rev the engine up to 1500 rpm in neutral and oil pressure is high.
-------------------------------
After the engine is up to operating temperature, rev to the same rpm (1500) and notice oil pressure is reading less.
-------------------------------
Therefore, we know the different oil pressure is due to viscosity differences, and the engine computer could use that.
Many readers here know OBDII engine computer fault detection algorithms are pretty sophisticated (EGR checks for example), so why not oil viscosity checks too, using pressure & temperature? I've never heard of an engine computer doing that, and wondered if any do, especially since fuel dilution in DI engines is a worldwide problem.
Last edited:
- Status
Similar threads
