MolaKule
Staff member
ATF Fluid Degradation as a Function of Temperature
by Permission of the author, MoleKule
The charts that every ATF manf. refers in terms of ATF degradation vs temperature were taken primarily from a 2004 study by GM engineers using Dexron-III (H revision) ATF as the test fluid.
Oxidatiion and friction degradation models were developed to predict the fluid's "End-of-Life" or EOL. These models were compared to lab* and fleet test data to determine model correlation, or "correctness."
Oxidation Degradation, End-of-Life (EOL) was defined as when the fluid's DeltaTAN increased by 2.5 during an ABOT test, which is much less severe the CEC test mentioned above. A TAN increase as you might remember is an indication that fluid acidity has risen to levels in which the fluid is no longer oxidatively stable.
Remember, this was a MINERAL-based ATF with none of the advanced, VII, FM or anti-oxidation chemistries that we have available today.
First temps stated are in C and F in "()."
First of all, the chart started at an ABOT temp of 140C(284) and went up to 160C(320).
The chart showed that the fluid at ABOT 140C(284) reached its EOL in about 1800 hours.
The chart showed that the fluid at ABOT 160C(320) reached its EOL in about 500 hours.
SO, a 20C(68) rise in fluid temp caused a 1300 hour decrease in fluid life.
Today's ATF chemistry is much more stable in terms of oxidation, friction, and viscosity.
If in doubt, have the fluid analyzed and or do a pan drain and refill every 30,000 to 50,000 miles.
This not only refreshes the additives, but also reduces particulates.
So how do OEM’s such as GM, Ford, Chrysler and others develop those Predictive models for the PCM, TCM, or Body Module?
In a nutshell, here is how the industry conducts studies in order to make predictions:
1.) An Hypothesis is put forth as to how the system under study behaves in terms of its physics (thermodynamics, mechanics, and chemistry),
2.) A predictive Mathematical model is developed based on the physics of the hypothesized system,
3.) Physical Tests, such as Lab and Fleet tests are executed and data collected,
4.) Feedback (data) from the Physical Tests are compared to the model to see how and if the physical data correlates with the model,
5.) If the correlation between the proposed model and physical data agrees, the predictive (mathematical) model is accepted,
6.) If the correlation between the proposed model and physical data does not agree, the predictive (mathematical) model is modified or redefined.
Correlation (close agreement) of the model with the actual physical data is important since the model should reflect real world physical data in order to develop TCM or PCM algorithms, such as the reflected in the Oil Life Monitors.
For example, the Friction Degradation model might look like this for a V8 powered vehicle with a 4-speed step shift Automatic:
TNS = 87800 – 2455*T – 6630*EPS + 1410*EPS^2 + 82*T*EPS
TNS = Total Number of Shifts,
T = Bulk (Sump) Fluid temperature in C,
EPS = Energy per shift in KiloJoules.
The Energy-Per-Shift model, EPS, might look like this for a 1-to-2 upshift in a V8 powered vehicle with a 4-speed step shift Automatic:
EPS = 0.3145*TPS – 1.4865
Where TPS is the throttle position sensor percentage from 0% to 100%; a 30% TPS is considered average for an up-shift from a standing start.
So these types of equations are compared to sensor inputs and algorithms developed to determine when a fluid needs changing and to turn on the OLM.
*One of the oxidation tests for ATF is the:
CEC L-48-A-00 Oxidation Test (DKA)
Let's make sure we differentiate bulk or TC oil temperatures from clutch and band face temperatures which can spike to over 400F for short periods.
Others are found in the ASTM literature.
by Permission of the author, MoleKule
The charts that every ATF manf. refers in terms of ATF degradation vs temperature were taken primarily from a 2004 study by GM engineers using Dexron-III (H revision) ATF as the test fluid.
Oxidatiion and friction degradation models were developed to predict the fluid's "End-of-Life" or EOL. These models were compared to lab* and fleet test data to determine model correlation, or "correctness."
Oxidation Degradation, End-of-Life (EOL) was defined as when the fluid's DeltaTAN increased by 2.5 during an ABOT test, which is much less severe the CEC test mentioned above. A TAN increase as you might remember is an indication that fluid acidity has risen to levels in which the fluid is no longer oxidatively stable.
Remember, this was a MINERAL-based ATF with none of the advanced, VII, FM or anti-oxidation chemistries that we have available today.
First temps stated are in C and F in "()."
First of all, the chart started at an ABOT temp of 140C(284) and went up to 160C(320).
The chart showed that the fluid at ABOT 140C(284) reached its EOL in about 1800 hours.
The chart showed that the fluid at ABOT 160C(320) reached its EOL in about 500 hours.
SO, a 20C(68) rise in fluid temp caused a 1300 hour decrease in fluid life.
Today's ATF chemistry is much more stable in terms of oxidation, friction, and viscosity.
If in doubt, have the fluid analyzed and or do a pan drain and refill every 30,000 to 50,000 miles.
This not only refreshes the additives, but also reduces particulates.
So how do OEM’s such as GM, Ford, Chrysler and others develop those Predictive models for the PCM, TCM, or Body Module?
In a nutshell, here is how the industry conducts studies in order to make predictions:
1.) An Hypothesis is put forth as to how the system under study behaves in terms of its physics (thermodynamics, mechanics, and chemistry),
2.) A predictive Mathematical model is developed based on the physics of the hypothesized system,
3.) Physical Tests, such as Lab and Fleet tests are executed and data collected,
4.) Feedback (data) from the Physical Tests are compared to the model to see how and if the physical data correlates with the model,
5.) If the correlation between the proposed model and physical data agrees, the predictive (mathematical) model is accepted,
6.) If the correlation between the proposed model and physical data does not agree, the predictive (mathematical) model is modified or redefined.
Correlation (close agreement) of the model with the actual physical data is important since the model should reflect real world physical data in order to develop TCM or PCM algorithms, such as the reflected in the Oil Life Monitors.
For example, the Friction Degradation model might look like this for a V8 powered vehicle with a 4-speed step shift Automatic:
TNS = 87800 – 2455*T – 6630*EPS + 1410*EPS^2 + 82*T*EPS
TNS = Total Number of Shifts,
T = Bulk (Sump) Fluid temperature in C,
EPS = Energy per shift in KiloJoules.
The Energy-Per-Shift model, EPS, might look like this for a 1-to-2 upshift in a V8 powered vehicle with a 4-speed step shift Automatic:
EPS = 0.3145*TPS – 1.4865
Where TPS is the throttle position sensor percentage from 0% to 100%; a 30% TPS is considered average for an up-shift from a standing start.
So these types of equations are compared to sensor inputs and algorithms developed to determine when a fluid needs changing and to turn on the OLM.
*One of the oxidation tests for ATF is the:
CEC L-48-A-00 Oxidation Test (DKA)
Let's make sure we differentiate bulk or TC oil temperatures from clutch and band face temperatures which can spike to over 400F for short periods.
Others are found in the ASTM literature.
Last edited:
