ATF Cooler tests and information on this site has been uploaded for consideration by anyone interested in an external ATF cooler installation. I'm sure many Honda forums will do the usual Google search and disagree with some, or all of the following - But will they supply data and measurement methods for a Honda 5AT when they discredit anything?
I've been dissecting published ATF temperature data recorded by owners on various forums going back over 10 years or more. Years ago when I initially read some of this data I had been influenced by the interpretation of results, since they mostly pointed to a similar conclusion, only now do I realise there were incorrect assumptions made. I don't blame these guys in any way, hindsight, more information on TC failure modes and AT flaws, my goal to accurately measure fluid temperature by inserting sensors directly into AT fluid streams, and most importantly, graph data against ATF cooler line pressure has added a few more pieces to the jigsaw of Honda 5AT and ATF cooler arguments normally biased around standard AT theory stated as 5AT fact, along with cooler efficiency, sensor installation, data logging deficiencies and installation folk law.
Some of this old data was published for 4AT and 5AT Honda odysseys as well as the Accord, the Odyssey had (has?) a standard OEM radiator tank (internal) cooler and some had an additional OEM auxiliary cooler mounted in front of the radiator. Of interest was the wide variation in temperatures when cruising, some reported ATF outlet temperature dropping down to 100°F or less (long periods of cruising - outside temps 60°F) and others at 150°F to 160°F with normal freeway driving and cruising. A few owners of various Honda models have reported temperatures down around 98°F at the TC outlet in cruise on the highway - these were all factory setups.
Keep this thought in mind for now: These low temperature readings - under certain conditions - may be incorrect in that ATF itself is not the only medium being measured, sensor readings are likely being influenced by sensor installation errors and something they had ignored and/or underestimated - most of this was speculated on and attemtps made to resolve errors, however these affect can cause errors in unforseen ways.
Temperatures around 150°F are likely valid and most often reported, so much for the ATF warmer theory - BTW the TC will lock in 5th at around 140°F - reported (and I've seen it) to be lower but it appears to depend on variations in road conditions etc? More data and time will answer this.
Measurements I have taken and numerous posted data sets indicate that the OEM cooler is too small to control temperatures over a short time frame. At cruising speeds, ATF "flow rate can be so low" that this small warmer/cooler cannot heat the bulk "volume" of ATF above the average ATF case temperature - cooler outlet temperature is utterly meaningless at low ATF flow rates. The transmission case, although coupled to the engine block, can hold ATF at around 150°F and possibly less under sustained periods of TC locked cruising, especially with an additional OEM front mounted AUX cooler found on some models, the case temperature and/or extra AUX cooler still overrides the small OEM coolant tank or Hockey-puck warmer/cooler under these conditions.
There are often reported 10°F to 20°F drops in temperature while stopped at traffic lights, followed by a sharp climb in temperature when moving away from traffic lights and repeating with each set of lights.
At idle in gear with HOT ATF there is very little flow, they realised low flow rate was the reason for some high cooler Delta readings, but flow rate is far less than initially realised and this causes the ATF cooler to artificially indicate a huge temperature drop, it's compounded by sensor placement on metal ATF pipes. These pipes are usually connected directly to the transmission and are located in the engine compartment, this is normally not a hot location compared to Coolant and ATF temperature. The transmission case can be around 150° or less and air temperature in the engine compartment can be around ambient when moving.
With a sensor attached to the outside of a metal ATF line, the line and sensor will be heated or cooled in four ways. 1. By a "decent flow rate / volume" of ATF flowing through the ATF line, which can override 2, 3 and 4, otherwise 2, 3, 4, override 1.
2. By heat exchanger from the engine compartment.
3. By direct heating or cooling of metal lines attached to the transmission or other metal object.
4. By air flow over a metal ATF line from fans and movement of air at vehicle speed.
At idle in gear with very low fluid flow rates, the temperatures reported by both sensors - TC outlet and Cooler outlet - are influenced by 2, 3 and 4 far more than 1. This is why those temperature can take such a dive.
Number 4 is the only one you can really reduce in this setup, you do this by insulating the outside of entire the metal line. The only way to eliminate 2, 3 and 4 is by placing a tiny K-Thermocouple head directly in the ATF stream so that the only variable that can heat or cool the sensor is direct contact with ATF to be measured. A metal or plastic housed ATF sensor such as typical low cost thermistor types, screwed into a metal fitting, like most, can be inaccurate over a wide temperature range and be affected by any "screw in" mounting hardware.
One thing that stood out was the time it took ATF to come up to temperature, with in tank cooler/warmers and cold radiator coolant keeping the ATF low until the engine heats enough to open the thermo, the remainder of the coolant then slowly comes to operating temperature. The OEM Hockey-Puck style of cooler/warmer is in a "small volume closed loop" of engine coolant that is heated to 170°F in around 3 to 4 minutes. As usual, the warmer is small and has almost no affect on ATF warm up time. For the same reason the ATF cooler has NO AFFECT, nil, zero, zilch, zip, on coolant warm up time. Again, the coolant reaches operational temperature in around 4 minutes in the Gen7. ATF takes almost the same amount of time to reach 170°F with or without the OEM warmer/cooler.
It's interesting that even a small external cooler "without a bypass thermostat" will delay ATF warm up time even further and can drop transmission temperatures too low "over time" at cruising speeds with the TC locked, especially in 5th - Again, very low fluid flow (engine less than 1900 RPM) and minimal heat generation by the transmission.
As mentioned previously - Reported sudden drops in ATF cooler outlet temperature at idle in gear at each set of lights, sometimes up to 20°F in 30 seconds. The TC outlet temperature also dropped "slightly" and this is the 2nd clue! The third clue was the sudden jump in cooler outlet temperatures of 20°F in around 10 seconds as the vehicle moved from the traffic lights.
Whilst in gear and stationary: It is impossible for TC outlet temperature to drop and almost equally so for the cooler outlet temperature to really drop ATF temperature by up to 20°F in 30 seconds. The TC is generating heat under this condition and because of low flow rate, higher heat than normal. The location and mounting of the sensors and more importantly, very low ATF cooler circuit flow rate under these conditions, skewed sensor readings.
Again, those readings are due to very low fluid flow in gear while stationary - lower that previously though - the TC outlet sensor was actually measuring the temperature of the ATF hard line - NOT the ATF. The cooler outlet was likely measuring a small volume of ATF sitting in the cooler and being really cooled at such a low (or no) flow rate plus convection cooling of the steel cooler line and remaining ATF in the cooler. Sudden jumps in cooler outlet temperatures of 10°F or more in 10 seconds can be due to TC temperature increase under very hard acceleration, but in this case, graph data appears to indicate it is due to a higher flow rate of HOT ATF through the ATF cooler lines when RPM increased, reflecting something closer to true ATF temperatures.
These same systems which can indicate ATF temperatures droping below 150°F cruising, also show 270°F at the TC outlet under load or driving in hilly terrain. Because of very long sampling rates times and sensor location, the spike in TC outlet temps would have been far higher than recorded - you need at least one or two second sample rates for all sensors. Without the probe directly in the fluid, you are flattening any temperature spikes. I can see this in my 0.1 second (100ms) sample rate graphed data before the cooler bypass opens - currently the cooler outlet thermo sensor is on a small separate aluminium housing (it holds the outlet pressure sensor) this is far better than mounting it on a steel line. This housing is small and insulated from everything by the ATF rubber line and is covered in foil insulation, but even this lags and in particular flattens tracking of "short duration" temperature changes compared to the TC outlet sensor in the ATF flow. These errors could be further exaggerated with brass, steel and metal encased sensors.
This confirms that:
1. You need a BIG efficient plate and fin cooler to cover adverse conditions - traffic and high loads - and it WILL control very high spikes, I have proven accurate 40°F cooler delta @ high fluid flow rates, not the normally quoted as acceptable? 5°F to 10°F delta.
2. Except in extremely cold climates, I personally no longer agree with the statement "You must leave the tank mounted cooler in circuit" as data just does not support that reasoning - especially the warm up theory, however cooling may have been passable (just) if not for 5AT transmission and TC flaws.
3. You must have a temperature bypass valve on "any size" external cooler. Even with the OEM tank cooler or G7 style cooler in circuit - this allows you to go to a BIG plate and fin cooler without any overcooling problems (if the cooler bypass thermostat is a quality device).
4. You may need fan assist when stationary in gear if there is insufficient air flow over the cooler or the cooler is too small, and especially if you bypass the OEM cooler, again use a big cooler.
Note on 2: I have removed the OEM "hockey puck" cooler / warmer.
Note on 4: In case you missed it, I have sufficient air flow through a BIG cooler from the AC fans. The climate control is always on in my vehicle - winter and summer. Again I have a really large plate and fin cooler AND it's mounted 3" in front of the AC condenser - bolted to the back of the front bumper - it works.
Some owners also reported shifting issues once ATF is hot - this is in part a result of low pressure from internal leakage and/or TC problem showing up with HOT ATF viscosity.
Question 1: Can cold/hot TC outlet pressure readings be used to indicate "some" future problems in a 5AT? specifically a blocked or partly blocked internal screen and classic internal TC ATF leakage from wear and heat.
Q1A: Yes I believe so! Question 2: Can (accurate) TC outlet temperature be used as an early warning sign of transmission failure?
Q2A: - Although I'm pretty sure it can, I don't have a definitive answer yet. More to investigate.
One thing I did verify under even normal (190°F) ATF conditions is a reported TC/cooler circuit bypass condition in the valve body. At idle, even in Park/Neutral, ATF pressure can drop to zero, this does not imply that ATF flow rate is zero - it clearly is not in my AT but it is the typical low value reported by transmission specialists who have studied and investigated Honda transmission design shortcomings to better understand why they fail and importantly, make needed modifications to rectify future problems after rebuilding the 4AT and 5AT - This is why a straight OEM rebuild can often fail again, especially under the same driving conditions driven by the same owner.
Bringing the engine up to 1200 RPM immediately blips ATF cooler line pressure, and as reported, the bypass valve shifts, and now you have higher cooler line pressure. In my case it holds after RPM drop back to idle, this translates to a higher flow rate. NOTE there was still ATF flow in my cooler circuit at what appears to be zero cooler line pressure, ATF @ 50°F has a viscosity around 160cSt, hot ATF has a viscosity around 12cSt @ 185°F and very little flow resistance at low flow rates (low pump speed) so there is little pressure required to flow enough fluid to cool the TC, and in my Auto transmission there are no spikes in TC output when RPM increases by shifting from D to N and increasing pump speed above 1200 RPM - If ATF spikes at this increased flow rate then it's a classic indication of "insufficient flow rate" in D with hot ATF while stationary - NOTE: You will not see this short spike with an incorrect sensor installation.
So, a variation of this Honda design flaw is there: What will it be like 5 years and many miles from now? What is it like at 220°F to 260°F as reported by owners and transmission testers/rebuilders on a hot day while stationary in gear in traffic?
Now that I'm aware of the future potential for this condition to harm my 5AT, by making three simple changes whilst driving I won't allow the TC/trans to overheat via this flaw.
This bring me full circle to the often quoted forum statements "I don't think you need an external cooler unless you are towing or in very hot climates" I know some people are trying to help, but this unqualified "I think" is starting to wear thin. Unfortunately that thought is usually based on a lack of understanding as to what is happening in the 5AT, quotes more likely sourced via Google and standard AT recommendations.
Driving in traffic or hilly driving conditions can overheat the TC in a 5AT without you knowing it, and the OEM cooler/warmer CAN NOT quickly correct this condition, it will get hotter and hotter with each hill or traffic light unless you make a corrective change - ignore or be unaware of this condition and soon you start noticing a flare with gear changes (the flared clutch pack is now permanently damaged or compromised) and ATF pressure goes too low in an already compromised older transmissions - especially with oxidised old fluid - causing more slippage followed closely by both flare and harsh shifts under different conditions, sound familiar?
The second forum statement goes something like this: "The transmission PAN or Sump is the correct and only place to read ATF temperatures, not the TC outlet" and this is also quoted by self proclaimed Hydraulic fluid experts on forums, this is totally misleading in a 4AT or 5AT. These same experts preach that a transmission produces virtually no heat at idle in neutral or under TC lockup while cruising, this is TRUE. But to make that statement, and then say you can't use TC outlet temperature to indicate transmission Bulk ATF temperature or that monitoring TC outlet temperature is meaningless is actually a contradiction of themselves. Why?
Because as I and others have proven in this Honda transmission, with "mid to high ATF flow rates", TC outlet temperature - IS - ATF BULK TEMPERATURE - WHEN - this transmission is in neutral and engine speed is above 1200 RPM (ATF pump speed with higher ATF flow). And also - outlet temperature - IS - ATF BULK TEMPERATURE - after 20 seconds or so with the TC Locked (if flow rate is not too low). There is almost no heat produced by a transmission under these conditions so medium to high ATF flow rates will quickly bring the TC down to bulk sump ATF temperature at its outlet. In lower gears under normal steady driving conditions ATF TC outlet temperature is usually within 7° of (correctly measured) sump ATF temperature.
Correctly monitored TC ATF outlet temperature can be critical to keeping the 5AT transmission & TC from another mode of premature failure. Possibly even a contributor, and a cause in some cases, for age and heat related 5AT failures. Again reinforcing why it's so important to refresh ATF regularly - especially Z1, as it oxides rapidly in this transmission and especially under TC and clutch pack spike temperatures.
NOTE: Because of Google data - misleading information from transmission cooler suppliers (referring to standard AT's) has been posted and some owners have even gone back to measuring only cooler outlet temperatures - this is just plain wrong! Again cooler outlet temperature is meaningless without an indication of flow rate and ATF Volume - and they have not measured anything that indicates ATF flow rate and have usually overestimated 5AT cooler line flow rate. These owners had previously measured TC out temperatures over 235°F with normal towing, now they only see temperatures of 195°F at the cooler outlet and assume the transmission is ok and the OEM cooler is all you need. This is a big mistake! The sensor connection to a metal line or block, at the cooler outlet, will mask even long spike temperatures that can start to damage the Torque converer and transmission seals.
It has been stated again and again: This preoccupation with cooler temperature Delta and even cooler outlet temperature alone is totally flawed, it cannot and does not reflect cooler efficiency without an idication of flow rate and fluid volume. However long term graphing with TC outlet, cooler outlet, verified bulk ATF temperatures with a meaningful indication of flow rate/volume when interpreted correctly and measured with reasonable accurately is indicative of overall cooler line efficiency.
Note: In the following I'm not talking about an overheated AT spewing ATF from the vent: Some people will state they can lift the bonnet and tell if the 5AT is running cool or hot - this is utter nonsense in these Hondas - the transmission case and even the engine block are at lower under hood temperatures - they have no correlation to internal ATF temperature. Two minutes ago the TC could have been over 280°F and you won't know it, clutch packs could have been over 300°F and you won't know it, bulk ATF PAN temperature will NOT reflect this if it was a short duration spike of less than a few minutes - with high flow rates and very hot spiked ATF that small OEM cooler may be dumping this hot ATF on gear sets, differential and various seals - again, most sensors mounting options and types of sensors construction will mask this.
IMHO with a correctly installed temperature sensor you can control a heating problem as soon as it starts, before it gets away from you, any change you make to gear selection or driving mode to correct this is INSTANTLY indicated in TC outlet temperature, not 5 or 10 minutes later in bulk ATF temperature.
Most of the time you can forget about ATF temperatures, but being aware of WHEN you need to be mindful could be the difference between a failure at 60K and 180K miles - but longevity of the transmission and making changes to extend it's life also depends on the condition of the transmission at the time of any change, the assembly tolerances, the age, the driving conditions it was mainly driven in, the climate, owner driving style, service history, and more.
Do you need an AUX cooler?
If you are in a coolish climate, don't drive in stop start traffic for hours, not sitting stationary in gear for 5 minutes at a time over and over in traffic, don't drive in hilly terrain, don't pull a trailer in adverse conditions or load the vehicle to the roof, then most likely not.
But just a simple trip over some extended hilly terrain for a quite day's driving with two people in cool weather can push the TC to over 250°F or more depending on the condition of the transmission, ATF and how you drive. At the very least, monitor sump ATF temperature if unsure.
From these tests and other owner data: A: I would not run lower viscosity fluid in hot climates.
B: If you are running Z1 or DW1 it's imperative to do regular changes to stop this inferior fluid loosing viscosity with age and internal heat (clutch pack) causing rapid fluid breakdown.
C: It's still important to keep the external ATF cooler line as free flowing as possible, especially when ATF is hot and engine speed is low - stop start traffic - cruising in 5th with TC locked and very hot ATF - and to this end, change the filter every few years or 30K miles.
D: An accurate K-thermo temperature sensor at the ATF outlet (in the fluid) is a brilliant way to keep an eye on TC and ATF heat problems before they leave you stranded.
E: A low cost pressure sensor ($22) can establish a base line pressure and alert you to low pump pressure or cooler line restriction before they become a major problem and leave you stranded.
F:On hot days in traffic: If you are going to be stationary for a few minute or so then shifting into Neutral will stop ATF and the TC from being heated by TC slippage with very low fluid flow. Note - ATF start to heat "the instant" you stop in traffic and STOPS "the instant" you take the trans out of gear (in P or N) as there is nothing to heat ATF in this condition. Of course without sufficient airflow over an external cooler there is also nothing to cool ATF when stationary - critically so if the OEM unit has been bypassed.