Couple more shots comparing drive/pass side trans fluid ports between the 3 samples.

Although there is no suggestion it matters, the drive side trans port nut is pot metal. The passenger side port nut appears to be same material as the heat exchanger - which *might be* brass.

Passenger side:

 

Here is the Video I uploaded when I pressure tested the Spectra, I had primarily done this to test if I could somehow "franken radiator" the OEM and the Spectra. After testing and making some calls to local radiator shops (none of which wanted to attempt the task, I set them both aside.

Short Video of Leaking Spectra seen above. I removed the upper and lower radiator hoses from the Ridgeline used 4 hose clamps to connect the upper and lower hoses together with a 1 1/8" socket. I wish now I had some pics of that backyard engineering, then again maybe not

 

Nice info OhSix.... thanks for sharing. Indeed looks like "clamping" approach is more robust with aftermarket design. Definitely less likely to get the SMOD phenomenon.

 

Great info OhSix. Thanks for the pictorial comparison.

 

Great pics; thanks!

To me, it looks like the Denso is a very significantly better (and more expensively-made) radiator, both as a coolant heat rejector and also the oil-to-water exchanger. Except for that exchanger connection...

 

I would be interested to know if the OSC rad has a stacked plate or a tubular exchanger.

Most certainly, the aftermarket approach to the exchanger spiggots dramatically reduces the chances of coolant being drawn into the transmission.

You may have a coolant leak, a transmission fluid leak, but it would take the exchanger tube itself to fail for coolant to be pulled in.

That said, there is no way that a tube will exchange heat as effectively as the OEM stacked plate design. Disappointing in my mind but it is hard to say how much cooling capacity is lost and if it will really matter.

 

I was thinking the same thing, but being a simple tube, and of course depending on the type of metal and gauge, it might well be more robust than the plate design. Certainly there are pluses for simplicity. And it kind of supports our homespun theory about the rad being a relatively small player in regulating ATF temps.

 

This thread was linked by another ROC (I think is was OhSix?) member and may indicate that the radiator cooler is actually doing the lion's share of the cooling work. . .

http://www.thenewx.org/forum/showthread.php?t=30287

 

I don't know enough about thermodynamics to understand all the reasons 'why'. Maybe we need to think about the exchanger more as a heat sink.

 

Decided to look a little deeper into the OTW designs. A couple more observations:

Had the idea to check out fluid capacity between the two designs. The idea was to weigh each design, pour water in calculate the delta. Info on that below. To ensure the cavities of both were staring as empty as possible, I settled on the idea of cleaning them with denatured alcohol.

When the OEM OTW was being handled, this fluid came out of it:

There was a fair amount of fluid inside but it emptied very slowly. The speed with which it came out of the OTW exchanger didn't seem to be related to viscosity. It seemed more like the flow was being impeded. I let it stand on end until no more came out. When loading it with alcohol, it seemed to flow into one side slowly and took an inordinate amount of time to appear in the port on the opposite side. This got me to wondering what's inside. Anyhow, once the fluid began appearing on the opposite opening, a little agitation made a few bubble appear in both ports. Once satisfied the cavity was filled completely, it was left to sit awhile. This is what came out:

Did the same with the Spectra OTW. One thing for sure, fluid flowed from one end of the chamber to the other very quickly, easily. Left it sit, this is what came out:

Rinse, repeat until fluid was clear. The OEM OTW exchanger never got any faster letting fluid run from one side to the other.

Once clean, both were blown out with dry compressed air and left for about an hour. Let the weighing begin:

Next post: wet weights and calculations.

 

Loaded with water:

Calculations:

Curious about why there was such an obvious difference between OEM and Spectra, it was time to look inside:

Well, there you go. Fluid movement in the OEM gets pushed thru a complex set of veins inside the stacked plates, while the Spectra gets pushed thru a sealed inner/outer tube with a grated wrap. Interesting, no?

And the port diameters are quite different too.

It appears to me Honda has purposely choked the ports and fluid passages, perhaps to slow it down for thermal reasons. It's doubtful the aftermarket has such depth of thought into there design(s). Most likely because they deploy a single design across everything they build - when Honda engineers pump pressure, fluid flow, thermal transfer of materials and a whole mess of other factors no one else is privy to.

What do you guys think?

 

Last couple observations.

While filling the OEM exchanger with alcohol, the black ring present on one port began to run. I noticed that before but paid no mind until I saw the black surface beginning to dissolve and run down the exchanger body. Started wondering if perhaps the heat exchanger is designed for a specific flow direction but I see no evidence of how that might be in the cross sections.

Then I remembered the Spectra had (what appeared to be) a pot metal nut on one of the ports, could there design be directional too? Perhaps, the Spectra might be set up that way to denote pass/drive side, but that's easy to ID by the ports angle, so I dunno.

close up of the chain mail style Spectra grating inside their concentric heat exchanger.

OH, and I took the hit for getting this stuff to the forum. See what I do for you guys?

 

OUCH!

I often wondered why the Denso, which has been out for over 2 years would still leak ATF.

This seems like from your chart, the 0.0041, ml of fluid seems small, but may explain why we (after the change) are getting more fluid out than the OEM spec?

 

Was surprised by that tiny amount of fluid capacity in both OTW samples. And a bit confused - but totally fascinated - by the flow constraint in the OEM exchanger. I have a buddy with a water jet, thinking he might be able to cut thru the core more smoothly than my crude non-ferrous blade on a chop saw. Would be interesting to see the composition of the passages undistorted.

This is about as clear an image as I can get with my crappy camera. If you stare at it long enough, you can see how all 88 star constellations would have appeared from the vantage point of Mare Nubium in the year of our Lord 918. I'm telling you, those Honda guys are magicians. LOL.

 

We do appreciate what you do, it is enlightening and an education < the best kind; Hands On... no pun intended.

 

It would seem that the Denso would slow the flow out of the Trans Drain plug? Just a thought

 

I very much appreciate your time and pictures dedicated to this. Thank you!

 

OK, the end of the geek rope has been reached. Taken this inspection as far as tools and imagination allow. Last night I was looking into the OEM OTW cross section and was seeing variations in the fluid passages that seemed odd. I really wanted a better idea of whats going on inside there, but cutting was distorting the surface and interior of the channels. After speaking to my pal with the water jet about cutting a cross section, it was determined turbulence would damage/distort the tiny veins inside the passages. So that was off the table.

Before leaving for the office this AM, a thought bubble appeared over my head: "cross cut the tube port!" That way there would be a clear view into unaltered passages. TADA.

These images were taken with an inexpensive AVEN USB microscope. What I thought I was seeing thru a jewelers loop turned out to be even more complex that imagined. Each passage looks to contain a wild staggered array of U channels. It occurs to me fluid passing thru these channels must have some pretty intense pressure behind them. And they are so small, any gunk in the trans fluid is definitely going to lodged in here.

Fascinating!

Edit: Adding this image:

For reference, here's a close up of the grating inside the Spectra OTW.

Geek, over and out.

 

My brain hurts.......

 

How do you think mine feels? LOL

One of those peel the onion things, only different. Obsession can be a wicked thing.

Wait till you see whats coming next. :act064:

 

I'm surprised those things aren't $15,000 each!

 

I've reached the point I might buy you one of those $75 ones off ebay

and Just drop ship it to your house...

 

Ive heard those refereed to as turbulators. I'm an IT guy in a high tech heavy manufacturing aerospace/marine environment, fun IT stuff, huge CNC, mills, water jet, etc, so that's where my personal level of knowledge drops off, but there is certainly lots of "turbulator" internet research available for anyone so inclined.

Very interesting, and appreciate the work! I have a marginally damaged Koyo thats not beyond repair (would have taken a number of hours straightening fins damaged in shipment) that was replaced under warranty. This has me considering doing a similar autopsy.

These latest pics are all great stuff! I was puzzling over why in the OSC pdf comparison pics you couldn't just look straight to the bottom of the open cooler inlet and see the bottom of what I previously imagined would be just a straight through pipe, or see a header where the stacked plates took off from. Looked like a funny strainer or something that I didn't understand. I appreciate more about the design now!

 

Just to be clear, none of the heat exchanger pics that o6 has done have been the OSC radiator. The stacked plate exchanger is the OEM Denso radiator and the brass tubular exchanger is the Spectra. I assume the OSC uses a tube type system. Do we know one way or the other for sure?

 

After close physical inspection of a new OSC rad & reviewing available info from their web site, tearing down the OEM and Spectra rads, and observing the construction of other replacement rads in the aftermarket, these are my conclusions:

- The Spectra (and others) - where the OTW exchanger is discretely attached to the bottom tank - is a far superior method for keeping trans fluid isolated from engine coolant.

- The OEM stacked plate design introduces significant trans fluid flow resistance which is not duplicated in the Spectra (and other) concentric designs.

- Flow resistance isn't accidental and (likely) includes all elements of thermal management, including materials, fluid chemistry, flow rate, "dwell" or duration of the fluid in the chamber, ETC.

- Based on images in OSC marketing materials seen here: URL it appears the presence of aluminum trans fluid ports visible from the outside indicates stacked plates inside. Follow the images in the link above and you'll see that copper/brass ports connect to concentric OTW exchangers, aluminum ports connect to stacked plates in 100% of the images provided.

- Because the ports seen in the new OSC rad are aluminum, I assume the OTW exchanger inside is stacked plate. Having stated that, there is no direct information available on the OTW inside the model 2830 for the RL.

- Assuming Honda optimizes thermal management much better than they accomplished fluid isolation in the Denso, it is highly likely aftermarket replacements are (to some degree) inferior heat management assemblies.

Why? From fewer/larger rows in the core to concentric vs. stacked plate OTW exchangers, the OEM and Spectra are very different designs. To what degree the aftermarket is an "inferior" heat manager remains unknown. In the grand scheme, small deltas in heat management might be incremental improvements.

Given there are many examples of high mileage RLs running around with aftermarket rads happily doing their job, this whole topic might be splitting hairs. Would your engine/trans serve 20K (or more) miles longer if you stay all OEM when replacing the rad? Maybe.

Given there is literally no reliable method to identify a pending failure of the OEM rad, I choose to assess the aftermarket, make the best informed decision available and run with it - secure in the knowledge that I've (all but) eliminated the potential for a SMOD event taking out my beloved RL.

 

I'd agree that the material of the fittings/pipes indicates the type of heat exchanger inside. Silver fittings indicate a stacked aluminum exchanger. Copper fittings indicate a concentric tube exchanger.

Regarding the resistance and potential dwell time in the exchanger, it would probably depend on the type of fluid pump used in the transmission. If it's a positive displacement pump, it's going to pump a given volume with each revolution or cycle of the pump. More resistance means more fluid pressure, but fluid volume should be the same. From that perspective, dwell time in the exchanger wouldn't really correlate with resistance. But, it's possible that the pump is not a positive displacement pump, I don't know.

If we assume that the OEM exchanger removes more heat than an aftermarket one (and that's a big assumption), then it'd probably be prudent to supplement it with an external cooler, if the vehicle doesn't already have one. The Ridgeline does, but something like the Pilot probably does not. Just food for thought for those who may be maintaining a Pilot/Odyssey/MDX as well as a Ridgeline.

 

ugg no end in sight, and replacement options in my mind are starting to go back to company reputation armed with some knowledge of common radiator designs, but all this is still great for anyone who likes to have a better understanding!

Its true that you can do all the research you want and still get a radiator that may be rebranded and sold under another name, or the result of an engineering change order. So, if you aren't close enough to the industry to know the whole production process its an educated guess as to what the best replacement might be.

The first pic is from the OSC website, and while the heat exchanger is SS, it shares atleast one internal O ring with a "threaded internal insert" similar to the Denso design that can result in SMOD.

The second is from my KOYO that arrived damaged and was replaced under warranty. Something made a two inch swath of bent fins. Probably fixable but it concerned me that there may be other structural damage. RockAuto was quick to replace. Packaging was sufficient, something in shipping breached the box itself.

You can see that the Koyo has a stacked plate, but also shares the internal O ring along with atleast some OSC designs. Makes me return to my thinking that possibly a inferior part made it into production at Honda and has certainly been remedied by now, and its convenient for Honda to just wait this out without acknowledging anything. The design still has inerrant flaws and could be vastly improved, but given all the other automotive systems, it may still be as reliable as many given that proper materials are used.. not corroding and expanding belleville washers.

In my view, breaking it down with the OEM, if the only the larger O ring failed, only anti-freeze would leak, but not mix with ATF, if only the smaller O ring failed ATF would leak but not mix with radiator coolant. Only the failure of the belleville washer which swelled and caused the failure of both O rings caused mixing of fluids. In looking at the Koyo, and looks like some OSC and other manufactures, but I can only vouch for the Koyo, that several counter clockwise revolutions of the internal fitting would have to take place before the smaller O ring would unseat from its landing before ATF would be lost. Its an O ring in a sleeve config, not strictly a compression seal maintained by a nut. Before that, however, the larger O ring would leak engine coolant but not mix with ATF. These kinds of connections are successfully used all over the place... again assuming proper material... and still maintaining the overall design could be easily improved which some like OSC have chosen to advertise, but not implement across all their designs.

 

... Makes me return to my thinking that possibly a inferior part made it into production at Honda and has certainly been remedied by now, and its convenient for Honda to just wait this out without acknowledging anything. ...

I posted a link the other day to a thread over at Piloteers.org where a 2009 Pilot suffered the same fate as many here with the SMOD. So I don't think it's safe to assume Honda has remedied the issue.

http://www.piloteers.org/forums/70-2009-2011-pilot/89866-2009-radiator-failure-same-2005-a.html

 

schwejo, that is very interesting. IME: shop floor control and supply change management/QA authority normally require part number revision to track and manage ECO introduced as a running change... *Usually*

But yah, I agree, documentation available to us little guys who WRITE THE @#$%ing CHECKS is sorely lacking.

Observation/question. The images you copied from the OSC pdf file are confusing. They are touting an additional plate in their OTW exchanger as a "feature" but the image of their competitor design has superior fluid isolation via the threaded port extending thru the lower tank. This is contrary to the OSC video claiming "braised in" fittings. Crazy. Based on the potential for SMOD, I'd opt for the competitor design!

On your Koyo w/stacked plates, is the fluid port the same as Denso? Meaning the port both isolates fluid and serves to clamp the OTW exchanger to the tank?

This is the Denso:

 

Observation/question. The images you copied from the OSC pdf file are confusing. They are touting an additional plate in their OTW exchanger as a "feature" but the image of their competitor design has superior fluid isolation via the threaded port extending thru the lower tank. This is contrary to the OSC video claiming "braised in" fittings. Crazy. Based on the potential for SMOD, I'd opt for the competitor design!

I think those images from OSC are for an older style of "cross flow" radiator, commonly used in domestic vehicles. In that situation, the exchanger is fixed to the tank with very flat nuts (like the large nuts on the OSC/Spectra/TYC radiator), and that seals the exchanger to the tank. The user could then either install the transmission cooler lines or not, depending on automatic or manual transmission. The exchanger does use female threads in that case, but they're not a flooded connection like Honda used. They're simply female threads into which a fitting can be threaded, like our replacement radiators use male threads onto which a fitting can be threaded.

 

I cant take any professional stance on the expansion force caused by a rusting washer, but in my mindset I was applying what I've seen in steel reinforced concrete and piers in Santa Cruz, CA where I've spent a lot of time.

No doubt my mindset could have been misdirected but a little research seems to indicate at least the possibility that a rusting washer could pull the OEM connection from its threads, and I'm not sure how much evidence would be left on the threads them selves if there is already not an over abundance of thread engagement.

Combining of fluids still takes the failure of both O rings and that is where my puzzlement comes from when the smaller O ring is more of a "sliding" application. Has to be a better or proper term for that, but its not the same compression application as the larger O ring, so in my mind something has to compromise the smaller O ring or pull it sufficiently out of its "landing" before any fluid mixing occurs. Still a puzzlement to be sure!

https://news.google.com/newspapers?...s?nid=2194&dat=19791123&id=uaMyAAAAIBAJ&sjid=8u4FAAAAIBAJ&pg=5936,1498239&hl=en
https://www.google.com/#q=expansion+force+of+rusting+steel

 

Points well made and taken schwejo. Following the google link, selecting images this photo was found

It reminded me of an issue in the spa of my pool a few years back. The gunnite surface allowed water to penetrate, which caused the rebar "bones" of the wall to expand. Eventually, rust stains appeared on the surface and small chunks of gunnite began to flake off as internal pressure increased. So there you go. If expansion pressure from rusting metal can bust concrete, it'll sure pull on the weak threads responsible for compressing the OTW cooler to both sides of the rad tank.

For reference, I took this image of the OEM washers on my 06 this morning. She's coming up on 10 years in the cold cruel world. Bought her in Feb 2014 but was able to determine she's been a So Cal resident since day one off the stealer lot. Spent the bulk of her life in Carlsbad, which might have slightly saltier ambient air than where she live today - 20 miles inland in north county San Diego. Obviously, this is a milder environment than some where washers have been horribly distorted by surrounding elements

For additional reference, these are the washers from Carsmak's OEM rad which was replaced in 2013. He's located in the high desert, very low humidity up his way. What I see in this image is the beginnings of corrosion between the inside of the compression washer & the outside of the spread washer. In his environment, it would have taken a few more decades for corrosion to be the root of SMOD. I reckon his rad would have died from material fatigue long before corrosion got to it.

Just for fun. Check out how many threads are holding this assembly together.

 

Can't... stop... thinking.... about.... SMOD

LOL.

Actually, comments about ORing failures and speculation over (other than the obvious) how coolant and trans fluid can mix, it occurred to me the chop saw was sitting in the garage with the metal blade just asking for another cut. SO....

The Belvile (sp?) washer is spring steel (which is one of the reasons it corrodes) and the spread washers is stainless so the blade on the chop saw wasn't gonna cut it (to coin a phrase) so in the micro generated image below the space occupied by both washers is approximated.

Not sure how this image will resolve, but, the arrows point to small gaps between the port nut face (responsible for compressing the OTW to the tank) and the small gaps on the OTW matting surface inside. The point is, when the port(s) are torqued to spec, they are a few dynamics at play. Both the ports and OTW exchanger are aluminum, as you can see in the cross sections, there is little distortion in the flat surfaces caused by the blade spinning thru them. Keep in mind this is a standard home owner level chop saw spinning pretty quickly (I'd guess 2K or so) and there is nothing special about the blade, other than it's designed to cut non-ferrous metals. Cutting thru these components was no problem - LIKE BUTTER - no bogging or nothin'. The point of that is, the material is soft and would be easy to strip or mangle the threads as pull O_Rings and washers into the desired shape and compression. Even though we are getting reliable service for many thousands of miles, this design sure does rely on 4 threads to keep our vehicles running.