I looked into this mod quite a bit a few months back. GMCTD (our wise sage) is an expert on how this works. I hope this is not inappropriate, but here's what he told me:
"Anytime a post is erroneous in content and scope, the required correction tends to make the poster look like an idiot, which is not the intent of the correction, by any means (sometimes) - I have posted detailed description of the 6.2\6.5L bypass coolant system on several occasions, and the material is still there for the reading - here it is again - you can reference some of the pics in Jason's 6.5 teardown thread for associative illustration.
The 6.2L Diesel engine, never known for overheating if properly maintained, was designed with a bypass-recirculation type coolant system - this design used a water-crossover\t-stat housing between the heads like many other dual-bank engines, but with an extra port below the t-stat on the engine side - this port was connected directly to the low-pressure inlet side of the waterpump impeller, such that the waterpump constantly pumped coolant thru the block and heads, drawing the return effluent from behind the closed t-stat - coolant was continually bypassed, even when the t-stat opened to pass hot engine coolant to the radiator - driver observing the coolant guage would see it sit at ~160degF, very very slowly creeping up to 190deg, then plummet like a shot back to 160deg when the t-stat finally opened, just as if the IGN was switched off - put a load on the truck, and the swings were quicker, but still swinging hi-lo-hi-lo - that system was used thru to the serpentine belt era, with the new ccw waterpump.
The 'new' 6.5L also used a bypass coolant system, but the turbo required an important change, required to maintain operable coolant temperatures: the gated-bypass system, where the bypass port was made larger, with larger waterpump impeller indicating greater btu output, but with a valve-plate extension on the t-stat, such that the bypass was full open when the t-stat was closed, but gradually valved closed as the t-stat opened, indicating even greater btu thru-put - driver no longer could watch the coolant guage needle move after reaching operation temps, only increasing as the load increased
Looking at either series block, you'll notice two holes at the front of the block either side of the timing cover, just below the cylinder deck - these are where the water pump feeds coolant into the water-jacket - also note a pattern of holes in the cylinder decks, identical side-to-side, front to rear, bank-to-bank - those directly around the cylinders are tapped head-bolt holes - most of the others are core clean-out holes, where the sand-casting is removed after the cast-iron block cools - you'll notice a larger rectangular hole at the rear of the each deck, near the edge, but none at the front: this is the water passage where the pumped coolant flows from the block into the heads - again, you'll notice no such holes at the front of the decks, effectively preventing fresh coolant flow directly into the heads - all the clean-out holes are blocked off by the head-gasket, with only small steam-holes to allow for air bleed-out and such, so there is little coolant flow into the head thru those - coolant flows from the front of the block, around the cylinders, thru to that rectangular hole, then into the heads at the rear - this is why the rear-most cylinder #8 runs hotter than the other seven, and the operating clearances are design-spec'ed larger due to greater heat-expansion.
Looking at the heads, you'll notice the holes in the deck sufaces are identical, front-to-rear, side-to-side, both heads, even to the large water passage, one at each end of the head - this is because GM builds one head for each engine: the head serves double duty, designed to bolt onto either cylinder bank, fully interchangeable - as in the block, the holes around the cylinders are for head-bolts - the two irregular holes on either end are for oil drain-back from the valve-train - two adjacent holes above each cylinder are for the valve push-rods - the remaining holes are core clean-out, with the rectangular hole at each end for coolant-flow - since the heads are interchangeable side-to-side, when the head is installed on the block the passage at the front is blocked by the blank cylinder deck, the passage at the rear matches the one in the block - again, the head-gasket blocks all the other holes to the water-jacket.
Now, what you cannot see is the large area inside the head-casting where the pre-cup nest and the valve-guides and the intake\exhaust passages around the valves take up critical space in the water-jacket, totally increasing heated mass while reducing surface-area exposed to coolant - area not only required for adequate cooling, but restricting normal coolant flow thru the heads - this area runs tremendously hotter than in those other engines, and is the total source of all over-heating in this engine.
Again, fresh radiator-cooled coolant is never, and is not, pumped into the heads at the front of the block - again, the last rearmost cylinder in all GM engines runs hotter than the other cylinders - even hotter in this block, #8 along with #6, sometimes #4, because of the average 650degF turbo\exhaust manifold temperature directly adjacent to those cylinders
To further combat the prevalent cooling problems in the 6.5TD engine, GM released the dual t-stat crossover manifold in '97, going back to the non-valved constant bypass system, utilizing dual t-stats with separate temp ratings to meter total flow, bypass and radiator-bound - they increased the fan diameter and number of blades - they again increased the impeller diameter in the new HO water-pump - they revised the coolant passages in the heads to promote increased flow thru\around the pre-cup\valve area block-off - they etc, etc..............
Supposedly there have been other changes, but I sorta passed-off the ball after switching to the Cummins side, so someone else will need to fill in those details.
Bill Heath worked with a PhD on the design of his coolant-bypass system, based on Bill's 6.2\6.5 knowledge and maybe some input from the 6.5 forums, and it works, tho somewhat expensive for the avg working stiff - it is not just three hoses tied to a can, for those backward-facing 'engineers'
The real problem is simple: the 6.2L was designed to cure GM's CAFE ills, never meant to be more than 135hp - the mistake was when GM increased the bore (to disguise the weak 6.2 heritage) and kludged a turbo onto the 'new' 6.5L engine with no other changes except the plastic TURBO POWER cover, in a desperate attempt to keep up with sales of Dodge\Cummins and Ford\Navstar turbo-charged medium-duty packages while the DMax resolution was being worked out (and a fine solution it is).
There you have it, in a nutshell, folks - that's why you need take particular care when considering snake oils, potions, and dual-electric cure-alls for your 6.5 ailments.
As we often tell you, and you just as often hate us for it, they never worked, they do not work, and they will never work.................
That's my story, and I'm stickin' to it."
