Saw an years-old thread on a popular forum come up, highlighting some remarks of mine covering a some obscure details of the T3 Vanagon cooling/heating systems, and realised I had forgot to add one other interesting observation to fill out the picture. But I’m putting it here, for those that care to learn, because I don’t work on those other guys’ plantations.
.
Here’s the resurrected remarks:
https://www.thesamba.com/vw/forum/viewtopic.php?t=671397&start=60
.
And what I would add: If you examine the 2.1 wbx’s thermostat housing, you’ll see the oil-water heat exchanger (OWHX) outflow ports right into the engine cooling thermostat well. At operating temp with the t-stat full open and the extension disc blocking the bottom opening in the well, both cylinder heads’ outflow is shunted to the radiator and heater loops. The heads’ combined hot outflow no longer provides the t-stat temp signal as it does during warm-up, instead the dominant t-stat thermal signal comes from the OWHX outflow, so the cooling system is then indirectly responding to oil temp, which will normally run in a range of 15-50ºF hotter than coolant temp and therefore be warming the coolant flow thru the OWHX. The radiator return flow is being mixed in with it in the t-stat well, so there’s also a response to ambient temps if they cause the radiator to overcool.
.
The way it’s plumbed, the OWHX outflow is directed like a jet at the t-stat wax element capsule. The open t-stat allows radiator return flow to swirl around the top of the well before flowing out the side toward the water pump. With the same pump suction bearing on both openings, the OWHX’s smaller aperture will increase the velocity of that current entering the well while the radiator return flow entering thru the large aperture afforded by the circumference of the t-stat main disc will be moving more slowly. So the OWHX outflow will have the dominant influence on the wax capsule. I conclude that by design the preponderant influence on the t-stat element both during warm-up and at full temp is the OWHX outflow.
.
Since oil takes 3-4 times longer than coolant to reach operating temp, this would have the effect of delaying full t-stat opening while the oil is still substantially cooler than coolant, and/or if the radiator return flow was overly cool. So it’s another feature working toward getting the entire engine evenly brought to operating temp faster and maintaining it, which has heavy benefits in lessening fuel waste, lowering emissions, and improving mechanical efficiency. Once at full temp, the OWHX outflow is there to keep the t-stat constantly open. Without it, radiator return flow is the only thermal signal. I don’t know if the t-stat would then oscillate open and closed, or maybe more likely it would find a median opening, but it would be responding more to radiator return flow temp than changes to engine temp, meaning the ambient temp will be inordinately influential, and the main purpose of a well-functioning engine cooling system is to get the engine quickly and evenly to operating temp and maintain that temp within a fairly narrow range regardless of ambient temps.
.
I could go on about the”wisdom” of removing the OWHX based on hollow arguments like “the 1.9 did just fine without an OWHX”, and my favorite “I want to reduce the potential points of failure” (another of many logical fallacies that are the coin of this particular realm). I won’t, but back over there the usual suspects probably will, but not from the same direction as I.
.
The lack of a direct thermal signal to the t-stat is also plainly the cause of the many reported cooling system irregularities in Vanagon Subaru conversions, prompting numerous plumbing redesigns, attempts at delivering a consistent temp signal, but none of the ones I’ve seen will ever attain a consistent and constant regulation at full operating temp. But then, Subaru’s own layout was a bodge as well, it only manages to work because the engine is barely two feet from the heater core. It’s also not even a full bypass system, so much for being “modern”. It appears to me, though, that on at least one iteration of Subaru’s factory cooling layouts on a turbocharged engine using an OWHX, they may have applied this same strategy, with their OWHX outflow plumbed separately to be directed at the t-stat element, but I don’t have one of those cars here to look at, so can’t be sure of that myself.