TheSamba.com :: View topic - SOLVED: How to get my Subaru swap to stop overheating?

Sat Nov 02, 2013 10:32 pm

<Edited title to be less alarming>

OK guys...I need some collective advice from The samba.

I swapped a naturally aspirated JDM Subaru EJ20 into a formerly air-cooled Westy.

It drives great and stays cool around town. It seems to stay pretty cool on level highway. It starts to overheat pulling up a hard grade on the highway, sometimes getting up to ~250F on my aftermarket temp gauge. After it starts overheating, it remains REALLY hot for several minutes even if I pull over and let it idle or rev it in neutral until the temp suddenly plummets down to normal operating temperature -- EVEN WITH THE HEATER SHUT OFF (see diagram below).

It has done this with two different thermostats.

I'm running the stock (NON-reversed) Subaru cooling manifold. I'm running a Subaru turbo water pump that has two heater hose inlets (kind of like a Subaru OEM version of a Tom Shiels adapter) to clear my exhaust:

Both inlets blow water on the engine side of the thermostat. I have one plumbed as a return from the heater core, and the other plumbed from a T in the heater feed line. Again, the issue happens even when the heater is shut off with a ball valve I installed in the line.

Cooling diagram:

Other things to note:
1. I'm controlling the radiator cooling fan with the Subaru ECU. It comes on when my gauge reads just over 220F and shuts off when it gets down to just over 180F, so I believe my gauge isn't grossly inaccurate.
2. Instead of a burp tank, I ran heater hose up to a snowmobile radiator neck that I mounted in the passenger side fresh air intake. It's the highest point in the system and like a built-in Libby bong. It seems to do a good job of bleeding the air out of the engine side of the system.
3. My engine used to get hot at idle but would cool down if I revved it. I thought this might be caused by the Subaru water pump having insufficient head pressure at idle to move water down the longer length of cooling line (because I'm using the stock Subaru coolant manifold and running more hose around the engine compartment). So I added a Davies Craig 8005 EWP-80 electric water pump to the top radiator hose and wired it to come on with the radiator fan. I verified that it runs during the overheating scenarios.
4. When it overheats, the radiator is cool and cracking open the bleed screw lets out very little air before coolant comes out (as if the water isn't moving through the radiator). The whole radiator gets hot when it's not in the throes of an overheating fit.
5. I haven't cut a lower radiator grille.
6. Water pump is new. Thermostats (The two I tried) are new, and one was Subaru OEM. Radiator is used.
7. I'm running a 16PSI radiator cap. I've pressure tested the cap and the cooling system. The cooling system holds 20PSI for a half hour with no leaks. No leaks driving around. The level in the overflow tank rises and falls slightly with the temp as you'd expect. It doesn't blow coolant out of the overflow tank or anything.

I need to fix this. It sure seems like the thermostat isn't opening when it overheats. The fact that the temp will plummet from overheating to 180ish also makes it seem like it's an issue with the thermostat not opening.

On the other hand, it's strange that the overheating happens primarily when pulling up a long grade...like the problem starts when the engine can't shed enough heat. However if it was just an issue of not being able to get rid of enough heat, I'd expect the temp to drop as soon as I pull over or ease up instead of remaining hot for several minutes.

I'm leaning toward getting an auxiliary electric heater pump from a BMW or Mercedes in a junk yard and setting it up to pump hot water from the T to the thermostat when the fan comes on ("a" in the diagram above) and see if that improves things...BUT I'm at my wits end, haven't heard of any other Subaru swaps needing this, and would love any insight from the collective knowledge of the Subaru Swap Sambanistas out there.

Thanks in advance!

Sat Nov 02, 2013 11:32 pm

OK...I think I might have figured this one out...

Now that everything has cooled down from our overheating adventure today, I checked the coolant level and it was low.

The radiator fill neck is an "H" of hoses:

That peeks up through the passenger side air vent:

If I had an air bubble big enough to displace all the water in the hose to below the middle of the "H", the thermostat wouldn't get any hot water flow and would be slow to open.

I suspect I had an air bubble in the system that got dislodged on my big outing today (maybe while pushing it up the hill at 70 MPH) and blubbed to the top of my filler neck. When that air bubble was heated to 220ish degrees, it expanded until it displaced the water in the "H" of hoses.

I'm thinking my next moves will be:
1) See if it overheats anymore now that I've displaced this bubble
2) If not, next time I have the cooling system apart, splurge for a proper burp tank.

If anyone else has and ideas, I'm all ears...sometimes it just helps to write and diagram everything out.

Thanks all!
-Andy

Sat Nov 02, 2013 11:52 pm

No bypass line?
No line feeding hot coolant to the t-stat?
Or maybe that's just not in your schematic.

Sun Nov 03, 2013 12:01 am

The line labeled "a" in the schematic feeds hot water to the t-stat (or at least should if I have sufficient coolant) :oops:

You make a good point though...maybe it would be better to have the burp line plumbed to the heater feed downstream of the bypass line -- instead of in the middle of the bypass line. Thanks!

Sun Nov 03, 2013 1:09 am

I'm afraid I can't help with your overheating problem, since the cooling system isn't the same as hundreds of other conversions have used successfully. Who knows what's going on in there?? I hope you can figure it out before the engine is toast.

On a similar note, every water cooled Vanagon ever made, with less powerful engines than the Subaru, rolled off the assembly line with upper and lower grill openings. Why do you suppose that is?? Good luck!

Sun Nov 03, 2013 8:29 am

I suggest that you compare your cooling system layout to one of Tom Shiel's from his web page. Either his concept 4 or 5 work great. If yours differs in any significant way I think you should examine the functionality of that difference .
I used his concept 5 in my 2.5 suby conversion with his thermostat bypass - rock steady 185 deg when driving - on a long uphill I might hit 195 and the fan kicks on - usually just happens in hot weather eg on a long bridge waiting in line for a toll booth or something like that.
No option for me in your poll.

Sun Nov 03, 2013 9:30 am

I have had the pleasure of working on most available conversions and different cooling system layouts.

First and foremost, a new radiator is a great idea at the time of the conversion.
A simple bypass loop is needed.
TS's thermostat housing extension works well.
Reversed coolant manifolds have shown to be more streamlined and efficient from my experiences.

Sun Nov 03, 2013 9:47 am

Thanks for the thoughtful replies all.

My system is very similar to Tom Shiel's Concept 5 on a non-reversed cooling manifold except:
1. No bleeder ball valve
2. The expansion tank is replaced with hoses up to the radiator neck and cap up high in the air intake. My filler neck has about 3 feet of 5/8 hose, which can only deal with about 35 cu inches/0.6 L of air before it starts getting air locked.

Definitely makes me think it would be worthwhile plumb in a proper expansion tank that doesn't have the air locking problems. I still think built-in bong was a worthy experiment, though.

Thanks.

Sun Nov 03, 2013 11:01 am

I'm kinda with John on this. I've been using Tom's Concept 4 with a non-reversed coolant manifold for 3 years/32,000 miles with nary a hiccup and hardly ever a variance from the needle on the LED except when stopped at a light/border crossing/work zone after driving 70MPH on the highway.

Did you use a new rad? Is it in the front?

Here's the answer: does the electric water pump you installed impede flow when not energized?

There are hundreds of SubaruVanagons bombing around the continent even as we speak, all of which are heating their occupants whilst keeping their engine temps at an even 188* F. and NONE of them have electric pumps f*cking up the flow to the radiator.

Just sayin'... ;)

Sun Nov 03, 2013 11:39 am

we have a similiar fill neck mounted in the rear pillar like you have in a Audi turbo 5cyl conversion works great. you fill and bleed the rad and go, very little burping..

I don't like your addition of an Aux radiator line pump.. not sure what you thought is/was a problem that that would overcome?
the front mounted radiator system is very easy on OEM coolant pumps. there's nearly no lift and what the pump pushes it gets right back inthe return to the pump.. it's not a household/farm (well/pond) water pump system. it's a fully recirculating (boiler) system.

you need to properly bleed the system, you can have your rad cap off, while the motor is running and assure a fully bled radiator in a very easy step.

your bypass to feed the T-stat in the heatercore return line via the H-pipe is a tad on the large size but there's plenty of cooling overhead to deal with that.
there are a few formerly aircooled vans running conversions w/o a lower rad opening.. at low speeds where the rad fan would come on it's unnecessary. and I do suspect it's not necessary at highway speeds as long as your grill opening is shrouded to direct all 'rammed' air into/thru the radiator and not let it wash around the sides of the radiator.

unless you have an AC condenser in front I'd say it should be fine.

Sun Nov 03, 2013 1:15 pm

I had the same issue on my Subaru swap into a 72 - It was OK around town and short jaunts. Long road trips and steep grades finally cooked it. Ended up being the vacuum advance in the distributor had quit (hole in the diaphragm) and the timing was not working as it should.

Take off the distributor cap and draw a vacuum (suck :) on the line going to the distributor and see if the plate moves freely.

Sun Nov 03, 2013 1:20 pm

Distributor on a subaru? What motor did you have- a early EA motor not EJ series?

Sun Nov 03, 2013 5:55 pm

It looks to me like you need a more direct bypass that does not require the coolant to travel as far. It's the flow of coolant directly from the coolant manifold to the thermostate that controls the thermostate.

As for the radiator why would you even consider not cutting out the lower access? Why is it not done already?

Mon Nov 04, 2013 12:09 am

First, a major THANK YOU for everyone's comments and ideas. It helps so much just to talk through all this stuff and hear everyone's ideas and experiences. It also lightened my spirit a lot to get reminded of all the stories of rock-solid 185F temp readings. I'll get there!
Gizmoman wrote: Ended up being the vacuum advance in the distributor had quit
Thanks for the idea...but yeah...there's no distributor or vacuum advance on the Subaru EJ motors. One thing to rule out, right? ;-)
danfromsyr wrote: we have a similiar fill neck mounted in the rear pillar like you have in a Audi turbo 5cyl conversion works great. you fill and bleed the rad and go, very little burping..
Thanks so much for the sanity check. The high-mounted neck seems like a much better idea than a burp tank mounted lower than the top of the radiator...It makes me feel a lot better to know that you have gotten this to work.
Jake de Villiers wrote: does the electric water pump you installed impede flow when not energized?
Great question...the manufacturer says "no" -- that it's fine to install as an aux water pump in an existing line and that it doesn't impede flow when off. I haven't noticed a big difference since installing it.
danfromsyr wrote: I don't like your addition of an Aux radiator line pump.. not sure what you thought is/was a problem that that would overcome?
After I initially did the swap in the heat of the summer, it wouldn't cool at idle -- the temp would just keep climbing while it idled but would drop down when I revved it. At the time, I thought it was because I was getting insufficient water flow at idle to push coolant through all the extra hose from the non-reversed coolant manifold. In retrospect, I suspect that was related to some other cooling issues I have since resolved, and maybe related to not plumbing hot water to the thermostat directly enough.
Syncrozilla wrote: It looks to me like you need a more direct bypass that does not require the coolant to travel as far. It's the flow of coolant directly from the coolant manifold to the thermostate that controls the thermostate. insyncro wrote: A simple bypass loop is needed.
I think you guys nailed it...That change will be totally easy to make. I'll make it and report back maybe after some more misadventures next weekend!

Thanks again to everyone out there!

Mon Nov 04, 2013 10:52 pm

What is the ID of that bypass hose going to the radiator cap? I run a 1/4 ID hose to my Subaru thermo bypass set up and it has always been steady. Too large a hose and you have effectively created a big shunt for hot coolant to return to engine without ever being cooled.

Tue Nov 05, 2013 6:48 am

syncrodoka wrote: Distributor on a subaru? What motor did you have- a early EA motor not EJ series?
Yea, I missed that detail - it was an EA. Still faster than stock though (till I killed it)

Tue Nov 05, 2013 7:03 am

we used (use) a Polaris ATV radiator neck when mounting in the rear intake pillar
personally my not always humble (but at times humbled) opinion is this would have been the best factory position
it eliminates all that song&dance with bleeding the system, it creates a fill and Air purge at significantly the highest point in the system.
you can run the engine w/cap off and pour in till the Radiator bleed screw spills clear coolant.

but I would have thought a simple pressure cap & burb in the front would have been ideal as well.

but hey what do I/we know?

I'm sure it would have been fairly easy to protect pedestrains from a venting cap with a redesigned plastic grill for back there.

Tue Nov 05, 2013 8:30 am

No subaru experience, but I would at a minimum wire up a light to the stock vanagon radiator fan switch. While turning on the radiator fan via the ECU is a way to do that, the radiator is in the next county on a vanagon. The idiot light connected to the radiator will tell you if the radiator is actually flowing and getting to temp.

Tue Nov 05, 2013 8:50 am

connorsvw2 wrote: What is the ID of that bypass hose going to the radiator cap? I run a 1/4 ID hose to my Subaru thermo bypass set up and it has always been steady. Too large a hose and you have effectively created a big shunt for hot coolant to return to engine without ever being cooled.

Great question/consideration. I my bypass line is 1/2" at the water pump. This is the factory Subaru Turbo water pump bypass fitting diameter and also the Tom Shiels bypass diameter...so hopefully it's not too big.

The line goes from a 5/8 Tee off the heater line, Through an adapter to a 1' hose, through the radiator cap, through another adapter to a 5/8 line and finally back down to a 1/2 line at the water pump. I'd think the flow would be governed by the bottleneck in the system -- the final 1/2" line.

danfromsyr wrote: we used (use) a Polaris ATV radiator neck when mounting in the rear intake pillar
personally my not always humble (but at times humbled) opinion is this would have been the best factory position
it eliminates all that song&dance with bleeding the system, it creates a fill and Air purge at significantly the highest point in the system.

I'm with you 100%, so why not give it a shot? :-)

damfromsyr wrote: I'm sure it would have been fairly easy to protect pedestrains from a venting cap with a redesigned plastic grill for back there.

Oh crap! I didn't think of that! Mine is even on the curb side at approximately face level! Kind of worse than getting splashed by a car driving through a puddle on a rainy day. :shock:

On the other hand, the chances of all this coming together for my one installation is pretty slim -- probably on the order of a toe shoe zealot loosing a toe after stubbing his foot on my wheel as I drive by ;-)

Tue Nov 05, 2013 9:17 am

I don't like the look of the filler neck thing in the pillar. An expansion tank works on everyone else's subaru conversion. The radiator being the high point is a good thing, any air winds up there and you crack the bleeder screw and pressure pushes it out. Better to have air in radiator than in air locked hoses.

I would try to make the system more similar to other subie conversions.

But one thing comes to mind. If you're intermittently overheating, do you have an internal head gasket leak? People sometimes think they haven't bled their system enough when that happens. They they think they have it fixed about the time it overheats again.

No matter how hard you push it up a long incline, at 4,000 rpm, the system can keep the temp right in the middle. Sounds like the coolant flow is messed up.