[STINick19]

My main concern would be if adding the additional coolant passage disrupts flow in a negative way. The pressure differential created by this mod should theoretically reduce flow to the other side of the engine. This could raise the average running temperature on the RH side of the engine (cylinders 1 & 3) by preventing sufficient coolant from flowing through the crossover pipe to the right cylinder bank. This will also impact the turbo temperature (and potentially its usable lifetime) since it receives coolant flow from that side. Reducing the flow to the RH side would ultimately reduce the flow through the radiator resulting in increased coolant temps (any coolant that passes through the new coolant path created by the kit will bypass the radiator). Which of course is probably why Dom tested various diameters

As I pointed out above, the average EGT increased by 18 degrees with the kit. This is largely due to the RH's EGTs increasing by 32 degrees. The LH's average EGT stayed relatively consistent (within 5.5 degrees).

This obviously proves my concern of causing the RH side to run significantly hotter, right? Not quite.. As you can see below, the between bank EGTs (RH vs LH) are actually much closer with the kit (balanced EGTs between cylinder banks = a good thing). However, the average EGT did increase which could be bad if that correlates to an increase in overall coolant temperature. Again, knowing the testing conditions and coolant temps would help draw a better conclusion here. We're just theorizing at this point.

Without kit:
Left side avg: 866.5 (832, 901)
Right side avg: 837.5 (848, 827)
Bank delta: 29 degrees (left side hotter)

With kit:
Left side avg: 872 (862, 882) - increase of 5.5 degrees
Right side avg: 869.5 (874, 865) - increase of 32 degrees)
Bank Delta: 2.5 degrees (balanced)

We have to keep in mind that we're analyzing EGTs in the before & after comparison rather than coolant temperatures per cylinder. EGTs can be influenced by a lot of factors (AFR, timing, etc.). Coolant temperatures would be much harder to measure per cylinder, but are easier to conceptualize with this kit and would provide a more accurate picture of the product's effectiveness. Any alterations to the tune and/or operating conditions can drastically impact EGT measurements.

You can certainly tell that Dom did his research, spent time developing this product and providing information to the community. I'm absolutely not questioning that. We're just missing some of the data and background info (that he probably has) needed to draw a conclusion.

It's unlikely that this kit would cause any real issues, but I was just analyzing the EGT data that Dom provided. I doubt there would be a problem considering that some twin turbo JDM models used this passage for cooling the 2nd turbo, but we all know Subaru's have other issues that they can be a bit slow to fix I'm sure that if there was a real issue we would have heard of it by now. So then then question just becomes the effectiveness of the kit as the theory behind it is logical.

Conclusion
Assuming that the testing conditions were controlled. We should be able to draw the following conclusions:

1. The EGTs are more consistent (balanced) between all 4 cylinders with the cylinder 4 cooling kit
2. The EGTs are more consistent (balanced) between the the RH & LH cylinder banks
3. The overall average EGT increased with the kit

Unfortunately, we need to have coolant temperature measurements to prove that this kit fixes a "huge imbalance" caused by "inadequate cooling". That's not to say that it doesn't offer any benefit. Without knowing the conditions of each of the test scenarios, it's difficult/impossible to draw any quantitative conclusions on this coolant modification. Time may tell if this helps prevent cylinder 4 failures.

Disclaimer: I am not an engineer or by any means an expert in thermodynamics or fluid dynamics.

 

[dlheman]

I am not sure whether the RH banks are hotter to balanced the LH, but I think the figures shown by Dom is to show how the delta between all cylinders are more equal to each other. I did saw a reduced in overall coolant temperature by about 5 Deg C (this is also confirmed by Dom via e-mail to me).

I think it is a pretty safe and worthwhile mod to do, but of course only time will tell if it truly protects the 4th cylinder.

 

[STINick19]

Thanks for the feedback. I'd almost expect the coolant temperature to increase at least a bit for 2 reasons:

• The coolant mod allows some amount of coolant to bypass the original path through the radiator (from cylinder 4 -> heater hose -> water pump housing inlet -> cylinder 2). While probably a minimal amount, it does reduce the percentage of coolant passing through the radiator.
• This mod claims to reduce the cylinder 4 "hot spot" which would mean that the coolant is absorbing more heat from cylinder 4.

Maybe there are some tuners or track-day junkies out there that have firsthand experience with the before & after results of this kit. I have a feeling that most have just purchased this kit as a "why not? it's cheap, additional reassurance".

 

[STINick19]

My thoughts on ideal testing.. which is very unlikely to ever happen for an $85 product :lol:

The before and after tests should only change one parameter - the cylinder 4 coolant kit.

• Tests should be performed with the same ambient air temperature, pressure and humidity (ie same car, same day, same dyno).
• Tests should be performed at steady-state loads for long enough time to achieve thermal equilibrium. Data logging should occur during steady-state thermal equilibrium. Again, use a dyno to control load and minimize outside variables.
• Tests should data log IAT, oil temperature, AFR, timing, etc. to to verify comparable testing conditions and rule out data manipulation.
• Ideally tests would monitor the impact on coolant temperature, coolant pressure, cylinder pressure curves, EGT, AFR, knock events, etc.
• Tests would use contact-style temperature sensors on each cylinder (block & head) rather than an infrared thermometer. Ideally coolant sensors would be adapted to each cylinder, but the contact probe is a less invasive option.

It would also be interesting to see a similar test performed with equal length and unequal length headers as we know that an ELH will help balance EGTs across the cylinders.

Dom's IR gun measuring the surface temperature of the cylinder head showed a 4-5 degree difference at ~1-2" from the fitting. With the thermostat open and cooling fans on at idle, the temperature difference was ~1-2 degrees at most. He claims to have measured coolant temps using a temperature sensor fitting, but doesn't provide any of that data.

Yes, I'm over-analyzing an $85 product. I love the theory behind this and the fact that it could make the EJ more reliable, but the data provided just doesn't produce a convincing argument for this product's effectiveness and/or potentially negative consequences.

 

[yamahaSHO]

I suppose the other question is... What is causing the #4 cylinder to be the most knock-prone?

 

[viper_crazy]

The line the cooling mod connects to I believe is under suction. The coolant exit port at the top of each head is also under suction. The right (USDM passenger) side head on turbo models basically already has this #3 cooling mod from the factory for turbo coolant.

In factory form, on factory single turbo Subaru's, there is inadequate flow around #4. Adding the #4 cooling mod allows the coolant to be sucked from two spots on the left (USDM drivers) side, essentially mimicking the right side.

The idea for the reason why #4 is so knock prone is due to heat. The hotter a cylinder is the closer to the auto-ignition (aka knock) point the cylinder runs and thus timing is decreased to reduce, or best case to avoid, knock.

Dare I mention this, but if an $85 bill is a hard pill to swallow, Subaru has a plug for around $5-$10 for their twin turbo applications. You could easily DIY this mod with a bit of ingenuity, however there is a disclaimer. Dom claims that the ID of his plug is the key to balanced temps across all cylinders. That's all I'll say about a DIY kit as this thread isn't about cost or DIY kits so feel free to PM me.

Sent from my SM-G955W using Tapatalk

 

[STINick19]

In factory form, on factory single turbo Subaru's, there is inadequate flow around #4. Adding the #4 cooling mod allows the coolant to be sucked from two spots on the left (USDM drivers) side, essentially mimicking the right side.

The idea for the reason why #4 is so knock prone is due to heat. The hotter a cylinder is the closer to the auto-ignition (aka knock) point the cylinder runs and thus timing is decreased to reduce, or best case to avoid, knock.

I agree that the logic makes sense, but where is the data that shows inadequate flow around cylinder 4 causing it to run hotter and that this kit resolves that?

If this is indeed true, why haven't Subaru engineers figured this out and fixed it over the years? It could certainly reduce warranty claims for a negligible increase in manufacturing cost.

Keep in mind that while cylinder 3 supplies coolant to the turbo, that coolant is still passed through the radiator afterwards for cooling. That's not the case with this modification.

I'd gladly pay more for the kit if more test data was provided.

Dom originally posted this on another forum:

Cylinder 4 fails first regardless of the intake manifold used.
Cylinder 4 fails first regardless of the exhaust manifold used.
Cylinder 4 fails first regardless of IAT, Intercooling, tuning etc.
Cylinder 4 fails first regardless of EMS or fuel setup.

*unless you intentionally add more fuel and remove timing from cylinder 4 of course. You'd have to give it more fuel than other cylinders and less timing.

That all proves a huge imbalance. What's the source of the imbalance? Inadequate Cooling.
Which I fixed.

My mod is proven to cool the head (allowing it to wick more heat from the chamber), as shown by both EGTs and Surface temps, that means it's just providing improved reverse cooling and aluminum head benefits. No other data should be needed. However, many customers have written me about more balanced temps, lower operating temps, near elimination of random knock events etc. I don't know what else to provide to the public.

Dom initially referred to this as the reverse coolant mod.. except that this mod doesn't "reverse" the flow of coolant. It creates a another negative pressure outlet for the normal flow of coolant. The pressure differential this creates will reduce the coolant flow to the other side of the engine (and through the radiator). This could raise the average running temperature on that side and impact how hot the turbo CHRA will run since it receives coolant flow from the same port on the other head.

"Reverse" cooling is where the coolant enters the cylinder heads first, then flows through the block's water jacket and out to the radiator. This was popular on old Chevy engines so that the heads would receive the coolest coolant first in hopes of reducing detonation (due to poor head design).

 

[dlheman]

The only real way to accurately test this mod is to connect individual EGT sensors and measure before and after.

The video which Dom shows a reduction of temperature on 4th cylinder when coolant is allowed to enter the 4th cylinder water jacket is good enough for me. So in my simple mind, yeah why not give it a go..

I wasn’t even expecting a reduced in coolant temperature but it did (now this could also be due to the increase of coolant quantity).

So yeah ;D

 

[RocketRobin]

If knock = pre-ignition, why does the ECU have a knock sensor?

 

[STINick19]

Interesting information on Cylinder 1 & 4's exhaust port design and theory on how it may influence cylinder head heat retention and cause higher EGTs for those cylinders even with an ELH.. let alone an UELH.

 

[STINick19]

I also wonder if the boxer engine design contributes to cylinder 4 failures at all. The downside of a boxer 4 engine from a balance perspective is that pistons don’t sit directly across from one another. The opposing pair of cylinders are slightly offset from each other along the crankshaft. This means they can create a torque that wants to rotate the engine about the vertical axis and create a rocking motion. This rotational moment could in theory accumulate at cylinder 4 and possibly cause deformation. The flat 6 design doesn't have this problem.

 

[awenthol]

For what it's worth (to those trying to DIY), the plug appears to be m28x1.5. Looks like they're making their own fittings, hence brass.

IMO, they did the right thing going with a 10mm ID, so youre not bleeding off too much of the flow to fully disrupt the system.

FWIW, I found a company making m28 to -10AN adapter... Would want to neck it down to -8. Link if allowed AN10 (M28 x 1.5mm) Fitting | Motorsport Industries

 

[EightTenths]

Not sure if anyone has mentioned this but cylinder 4 cooling hose is on Subaru Japanese race engines according to my builder. I ran one for 10k miles on my 2017 STi without any issues.

 

[Leemanfor]

I will be upgrading to this here this winter, when I pull my engine out for reseal and new clutch. It definitely helps cooling

 

[SonicBoom]

Had it installed recently, along with downpipe, AOS, One-Step colder plugs and ECBS with Protune. Can't really say if it helps or hurts, but I figured for $85... Why not?