[Cooling System Upgrade] 1-BAR Radiator Cap

^Ok, I will rephrase that the radiator will not be able to keep up with the cooling requirement alone in a high-demand situation as much as it does under normal driving conditions. I stand behind my logic for the rest of my statements, and I do believe the M54 operates below 1 bar unless you push it too much to its limits.

Well, I do agree with your facts even if I disagree with your conclusions. But, help me understand: why would a 1 bar cap help anything at all, even if the system does stay <1 bar under normal conditions? By your own logic, since neither a 1 bar or 2 bar cap would vent, nothing has changed overall. I just think that a 1 bar cap will vent, or BMW probably would've put a 1 bar cap on in the first place.

To further compound your thoughts on the radiator in high-demand situations, high rpms means more rpms at the pump, and thus more flow. But, higher rpms over a sustained period almost always means more air flow through the radiator, and thus more cooling/delta T/whatever (and lower system pressure, lower thermal capacity of the coolant, etc). Not particularly relevant with regard to cap pressure, but that's what the thermostat is for--to hold temperature as close to constant as possible to prevent venting from overtemperature or inefficiency from undertemperature.

Originally Posted by lcoleman

Well, I do agree with your facts even if I disagree with your conclusions. But, help me understand: why would a 1 bar cap help anything at all, even if the system does stay <1 bar under normal conditions? By your own logic, since neither a 1 bar or 2 bar cap would vent, nothing has changed overall. I just think that a 1 bar cap will vent, or BMW probably would've put a 1 bar cap on in the first place.

To further compound your thoughts on the radiator in high-demand situations, high rpms means more rpms at the pump, and thus more flow. But, higher rpms over a sustained period almost always means more air flow through the radiator, and thus more cooling/delta T/whatever (and lower system pressure, lower thermal capacity of the coolant, etc). Not particularly relevant with regard to cap pressure, but that's what the thermostat is for--to hold temperature as close to constant as possible to prevent venting from overtemperature or inefficiency from undertemperature.

Please pay attention to the bolded statements below:

Originally Posted by SeanC

Just did a quick calculation. Apparently 1 bar cap is good enough. Take it for what it's worth:

Assuming 96 degree Celcius coolant temperature, which is pretty much the max, the pressure of an ideal gas (e.g. air) is 1.3516 atms. That means you'd need a cap capable of holding 1.3516 bars at the minimum. This would be called a 0.3516 bar cap. This would, however, increase the boiling point of the coolant by a mere 8.5 degrees (Celcius):

So %100 water as your coolant would boil at 108.5 degrees Celcius instead of 100 degrees.
%50-%50 mix would boil at 114.5 degrees Celcius instead of 106.

These numbers are ok theoretically, but they are too close for comfort in real life. Locally, temperatures could be higher than these (e.g. engine block), and your coolant will turn into gas whenever this occurs. Since gas occupies more space than liquid, you'd be replacing your hoses quite often. I am assuming the engine block can take a beating.

On the other hand, 1 bar cap increases the boiling point by about 24 degrees Celcius. So for 100% water, you'll get 124 degrees Celcius as your new boiling point. For 50-50 mix, it will be 130 degrees. Similarly, 2 bar cap raises it by 48 degrees, so your new boiling points for 100% water and 50-50 mix will be 148 and 154 degrees, respectively.

Since the operating temperature of your engine doesn't change, using a 1 bar cap will lower your chances of blowing the head gasket. You'll probably see your car smoking before the temp gauge hits the maximum. But there is more chance of introducing excess air (gas) into the system, which could require you to use heavy duty hoses in lieu of what comes from the factory.

Originally Posted by lcoleman

Thinking off the top of my head here...pv=nrt would probably only work for a static system. The entire cooling system has a pressure differential (consider the points directly before and directly after the water pump) which causes flow. Also, the volume of the system is fixed, and the volume of a given amount of air changes depending on pressure, no?

Regardless of the science behind it, I agree with TxZHP04 in post #13; all this modification will do is introduce a leak as the "mean system pressure" in the ET exceeds the limits of the lower cap. Changing the cap will not change the operating pressure of the system, as it simply bleeds excessive pressure (which only happens if the T gets too high, right?). Think more like underdrive pulley if you want to reduce the pressure...and have a head gasket handy.

Long story short: waste of time.

PV=nRT (or PV=NkT for those who prefer that form) is applicable to gasses. So as the temperature increases, the pressure exerted by the gas in the ET will increase. So while the flowing water is not applicable to the ideal gas law (as it is not a gas), it still has relevance. Remember Bernoulli's equation 1/2(rho)v^2 + (rho)gh + P = constant... The "P" value will increase as a result of the pressure exerted by the gasses.

Originally Posted by SeanC

The equation you wrote is a special case of the ideal gas law. Not its simplified version. Ideal gas law does NOT have a simplified version.

And it's not steam, but air. Steam won't be produced until you boil your coolant. Do you use steam in your expansion tank to fill it up 100% or just close the cap?

Please refrain from posting stuff that goes over your head

Sean, what WDE46 wrote is indeed a simplification. Since we are assuming volume and n are constant, we can write the equations for the 2 scenarios as follows --

P1*V = n * R * T1
P2*V = n * R * T2.

Rearranging, we get:
P1 / T1 = nR/V
P2 /T2 = nR/V

Therefore
P1/T1=P2/T2

Obviously this only applies when n and V are constant, but I think it's pretty obvious that WDE46 understood that.

Originally Posted by SeanC

Please pay attention to the bolded statements below:

Sean, I disagree with with your statements about the coolant boiling around areas of higher temperature. Remember, the fluid will be flowing a lot slower around the block - there's a much larger volume for the fluid to fill. Slower fluid = higher pressure = higher boiling point.

There are only a couple ways the cooling system will actually reach 2 or 3 bar. Either A) Coolant starts to boil and the total molecules of gas increase or B) More fluid enters the expansion tank and reduces the volume the gasses can occupy (this will happen when in situations where the heater core closes, thermostat closes, or if the system is overfilled).

If coolant starts boiling, you're in trouble - the engine will overheat drastically anywhere the coolant turns into a gas. So I highly doubt the system is engineered to even allow that possibility. I must therefore conclude that the cap is to allow ventilation in the event the system is overfilled.

Originally Posted by Zell

Sub'd as well, but I believe German Auto Solutions has been documenting this for some time now.

German Auto Solutions is testing a 1 bar cap. They are close to production, which will be a 1.2 bar cap as a sefety margin. They have done a fair amount of real world and dyno testing.
Also, the way this guy works so hard to guarantee his products work, are safe, and a good value, I would believe him before any faceless company.
With his DISA kit, he sold some initial units a higher price, and then found some better pricing for come of the items in the kit, and reduced his price. Then, he went back through his books, and refunded the difference back to his original customers. Who else does that?
This guy is the real deal, honest, and doing it for the love of the cars. He says that he is taking a loss on the kits, but does not want to raise the price. He just wants to get more customers. He is me, with a machine shop at his disposal, and probably much smarter, and he is applying his skills to perfect one small unit at a time, after work hours. and he is backing it up with a really good warranty
He is also prototype testing a new CCV system. I can't wait to see what he comes up with. I hope it does not beat him. That thing is a pain.

Sounds like a bad idea. Decreasing cooling system pressure increases the chance of localized boiling and cavitation. Localized heating leads to detonation and warping. Cavitation leads to reduced cooling and increased coolant pump impeller wear.

That's a lot of pretty bad negatives just for a cheap fix for sporadic expansion tank failures.

Originally Posted by bryce-o

167 degree coolant temp today with 78 degree ambient

Your thermostat is broken if that is your coolant temperature.

Causing cavitation in a liquid is very difficult. It requires a rapid change in pressure that just wont be experienced in a car's cooling system.

Having a 1 bar cap will have some negative effects. You will no be able to keep as much coolant in your car, and it boil at a lower temp.

These could both lead to problems, however weather either of these issues rise to a level where they could cause damage... I dont know.

I am only going to say this once!

As everyone knows raising the pressure on a liquid increase its boiling point, that is the main purpose of a pressurized cooling system.

The thing with the E46 the cooling system uses an expansion tank so the coolant is pressurized against an air pocket long before it even gets close to the 2 bar pressure cap.

The pressure release in any cap is more of a safety valve or mechanism, just look at any water heater pressure release valve.

1 Bar is approximately 15 psi, which is about where most domestic cars run that do not use an expansion tank system.

2 Bar is approximately 30 psi, which is totally absurd for any car using an expansion tank system.

Just look at all the expansion tanks that burst the minute new BMW owners check their coolant and top off the expansion tank like it is a domestic car with a radiator. This is when the pressure cap is pushed towards its limit and just looks what happens, BOOM. There is no way most of the parts in the E46 cooling system could survive 2 Bar on a daily basis. Also I would hate to have the heater core blow under 2 Bar and dump scalding hot water on the occupants feet.

You can also squeeze the upper radiator hose on the car and feel there is nowhere close to 2 Bar of pressure in the cooling system, matter of fact there is probably nowhere close to 1 Bar in the cooling system under most situations. Almost anytime the pressure cap is removed from the expansion tank there is very little pressure built up.

So until someone has connected up a pressure tester and logged the cooling system pressure under daily and track conditions and can prove these engines even get close to 1 BAR, I think all this concern over not using a 2 Bar cap is a waste of air and mental capacity, assuming there is much mental capacity to even waste!

Originally Posted by SeanC

It's not a waste of time and there is a very simple justification for why ideal gas law should work here.

You are forgetting that the liquid in the system also expands when it is heated, so, the volume of gas is not a constant.

Originally Posted by SeanC

The equation you wrote is a special case of the ideal gas law. Not its simplified version. Ideal gas law does NOT have a simplified version.

And it's not steam, but air. Steam won't be produced until you boil your coolant. Do you use steam in your expansion tank to fill it up 100% or just close the cap?

Please refrain from posting stuff that goes over your head

Yeah, you're right aobut the steam. No need to be a dick though. I know this stuff, but I will admit I am a dynamics and kinematics expert, not a theromodynamics expert. You apparently just took Chem 101, though. Also, that equation is "simplified". It is two ideal gas equations combined into its simplified version, eliminating the common constants.

EDIT: You are kind of right about the steam. The air in the ET will be at 100% relative humidity, which may be significant.

Originally Posted by TerraPhantm

PV=nRT (or PV=NkT for those who prefer that form) is applicable to gasses. So as the temperature increases, the pressure exerted by the gas in the ET will increase. So while the flowing water is not applicable to the ideal gas law (as it is not a gas), it still has relevance. Remember Bernoulli's equation 1/2(rho)v^2 + (rho)gh + P = constant... The "P" value will increase as a result of the pressure exerted by the gasses.



Sean, what WDE46 wrote is indeed a simplification. Since we are assuming volume and n are constant, we can write the equations for the 2 scenarios as follows --

P1*V = n * R * T1
P2*V = n * R * T2.

Rearranging, we get:
P1 / T1 = nR/V
P2 /T2 = nR/V

Therefore
P1/T1=P2/T2

Obviously this only applies when n and V are constant, but I think it's pretty obvious that WDE46 understood that.

Terra, I know the algebra, but thanks for laying it out for those who may not. Technically it is a called special case of the ideal gas law when n and V are constant. It is not called a "simplified" version. If you insist on writing it for 2 different conditions, you need to have the knowledge of 3 of the 4 unknowns in the form that you write it (1 equation 4 unknowns). That means you first have to use the PV=nRT twice to get to those 3 unknowns (actually you'd know all 4 by then). How's this for simplified?

Originally Posted by TerraPhantm

Sean, I disagree with with your statements about the coolant boiling around areas of higher temperature. Remember, the fluid will be flowing a lot slower around the block - there's a much larger volume for the fluid to fill. Slower fluid = higher pressure = higher boiling point.

There are only a couple ways the cooling system will actually reach 2 or 3 bar. Either A) Coolant starts to boil and the total molecules of gas increase or B) More fluid enters the expansion tank and reduces the volume the gasses can occupy (this will happen when in situations where the heater core closes, thermostat closes, or if the system is overfilled).

If coolant starts boiling, you're in trouble - the engine will overheat drastically anywhere the coolant turns into a gas. So I highly doubt the system is engineered to even allow that possibility. I must therefore conclude that the cap is to allow ventilation in the event the system is overfilled.

Good thoughts, but you need to quantify how much fluid velocity will affect it's temperature, hence its boiling point, in order to convince me. I never said my calculation accounted for everything in this system. It's not comprehensive by any means. I am well aware of a lot of other things that might affect, but I am not an automotive engineer. This is the best I can do with my current knowledge of the matter

Originally Posted by dknightd

You are forgetting that the liquid in the system also expands when it is heated, so, the volume of gas is not a constant.

I am aware of a lot of other things buddy, this being just one. I mentioned about it somewhere above. My calculation is just an approximation, and I believe I took into account the most important factors, and my conclusions make sense for what they are. If you can do a better calculation, I'm all ears

Originally Posted by WDE46

Yeah, you're right aobut the steam. No need to be a dick though. I know this stuff, but I will admit I am a dynamics and kinematics expert, not a theromodynamics expert. You apparently just took Chem 101, though. Also, that equation is "simplified". It is two ideal gas equations combined into its simplified version, eliminating the common constants.

Sorry if it came that way. I have to say that I have a Ph.D. in theoretical physics. While my Chem 101 may not be as fresh as those who just took it, I know a thing or two about thermo

Let's try that again. That equation is not "simplified". You cannot combine two gas equations that "easily". If you do, you'll end up with more unknowns (4) than you have equations (1). The way you wrote it applies to systems where you have 3 of those 4 unknowns beforehand. See my reply to Terra above. I hope it's clear now.

Let's try that again. That equation is not "simplified". You cannot combine two gas equations that "easily". If you do, you'll end up with more unknowns (4) than you have equations (1). The way you wrote it applies to systems where you have 3 of those 4 unknowns beforehand. See my reply to Terra above. I hope it's clear now.

Thanks for explaining basic algebra to me like I'm a 5 year old.

No we have 1 unknown, and that is P2 in my equation. We have P1,T1, and T2 through the use of our coolant sensors and barometric pressure. There is also the assumption of constant volume, which may or may not be poor, depending on how much the water actually expands. This is where you use steam tables to solve for everything (empirically determined numbers with interpolation). I'm not writing a paper here, I thought I could skip that part of showing my work...

Good thoughts, but you need to quantify how much fluid velocity will affect it's temperature, hence its boiling point, in order to convince me. I never said my calculation accounted for everything in this system. It's not comprehensive by any means. I am well aware of a lot of other things that might affect, but I am not an automotive engineer. This is the best I can do with my current knowledge of the matter

I am one.

You're clearly a smart guy, and we both know what the other is talking about. Let's stop arguing semantics. I also see that we both enjoy math which is hilarious. I always find myself making things a bit more complex than they are, just because I want to use my math.

Originally Posted by WDE46

Thanks for explaining basic algebra to me like I'm a 5 year old.

No we have 1 unknown, and that is P2 in my equation. We have P1,T1, and T2 through the use of our coolant sensors and barometric pressure. There is also the assumption of constant volume, which may or may not be poor, depending on how much the water actually expands. This is where you use steam tables to solve for everything (empirically determined numbers with interpolation). I'm not writing a paper here, I thought I could skip that part of showing my work...

Why do you need two sets of P and T? One set is not only redundant, but will also bias the results. For example, barometric pressure P1 is not the same everywhere, and what value of T1 you choose? It is arbitrary.

Originally Posted by WDE46

I am one.

You're clearly a smart guy, and we both know what the other is talking about. Let's stop arguing semantics. I also see that we both enjoy math which is hilarious. I always find myself making things a bit more complex than they are, just because I want to use my math.

That's what I am trying to convey to you. The situation is simpler than you think it is. You only need a single temperature value to calculate the pressure in this case.

Originally Posted by SeanC

Why do you need two sets of P and T? One set is not only redundant, but will also bias the results. For example, barometric pressure P1 is not the same everywhere, and what value of T1 you choose? It is arbitrary.


That's what I am trying to convey to you. The situation is simpler than you think it is. You only need a single temperature value to calculate the pressure in this case.

I am trying to work with the same methods an engineer designing the system may have used. So I use a minimum temperature and pressure (room temp/1 atm),then I use the operating temperature. Then I solve for the operating pressure. I just used some more assumptions, but they are perfectly valid for most cases. I didn't have to look anything up to do it either.

I used 101.325 kPa for P1 and 293.15 K (20C) for P2 then 373.15 (100C) for T2. It gives me 128.98 kPa, which is close to your number, so our math checks out. I just don't see how you got the 1.35 atm number without using the equation I used. If you used the gas equation you would need to know volume, which we don't. What am I missing here?

^Yep, we both got surprisingly low value for pressure at operating temperature. Makes the use of a 2-bar cap even more of a mystery in the e46. Even more so given that the e90's use only a 1-bar cap. Wonder what exactly is different between two systems.

Originally Posted by SeanC

^Yep, we both got surprisingly low value for pressure at operating temperature. Makes the use of a 2-bar cap even more of a mystery in the e46. Even more so given that the e90's use only a 1-bar cap. Wonder what exactly is different between two systems.

I'm glad we have come to the same conclusion. I can think of no engineering justification for the 2 bar cap . I also don't think a 1 bar cap will help with the longevity of the ET. I think it WILL help prevent a headgasket failure in the event of an overheat, but in day to day operation, the system never runs over 1.4 bar and that is generous. I will continue to try to justify a 2 bar cap, but I don't think I ever will.

EDIT: I may try to calculate the temperature necessary to generate 3 bar pressure to trigger the 2 bar cap. I just got over 500C for an ideal gas, but at that temperature there will be lots of steam in the ET and the water will have expanded, reducing the air/steam volume. I'd have to reference one of my books for this, though I might have a matlab program for it.

Originally Posted by SeanC

Terra, I know the algebra, but thanks for laying it out for those who may not. Technically it is a called special case of the ideal gas law when n and V are constant. It is not called a "simplified" version. If you insist on writing it for 2 different conditions, you need to have the knowledge of 3 of the 4 unknowns in the form that you write it (1 equation 4 unknowns). That means you first have to use the PV=nRT twice to get to those 3 unknowns (actually you'd know all 4 by then). How's this for simplified?

I meant it's a simplification in that it's an algebraic simplification by definition. If you assume volume and molecules of gas are constant, you can simplify the equations by canceling the terms out. I realize that it is a special case and doesn't apply in every situation. No one is saying that the ideal gas law always simplifies to that form.

Originally Posted by WDE46

I'm glad we have come to the same conclusion. I can think of no engineering justification for the 2 bar cap . I also don't think a 1 bar cap will help with the longevity of the ET. I think it WILL help prevent a headgasket failure in the event of an overheat, but in day to day operation, the system never runs over 1.4 bar and that is generous. I will continue to try to justify a 2 bar cap, but I don't think I ever will.

I think it's time one of us measures the pressure and see if the theory matches the reality. The only thing I can think of is that if you reach atmospheric pressure while the heatercore is running, and then subsequently shut off the heat, the coolant level will rise a bit and decrease the volume for air. I don't know if that would be enough to bring the pressure to 2 bar (let alone 3) in any circumstance though.


I should probably mention - I'm just a senior in college, not majoring in engineering or chemistry, so my knowledge will have gaps compared to both of yours. But nonetheless, this has been fun

Originally Posted by SeanC

^Yep, we both got surprisingly low value for pressure at operating temperature. Makes the use of a 2-bar cap even more of a mystery in the e46. Even more so given that the e90's use only a 1-bar cap. Wonder what exactly is different between two systems.

So, all this theory is interesting, but I wonder if something is being missed in practice?

I have three other cars that have 15 PSI relief caps. Once heated up, I can squeeze on a radiator hose and tell that the system is pressurized, but not how much. On the BMW, when I squeeze on the hose (once heated up), there is noticeably more resistance to squeezing than the other cars. To me, this implies that it is indeed pressurized to a higher pressure than the other cars I have.

We can speculate about this, but I suggest that it is easy to find out for certain. Just tap a small pressure gauge in and find out what pressure it goes to. It seems like the bleed screw on the upper radiator hose would be a good place to do it. It would be a sacrifice of a $25 hose, so not too bad.