Coolant pressure inside engine 1

[jcd06]

It begun when a friend started talking about replacing the stock water pump by an electrical aftermarket pump on his CA18DET engine.

One of the elements in the manufacturers installation guidelines struck me: remove the thermostat.

This brought me to think about what happens inside this liquid system.

In the case of this engine, and I think for most production car engines, the path of the cooling circuit is: pump – engine block – cylinder head – thermostat – radiator – pump.
The thermostat is open when the engine is at normal working temperature.
How much its valve is lifted from its seat depends on the design and the momentary situation. Obviously this will be different at idle or when racing, when freezing outside or in hot weather.
I gathered some figures from a couple of random service manuals and I found values of 8-12mm when heating it in a pot of water.
It is not hard to imagine that the thermostat being open only that little will cause a flow restriction in the coolant path.

Many literature mention the pressure generated inside the engine and the head because the pump is pushing its liquid against the restriction caused by the thermostat.
This pressure is supposed to be beneficial because it further rises the boiling temperature of the coolant.
Especially inside the head, where the coolant runs sometimes very close to the hot exhaust gas, only separated from it by less than a centimetre aluminium wall.
As I didn't find real-life values of this pressure, I decided to measure it myself while I already planned to measure some other engine parameters.
I took advantage of the threaded hole for the dashboard gauge temperature sender in the thermostat housing to plumb-in a pressure transducer.
Link to the graph

The signal was logged and plotted against RPM.
As you can see, pressure went well over 4 bar.
Mind you, this was at around 5°C ambient. When the weather is hot, the thermostat will open further to achieve the same coolant temperature.
When the thermostat opens further, the pressure generated by the pump will decrease.
Also, this is a fairly high-revving, high hp/cc engine.

Anyhow, this 4bar is a rather high value and I must say that it surprised me.
Installing an electric waterpump and removing the thermostat, you lose this advantage completely.

I was under the impression that nowadays some cars are OEM equipped with electric waterpumps.
Am I wrong and are these just secondary pumps for accessories circuits like heating or turbo core?
Or is the pressure that I measured not really necessary and to be considered completely as parasitic in terms of power consumption?
Or are these primary electric water pumps (if they really exist) fitted to low power engines with a much better thermal design of the internals of the cylinder head casting?
After all the CA18DET was a design from the eighties.

Also worth noticing is that the curve is not flattening off at high revs. Am I right to assume that the pump is not (yet) cavitating?

I would appreciate to know your professional opinions about this.
Thank you for sharing your knowledge.

Jean


When we all think alike, no one thinks very much. ---Walter Lippmann

 

[MiketheEngineer]

I don't think the thermostat helps raise the boiling point - because when it opens - it can't help. My undestanding is that electric pumps are somewhat more efficient and can be better controlled.

My $.02

 

[MikeHalloran]

The curve you measured is the system resistance curve for the cooling circuit, seen from the water pump's perspective, over a range of RPM.

An electric pump with no fancy controls will work at one RPM only, just one point on that same curve, and probably not a high RPM, so you wouldn't see 4 bar there. The electric pump is more efficient, when and to the extent that the mechanical pump is spinning faster than it needs to in order to meet the engine's needs.

Drag racers use electric pumps to save weight, and to save that excess power needed to run the mechanical pump at higher speeds. They are also relying on the thermal inertia of the block and the coolant, so the electric pumps sold for drag racing might or might not be large enough to cool the same engine on the street.

I'm not aware of any street cars with an OEM electric pump, but I don't keep up with that stuff like I used to.


Mike Halloran
Pembroke Pines, FL, USA

 

[patprimmer]

There is a company in OZ (Davies Craig) that makes electric pumps and I am 99% sure some are supplied for OEM.

http://www.daviescraig.com.au/Electric_Water_Pumps-list.aspx

As they are controlled by a switch sensitive to coolant temperature there is no need for an additional thermostat.

If the thermostat does offer substantial resistance to flow to the extent that it builds that much pressure before the pump flow stalls, then there would be a substantial drop in pressure at the suction side of the thermostat, likely inducing cavitation at the pump if the temperature where above moderate, but then the thermostat would be opening anyway.

I think your over thinking this. Just fit according to the manufacturers recommendation.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[JayMaechtlen]

Wait a minute- if you remove the thermostat, is the electric water pump turned off completely until the temp hits set point?
Is it variable speed, or just on-off?

At least the thermostat is an analog device, doesn't just slam open and closed.

Jay Maechtlen

http://www.laserpubs.com/techcomm

 

[NickJ67]

With the mechanical pump, where there is no control of the pumped flow (it just varies according to engine rpm), the thermostat is there to proportion flow either though the radiator or recirculate through the engine passages according to the coolant temperature and thus regulate the engine temperature. It should have no effect on pressure except indirectly via temperature.

With the electric pump comes the ability to control temperature via total system flow instead switching the radiator in and out of circuit, so if operated by the pump manufacturers controller it will run at a low rate when the engine is cold and speed up when warm/hot so in theory it will only flow what is necessary to get the cooling needed. This is why the thermostat is no longer needed. I imagine that the placement of its temperature probe is fairly important.

As a pump engineer I have a problem with your pressure figures as I just cannot see a small diameter centrifugal pump generating 4 Bar differential or anything near that. I suspect the decimal point is displaced and the true figure is more like 4m (0.4 bar).
The system pressure should be a function of system temperature and not the pump flow. Ultimately it is controlled by the pressure cap on the header tank, which I'm quite sure will release at considerably less than 4 Bar. Probably more like 1.5 - 2 Bar.

I do agree with your comment that the steadily increasing pressure shows that the pump has not reached cavitation point. That could change when the coolant heats up though; one reason why the system is run under pressure is to delay the cavitation onset and is in fact why a faulty pressure cap can cause overheating, especially at higher rpm.

As Pat says, if you fit according to the manufacturers instructions then it'll probably work just fine.

Cheers

Nick

 

[patprimmer]

Nick.

I think from memory, that an open thermostat in a typical car probably flows about the same as a 1.5" pipe. maybe a little less. With the thermostat closed the bypass is more like a 5/8" pipe on a fitting with more like 1/2" ID so there is a substantial difference.

I was inclined to say impossible to a 4 bar difference between pump inlet and pump outlet, but I'm not a pump engineer and didn't really know. It just sounds impossible for an automotive water pump to generate 45 psig in a system with typically a 15 psig pressure vrelief valve.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[ivymike]

I don't want to fire up the work computer on a weekend, but from memory I believe some centrifugal coolant pumps can sustain a pretty good flow w/300-350kPa pressure rise across the pump. The rad cap establishes a minimum for the system, roughly (there is some pressure drop from the top tank to the pump inlet). If I remember on Monday I'll pull up a pump curve and double-check...but the curve I know I have handy will be for the pump on a 300 L engine, so perhaps not as interesting to you anyway (although it's still just spinning metal vanes inside a housing).

 

[JayMaechtlen]

Well, the thermostat is around 1 1/2 inches in diameter, maybe.
The hole in it that passes coolant is smaller than that, and has parts in the flow path that would cause some restriction.
It would have to be a big thermostat or a long pipe for the flows to be equivalent.
Since the pump is working against a restriction, and the pressure relief is downstream from the restriction, of course the actual pump head can be higher than the pressure relief setting.





Jay Maechtlen

http://www.laserpubs.com/techcomm

 

[BrianGar]

''This brought me to think about what happens inside this liquid system.

In the case of this engine, and I think for most production car engines, the path of the cooling circuit is: pump – engine block – cylinder head – thermostat – radiator – pump.''

You forgot about the cabin. The cabin is on the engine loop also. If your pump is to be on/off - temperature controlled then you will have no heat in the cabin once the matrix is cooled by fan. I would think very hard as to where to mount said temperature sender too.

Also, I still think removing the stat id a bit daft, on some housings, they have to be there in order for the coolant to flow correctly.

First image is what I call a single plate stat, It is from a Subaru, it is a basic, 'stop the flow from housing out to hose' type

http://images2.carpartsdiscount.com...ostat_77c_171f_oem_21200_aa072_21200aa072.jpg

Second Image is a Vw Mk2 Gti stat, and what I call a double plate, not only does this open the flow to the rad hose, but it also uncovers a port machined into the block and re routes coolant. If you remove a stat from a Gti it all goes a bit haywire in terms of circulation.

http://images.whisystems.com/smartpages/partinfo_resize/IMC/078121113FMY.jpg

I cant remember exactly what happens with the two plate - unbelievably since Ive studied these engines in great depth. Perhaps Ill remember in the morning, its 5.40am here at the minute.

Check your exact coolant path out before you do anything, thats all,

Brian,

@ Pat, I just got reminded of being in a pit some time ago, the head mechanic was screaming down the phone ''We need another Craig David pump'' it was hilarious, and tense at the same time as he was paying, and a right old grump.

 

[ivymike]

my memory was a bit off - at 2250 pump rpm, the pump I was thinking of delivered 3000lpm at 225 kPa pump rise.

 

[jcd06]

Pat,
For its pumps Davies Craig recommends a controller.
It seems to feed the pump with a combination of pulsed and proportional driving, dependent on the operating point.
They have a very good reputation in the tuning world. I would expect nothing less than this stuff to work when fitted as per manufacturers recommendations.
I do not want to question the idea that 2 flow + 2 pressure can be replaced by 3 flow + 1 pressure. That is basically what they do as they keep the system pressure, they do not rely on any pressure bonus generated by the pump but rather rely on more flow.
But perhaps I'm over-simplifying this one
I measured the pressure just upstream of the thermostat, while the rad cap is downstream.

Nick,
I was at least as sceptic as you about this 4bar
They are real. I tested the sensor and measuring chain with compressed air to double check.
On the graph you can see I have been driving the engine till 3600rpm while warming up, then up to 5400rpm and when it was warm enough (and reached system pressure) till redline.
To know the differential pressure generated by the pump, the system pressure (0.8bar) should be subtracted.

Brian,
The thermostat in this engine is a single valve type, it does not block any bypass when it opens.
But your remark made me wonder about the actual hydraulic layout.
In the standard car setup there are 3 fixed bypasses.
A first one that heats the throttle body and a cold start air regulator thingy.
Smallest diameter in this circuit is about 5mm.
A second one that was feeding the turbo core; also around 5mm.
It is blocked off because the turbo was replaced by a supercharger.
A third and larger one supplies the cabin heater and the oil/water heat exchanger.
This bypass is connected halfway the engine almost at the rear of the block, roughly half a meter upstream of the thermostat.
During the test, the heater control was at minimum, which is really a closed tap in this case.
The heat exchanger has been ditched since years, replaced by a front mounted oil cooler, and that water path is blocked.
That was -roughly- a 10mm inside diameter hose.
Having this blocked off will be the cause of the pressure to rise that high inside the engine.
So my measurement turns out to be in no way representative for the standard situation

I will install a hose with a fixed restriction inside in order to achieve the same bypass as with the heat exchanger and then repeat the measurement.
This will not be before spring unfortunately.
Also with another ambient temperature the thermostat will not be open the same amount and this will influence the pressure, so do not expect a back to back measurement.
Anyway, there will be pressure generated by the pump and I am still wondering to which extent the designers have been counting on that when they designed the cooling system, this was basically the underlying question in my first post.

Unfortunately nobody showed up and said hey I designed the cooling for the new [brand] engine.
I’ll try my contact person inside BMW but he’s commercial so not sure if this will pay off.


When we all think alike, no one thinks very much. ---Walter Lippmann

 

[BrianGar]

Jc, I remembered it, give me a few hours and Ill do a decent post. I have a few Interesting links for you too, so hold your horses. Interesting, Mike? @ Craig pumps often advises the stat to be left in, but two to three 5mm holes drilled in it. Since as you mentioned, the controller outputs a square wave, and is never in a position when the pump is OFF fully i still think it is madness as mentioned to run with the stat out, even more so if you have performance in mind, and when you would want to be running as hot as possible, as soon as possible.
The twin plate stat is better known as a bypass stat - I should have mentioned that, it was late when I wrote the last post.

Brian,

 

[140Airpower]

Pat, do you have any idea which OEMs use electric water pumps in what models and for what?

 

[140Airpower]

JC, what was the water temperature during your test? I would like to see the same graph for when the thermostat is fully open.

If the thermostat is removed, high pressure (probably lower than 4 BAR) will appear at the radiator cap and may push water past it.

I'm sure you have no idea how many liters/min or gpm are flowing? With an engine-driven pump and unlimited power this will be proportional to rpm up to high powers (like over 10hp) unless there is belt slip. With an electric pump this will be lower for an increased pressure head as the electric pump is power limited (very low power, like 1/4 hp).

Cavitation will lower efficiency and will erode the blades, but won't stop the pumping. The flow increase with rpm should fall off the straight-line, but may not stop rising unless it's really bad.

 

[ivymike]

If the thermostat is removed, high pressure (probably lower than 4 BAR) will appear at the radiator cap and may push water past it.


The pressure might go up a little bit due to increased flow through the radiator and water lines from radiator to pump, but the main effect will be that pump outlet pressure drops (in some instances possibly leading to pump cavitation). As mentioned above there are also systems where removal of the tstat will give unfettered bypass flow - in those systems I would not expect an important change in pressure at the radiator cap (and much lower pump outlet pressure).

 

[140Airpower]

Hi Mike. I think you're right. I guess it will depend on the system and what is meant by removing the thermostat. Taken literally, the problem you mention will happen in, I guess, most systems. I should have said "remove the restriction of the thermostat". In that case, flow will increase and then the radiator is likely to be the most important restriction. At some flow rate the pressure at the cap will probably open it. Note that for current aftermarket electric pumps, there is almost no possibility of pushing water past the cap because they will never have enough power to make the flow that will cause the required pressure head at the cap.
This is not to say that electric pumps can't be powerful, but they would require the electrical system of a hybrid (especially higher voltages) to run the pump for a decent size engine. That is where I would expect to see electric pumps employed.

 

[jcd06]

Brian,
The drive strategy for the Craig pumps is explained in an instruction sheet on their website.
Depending on the difference between measured water temperature and setpoint, it toggles between on and off at 6V with various duty cycles or it ramps up to system voltage.

140Airpower,
As I did not log the absolute water temperature, I can only tell that it was at normal operating temperature, which is confirmed by the voltage over the temperature sensor being stable.
This is the same graph but taken into account only the data around the redline pull, but to be honnest I don't see what new insight it brings:
link to the graph

At 5°C ambient, the thermostat will not have been open completely.
Still I believe that in this case, pressure right upstream of the thermostat gives a good figure of the water flow as it is proportional to it.

Just as Mike also pointed out, you are right that with the thermostat removed the radiator will be the next restriction in line and obviously it will get pump induced pressure in its inlet tank.

BMW seems to use electric pumps on some engines:
Link
but I have no feedback from that side yet.




When we all think alike, no one thinks very much. ---Walter Lippmann

 

[140Airpower]

JC, since it was at normal temp, I am more impressed by the 4 BAR pressure. The open thermostat turns out to be quite a resistance. Getting rid of the thermostat would greatly reduce the power required to circulate the coolant.

BMW seems to have reduced the required power a lot. They do use an electric pump. I would like to know more about their cooling system.

 

[Compositepro]

An open thermostat will allow more flow and cause the pump to use more power than when the the thermostat is closed. This is the characteristic of a centrifugal pump.