Electric waterpump project
Turbo Mini Forums

Wow, this thread has exploded!

Thanks for all the comments, it's good to hear some of you have had some very good/noticeable real world experience with similar setups. In response to some comments / questions raised:

The 10BHp mentioned is the combined effect of pump & fan.

As others have stated the fan is only useful up to low vehicle speeds, where little or no air is forced through the radiator. This in itself poses a big compromise in fan design since the fan should be able to flow enough air at idle to cool the engine in a traffic jam....whilst at higher revs / speeds over 30 mph the fan absorbs some serious power accordingly without any actual need for its assistance due to ample 'driving wind'.

The mechanical water pump actually poses the same challenge, give ample coolant flow at low engine revs....resulting in compromised behaviour higher up in the rev range such as cavitation and power absorption. This can be combatted by different sized pulleys up to a point, as high rev optimization limits low rev flow / makes the engine more prone to overheat.

This is the beauty of an electric pump and fan. When necessary the pump already delivers its maximum flow at idle, without leading to cavitation / unwanted power absorption at max revs since the speed at which it runs is totally unrelated to engine revs. Same goes for the fan, it is only activated when necessary (low vehicle speed) and again can be run at maximum capacity at idle.

It is the combination of the two that gives the cooling system more of a chance to cope at idle/low speeds than say a mechanical pump and electric fan setup that has to deal with less coolant flow under these circumstances. -sorry for some more science: Heat transfer is directly related to coolant mass flow rate says the 2nd law of thermodynamics Q=MxCpxdeltaT-

As far as required 'thermostat pressure' goes, when coolant is heated its pressure rises....the old school blanking sleeve serves another purpose, as far as I understand it is there to provide enough resistance for the coolant to flow towards #3 and #4 cilinders without taking a direct short cut fro pump entry to thermostat housing. Even in this confguration we know #4 cooling is somewhat flawed. Dry decking completely deletes this requirement.

Will it give 10HP? We shall see, I have often seen 5-7HP quoted for the fan alone and have no reason to question KAD as a company nor its dyno....

....as said to be continued

Edited by Evoderby on 24th Oct, 2013.

Good effort.

The amount of heat that the engine needs to lose to maintain a safe operating temperature is directly related to the output power at the flywheel.

So, assuming a flat torque curve for now, if you are revving to 8000rpm you need to lose twice the heat that you were at 4000rpm. Now, we know that torque curves are not flat so the ratio is probably around 3:1.

A centrifugal pump characteristic changes with speed in so much that the flow increases with speed, but the head goes up with the square of the speed. That's why the power goes up in a cubic ratio. By the way, 2^3 = 8. So doubling the speed of the pump will increase the power required to drive it by 8 times.

BUT, due to characteristics of the pump and system, you will get less than twice the flow at 8000rpm when you actually need more than that.

I believe that the mechanical pump is the best solution, but you are correct is saying that you can show on a dyno that a power saving is available by using an electric pump. However, the power to drive the electric pump then has to come from somewhere. If it has to come from the alternator then it's just a matter of comparing the efficiencies of the power source, drive and pump. That's another subject altogether.

I think you take on the "old school" sleeve is incorrect. A restriction will not influence upstream flow at all. The sleeve is used to increase pressure at the cylinder head to prevent localised boiling of the coolant.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Nice to see some thought/science behind the idea, rather than 'it'll be shit'.

I will reserve judgement until you have had the car back on the rollers and ran it in the real world. I do think the results will be questionable though, as you have changed multiple things (dry deck, rad, fan etc). It would be good to see back to back tests with the mechanical pump and the electronic pump. That would tell you the difference without any argument.

For me personally I wouldn't be doing it for power gains, I'd be doing more for control of the coolant flow and engine temperatures, especially important when boost and detonation is concerned. Up until a point I was gonna stick with a decent side mounted rad, but having seen so many ways of having the cooling system I haven't finalised anything yet. Although having a clubman fg front with no inner wings gives me more freedom to place things than a steel roundnose for example. I quite look forward to seeing the results, It would be nice to be able to see the before and after effects of the temperature in different locations around the engine too, especially around 4 with and without the dry deck.

If nothing else, getting rid of the mechanical fan makes it quieter.

Both my turbo cars run without a mechanical fan and the thermo controlled electric one only cuts in on a hot day after standing in traffic for at least 10 mins.Fans do use up 2/3 hp. depending on size and speed.

From my own back to back testing with mechanical pump and fan vs a similar install as you there was only 4hp difference. This is on a far more powerful engine than yours so (and I stand to be corrected) Im guessing your lower powered engine will not be able to take advantage of the lesser friction and drag as mine therefore giving less power gain.
Tbh, I dont really believe much of what KAD have told me over the years

I seriously doubt it!

The more I think about this, the more it interests me.

If we just forget about pumps for a minute and just consider what is required to cool an engine in a steady state mode. Say it requires X litres/second of water circulating to keep it at a steady temperature at at 50hp at 4000rpm, then it will need 2X l/s to keep it cool at 100hp at 8000rpm.

Now all pumped system have to work against the restriction in the system, usually friction losses. These pressure losses vary by the square law. If you double the flow, the losses quadruple ie. the pressure needed to move the water goes up by a factor of four.

Power required to pump water is a function of the flow times the pressure. Hence doubling the flow needs 8 times the power.

Now, as I said above, the characteristic of a centrifugal pump varies with speed in exactly the same way. Flow goes up proportional to speed whereas the head goes up with the square of the speed.

So a mechanical pump is the best means of cooling an engine under steady state conditions.

Where an electric pump will have an advantage is when it does not respond to transients. So you could size the electric pump to maintain a steady state under cruise, but use the capacity of the cooling system to absorb the extra heat generated by short term accelerations or dyno pulls. I'm guessing that under duress the electric water pump could use stored energy from the battery as well as the alternator, making further savings at the flywheel.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

I tried the car without the mech fan, and it wouldn't keep cool at speed which isn't what you would expect when reading up.
I showed a 2 hp increase but the reading wasn't on the same day.
I put the mech 2 blade fan back on

I run a supercharger and I don't care the TB is on the wrong side.
VEMS + 12 PSI + Liquid Intercooler = Small Bore FUN!

Thanks for your indepth input Paul! -BTW I already corrected the 2^3=8 in my initial post -

Indeed the power to drive the electric pump has to come from the battery/alternator. When looking at the Meziere pump it comes fitted however with a 20A fuse so let's assume 15A is a normalised max output figure. This equals 200 Watt.

Now let's assume 50% alternator efficiency, this means 200 'water pump Watts' become 400 'crank Watts' or 0.4 Kw = 0.53 HP

Accordingly, the fact whether the KAD dyno mule had an external electrical power supply or had to rely on alternator power is negligible in the light of the reported power hike. It also points towards the apparant 'overspeeding' with a factor 2 - 3 of the standard water pump.

I do agree that pump flow is directly proportional to pump speed, and that given your flat torque curve this exactly matches the rise in heat load between 4000 vs. 8000 rpm. However 'deltaT' in the equation as achieved by the radiator is not a steady value and not so much affected by mass coolant flow, as mass AIR flow. This is why mechanical water pumps 'overspeed', because they have to provide enough coolant flow at idle / city traffic with little or no airflow over the radiator.

Without such requirement, as you say, a mechanical waterpump is perfectly suited for steady state situations with enough airflow over the radiator. Just select a pulley size that supplys enough flow and uses the least power to spin the pump.

As far as the thermostat blanking sleeve is concerned....additional pressure created by a restriction has a very limited effect on boiling point, in either case it's safe to say nucleate boiling will occur in the cilinder head (as it does with thermostat in place). This gets us in an area where I haven't been able to find any ready made scientific answers, surface tension (reducers as water wetter) and coolant flow / speed seem to be the most significant elements in coolant system performance / control over localised hot spots rather than pressure (???)....as it focuses on reintroducing gasses into the coolant stream most effectively rather than tryning prevent what will happen anyway.

BTW the latter is why Evans waterless coolant sucks!

Edited by Evoderby on 25th Oct, 2013.

You'll be fine using that pump for cooling, i drove for a steady 110mph for about 4 hours (on a nice straight track) :) and it didn't go above 85 degrees. That was using the same pump, chinquecento rad and an 80 degree Shuttle type thermostat. Couldn't comment on power losses, it definitely takes a fair amount of current to run, but to keep the engine at a steady running temp is worth more in my eyes than power gains.

https://www.facebook.com/pages/Fusion-Fabri..._homepage_panel

www.fusionfabs.co.uk



1/4mile in 13.2sec @ 111 terminal on 15psi

The power and what makes the power is irrelevant - just trying to say that while its a worthwhile thing to do, it wont make as much difference as you are hoping. Im not trying to piss on your bonfire - just pointing out my experience that the power gain is little and I believe that its not a case that it will be the same for every engine, I think it will be less of an advantage on a lower powered engine because its not a case of gaining power but claiming back lost power - if you can understand that, Im not very good at explaining myself and rereading my previous post it may have seemed a little abrupt.
I think Im gonna shut the fuck up now, so much for trying to help

Edited by paul wiginton on 25th Oct, 2013.

I seriously doubt it!

Paul, as said before I appreciate all comments....positive or negative.

Since you mention to have performed some back to back testing on the subject I am genuinely interested in how you tested this. I know for a fact that when my project is finished so many other little things will have changed that 'after' testing on a rolling road will not be fully representative of the pump and fan alone....I think that is the beauty of the KAD test, a reputable firm testing the same engine under the same conditions, without anything to gain in terms of marketing a bespoke KAD part.

Anyway, feeding power through a set of gears to drive the wheels does have a marked relationship between the power going in and power going out. When making improvements to the efficiency of those gears more HP wil be 'recovered' in absolute terms when feeding 200HP through those gears compared to 100HP. Accordingly, I can see where your thoughts are coming from.

With a water pump and fan this relationship however doesn't exist. Trust me, this is pure scientific fact. The only thing that determines the power usage of those devices is the RPM at which they are spun, doubling pump RPM requires 8 times the power to drive that pump.

This indeed means that engines making max power higher up in the rev range can have more benefit of having an electric pump and fan than those that make power lower down the rev range. Where a turbo mini previously making 150 HP @ 5500 rpm may only see a 5 HP 'increase', a balls out short stroke 998 making 130 HP @ 8500rpm will see 20HP according to the KAD test.

Now the (KAD) numbers may be wrong, the relationship between pump speed and required power is simply fact....sorry for the long answer

If my tired brian is working right, is the cooling system on a Mini/A-Series not a pump assisted thermos-syphon system ?

I know I've driven many miles with a snapped fan belt with no problems at all as long as my speed didn't drop below 30-odd mph and I didn't go above 55-ish.
Now that was a 998 not a tricked out turbo beast, but the cooling system still works on the same principle.

Rambling thoughts:
At tickover and low rpm the pump and the mechanical fan are the crucial elements of the system.
As rpm rises, the pump will become inefficient and possibly cavitate, so the thermo-syphon is more important.

I suspect that this is why the power increases are not as great as they could theoretically be.
An electric pump has to pump the required amount of water at all speeds - I'm not conviced that it will allow a thermo-syphon to occur - so it never gets a free ride.

Metric is for people who can't do fractions.

The other thing to consider is the NPSH (Net Positive Suction Head) requirements at the pump.

An explanation of NPSH can be found here:

http://www.engineeringtoolbox.com/npsh-net...head-d_634.html

Water temperature has a big impact on the NPSH available. At 80 Deg C, the pumps needs about 0.5 bar extra pressure at the inlet to operate effectively. Also NPSH required goes up with the square of the speed.

In all probability, the mechanical pump will suffer from insufficient NPSH well before it gets to 8000rpm. I don't like to use the word cavitate as it means different things to different people. I prefer to say the the pump fails on suction. Basically, the pump causes a breakdown of the water column at the inlet of the impeller and cannot get hold of the water to move it.

What would happen, as engine speed increases, is that the water flow will reach a peak and the pump will not be able to push any more. In these circumstances the pump will not absorb the sort of power expected from the cubic law relationship.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Whenever buying a water pump, I've always opted for the large impeller type, not sure if it helps or not with water flow.

Paul, I think you've expressed what i was getting at a bit better than I did.

If the impellor is spinning in it's own froth but not really pumping, it probably isn't using as much power as the sums would suggest.
It will also then be acting as an impedence on the water flow which needs to be much higher than the system can provide.

An electric pump ought to be able to supply the required volume of water at sustained high rpm and maintain a sensible temperature, but the power difference won't be as great as theoretically suggested.

Metric is for people who can't do fractions.

I think the gains may be more due to the removal of the mechanical fan, but I doubt a 10bhp gain can be made on a 998 from it, maybe a high revving engine

There must be an optimum flow rate that water passes through the engine for optimal cooling. As surely too fast and the water will not have time to absorb the heat from the block. Same really with the speed of water through the radiator. Does anyone know what speed this is?

Team www.sheepspeed.com Racing

The controller I bought with my pump varies the speed according to demand - engine too cold slows the pump, rising temp speeds it up - and acts as the thermostat.
Hadn't thought about it going too fast too cool efficiently.

Having said that, the boiler feed pumps at work chuck water in at 500kgs a second, and they don't seem to over cool !

Metric is for people who can't do fractions.

I'm positive this is a fallacy,

as far as I'm aware the heat transfer obeys the law,

Rate of heat transfer = Mass flow rate x specific heat coeficient (in this case water) x (fluid temp out - fluid temp in)



I suspect that this idea has come about by someone measureing the temp across a radiator, or heat source and varying the flow and observing the change in temp rather than the change in efficiency... or somthing

Joe, this is indeed the relevant equation. The delta T part is the element least intuitive to comprehend I find:

In a steady state condition with our engine producing say 125 HP we have a constant heat load. With a constant pump flow, and constant coolant heat index (say water) , the formula tells us a certain delta T (temp difference radiator inlet vs. outlet)is required. Say 10Deg Celsius.

This is where a large nice and shiny radiator comes in vs. An old and battered item full of grime and slime. To gain the 10C delta T the nice and shiny large frontal area radiator is able to keep the water temp going in much closer to ambient cooling air temperature than the old grimy one.

Or in other words the super duper radiator keeps the average water temp cooler, whilst the old grimy radiator has to rely upon higher average water temps in order for ambient temp air flow to achieve a 10C difference. This all goes well until the water starts to cook. At this point in time the water goes from liquid to gasious state drastically reducing its specific heat index by a factor of 2, leading to a negative chain reaction.

Increasing mass coolant flow 'alleviates' the previously required delta T making it possible for the old and grimy radiator to cope without inlet temps dangerously close to boiling, or the super duper rad to handle even more HP. This up to a point where the flow speed doesn't induce cavitation, making the coolant go from liquid to gasious state thereby reducing its specific heat index. Literature suggests around 7 feet per second inside radiator tubes to be ideal, 10ft+ per second to be avoided.

I can help with the heat load:

http://www.turbominis.co.uk/forums/index.php?p=vt&tid=486700

You will still need a U value - say 650 W/m^2 K ???

http://en.wikipedia.org/wiki/Heat_transfer_coefficient

EDIT Cp for water = 4.1813 J/cm^3 K

Edited by Paul S on 28th Oct, 2013.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."