Electric water pumps

[Swiss_Tony]

Has anyone had any experience of electric water pumps?

 

Thinking of reliability etc as well as how well it works.

 

TIA

 

James

Su77on Se7ens

 

[EFA]

You really are worried aren't you....

 

Fat Arn

 

Slay the K.

 

[Peter T]

James, are you still taking those pills 🤔

I think you need to consult your shrink 😬

 

Founder Member of The Twisted Prop Club.

[Swiss_Tony]

Arnie

 

 

James

Su77on Se7ens

 

[Swiss_Tony]

Pete

 

I am having a triple layer Nomex straight jacket specially made 😳

 

James

Su77on Se7ens

 

[normalbloke.29]

I have purchased one and a controller for my Caterham too.As the engine and box are out at the mo, I can't give you any feedback until spring! Apologies, but the Ferrari 550 Prodrive team used the Davise Craig kit to good effect.

[Graham Perry]

Plenty in the archive

[simonbell]

Birkin were running one in their duratec IRS car. You can set the temp you want to run at on the controller and it varies the speed of the water pump to maintain the temp. Apparently on cold start the pump is static, when the water reaches a certain temp it rotates once to move the water round a bit, waits for the next batch of water to come up to temp and pulses again. Once all water in system is up to temp it starts to circulate continuously at low speed and speeds up and slows down as necessary. I`ve driven the Birkin with it fitted and it seemed to warm up much quicker than the standard system and once up to temp the water temp was steady as a rock no matter what the driving style.

 

I`ve had people tell me they are great, others who wouldn`t touch them with a barge pole. However I`ve yet to hear a good argument against them.......maybe someone here will put forward that argument.

 

I`d be interested to hear what people think.

 

Have engine.......need car

I`ve seen the future.....and it`s powered by duratec

[Bob Corb]

Apologies for the technical answer, but here goes...

 

The main problem is that very little info is given on the EWPs flow vs pressure capability. The rated flow of 80l/min is nowhere near as high as a normal water pump. A K series for example flows about 180l/min and a I4 duratec even more. The high flow rate is needed to induce a high heat transfer coefficiecnt (HTC) out of the cylinder head into the coolant and to keep the water temp rise through the engine very low. If you drop the flow rate then the HTC drops too.

 

The heat rejection from the engine wont change, so what does?

 

Q (heat flow to the coolant) = HTC * A (area of the cooling jacket) * dT (temperature difference from metal to coolant).

 

This means that although the coolant out temp may be maintained at 90°C the metal of the cylinder head must be much hotter as the engine is rejecting a similar amount of heat from the combustion process and friction. Also the coolant entering the engine must be much colder as there is less of it to absorb the heat.

 

So, okay you save a load of power by not driving a mechanical pump (although you still drive an inefficient alternator which drives an inefficient electric motor...?) but the engine internals will reach temperatures which they are not designed to cope with. The structure will lose mechanical strength resulting in cracking and distortion. This may take a while to be noticed (hence its used on race cars which are rebuilt regularly) but on road cars the electric pump concepts which are being investigated are much higher capacity and usually run on more efficient 42volt electrical systems.

 

At the end of the day I wouldn't bother with the EWP on a road car which you want to be reliable.

 

BC

[Miraz]

I've been running an EWP for a year now - I don't run a controller, the pump runs flat out the whole time. I've moved the thermostat into the top hose of the radiator to keep the pump in the primary cooling circuit.

 

This setup appears to work without any significant problem even in very high ambient temperatures - I agree with most of what Bob says, I monitor the temp gradient across the engine to check the cooling system performance and typically I'm looking at less than 5 degrees difference - I see two major benefits from the system.

Constant load - there is a constant electrical load on the engine, one less thing to accelerate when you blip the throttle.

Cool down - more importantly I can continue to run the cooling system after the engine is shutdown

 

I would stay well away from the electronic speed controller setup as there is the potential for significant engine damage - the system was designed and tested on iron block V8's with massive thermal inertia and large internal coolant capacities - the K-series only has an internal coolant capacity of around a liter.

[Bob Corb]

I had a better think this AM and did some maths, I'd estimate that the metal in the cylinder head would increase in temp by about 15°C which may have an effect on the maximum spark advance which can be withstood before onset of detonation so you may lose some generated torque (v marginal).

 

I also checked on the power consumption of the stock Rover pump. This will be somewhere in the region of 1 to 1.5kW (thats from a former employee of the pump manufacturer). The EWP needs about 100W (manufacter quotes 7.5Amps max at 14.5V = 108W) so you can see that the maximum flow potential of the pump will be a lot less) but your crank power may increase by 1.3kW (1.75BHP) by reduced losses.

 

The idea of the controller switching the pump on and off from a cold start sounds disastrous. The engine would heat up rapidly at first and then the slug of cold water would chill it down again. There is a durability test which is designed to fail the head gasket and that is exactly how it works. Not good. Also your heater wont work very well either! Miraz scheme of a top hose stat is much better than the compromised rover bottom hose version.

 

The big plus is the run on capability, but you can add a small run on pump yourself (many cars have them, K series in Exige with A/C for example) this will just keep water running through the engine to stabilise temperatures.

 

A more simple solution to the power increase by reduced pump loss issue is to fit a smaller more efficient impeller. I reckon a batch of 25 could be made for ~ £100 each which would give about 1BHP increase?

[Graham Perry]

Bob, Interested to know your thoughts on Cavitation with a stock pump, as I understand that this can be an issue. My steel crossflow revs to 9000 plus and during a hard session, the water temp would go through the roof with the stock mechanical pump. Now that I have a electric water pump it never gets anywhere near the old temperatures when driven very hard. I do use it on perminant pumping when giving it some though. Presumably I could put some of this down to cavitation with the mechanical pump ?

[Bangers quotNquot Mas1697456953]

I was waiting for Bob to appear!!!

 

Happy new year!!

[Matt Seabrook]

Don’t want to start a fight but something does not seem quite right to me. You are saying Bob that it is partly down to the flow of the EWP that is causing the problem. Well if you had a car towing a heavy trailer doing a steady 40-50ish mph at aprox 2500rpm would you get the 180lm that you say you are going to need? If you don’t get high flow rate then what you are saying in a short space of time you are going to have a buggered engine due to poor heat transfer.

 

On the other end of the scale if you have a high revving K how do you slow down the pump speed as you are just going to turn the coolant into lots of tiny bubbles? Like I said I am not looking for an argument just trying to understand both sides of the issue.

[Bob Corb]

Graham

 

I've not encountered a pump with that degree of cavitation but I wouldn't be surprised. I guess the drive ratio and water pump you were using were standard crossflow? If you'd slowed the pump down I reckon the flow would increase and temps reduce. If it were available a smaller diameter rotor would help (reduced tip speed).

 

Matt

 

At low speed the EWP is fine as you can have full capacity (probably ~50l/min) which is just the same as the std K mech pump would give (@ 2500rpm). I admit now I got my figures wrong. The max K flow is about 130l/min I just checked (@7000rpm I dont know what it does at higher speeds). The problem is at max power when the engine kicks out ~35% of crank power as heat. If you still have just 50l/min (EWP) then the metal will have to get hotter to transfer the heat. This will have an effect on durability. I'm probably just being picky as it wont destroy its self overnight, just not as strong in the long term. But you will only recoupe a KW or so and you may lose that from having to back off the spark advance...

 

It doesn't appear the the K pump is as susectable to cavitation as the R500 appears to cool on track....

 

BC

 

 

[Matt Seabrook]

Thanks Bob still having trouble getting my head round this one but thanks all the same.

[Graham Perry]

Thanks Bob, I forgot to add I was using a larger mechanical pump pulley to compensate for the revs. It is an interesting subject, and I suspect the debate will continue (I remember Peter C saying that he was doubtful about their effectiveness). I can only speak as I find, and I think my car runs much cooler with it when blatting hard or competing. I do agree though that the Controller needs refining as it can provide a 'shock' to the engine when it engages, moving hot water out of the block and letting cold in. On mine I think the difference is about 15 degrees measured at the thermostat in the head. I have not so far had any head gasket troubles though, although it is an all iron engine.

 

Edited by - Graham Perry on 6 Jan 2004 18:42:30

[simonbell]

The EWP by Davies Craig has a variable flow of between 20 and 92 l/min

 

I thought the whole point of them being variable was to overcome the cavitation at high rpm`s which se7ens spend most of their time doing.

 

I`m interested in all this as I`m currently considering using one with a duratec. If they are generally seen as a bad idea I will have to rethink.

 

Have engine.......need car

I`ve seen the future.....and it`s powered by duratec

[Chelspeed]

> I thought the whole point of them being variable was to overcome the cavitation at high > rpm`s which se7ens spend most of their time doing.

 

Nope, the opposite. The mechanical pump on the engine is variable speed, it changes speed as the engine revs change. It has to be designed to pump enough water at tick over to keep the car cool in mid summer in spain. This means that at full design revs, or more if you've tuned the engine, there is a serious risk of the pump cavitating, just when the engines working hardest.

 

Whereas with the electrical pump you can run it at constant speed if you want, as Miraz does. Or if you want to get more sophisticated you can vary the speed but not vary it as much as the standard pump, ie you would run it a bit faster than the mechanical pump at tickover but slower at full revs so no risk of cavitation.

 

I've had an electric pump all season and it's been fine. Well pleased. Just trying to find a shorter cam belt so that I can leave the normal pump off. At present I still have the pump fitted with the impellor ground off.

[Bob Corb]

Simon

 

When we looked at the EWP we weren't convinced that it would ever deliver 90l/min at full chat and gave it a wide berth. Compared to the K's 130ish l/min it isn't much of a drop and people here find that it works fine(probably because if carries on pumping at full chat when you've backed off). The Duratec has a much bigger pump and has fantastic cooling capability which gives the engine good combustion characteristics. I know the cooling of the Duratec is well up to the demands of high power outputs although I cant say about the high speed cavitation issue. Worst case again you could slow the pump down a bit.

 

On the K series the EWP has some benefits post shutdown, and during normal running works acceptably. On the Duratec again post shutdown the same benefits exist but the EWP does not have the capacity to provide the same cooling capability. I'll leave the stock pump in place, Ford (Mazda) know what they're doing. Just get your ECU to run the variable thermostat if poss and you will be able to optimise temp for performance or fuel economy.

 

BC

[normalbloke.29]

Bob, what does the standard YB pump flow, out of interest?

Thanks for your input.

[Bob Corb]

I'll get back to you on the YB...

 

The duratec is over 200l/min, phew.

[simonbell]

Bob

 

Thanks for that!

 

Have engine.......need car

I`ve seen the future.....and it`s powered by duratec

[Bill Shurvinton]

Err Bob,

 

its Qdot, not Q.

 

Now looking at it from empirical evidence. If Geoff is getting around 5 degrees deltaT, then this sounds as if the flow is almost perfectly matched to the heat output of the engine. The clever people who first worked this out give 10 degrees as a rule of thumb for max temperature rise as the coolant passes through the engine.

 

Also it should be noted that additional spark advance IMPROVES heat transfer as the flame is exposed to the head for longer. So retarding the ignition is the wrong thing to do unless you are already on the point of detonation.

 

The assumption that is troubling me though is that the pump was sized based on engine coolant flow requirements at 7000RPM. This both sounds and feels very very wrong. Lets face it cars rarely overheat at high speed/load, but pile into the alps out of spain at full welly and then stop suddenly for the border crossing and see what happens. I would therefore contend that the pump is sized to not overheat when idling in traffic in nevada summers and to handle towing a caravan. Provided it doesn't cavitate at very high RPM then its good to go. Might be wrong, but that sounds eminently sensible to me.

 

 

[Bob Corb]

Bill

 

I agree entirely, trailor tow incline is the critical test. The EWP will provide enough flow for a modest engine to pass this test. But, as tests are not run at high engine speed, the engine doesn't produce massive amounts of heat. Take a 170BHP I4 engine, at 3000rpm WOT will produce about 17kW heat which with a coolant flow of about 80l/min will give a sensible coolant dT of about 3.5°C. At 7000rpm it will produce 45kW which still with 80l/min will give about 9°C dT which is getting a bit toasty when you consider that the same engine on its original pump had a dT ~3°C. Okay, so it doesn't boil over then that must be fine...not strictly true.

 

When an engine is developed a team of engineers developed the water cooling jacket to give high heat transfer coefficients to keep the metal temp as close to coolant temp as possible to prevent cracking and distortion. This requires a high velocites of cooling water are applied to the critical areas of the head structure. Inevitably this requires a fairly beefy water pump in a lot of cases. If this high velocity flow is compromised then the heat transfer efficiency drops and a higher metal temp is required to transfer the heat therefore all the metal in the head will be hotter. Also the pressure in the head will be much less as the flow used to generate a high pressure drop through the system therefore the localised boiling point will be lowered leading to pockets of boiling, which then results in cavitation (steam bubbles collapsing causes cavitation too) again causing structural damage.

 

I guess Grahams crossflow lives as the 2 valve head has more cooling area and quite thick metal sections (less valves etc). A good example of this cooling issue is the ford focus ST and RS. On the ST the engine has been made durable on the standard pump giving about 130l/min max. But the RS with just 30BHP more suffered badly during development with head cracking, the solution (or part of it) was to install a much larger water pump (rumoured to give about 250l/min). I'm sure if the RS was released into production on the stock pump then very few people would notice until several years down the line but modern quality standards mean that engines must be tested to much tighter standards and cannot be released in to production until all tests are passed.

 

If Grahams Xflow produces just 5°C dT at a steady Vmax condition then I'll eat my hat. 5°C on track where it runs at variable speed at not continual WOT maybe. 5°C would equate to an average heat to coolant of 26kW which roughly equates to 100BHP at the crank. Back straight at the ring...?