Electric Water Pumps (EWP)

[madjak]

I figured I'd start a thread on EWP's so that everyone can contribute to it. There are quite a few threads out there touching on them but I'll try and summarize coolant theory, pros / cons of EWP and installation.

The idea behind running an EWP is mostly two fold. It frees up horse power and it can give more effective cooling at idle or when the engine is shut off. For these reasons, I think you'd only look to install one if you're competing in motorsport. So the general gist is to remove the insides or blank off the water pump and then mount a EWP somewhere in the coolant path to circulate the coolant. They are either controlled via an external controller or via an ECU that uses a coolant sensor to control the speed of the water pump.

There are a couple of key benefits

1. More consistent coolant flow through the head

So the idea here is that you can keep a fast flow of coolant regardless of the RPMs of the engine. This keeps the head temp even across it's length. A mechanical pump will push more coolant through the engine at high revs, however there can be so much flow that the coolant at the pump can cavitate and cause air bubbles. At idle the stock water pump slows down a lot and is circulating far slower than an EWP can. With an EWP, if the coolant temp is high, the water pump will cycle the coolant at it's fastest rate until that coolant temp drops below the required threshold regardless of if the engine is being dríven or sitting in the pits idling. With an EWP coolant speed is only linked to coolant temp.

There is also the option with EWP to reverse the flow of coolant. So instead of entering the block at the mechanical water pump, flowing through the block, up into the head and then out the front or back (dependent on Coolant re-route) to the radiator, the coolant runs through the head first, potentially entering at both the front and the back of the head. If you reverse the flow of coolant through the engine, it would be advisable to do this so that any air trapped up in the head can be extracted through and out the block. The benefit here is that the head is getting the cool water directly from the radiator before it's heated by the block reducing reducing temperatures significantly across the whole head.

2. EWP frees up HP

This is really only true at high revs. The mechanical pump saps a lot of power at 8000 rpm as it's being dríven at a faster rate than it was designed. The power it uses goes up exponentially with revs until the coolant starts to cavitate. The amount of power the water pump uses is debatable, but I think it's probably in the range of 2-10HP. The EWP water pump can use up to 8 amps of power to drive, which is basically taken from the engine via the alternator or battery if the engine isn't running. That means approx 0.3HP of power taken from the engine via the alternator (assuming 50% efficency). So all up there is some gain to not running an mechanical pump... especially if you are reving the engine a lot.

3. No thermostat

In order to run a EWP you need to remove the thermostat. Instead of the thermostat here are a few different methods to get heat into the engine quicker. One of those is using a solid state relay on a PWM to rapidly pulse the EWP making it turn slowly. Alternatively you can pulse the EWP in 5 second brackets until the coolant reaches a threshold and the EWP turns on full time. This is basically what the Davies Craig controller does. Electronic set up and control will depend a lot on what ECU you are running.

There is claimed benefit of the thermostat in that it creates a restriction that causes a build up of pressure in the head. The theory is that this leads to less localized boiling or hot spotting. Personally I'm not sure on this claim given that when the thermostat is present the coolant is flowing slower in the first place. I think the speed of coolant flow through the head helps drive down the temp delta across the head which would give a greater benefit than the 2-3 PSI of increased pressure due to the restriction caused by the thermostat. In my track car I don't run a thermostat for this reason.

4. Failure rate

Switching from a mechanical belt dríven water pump + thermostat to an electrical water pump as pros and cons. I'm not sure which is more risky but given that stock water pumps don't fail that often I think the EWP probably has a higher likelyhood of failure. The EWP itself has apparently a very low failure rate, so I see more risk in the electrics failing. The good thing about electrics is that they can be set up to display a warning light if failure occurs, subject to the correct setup, whilst a mechanical water pump gives no warning.


Overall, we don't see many installations of EWPs on MX5s. As a result we don't have a lot of information on how effective running the coolant in reverse actually is and if there are any issues like trapped air bubbles etc. EWPs are used commonly on racecars as the benefits of the extra HP and overall better cooling control, especially in the pits is seen as worthwhile. Over the next few months as I install mine, I'll document what I find.

[madjak]

The starting point for me is to replace the stock water pump with a new billet blanking plate.

I spent a few nights on the milling machine making my first prototype. This one is a bit different to what you'd run on a standard engine in that I've moved the alternator to mount on the hot side of the car, and I've also moved the passenger side timing belt inwards 2mm to gain a bit more belt tension lose due to my very decked head. the plate came up reasonably well, however it's heavier than I like.





I'm planning on running silicon hoses on the output to and from the radiator to the EWP and 2 x 12AN (or maybe 16AN) lines to the front and back of the head. Davies Craig recommends 16AN sized fittings or 2 x 12AN. Looking at the ID it's pretty small so I'm not 100% sure on this sizing.

[Magpie]

Appreciate the post. This was on my build plan, however never made it to the final build. The main reason was that the failure of the mechanical water pump was seen as unlikely.

To assist with wiring in a Solid State Relay this link will help http://www.haltech.com/wp-content/uploads/2011/01/Solid-State-Relay-HT030202.pdf. As indicated by Madjak the SSR has the advantage over a normal mechanical relay in that it can be switched on/off much faster and as it has no moving parts it is not prone to wearing out.

[NitroDann]

Ive done some stuff that even haltech said "no idea if that can be done" with ewp.

Ill get a post up tonight.

Dann

[madjak]

Yeah the PWM control is the one aspect I haven't really sorted yet and would really appreciate some more information on. I think with a Haltech you can easily run a solid state relay and pulse it to vary the rotation speed. Set up a 2D table for PWM % vs coolant temp.

For me it's not actually a big concern as I'm happy to run the EWP at a constant speed regardless of engine temp. I will have a switch in the cockpit so whilst I'm warming up the engine I could always mechanically pulse the EWP but given I'm not moving the car whilst it's warming up, the radiator does very little cooling anyway. The engine won't idle by itself whilst cold so I need to sit in it anyway to get it up to temp.

[madjak]

I picked up a CNC machine to have a play with milling (it's a loaner). This is a billet water pump replacement for a stock engine (idler and alternator stay in the OEM location). Magpie, I'll make you a nice shiny metal one if you want.

[NitroDann]

Yeah the PWM control is the one aspect I haven't really sorted yet and would really appreciate some more information on. I think with a Haltech you can easily run a solid state relay and pulse it to vary the rotation speed. Set up a 2D table for PWM % vs coolant temp.

Its not, ill explain soon.

Also the Craig Davies controller isnt just pwm, its clever-er than that because pwm control is awkward and difficult.

Dann

[beavis]

Very interested. I'm just gonna hang back and watch where this goes.

[Magpie]

Madjak yes please

Also watching, especially for NitroDann's post as I was going to use the Haltech PS1000 to control the SSR.

[madjak]

Dann, this is the thread I found:http://forums.haltech.com/viewtopic.php?f=10&t=6357&start=30

Also it's a 3D map, so Manifold pressure vs CTS vs PWM %

information of interest:

http://www.haltech.com/wp-content/uploa ... 030202.pdf
Wiring Diagram

http://www.ebay.ca/itm/Haltech-Solid-St ... 27c90059f6
Relay Part

http://www.ebay.ca/itm/Pull-Up-Resistor ... 35b4f43622
Pull Up Resistor Part


Here is a screen capture of the map

There is a ZERO Spot at 0 Vac when cold so that the pump does not turn when the car is cold but the key is on.
This lets you upload and edit a map without killing the main battery.

I found that I needed to have an aggressive slope for when on Boost to offset the heat that is generated.

Cruising requires very little cooling.

You Can Not use this in conjunction with a mechanical thermostat. Please don't be insulted but I need to stress this to be 100% sure.
The Variable pump IS the Heat Control

Note: The Power that feeds the pump and fans is on a third relay that is activated by the ignition "RUN and START" positions. If you do not do this the pump and fan will go to 100% with the key out.

(Standard BOSH) when there is no power, a Solid state relay will go 100% On.


The Pump Relay is triggered off of DP#18 50Hz Never go below 2% Duty Cycle because you do not want the pump to stop turning, It will make hot spots and kill the motor FAST
I have a second relay controlling the fans. DP#17 Set to come on at 90C 0% and and 105C 100% Setting at 40Hz

[speed]

I like where this is going

[zossy1]

You don't need to remove the thermostat to run an EWP. Ask me how I know.

BTW, chalk me up for a water pump blank off plate please! Happy to pay fair cost


Sent from my iPhone using Tapatalk

[NitroDann]

So the haltech will turn a SSR on when the engine is turned off. So when you walk away from your car it will be on full blast.

PWM control of this stuff is not close to linear whatsoever. On one setup it was stopped at 92% DC and full ball at 95% DC, not linearly between those 2 points either.

Davies craig uses voltage switching, utilising a 6v circuit for low speed.

Dann

[3gress]

not as pretty as yours madjak though still effective with the ewp80 and controller setup.
Never more than a 5*c variance in temps on logs.

[madjak]

Looks fine... I'm just using this as an excuse to make billet bits! Plus I can move the alternator and idler in at the same time.

Is that a 16AN or a 20AN output?

This is the standard billet plate I'm making as a test. It all bolts up fine and looks good. I've added an extra support rib across to the idler pulley just because the tab it mounts too is cantilevered a bit. I also didn't bother drilling all the spring and water pump connection flange holes in this model.


You can see the size difference between the stock MX5 alternator and a Suzuki one. My 2yr old is helping 'fix' my motor. If every I use this block I'll have to carefully check all holes for any bolts / nuts / screws that might have found their way in there!


I'm cutting the Aluminum version now. When I get home one side should be finished. Hopefully the bit doesn't chatter out or something.