EWP (electric water pump)
 

I'm setting up my engine to use an EWP instead of the mechanical pump. I'm doing this for a few reasons. More HP at high revs, better cooling in pits / idle, just cause I like making stuff and experimenting.

First task is to either blank off or replace the stock water pump. Since I like to over engineer stuff I figured I'd make a billet blank off plate. Not only does it look better, I can design it to move my alternator to the hot side via some custom mounts.

I started with a waterjet cut 22mm plate.
https://www.miataturbo.net/attachmen...ine=1462852100

Then machined down the idler pulley and tensioner to 20mm. (I've moved the idler in 2mm so I can get some more belt tension due to my decked head). I also machined down the center to 12mm to take out some weight and for the mounting bolts to clear the belt / pulleys. This was done on my milling machine.

https://www.miataturbo.net/attachmen...ine=1462852100

With a bit of filing and sanding most of the machining marks came out. The center mount is for the Davies Craig fittings. I need to figure out if I'm going to use silicon hose or go all out with 16AN fittings. The back of the output hole has a routed 10mm radius to help the waterflow.

https://www.miataturbo.net/attachmen...ine=1462852100

All up it came out pretty good. You can see my small Suzuki alternator which I picked up from a wrecker. It even had the right 4rib pulley.

https://www.miataturbo.net/attachmen...ine=1462852100

Next task is to make up a couple of plates to feed water into the front and back of the head.

are you going to do a "reverse" flow? And why are you putting the alternator on the hotside?

Yeah I'll run reverse flow. I'm thinking of running a 16AN line to the back of the head (1" ID) and a 12AN to the front (3/4" ID). I think you need to run intakes on both ends to reduce the chance of air bubbles being trapped in the head.

The alternator is on the hot side so that I have better access to the oil lines / filter and also to move some weight to the passenger side (I'm RHD). As a bonus, I figure the existing alternator spot is ideal for a low mount Rotex.

I am very interested in watching this! I always thought electric water pumps were a great idea.

I'm not planning to re-dyno the car so I won't be able to confirm any power gains. Supposedly at 8000+ RPM the stock water pumps can sap up to 10HP or the other theory is that they cavitate and don't circuilate the coolant effectively. I'm not sure on either claim but my guess is that there is 2-4 HP in it maybe.

I already don't run a thermostat so my cooling system won't really be any different except the coolant will cycle faster at idle and slower at high revs. Plus I'll be able to run the pump after the engine is shutdown and also pulse it when the coolant is cool... not that warmup is much of a problem for my car.I'll make a decision soon on whether I run silicon or AN hose but it will be a while until I plumb it all in as I'll have to wait for shipping.

Nice! I have been running a Davies Craig EWP 115 for 5-6 years now. It is awesome. Will you control the pump with your ECU, or use a dedicated EWP controller? I resisted using the EWP controller (also Davies Craig) at first, but it is a very good solution. The old style controller I have is just OK - the new display version seems better and I hope is a bit more robust (I have been through one, replaced under warranty). The EWP controller can vary the duty cycle on the pump, and its speed (voltage). The Davies Craig system is self-contained, meaning that the pump controller has its own thermostat, and also controls the cooling fan. My ECU no longer controls the pump and fan. When I come off the track hot and shut down, the pump and fan will continue to run until the water temperature reaches its set point. The set point can be adjusted by pushing a button on the controller. I never have cooling issues, except that on very cold days, my radiator & cooling ducts and vents are too effective (so I have to block them off).

I installed the EWP for the same reasons as you - high revving motor, potential pump cavitation, better cool-down, efficiency. I took the opportunity to completely remove the filler neck, re route the coolant hoses, and relocated the top coolant port on the radiator to the driver side, so the front of my engine is completely open, unblocked. Your setup looks very clean, too!

WIth such an efficient cooling system, radiator shutters (for aerodynamic improvement) could be a legit option! :)

Small update. I used to own a small 4-axis CNC machine that I sold off years ago. I still talk to the guy who purchased it as he's involved in motorsport and he said I could borrow it for a while. The machine has a bed size of 650 x 450 and a working height of up to 150mm. Plus it has a nice little rotary axis for smaller work.

Here is my first test cut from the machine. All up around 4 hours of cutting time into chemiwood. I'll pick up some aluminium stock today and make it for real over the weekend, which should take around twice as long to machine.

https://www.miataturbo.net/attachmen...ine=1463623655

https://www.miataturbo.net/attachmen...ine=1463623655

https://www.miataturbo.net/attachmen...ine=1463623655

Very nice, keep the updates coming!

Very cool! I had welded a fitting on the pump where the pulley was. However having a couple of CNC machines now and designing my own drysump, I am most likely going to do something from scratch as well. This is the previous/current piece:

https://www.miataturbo.net/attachmen...ine=1463681071

https://www.miataturbo.net/attachmen...ine=1463681071

The billet aluminum water pump plate is now finished. All up the surface came up pretty nice and very smooth, with a few marks on the sides where I initially had the bit too far out of the collet causing chatter. The photos make it look rough but it's just the machining marks in the surface. I'll probably get this part anodized to protect it from the coolant.
It all mounts up fine. Next part to machine is a manifold that connects to the Davies Craig water pump and outputs to 2 x 12AN lines for coolant to the head.


https://www.miataturbo.net/attachmen...ine=1463972963

https://www.miataturbo.net/attachmen...ine=1463972963

https://www.miataturbo.net/attachmen...ine=1463972963

https://www.miataturbo.net/attachmen...ine=1463972963

Damn fine-looking work.

Now I want one....

soooo sexy. I now want to bug the machine shop guys downstairs to cut one for me. Would you share the files?

So how does this work? What pump do you buy? How is it mounted if it's not part of the milled peice you are showing here?

An electric pump. Probably Craig Davis, which has been mentioned elsewhere. Its electric. So mount it wherever you want.

Serious question:

The tab to which the upper alternator mount attaches seems kind of long and skinny, like it might tend to flex and eventually suffer fatigue failure. Is there a reason that part was made like that?

Good question... when I made the 3D model I figured 6mm would be thick enough. Now I have the part it feels and looks too thin so I'll add a triangular brace on the bottom. Its probably strong enough but since I cut it with an end mill there is a potential stress point where the tab joins the bulk of the part.

I find its hard to judge size and thickness when modelling in 3D. I really need to hold the part in my hand, tap it and try and bend it before I get a good sense of where it's overdesigned or a potential weak point. Thats why custom design is so time consuming and expensive.

I'll probably get a run of 10 of these made at the machine shop that leant me the mill. I need to mount one up and test it all first to make sure it works as it should.

Im currently sorting out the other parts I need to finish the system. I'm not sure if anyone has run the coolant in a reverse flow direction so it will need some testing to see if its effective or not.

That's kind of what I was thinking- that a couple of ribs running along the backside, along with a radiused join, would seem to be called for.

But nitpicking aside, this is phenomenal work.

Its easier when you use inch. With small stuff, 6mm seems so big, 1/4" seems like its not enough. On the other hand 20mm steel plate seems pretty small and then you remember what 7/8" steel is really like.

I don't know how you guys can even work in inches let alone design stuff! I suppose it's one of those things that you have to spend your life working in so you can memorize the sizing and learn the relationships. I find I tend to convert it back to decimal inches whenever I'm drawing things which makes it suit my more metric programmed brain.

For example 27/32" I have to spend a lot of time figuring it out if it's larger or smaller than 7/8" or 13/16".

Very nice work! Oring'ed AN fittings make life worth living.

By designing in decimal, not fractional numbers. Fractions are convenient, I guess, for certain things.

I may be interested in 1 or 2 of these if you get them all made. Well this depends on how much shipping from down under is.

https://www.miataturbo.net/attachmen...ine=1464117161

It would probably be cheaper to get it machined here than ship it lol

Shipping isn't too bad, at around $15AUD for up to 2KG or $25AUD for 10KG. It's actually far cheaper than getting stuff sent from USA to Aus! There seems to be a bit of interest both locally and overseas so I'll get some quotes from the machine shop to get 20 made. With the low Aussie dollar it might be cheaper than making them in the states anyway.

Won't the water flow in the reverse direction though? Or is that just the toilets?:rofl:

The sad thing is that our modern toilets are so water friendly now that you can't tell if the myth is correct or not. There is no swirl!

How many of you, if you were looking to run an EWP would want to retain the heater? I think it could be accommodated by a small offshoot loop or even redirecting the coolant to the back of the head through the heater core. This would have the very small benefit of cooling the head even more by radiating heat via the cockpit.

I made a cool CNC machined 'Y' manifold that directly attaches to the Davies Craig EWP. It's made so that I can weld on 12AN bungs on each face with the one facing upwards going to the front of the head and the straight output going to the rear of the head. It's designed the way it is so that it would fit into a small round block of ally I had lying around. I could do this with a 16AN tee fitting and then stepping down to 12AN outputs with a reducer but the total cost of the fittings added up a bit and I would prefer a smoother turn than the 90degree that the tee would do.

I hour on the lathe cleaning up the flange and 5 hours on the CNC mill... here is the end result:

https://www.miataturbo.net/attachmen...ine=1464422287
It will have 2 x 12AN bungs welded to the flat faces. I'll probably also drill a 1/8NPT hole in the side to mount a coolant temp sensor.
https://www.miataturbo.net/attachmen...ine=1464422287
Bolted onto the Davies Craig EWP.... looks like it was made for it!

https://www.miataturbo.net/attachmen...ine=1464422287
A bit of porting with a drill and carbide bit. I really need to get a die grinder! I even added the 'o' ring channel from the Davies Craig flanges.
https://www.miataturbo.net/attachmen...ine=1464422287



https://www.miataturbo.net/attachmen...ine=1464422287

I'll drop this manifold and also the two head flanges off to the welder to attach the 12AN bungs. I'm just waiting or the AN hoses and fittings to rock up now before I can install it on the engine and test.

With the reversed cooling, are you doing anything with head gaskets and whatnot? Obviously Mazda had a couple different gaskets to address hot spots in the engine. While you are introducing cooled water to both sides of the head, I would think you still need some type of means of controlling flow so that it actually works as intended.

I don't have any experience with reverse cooled engines other than the time I owned a Z28 that had an LT1 which Chevy retrofitted reverse cooling on. Then with a clean sheet of paper on the LS1, they went to a standard cooling arrangement. That decision made me think twice about the benefits of reverse cooling although I have no inside information about why it was done. Obviously, bean counters could have had a ton of leverage on the decision.

Tbh, I don't know if there will be any issues running reverse flow on a BP. I haven't seen anyone do it yet. The gear I'm making can by used in reverse if required but I think one of the benefits of EWP is that you can change the coolant direction. Really I doubt there is any issue with the head unless air bubbles get trapped inside. I plan on running a few tests where I run the coolant through the head in the reverse direction (whist the engine is off) and then reverse the direction of the flow and see if any bubbles come out of the head. If there are I know there is some trapped air.

In terms of head gasket, I figure if most of us run coolant reroutes on the same head gasket I doubt changing the direction will change much anyway. In terms of hot spots the coolant should be cooler as it's coming directly from the radiator so I can't see how it would ever be worse than OEM.

FWIW Evans waterless coolant supposedly has benefits for reverse flow. I suppose it's because it doesn't produce steam bubbles.

10 years later, I still haven't seen a test comparing detonation resistance...

Well i run N/A with E85. Detonation isn't a problem for me. Im still interested if it improves cooling.

(guttedmiata???)

dude with the black carby NA8 drag car has a reversed flow setup, ask him.

Posting stuff from Melbourne to Sydney is costing me $15 at AusPost.
I hate the aussie postal service.

Installed... I decided to finish a few projects and instead of porting my head so I figured I'd bolt up the EWP. I still need to wire it up and fill with coolant.

https://cimg4.ibsrv.net/gimg/www.mia...7d653a40e4.jpg

https://cimg5.ibsrv.net/gimg/www.mia...05e59824cb.jpg

https://cimg6.ibsrv.net/gimg/www.mia...22ede9649a.jpg

I have room still to mount my small alternator on the hot side in front of the headers. It clears up room under the intake so I can get better access to stuff.

This is so cool.

I've just come back from running my car at a weekend event and the EWP ran flawlessly the entire weekend (practice, sprint, track). My tuning map is set up to cycle the coolant slowly (eg 20% PWM) until the coolant temp hits 175F (80C) then rapidly ramp up to 100% by the time it hit's 195F (90C). I also monitored the head temp via a k-type thermocouple and hooked up to an analogue output amplifier to my data logger. Lastly I have it set up so that when I'm back in the pits, the ECU will keep the EWP and thermofan cycling until the temp hits 60 degrees even with the engine off.

The coolant temp rises quickly on idle, lifting up to around 150F within a couple of minutes... it's probably more rapid on my car than stock due to the small radiator and high idle speed. Once out on track the coolant rises rapidly to 175F and then plateaus off at 180 - 185F when at full load on the track. With the engine off the coolant drops back down to under 140F within a minute or so.

Engine wise, I really do think there is some extra power there from 7000 upwards but it might be my own bias effecting me. The engine seems keener to rev and I think I'm shifting faster than before. I dropped over a second off my previous PB at the track but I've made other changes such as adding a rear wing and changing the alignment slightly. Still I've very happy with the end outcome.

The thermocouple head temp sensor should be a default install on any race car with aftermarket ECU. It's cheap, works, and will still measure temp if the pump stops or there is coolant failure. BTW, the head temp trails the coolant temp by around 20F in both the up and down direction.

This is great! Where do you have the thermocouple mounted and are you just using a generic K-Type?

This is great, thank you for the effort and the time spent to come back with results.

Yeah its a standard k-type bolt on sensor. $10 on ebay and they sent me two! This is bolted to the back of the head next to the coolant neck in one of the spare bolt holes.

https://cimg8.ibsrv.net/gimg/www.mia...bc1aebf297.jpg

The sensor then runs to an analog adafruit k-type thermocouple amplifier and from there to my ecu. I think the amp was only $15 but you can also get 4 channel ones if you want ro run EGT sensors or other temp sensors. All up its a very cheap and easy setup.

Accuracy wise I'm get a stable reading when the engine isn't running. The ecu is needing to measure tiny changes in voltage as the temp range of the k-type is -270 - 1250 C (over 8 volts) so the temp range we are measuring for coolant is within a few tenths of a volt. When my engine is running I'm seeing fluctuations around 2-3 degrees but then all my sensors do the same so maybe I'm getting noise on my sensor earth or something.

Still even with the noise the logs the temp measurements are very useful.

That sounds good. I already got an 8-channel EGT which only 4 are being used. So I will use the 5th for this perhaps.Thanks!

Links to the items I mentioned above:

K-type Thermocouple: Thermocouple Sensors K type with Washer eBay
Adafruit Amp: https://www.adafruit.com/product/1778
Quad Channel Amp (Unproven) : Quad Channel K Type Thermocouple Amplifier Conditioner Analog Output AD8495 eBay

That would be me. My setup has a few flaws. so I will be building a new system from scratch. This thread has given me a few pointers on how to improve it. I will be using some of the parts I already have, but will definitely be using a different pump. I used a CSR pump which is way to heavy and bulky.

Hi Madjak
Do you have any of the billet water pump adaptor plate left to sell?
Cheers
Ado