Electric Water Pump How to.

Howard Jones

Randy, I live 50 miles from ya so
If ya want you can come and get it. Bring a 6 pack of your favorite beer. Otherwise PM me your address and I'll check with UPS and let you know.

I have the backing plate also. You will need a couple of gaskets and some hardware but otherwise, I think it is good to go.

Free homemade fraction of zero elegance!!!!!


  • HPIM2614.JPG
    897 KB · Views: 158
  • HPIM2615.JPG
    849.6 KB · Views: 156

Randy Folsom

Randy, I live 50 miles from ya so
If ya want you can come and get it. Bring a 6 pack of your favorite beer. Otherwise PM me your address and I'll check with UPS and let you know.

I have the backing plate also. You will need a couple of gaskets and some hardware but otherwise, I think it is good to go.

Free homemade fraction of zero elegance!!!!!
Joel. I have spent a little time thinking about your question. Here my thoughts. The "head" issue is so fundamental that I think you would be best served to mount the pump as low in the chassis below the coolant volume as you practically can, Then as close as possible to the engine cool side inlet as possible. If this is a location at the rear of the front radiator mount/enclosure rather than directly into the radiator then closer is better from a design standpoint. A couple of feet would be worth doing IMHO. Inches? Not so much. Might you have a location back in the engine room but not right at the engine inlet? The effective full length of the sidepod would be worth it I think. However, don't add a lot of turns to the system or raise the pump to save a few inches in length. Straight and smooth is best but LOW in the main thing here. Overall system "head" is also improved by raising the coolant volume in relation to the pump. How? Place the header tank as high as possible. A bigger header tank with more volume is also useful.

Looking at your pictures again. If it was me I wouldn't change that very nice installation unless I found I had an issue. Really nice work. Maybe my comments would be better considered in the planning phase of a build for the next guy. But it's always a but isn't it? It has all got to fit in there somewhere and you just do the best you can.

Ok, based on Howard's comments, I thought it worth a try to find an alternate location for the EWP. Turns out there's a little opening just under the rack & pinion big enough for the pump with a direct connection to the pipe going back to the engine. The pipe to the left in the picture below is bringing hot coolant from the engine to the radiator.



It's a tight fit but does lower the EWP 6 inches in vertical height from the location where it was previously. With this placement, the EWP is at the absolute lowest point of the coolant system and thus it should maximize the head pressure going into the pump. I do think if the pump is located here that a skid plate for it would be wise to keep road debris from damaging the EWP and/or taking out it's electrical wiring.


The other recommendation for an expansion tank with a line to the EWP inlet can be done by welding a bung on the aluminum tube going to the pump and mounting the tank somewhere up high under the front clam shell. The harder thing is bringing a -4 line from the top of the engine to the expansion tank. As you can see, there's lots of coolant piping, AC piping and electrical cables running in the chassis bottom channel but there should still be room to stuff a -4 tube in there.

Hopefully moving the EWP to this new location will avoid a potential overheating headache down the road so it should be worth it.

Howard Jones

That, my friend, is as low as it goes. Well done! And yes skid plate is required. It also appears that is is located in a spot that it can be easily serviced should you need to do so. While you are at it put a bung and a plug in one of those low coolant pipes so you can drain the system. You can just see mine in the lower coolant tube at the bottom left.


  • HPIM2392.JPG
    85.9 KB · Views: 139
So now to the next dilemma for my EWP coolant system, where to locate the expansion tank? For a car with the EWP located up front, if the expansion tank is also located up front, then there's fewer coolant lines that have to run the length of the car. Given that the highest point in the front clip is lower than the highest point in the rear clip, the challenge is that the expansion tank can't be mounted as high in the front.

I took some measurements on my chassis relative to the bottom of the car. Vertical space upwards where an expansion tank could be mounted:
  • front clip = 24 inches
  • rear clip = 34 inches
  • vertical distance from bottom of chassis to top of swirl tank fill cap on engine = 26 inches
So if the expansion tank is mounted up front, the coolant level in it will be 2 inches lower than the engine coolant upper level. I think this would be common scenario for a car with front mounted EWP unless the engine is dry sumped and mounted very low.

I think it will still be better to mount the expansion tank up front than in the rear. This would require only a single, small line to be run from swirl tank on the engine top to expansion tank versus two lines (radiator top to expansion tank, and coolant feed from expansion tank to EWP intake pipe) if expansion tank is in rear. In both these scenarios, the lines must go down before they go back up so any air traversing them will need to be forced through. So I'm thinking the 2 inches difference in vertical between swirl tank top and expansion tank top is likely not significant.

Does the logic make sense here? Am I misunderstanding something here?

Howard Jones

Both my cars have the expansion tank mounted as high up in the rear engine room as possible. There is an AN-8 hose running from the inlet to the pump (cold coolant return) to the bottom of the expansion tank. I think you should run the expansion tank hose all the way to the front of the car where the pump (low-pressure side) is located because that is the way it works best. You could hard line that pipe with 1/2 aluminum tubing. That's what I would do.

However, having said that. If you are constrained by available space then putting the expansion tank up as high as possible in the front of the car will probably work fine. It's just not theoretically optimal. But a lot of decisions about building your own car work out that way. You do have a unique design. Don't beat yourself up over this. The main thing is the pump location and you have that pretty much perfect.

I walked out to the shop and measured my SLC. From the radiator cap down to the inlet of the pump is 22 inches. If I take into account the lower coolant level in the expansion tank, it is about 18 inches. So your car at 24 inches would be quite a lot better than that. My SLC coolant system works pretty good, but this serves to illustrate the problem with very low, mid-engined prototypes very well.

The thing to keep in mind is that the most important element is an effective bleed system. All of the rest of this just makes the bleed system more effective as well as improves the entire coolant system's effectiveness.

This thread probably started out somewhat theoretical. Best-case scenarios rarely present themselves, especially when you are trying to cram all this stuff into a small, low, and mid-engined car. I think your real-world example is very useful and adds a lot to the discussion. Keep up the good work.
Here's a design criteria I have.

That first quart of antifreeze you pour in needs to have a nice direct path to the pump's input, regardless if it's the OEM mechanical pump or an electric. This is so when you're adding coolant, what you put in gets pumped where it goes, and air will be expelled. The pump can't effectively pump air, so a reservoir of fluid on the pumps intake (expansion tank) is a good idea.

With a bleed from the top of the engine to the expansion tank the pump can expel the air from the engine to the tank even with the thermostat closed. Given these criteria, I think the expansion tank in the front isn't a problem, and likely better if you have an electric pump at the front.

While we're noodling theoretical...
My thought on locating the expansion tank up front in this scenario is basically about simplicity. Making the feed from expansion tank to EWP intake a straight shot downward of about 18 inches instead of having it go horizontal for about 6 feet. I worry less about the "bleed line" feed from engine top swirl tank forward to expansion tank because it's likely the hottest, high pressure part of the coolant system will push air bubbles down a tube if anywhere. If air does get trapped in swirl tank, I can always pop off the cap and add coolant till swirl tank is full.

A related question, has anyone used one of these combo expansion tank/recovery tank designs?

Again, thinking about simplicity that maybe a combo unit will work better than trying to integrate two separate tanks.
If the water passing through the expansion tank is hot enough to boil over under pressure, it'll boil off the contents in the recovery tank not under pressure. So it's not to useful as a recovery can.

I think maybe this part is for show cars that are not driven beyond garage to trailer.

Howard Jones

The one you posted a link to is pretty small. I think the expansion tank ought to be about 2-3 quarts of capacity. You will not want to run it full so if there is going to be a few pints of air at the top then you effectively have about a 1/2 gal or a bit less of coolant in the expansion tank. That's about right.

Recovery tanks will return "overflow" back to the system when the car cools but its location isn't critical, just convenient and sealed to the system. You will need to experiment with levels in the expansion tank and the recovery tank to find a nice transfer amount between them when you run the car I think. Did the original car have one? Back in the day many old cars just vented to the road.

As far as the overflow tanks go, the other kind of additional tank, they are required on track cars to keep fluids off the track. They are not technically necessary on a street car. You could leave it off and run the radiator cap blow-off line out to the atmosphere. In any case, the catch tank is optional as far as cooling system performance is concerned.

3 quarts of fluid is 174 cubic inches. A container 3" X 7"X 9" is a bit more (189 ci) than a gallon and a half. I would use a 1/8 inch wall thickness sheet.
Last edited:
If the water passing through the expansion tank is hot enough to boil over under pressure, it'll boil off the contents in the recovery tank not under pressure. So it's not to useful as a recovery can.

I think maybe this part is for show cars that are not driven beyond garage to trailer.

I guess the basic question to answer for this type combo tank is on a mid-engine street driven car (assuming a well designed coolant system), how often does boiling temperature coolant actually make it to the expansion tank? Especially when the expansion tank is mounted in the front of the chassis. If coolant is boiling temperature when it exits the engine, it's going to have lots of opportunity to cool below boiling as it goes through an aluminum tube all the way to the front of the car. This would apply to both the main coolant return to radiator (and then the bleed line from top of radiator to expansion tank) and the bleed line going directly to expansion tank from swirl tank top on engine.

My guess is that coolant temperature in the expansion tank will be well below the coolant temperature in the main system. The volume of the coolant regularly cycling through an expansion tank shouldn't be enough to maintain a high temperature there either. If so, then coolant temperature in a siamesed recovery tank should not be above boiling, should it? I could see it being different in a track car but if a street car often has boiling temperature coolant going on, then you have issues with the coolant system design.

I'm dealing with very compact spaces and am looking for every opportunity to save space with components. That is, saving space without compromising the function.
The pressurized water in the cooling system doesn't boil until 230 + degrees. (off the top of my head and not looking a vapor pressure graph for half and half antifreeze).

220 degree exiting the engine isn't at all unheard of in traffic with A/C, and flowing water has a lot of BTU to keep that tank at full temp. If the tank is fed by hot bleed from the top of the engine, it's full max temp. You did insulate your tubes, didn't you?

Siameseing the tanks seems like a bad idea if you want to recover that coolant. If you're pressurizing to the point of venting, you're 230 degrees unless it's purely from expansion, but the air in the tank is supposed to allow room for that as air is compressible.

Ian Clark

The cooling systems of mid engine race cars and road cars for that matter, are well known for being difficult to fill, remove trapped air, overflowing header tanks and occasionally (never at a convenient time) runaway boil overs.

It seems that with a few modifications the Moroso Overflow Tank that Joel is looking at has potential. The pic attached is for informational purposes only but you get the idea...

The main header tank stands off the rad hose leaving the engine on it's way to the radiator. The standoff connection must be at the lowest point possible as in floor level before the hard piping travels through the center tunnel to the front of the car as in the SLC or GT40 application.

The coolant level in the header tank after hot/cool down cycles level out to about an inch above the bottom of the tank. The result is 3 quarts of expansion capacity, should be adequate, yes? I've had good results with -4AN bleed lines, -8AN for the standoff hose.



  • Modified Moroso Header Overflow Tank 63772 dwg.jpg
    Modified Moroso Header Overflow Tank 63772 dwg.jpg
    136.4 KB · Views: 129

Howard Jones

I think that the temperature of the coolant varies quite a bit from the top of the engine (heads) to the outlet of the radiator (coolest) so I'm sure that the expansion tank coolant temp is less than the max temp but hotter than the radiator outlet. Somewhere in between. This somewhat illustrates why running very high design coolant temps like in most modern cars is a bad idea in a 500HP+ high-performance car. If you can get the design temp (what the gauge says when the system is stabilized) down under 200 F then that gotta be a good thing. That should be the maximum coolant temp in the heads. My SLC runs at about 190-195 on track on very hot days.

However, given your need to prioritize the components' location and limited space you might be better off placing the expansion portion of the tank at its most efficient location and then placing the separate recovery part wherever you can remotely.

Cardboard and a hot glue gun are your friends. This is a very cheap way to mock up things like tanks.
So I'll dabble back in the theoretical with this question (that is unless someone here has actual field experience with this scenario). The scenario is a mid-engine car with an "adequate" coolant system and then it's changed to go from an "inline" physical thermostat to a sensor/electronic controller coolant temperature approach. I think this scenario almost by definition would also include changing the coolant pump from a belt driven variety to a electronic variety unless there's some variable speed belt drive out there (which I've never seen used on traditional water pumps).

So the question is, can the stabilized coolant system temperature be maintained at a lower level with the sensor controlled approach versus the physical thermostat approach? My gut tells me yes, because there should be less coolant flow restriction and thus a larger volume of coolant going through the radiator. While a physical thermostat orifice opens with heat, the opening will never be as great as having no physical thermostat present at all. I guess the answer to this is also dependent on the radiator/heat exchangers ability to remove more heat as the volume goes up from a baseline amount. In my cars, I use two pass radiators (going from single to two pass eliminated an overheating challenge in my Cobra) which do seem to have ability to remove more heat.

I'm raising this question because it was sort of inferred in answers to my earlier expansion tank question that coolant system temperatures would/could be near or above the boiling point of the coolant. I guess I'm questioning that assumption in the "normal" case for a coolant system that uses a EWP and a programmable controller to maintain a stabilized system temperature. Yes, there will likely be "exception" cases where something is off but the normal case should apply the vast majority of the time. It seems to me that with the proper settings in the controller (for both EWP and cooling fan) and an adequately sized and outfitted heat exchanger (i.e. proper shrouds, air flow, etc.) that the system should be able to maintain coolant temperatures below boiling in the "normal" case.

Isn't that why we're swapping out traditional water pumps for EWP anyway? I'm mostly asking this question to make sure assumptions are aligned with the proper scenario/use case and to make sure I understand the dynamics for what's really going on.

Ian Clark

Wow, you've really opened a thermodynamic can of worms... luckily there's a lot of very smart people on this forum to chime in:)
I do have some empirical data I can share.

The one place I had on my manifold for a temp sender, I used for the Megaquirt EFI. I drilled and tapped the boss at the BACK of the manifold for the gauge in the car. Using the computer to watch the real temp and seeing what the car's gauge read was enlightening. The engine quickly warms from cold to 185 and stays there unless it's real hot outside and/or I'm on it hard, then it'll climb to 190 or so. I have the EFI set up to turn on the fans at 195, and it cycles at that temp, and only when I'm sitting still. So I'm confident the engine doesn't get more than a few over 195 ever.

But the gauge in the car is a different animal. It's showing me the temp of the water that has passed through the block, (427 Windsor) but just entered the head where most of the heat is. When it's cold outside and I'm highway speed at low RPM I've seen it as cool as 120 degrees at that point, but when it's hot outside, idling it's showing 190-195.

So I would estimate when 50 degrees outside, at highway speeds, water coming out of the radiator is around 100 degrees, or a 90 degree drop.
When it's closer to 100 degrees outside, I'm only seeing a 15 to 20 degree drop. There's 2 reasons compounding to cause this huge disparity.

When It's cold, the radiator has a 140 degree temp differential to move the heat, so it's far more effective. Also, this makes the thermostat feather the flow way back through the radiator, so it has more time to lose heat. When the engine's inlet temp is 100 degrees, it can't take anywere near the flow rate to maintain 190 degrees out.

This is speculation, but I think a OEM pump at 6000 RPM will move more water that the electric can. I seem to remember it goes up to 5400 RPM, and has a smaller impeller. But The electric can run 5600 RPM when the engine is running 2500 RPM. Unless your at LeMans, I don't see anyone sustaining over 6000 RPM for any long duration.

Honestly, I don't think I would eliminate the T stat all together, but maybe put a 165 degree in with a 1/8 bleed hole to get to temp quicker and let the pump modulate from 185-190 degrees, then from 190-200 modulate the fan. Then you maye need some sort of PID control system with all of these interdependent loops and latencies to keep it from oscillating wildly. That's part of the reason I have the fan bang bang, on or off with a 5 degree hysteresis.

My 220 degree case was OEM mech pump, stock 195 degree Tstat, you're kicking it hard, then pull up to a light and idle for 2 minutes. At that point the metal in the head and block is 250 or so degrees, and the water comes almost to a stop with 800 RPM water pump impeller, and the water coming from the radiator is 175 degrees. For a minute or 2 the heat from the head and block will soak into the water quicker that it's getting changed. Let it idle another 2 minute, the fan will have the slow water in the radiator returning closer to 150 degrees, the engine is no longer producing much heat, and the heads and block cool down 10-20 degrees and all will settle back to a happy equilibrium.