Electric Water Pump How to.

Howard Jones

Supporter
I have started this thread in an effort to make available all I have learned about this subject in an easily searchable place. Of course, I don't propose that I know it all or that everything to be known is in this original post but I think that a central location for this subject with a simple search title will be useful to many. So here it goes.

I have a Davies Craig EWP 150 with a DC controller installed in my SLC. I have run this pump since the first time I started up the car apon completing the initial build. The engine in the car is a SBC making 430ish HP at the flywheel. At first, I used the standard RCR radiator and the as-delivered fans that RCR included in the kit. More on this later.

I learned how to efficiently bleed a V8 in a mid-engined car when I built and developed my GTD40 so on this car I installed bleed ports at the rear of both heads. "T'ed" them together and connected the single line (all AN-4) to the expansion tank. I also ran a separate line from the front of the car back to the same expansion tank. At the radiator, I used the two ports at the top of the radiator on each side to again "T" them together and connect them to the single line going back to the rear of the car. So the bleed system ties all the highest points in the car together at the expansion tank.

The expansion tank MUST be located at the highest point available in the car. This is a very important point. The expansion tank connection to the cooling system (expansion line) MUST be connected to the lowest PRESSURE point in the system. This will always be at the pump inlet return from the radiator right before and as close to the low-pressure side of the impeller as possible. This will also be the lowest-temperature coolant in the system. This is also a very important point.

Pump location: the pump should be placed so that the highest pressure available is directly applied to the hottest heat-generating point in the system, This will always be the ports that the previously removed mechanical pump used to deliver cool water to the engine. Engine design is different from engine to engine but in any case as close to the engine cool inlet as possible. The pump inlet position should be as low in the car as possible to place as much of the total water volume as possible ABOVE the pump. This is known as "head". Think of a damn and its electrical generation turbines. The turbine inlets are always BELOW the dam for just this reason.

Head is the biggest problem to overcome in a mid-engined car. Especially a very low car with the water volume also physically very low. And this is the first real reason to use an electric pump. The pump in comparison to a traditionally located mechanical pump can be placed much lower thus improving head. In both a SBC and SBF with a mechanical pump, MOST of the water volume is below the pump inlet. The radiator volume is lower, as well as the transfer pipes and nearly all the water in the engine with the exception of the water in the heads. This is exactly wrong and why an electric pump must be located as low as possible.

The second reason is that a mechanical pump is driven as some derivative of engine speed. Its impeller is always either running too fast or too slow in a road race application. Here the rev range of the engine works against the efficiency curve of the impeller. The engine rev range is far wider than the effective impeller rev range. Fluid pump impellers typically don't like to be driven faster than about 3000-3500 rpm and need about 1000-2000 rpms to enter into the linear and most efficient portion of their performance curve. This can be addressed in a street car with pulley ratios where engine speed is usually just off idle and nearly always below 3000-3500 rpms. Just select a pulley ratio that keeps the impeller speed at 1500-3000 rpms and all is well. But if you spin it too fast, you get cavitation. Too slow and it's not really efficient. However, in a street engine power levels and thus heat generation are usually pretty moderate and you can get away with it. After all our V8s are making far less than 50hp at idle and usually less than 100hp at cruise speed. This is a lot less heat load.

Run the engine on track at 60-100% of its maximum power and use a rev range of 2800-6600 rpms for a 1/2 hour continually and it just will not work. Especially no matter what pulley ratio you select the impeller is faced with a 4000 rpm use band and it spends too much time out of its best efficiency range. An electric pump solves ALL this by running a set speed that is calculated to turn the impeller at exactly its most efficient rpm. On/off duty cycle is used to accommodate changes in heat load. This is exactly what the control box is doing on a DC system.

Note: I tried to locate my pump up at the front of the car for packaging reasons. It did work there but it DID work better in the rear of the car attached directly to the engine. I suspect pressure loss over the pipe distances was the reason. Remember a mid-engined car has an effective coolant system length that is many times longer than a front-engined car. A point to remember is every 90-degree turn in fluid flow is an effective 3-foot increase in system total length for a 1/12-inch diameter pipe. Our mid-engined cars have an effective coolant path length of at least 30 feet. In reality, it is quite a bit more than that.


So an electric pump has huge advantages. But are they reliable? I say yes. The market has driven the development of EWP to the point that there isn't much concern with reliability with modern designs. Many OEM's are using them in production cars with no real ill effect on warranty costs.

So do I use one with a mechanical pump? You can....... but I would recommend that you bin the mechanical pump and design the system correctly on the first go.

Do I keep the thermostat? Not on a DV system, The controller is the thermostat. Other pumps recommend you use the thermostat to control system temp because their pump is only a pump. That's not bad it's just the way their system is designed to work. Install the pump using the manufactures instructions and you will be fine.

Who makes good pumps? Davies Craig, Stewart, and Meziere are all good. There are also OEM pumps that would work as well I suppose. I haven't looked into them as thoroughly.

OH....Drag race setup with the EWP simply replacing the Mech pump. In a word, No. It negates all of the placement advantages discussed above. Because of the length of the system AND the Head issue don't use one of those in a mid-engined car.

Pitfalls: Place the sensor for the controller at the top of the intake manifold where the coolant is hottest. This is also very important

Very high HP: When I was talking to an engineer at Stewart Racing pumps, he described how EWP's are used in LeMans Prototypes. They use two pumps. One at the engine coolant inlet as I describe above and a second one at the radiator hot side inlet. This effectively divides the total system length into two parts allowing pressure to remain higher throughout the system and thus flow rates. It also provides for a redundant system in the event of a single pump failure.

So how does it work for me? I have run my SLC at COTA on days when the temperatures were too hot for me to continue to drive the car in the afternoon session. I seriously believed I was doing something that might be dangerous. I'm taking air temps in the sun of 106F and humidity levels above 80%. THATS FUCKING HOT! The car ran solidly at 191F all day exactly on the controller's default setting. I have no doubt that if I had set it to 180 or 185 it would have been just fine.

Another important point for track cars. OIL COOLERS!!!!! You need to control oil temps just like you control coolant temps. Oil temps should be held above 225F and below 275F. Install a cooler that will achieve that temperature range on a day as I describe. Oil system capacity: you need at least 7-8 quarts systems, Less than that and it gets hard to cool the oil for 30 mins with 400-500HP on a 105F day.

Remember the air temps the cars see are about 4-10 inches off the asphalt. I have aimed a laser temp gauge at the track and seen track temps above 150F on more than one occasion. IN MID MORNING! God only knows how hot the track is at 4pm in Texas in August....................... That is what is going into the radiators.

Radiators: I bought a custom-built C&R radiator for my SLC early on in this process. I am sure it works very well. However, I am not convinced that if I had run the RCR standard radiator it would not have worked well also. Do everything else right first before you buy a new $1200 radiator and huge fans, By the way, I am still running the RCR standard fans. They work fine for cooling the car as I leave the track and drive slowly back to my pit location.

Fans: You can overcome a badly designed ducting system (nor none at all) by installing huge fans, What is happening is you make the radiator believe it is going 60-70mph through the air all the time with the fan airflow. If you need to run the fans at highway speed to cool the car you need to work on ducting. There is no reason to run the fans at 65mph at cruise IF you have a good ducting system. After all the engine only makes 50 -80 Hp at part throttle when you are cursing along at 70mph or so. There is plenty of airflow volume at the nose inlet to cool that. BUT it needs to be forced to flow THROUGH the radiator and out of the bodywork.

Last thing: Ducting! Every effort needs to be made to ensure ALL the air going into the front of the car is forced to go through the radiator and then forced to leave the bodywork out of the top of the car. This may be the single cheapest way to make the coolant system work well after an effective bleed system.

Oh, and ducting will solve all of your hot interior footbox problems at the same time.

Be cool buddies!
 
Last edited:

Howard Jones

Supporter
Well, you are welcome, I hope this can help somebody else. If anybody has something to add, go for it. Some information on OEM EWPs might be useful if it is current. The latest equipment is pretty good. I know BMW is using EWP on their cars.

I have used the term "head". What I am referring to is the NPSHa or "Net Positive Suction Head available". What is important about this is the higher the HPSHa the wider the operating speed band the pump can be run and not experience cavitation. This subject of fluid dynamics is frankly beyond my educational experience. However, the problems presented by a very low, high HP, mid-engined car with its very long coolant system length require all the efficiency that can be squeezed out of the component parts.

The last variable I have not discussed is coolant temperature. The hotter the coolant is, the more sensitive the pump is to cavitation.

I believe that it should become apparent that everything is working against us. However, attention to detail will serve you well, the good news is you can make it work!
 

Randy Folsom

Supporter
Howard, thank you for all that great information. As always, you are a wealth of real world knowledge. I am in the process of sourcing my cooling system and plan on using a DC EWP150, maybe two. Do you have a replacement recommendation for the mechanical water pump. A salvage yard pump with the impeller removed and the shaft hole plugged will work, but I am hoping for something a bit more elegant. Also wondering about heater hose connections. Thinking to tap into the radiator hoseat the front of the car or not even bothering with heat.

You also mentioned the importance of an oil cooler. Can you recommend a remote filter and cooler setup?

Cheers, Randy
 

Scott

Lifetime Supporter
I agree with everything that Howard has said above, but I'll add the following:

1695491898128.png

The plan is not use a thermostat
  • A shroud is important if you're using fans.
  • I also have a custom C&R radiator which is a nice piece of kit. I also had them also provide proper vibration isolation and a shroud with SPAL drop-in fans.
  • You should wire the car so that you can run the fans and water pump when the engine is off.
  • I also recommend an engine oil cooler. Don't forget to plumb a thermostat.
I suggest a remote pump located as Howard suggests; as low as possible and pushing into the engine. If you go remote, I highly recommend that you look at Pierburg pumps. Pierburg produced the world's first series product electric water pump and they supply OEMs like BMW and others. My guess is that they produce more pumps in a week than all of the aftermarket guys combined in a year. Their larger pumps are designed to bolt directly to the engine block and feature rubber vibration isolation, so I would expect them to hold up well mounted to a chassis. I have seen these pumps in multiple endurance SL-Cs and LMP cars. The historical challenge with OEM pumps was controlling them if you didn't have a MoTeC or similar type of system. Fortunately, Tecomotive makes a really nice controller that fits in the plam of your hand. While you don't need to purchase the pumps from them, their site has specs for the pumps. (Tecomotive - tinyCWA | Tecomotive Store). I'm running a CWA400 controlled by MoTeC on my SL-C.

1695493355679.png

The Tecomotive pump controller (orange) is tiny!

1695493081866.png

Pump is mounted as low as possible and immediately before the engine. The pump cap was modified to provide both inlet and outlet manifolds

Removing the mechanical water pump reduces weight and provides the opportunity to change belt routing. In particular, those of you with superchargers can eliminate belt slip by increasing the size of the crankshaft pulley and decreasing the size of the super charger pulley. In my application, I had ATI fabricate a super damper that was 32.4% larger than stock while also increasing the stock 8-rib belt to a 10-rib belt.

1695492285862.png


Howard is correct to point out that you need to watch the length and number of turns in the system. Having looked inside a couple of mechanical pumps, flow doesn't appear as smooth as you might hope due to packaging concerns. This isn't an issue in stock form, because the system is much shorter with less turns. IMO, you can achieve better flow with aftermarket parts. For example, for an LS you can purchase inexpensive O-ring-based adapters that allow 4 swept 90 degree AN fittings (2 ins and 2 outs) to be added (see picture below). Each pair can be easily "Y'd" together. This provides less resistance than the stock pump and some of the fancy billet adapters that I've seen.

1695492598796.png

Cooling system adapters.

If you don't want to remove the pump due to serpentine issues, you might consider laser cutting a plate and using some standoffs to mount a pulley to replace the water pump pulley. This would likely reduce weight while also allowing you optimize coolant flow.

You might want to consider a coolant swirl pot. According to Carrol Smith in Engineer to Win: "I still consider the water system de-aerating swirl pot (as described in PREPARE TO WIN) to an absolute necessity on any racing car." In my case, I used a swirl pot to combine the two cooling outlets on the block, so it serves multiple purposes.
1695492842235.png

Coolant swirl pot. It's not obvious from this side, but the inlets and outlets are tangential. The two bungs under the inlets are for monuting.
 

Howard Jones

Supporter
Scott. Ya, that's an LS-type engine with a heater and thermostat. Good work. Now it makes the difference very clear. I have never had a heater in either of my cars. In California and Texas, the problem is heat not cold. If you do use a heater and you have bleed issues you might think about a bleed line from the heater core also. I have never done this but I have often wondered if this might be necessary. Especially if the heater core is high up as it might trap air in it. I don't know but its something to keep in mind. Thanks for the really nice illustration.

A shroud is always a part of the ducting system. When I say ducting system I mean everything from the nose body opening to the radiator inlet face or oil cooler and from the outlet face all the way out of the car until the exhaust airflow clears the bodywork. Sealed tight. I agree, that oil thermostats are a good idea for street cars.

One thing to also keep in mind is two 45-degree bends with a short length of straight pipe between them is less restrictive than a 90. Long-radius 90s are better than short 90s and a no-radius 90 should be avoided if possible. Sometimes packaging restrictions dictate things and you can't do anything about it.

Randy, Oil system components can be sourced from Pegasus Racing supply. They pretty much have everything you will need and their prices are competitive. The one thing I do recommend is don't use a cheap cast filter adapter and block adapter. They crack and leak. The good ones Moroso and Canton make are the type I like. I also only use Earls or Airoquipe hose fittings for oil lines. The cheap ones break and pump all the oil out and onto the track.............ya.............. ask me how I know that. The low-pressure oil idiot light set at 30psi (recommended) saved the motor but not the $1500 track cleanup bill. That is a whole pickup truck full of kitty litter bags...............

A gutted stock pump can be used and will work. I like the idea Randy has and less expensive components can be used like two flat pieces of 3/8 aluminum cut to shape with threaded holes to accommodate a quality radiused adapter. Davies Craig makes a ready-to-fit front adapter set for LS and most usual engine types. I used one on my SBC in my SLC.

Oh, You cannot run two pumps off one DC controller. At least that's what they told me when I asked. But what I would do is install the second pump without a controller up in the nose and wire it so that it was always on. Now the controlled Davis Craig pump in the rear can regulate coolante temp with its duty cycle system. I haven't done this but I think it would work. However, I do not think it would be necessary to use two pumps unless you had a huge heat load (very high HP race car). Steward recommended that if I wanted to use two pumps then to use two of their inline pumps and run them in the on state all the time in a race car. If you did this the only way I can think of to control coolant temp is to restrict airflow through the radiator with some form of blocking system tuned to the ambient air condition. This does not lend itself to a street car I would think. On a track car, it would be experience-driven.
 

Randy Folsom

Supporter
Scott, Why two AN fittings on each port? Seems to me one larger fitting would be easier and less likely to fail. Is that a characteristic of Chevy motors? There is one port on each side of my 302. They angled with openings about 1-1/8” x 7/8”. As John mentioned, I am thinking to mill some billet aluminum blocks to accept 1” AN fittings at angles that match the angle of the ports. That way I can use one 45 deg radiused bend instead of 90 deg. I like the idea of using a swirl tank to split the pump 1-1/2” output to two 1” fittings, but worry about the additional load. That’s gotta be like adding 5’ of hose.

Howard, I wonder if the controller could be interfaced to two pumps via a relay. In any case, if I follow the guidance presented here, then one pump is probably sufficient since I’m not competing in any endurance races where redundancy is important.

I am planning to get a good quality remote oil filter with fittings for temp and pressure sensors. I have given up on anything remotely budget minded. I have already far exceeded my anticipated spend and I have not even received the car yet.
 
Last edited:

Scott

Lifetime Supporter
Radiused 90 degree fittings are preferrable to simple right angle AN fittings.
Neil, I'm not sure what you're referring to. Those fittings have as large of sweep of any fitting that I've seen with that ID. In addition, those fittings are a lot more exotic than AN fittings. They have an O-ring-boss (ORB) connection on one side and a Wiggens clamp on the other side.

One thing to also keep in mind is two 45-degree bends with a short length of straight pipe between them is less restrictive than a 90. Long-radius 90s are better than short 90s and a no-radius 90 should be avoided if possible. Sometimes packaging restrictions dictate things and you can't do anything about it.
I completely agree. The fittings are long radius because two 45s wouldn't fit.

[TWO PUMP CONFIGURATION] I would do is install the second pump without a controller up in the nose and wire it so that it was always on. Now the controlled Davis Craig pump in the rear can regulate coolante temp with its duty cycle system.
It might work, but it sounds less than optimal. If the second pump running full tilt all of the time increases the amount of circulated water vs it being off, it will take the engine longer to come up to temp which isn't good. In addition, I'm not sure why you'd run the pump full tilt. Keep in mind that the CWA400 can draw up to 35.5A.

Steward recommended that if I wanted to use two pumps then to use two of their inline pumps and run them in the on state all the time in a race car. If you did this the only way I can think of to control coolant temp is to restrict airflow through the radiator with some form of blocking system tuned to the ambient air condition. This does not lend itself to a street car I would think. On a track car, it would be experience-driven.
This makes no sense to me. Maybe these race engines are rebuilt so frequently that there is no need to warm them up. Keep in mind that a mid-engine car with the radiator in the nose and no thermostat is going to take a while to come up to temp. Maybe the use case is drag racing? While I've covered part of the radiator core in my Cobra to maintain optimal temp in the winter, that's 1960's technology which is appropriate for a car that's an aero dynamic brick. Unless you have some intelligent (from an aero perspective) way of reducing mass air flow through the radiator core, it seems to me that you're going to increase drag and decrease downforce on the nose (assuming you have an engineered radiator outlet).

It seems to me that those two manufacturers are recommending these solutions because their controllers don't handle a two-pump configuration. I haven't seen a dedicated two-pump controller, but this is trivial to configure in a modern PDM or ECU.
 

Scott

Lifetime Supporter
Scott, Why two AN fittings on each port? Seems to me one larger fitting would be easier and less likely to fail. Is that a characteristic of Chevy motors?
On an LS, each side has an inlet and an outlet, so the adjacent ports can't be combined. There must be four discrete connections and an LS mechanical pump just does the routing internally. While more connections results in more potential failure points, I'm using high-end hose and Wiggens fittings that were designed for aerospace (i.e., they have standard and zero atmosphere ratings) and have trickled down to motorsport.

I wonder if the controller could be interfaced to two pumps via a relay.
You don't want to do that for the reasons listed in the above post.

In any case, if I follow the guidance presented here, then one pump is probably sufficient
I agree
 

Howard Jones

Supporter
I have used the term "duty cycle" to describe the method that the DC controller uses to regulate coolant temp. This is a bad choice of terms. The Davies Craig controller varies the speed of the pump not an on-off/time method to control flow rate and thus coolant temperature.
 
Pump location: the pump should be placed so that the highest pressure available is directly applied to the hottest heat-generating point in the system, This will always be the ports that the previously removed mechanical pump used to deliver cool water to the engine. Engine design is different from engine to engine but in any case as close to the engine cool inlet as possible. The pump inlet position should be as low in the car as possible to place as much of the total water volume as possible ABOVE the pump. This is known as "head".

Thanks Howard for starting a thread on this important topic.

As with most all mid-engine cars, I'm dealing with a trade-off situation with my homebuilt Miura. The trade-off deals with the EWP pump location. If the pump is located in the front of the car (next to radiator coolant outlet), the pump can be located very low in the chassis. If the pump is located near the engine, there is no available space to locate the pump low in the chassis, thus it would need to be located vertically about the same height as where the mechanical pump (now replaced with a hose inlet) was originally positioned.

So the question becomes, is maximizing the "head" at pump inlet by keeping it low in chassis more important than maximizing the coolant pressure at the engine? I know coolant pumps push better than pull/suck (especially if pulling with little head pressure), so logic tells me the better trade-off is to put the EWP up front and low. Does anyone have practical experience that gives insight on this particular trade-off?
 

Brian Kissel

Staff member
Admin
Lifetime Supporter
I had the same questions as Joel. I have seen radiators that have a pump built right into them, so up front sounds like a possibility. I would like more thoughts on this as well.
Thanks
Regards Brian
 

Scott

Lifetime Supporter
If it were me, I'd locate it in the nose after the radiator and take as much care as possible to minimize constrictions on the way to the engine. This maximizes pump head and relocates a small amount of weight forward and low. The endurance SLCs built by Agile Automotive had the water pump in this configuration. To my recollection they were running a Pierburg CWA200 and a LS3. If you're worried about capacity or pressure you could go with a CWA400 which has roughly twice the capacity for about $110 more and let the ECU figure out how hard to drive it.

The big question with the pump in the front is how you're going to handle engine warm up. With a mechanical pump there is a thermostat that takes the radiator out of the loop until a certain temp is reached. A mid-engine car has longer coolant lines so there is even more of a need for a proper heat up circuit. If you look at the diagram in my post above, you will note that there is a 5/8" by; pass line in the engine compartment that purpose.

You could run a similar line from the engine to the pump inlet or you might be able to tap into the radiator supply line so that you don't need to run a separate line all of the way from the rear of the car. The point is that you want little to no water flowing through the radiator until the engine has hit the temperature where the thermostat would normally open.
 

Chris Kouba

Supporter
The big question with the pump in the front is how you're going to handle engine warm up... [snip] The point is that you want little to no water flowing through the radiator until the engine has hit the temperature where the thermostat would normally open.

Would you still not be able to locate a t-stat in the block which would send the pump controller the block temp? The only fluid movement would be thermally induced, density driven flow, which would be somewhat contingent on the pump impeller turning.
 
Last edited:

Scott

Lifetime Supporter
Would you still not be able to locate a t-stat in the block which would send the pump controller the block temp? The only fluid movement would be thermally induced, density driven flow, which would be somewhat contingent on the pump impeller turning.
Chris, I'm no expert, but here's my thoughts. Engines have hot spots and you need a fair amount of coolant flow to keep temps even and prevent flash boiling. The mechanical pumps that I've seen have an internal 5/8" bypass which is critical when the thermostat is closed. IMO, without flow a temp sensor isn't going to do much so you're going to need to pulse the water pump in a warm-up mode. If the pump is in the nose and there is no bypass line, then the flow includes all of water in the system; radiator, side pod lines, expansion tank, etc. Keep in mind that the later two are larger in a mid-engine car.

No by pass will work, but it will result in a longer warm up time and more engine wear. This may not be a concern, but if you have a $40k motor or start your car in the winter, it may provide some concern.

I forgot to mention the expansion tank line. As Howard mentions above:

The expansion tank MUST be located at the highest point available in the car. This is a very important point. The expansion tank connection to the cooling system (expansion line) MUST be connected to the lowest PRESSURE point in the system. This will always be at the pump inlet return from the radiator right before and as close to the low-pressure side of the impeller as possible. This will also be the lowest-temperature coolant in the system. This is also a very important point.

Since the expansion tank must be mounted in the engine compartment to achieve the desired height, you need to run the expansion line to nose. This is one of disadvantages of putting the pump in the nose.
 
Last edited:
Another option, many of the OEMs that still use a belt driven waterpump have a newer design thermostat that rather than just blocking the flow to the radiator, they divert it back to the water pump intake giving much better closed thermostat circulation than just relying on a bypass, and more positive flow to the radiator because the bypass is closed when the Tstat is open.

Here's one for a Duratec V6 https://www.ebay.com/itm/186006714374

The bottom end would bolt to a simple plate with a hole where the Tstat sits, and a nipple welded to it.

I'm planning to just use the 90s 3/4 truck timing cover and heavy duty water pump, and a restricted bypass, but I have this on a Kelmark, at the top of the coolant reservoir. If I remember correctly, hot water from engine goes to the middle hose. If it's below 190, it's sent to the pump intake. Once it exceeds 190, it's sent to the radiator. Then the radiator also returns to the reservoir.
 
Would you still not be able to locate a t-stat in the block which would send the pump controller the block temp? The only fluid movement would be thermally induced, density driven flow, which would be somewhat contingent on the pump impeller turning.

For the Coyote engine being used in my Miura, the temp sensor for the coolant/pump controller is located on the engine coolant outlet just prior to the fill cap which is vertically the highest point in the system. I put the temp sensor there thinking the coolant while warming up to temp would thermally rise up and thus tell the controller when to start moving the coolant during the engine warm up process.

I did plumb in a second, much smaller pump (visible in lower portion of picture), to push hot coolant directly to the heater core for when the interior heater is in use. If some circulation is required during engine warm up, I'm guessing this pump could be turned on to provide it. But now saying that, it probably wouldn't help move coolant by the temp sensor as it's drawing from a different port than the main engine coolant outlet.

P1170168.JPG


The EWP is currently attached directly via a hose to the radiator lower outlet.

P1140963 - web.jpg


As you can see, I put a bleed screw at the highest point of coolant system in the front of the car. I was thinking that once air was bled from the system, no more air would accumulate here. Given the earlier discussion pointing out having a line run from the top of the radiator back to an expansion tank, I'm guessing I'm making an inaccurate assumption and air will continue to accumulate here. Should I change this bleed screw to a dedicated connection to a line running back to an expansion tank? If so, there's no way to run a line "uphill" from the front to back of chassis in this car. Is it OK if this "bleeder line" goes "downhill" before reaching the expansion tank?
 

Howard Jones

Supporter
I wondered if it would work too a long time ago. I think what is happening is the coolant forces the air down the bleed line to the air pocket in the expansion tank. After all, it's the coolant pressure that forces the air out if you open the bleed valve. It is kind of the same thing, except the bleed valve is always open and is connected to the remote expansion tank via a hose. It may be that the small diameter of the bleed hose increases pressure in that hose enough to force the air down the line instead of accumulating at the top of the radiator. I use AN-4 size lines and fittings for bleed lines throughout the bleed systems on both cars.

I have run the bleed line that runs from the radiator to the expansion tank as high in the car as I could, along the top of the nose box, through the side pod, and in the rear engine room. It just seemed to be the best way to do it.

What I do know is that both cars, my GT40 and my SLC, use very similar layouts of the same system. When I drain the coolant out of the car for some reason or another all I do to refill it is pour fresh coolant into the fill/radiator cap port on the top of the expansion tank until it is full, start and run the car for a few mins until the coolant temp just starts to rise, then shut it off check and refill expansion tank to half full, cap it and it's done. All the air has accumulated in the expansion tank and does not require any further bleeding. This is why I call it a self-bleeding system.
 
Last edited:

Howard Jones

Supporter
A bit about fluid pumps used in combination: Two pumps (we will use the same type and performance for this illustration) when installed in a system in series will not increase total flow in quantity (gallons/time) BUT two pumps in series will increase the total pressure (head) of the system.

If you run them in parallel they will double the total flow but not increase pressure (head).

What I have suggested as an option for a mid-engineered very low-profile car is a series installation. The reason for this would be to address the possible issue of inadequate head (pressure) in such a configuration. As an aside, the reason I like the idea of the pump (in a single pump plan) discharging directly into the engine is to use as much of the available pressure right at the heat source. System pressure in a fluid system will drop over distance, in relation to the resistance of the system to flow. Things like tight bends, less than smooth surfaces inside the pipe or other components, restrictions in tube cross-section, unnecessary increases in pipe diameter, or long distances of equal diameter size pipes.

So......when to use a second pump? First, what is the required volume (flow rate) in one of our cars? Well, all I know is that the three different manufacture examples of pumps that seem to work all have a published flow of 50-60 Gal/min. So if you had an installation that seemed to have insufficient (head) pressure then you might be able to solve that with a second pump in series. However, all of the fundamental design elements mentioned in the original post should be addressed first.

What does a low "head" problem look like in one of our type cars? I think that is a car that boils over a lot or runs very hot despite having a good bleed sys, a properly sized radiator, and moderately effective fans as well as good radiator ducting. Usually, boilover in an otherwise nominal system (no stuck thermostats, bad fan, or low water) is related to steam generation in the heads. Maintaining high coolant pressure in the heads is the first thing that must be done to prevent stream generation in the heads. Adequate "head" is the fundamental design parameter at work here. Again, it is also why a proper bleed system is so important.

And again our low and long coolant system cars with high heat loads and minimum "head" is the perfect storm.

So to not be negative. The one thing we do have going for us is a relatively larger coolant volume as compared to a front-engined car, There is quite a lot of coolant in these long pipes in the side pods as well as the ability to install a relatively large volume expansion tank. (up nice and high).

The three pumps listed below all have about 50-55 Gal/min flow rate



"Davies, Craig’s range of patented Electric Water Pumps (EWP®) are available in flow rates from 90L/min (23.8 gal/min) up to 162L/min (42.8 gal/min) at 12v & 24volt and in Nylon66 / Alloy housings." I believe that the EP150 is the one rated at 48.2 gal/min.
 
Last edited:
Back
Top