Ulix

True Classic
As I have mentioned, I have removed the water pump from my engine and installed a Davies Craig electric pump. I will go into detail about this mod later, but I want to share an observation I made in the process.

The engine is a 1.9L with high compression and a bunch of mods.
I am still in the process of running it in and setting it up.
But extented drives so far have shown that the engine is very cold blooded. With the electric pump, the coolant temp usually runs around 160°F even when pushed. Granted it is only about 55°F outside, but the pump is only running it its lowest setting at that point (6V, 10 sec on, 30 sec off), so there is alot more pumping flow rate to be had.

My point is, if you have temperature issues, and you have already recored the radiator (which should be the first step), it would be worthwhile to look into an auxillary electric water pump to increase the flow rate when things heat up. There iseems to be a lot of potential there.
Note that I haven't tested this on a really hot summer day yet.

(At this point, I would not recommend a full conversion like I did for a street car, because the swap removes the thermostat, so the engine takes a long time to heat up. Also, the control algorithm and the cycling of the pump results in large variations and jumps in coolant temp.)
 
I ran a similar setup on the Rotary engine X I had awhile ago. It had no problem keeping this engine cool. Trust me, if an electric pump can keep a 13b motor making 350hp cool, then it will probably be overkill for a standard motor. Now, Ulix has a good reason to move to electic given how he uses his car, however if you have a street driver and it is overheating, please look at the other issues first. Have the Rad cleaned, check lines, check pump, check tstat. Bleed correctly. On my Uno Turbo, I am running the stock rad and I can beat the crap out of it and never see it above 190. Just my .02 on this.
 
Some racers report problems with cooling associated with cavitation induced by sustained high RPM operation. The easiest way to address this type of problem is to use a water pump with the pressed steel type impeller, then trim every other impeller blade back (shorten it) to decrease the water pumps capacity. Of course an electric water pump could be much more effective in racing applications as long as you have enough power to run it.
 
Agreed, the stock system is adequate when everything is in proper condition, on a stock engine, used for normal street driving, under typical conditions. But given the long cooling pipes from the back of the car to the front, the limited size of the rad, and the ancient design of the system (for example the small size of the pump and the relative heights of the rad/pipes/block), it would make perfect sense the stock system is somewhat on the edge of its limitations with regard to flow capacity. So the use of a larger capacity/higher flow-rate pump seems appropriate when running a modified engine or if used for more than normal street driving. Although 'fastx19' seems to be proving otherwise with his UT engined X, but the UT engine might have a better pump compared to the X engine (I'm not sure)?

I wonder how much retaining the stock pump hinders the effectiveness of adding a auxiliary ("booster") electric pump into the system? The two pumps might be competing or fighting one another.

Also, could the T-stat be retained with the use of a full electric pump (eliminating the stock pump), instead of relying on the e-pump's program to control temps?

I seem to recall some old discussions on ideas like running the stock pump faster (pulley size), but that might cause cavitation. Could a different impeller with greater capacity be used in the original pump housing? Mike, is there a specific steel impeller you are referring to?
 
As I have mentioned, I have removed the water pump from my engine and installed a Davies Craig electric pump. I will go into detail about this mod later, but I want to share an observation I made in the process.

The engine is a 1.9L with high compression and a bunch of mods.
I am still in the process of running it in and setting it up.
But extented drives so far have shown that the engine is very cold blooded. With the electric pump, the coolant temp usually runs around 160°F even when pushed. Granted it is only about 55°F outside, but the pump is only running it its lowest setting at that point (6V, 10 sec on, 30 sec off), so there is alot more pumping flow rate to be had.

My point is, if you have temperature issues, and you have already recored the radiator (which should be the first step), it would be worthwhile to look into an auxillary electric water pump to increase the flow rate when things heat up. There iseems to be a lot of potential there.
Note that I haven't tested this on a really hot summer day yet.

(At this point, I would not recommend a full conversion like I did for a street car, because the swap removes the thermostat, so the engine takes a long time to heat up. Also, the control algorithm and the cycling of the pump results in large variations and jumps in coolant temp.)

Have you considered making your own controller for the pump and or going with one of their smaller pumps?

I have played with a pwm circuit to manage the fan and an electric pump using an Arduino but do not have enough confidence in the unit yet to actually place it in service. (I have cadged assistance from one of my coworkers for programming as I am not adept).

Just a thought.

Thanks for bringing this up, it is an interesting area creating control systems that replace purely mechanical systems.

Getting to a narrowly controlled solution is really going to be critical to getting consistent performance out of the engine. Having relatively wild fluctuations in temp is not good for combustion or oiling of the cylinders.
 
but the pump is only running it its lowest setting at that point (6V, 10 sec on, 30 sec off),

First, I would say you absolutely need a thermostat in the system. A Beta dual action stat functions nicely in my K20 X and keeps the temps very stable.

Second, I would say a 10 second/30 second duty cycle (or any timed duty cycle) would be liable to cause very erratic cooling. The water has to spend enough time in the radiator to shed the heat. If you run for 10 seconds, just how far exactly did the hot water from the engine get pushed? Up to the radiator? In it? Past it? Seems like it's a recipe for having lumps of hot or cold water at various parts of the system.

In short, I would be looking to upgrade to a thermostat and continuous flow. How to make the water pump push the water around in both modes (stat open and stat closed) might be tricky...

Sounds like a neat setup though, I'd be interested where this goes.

Pete
 
I agree with Pete. I use an electric pump on my track car running at full speed continuously with the standard thermostat in place, original pump replaced by pipe going straight on to side of block. Car warms up quickly, usually runs at 90 degrees C on track, but can get a bit hotter on a very hot day, then I need to fit air deflectors in front of the hood ducts to improve air flow.
 
Sorry I used some jargon, pwm stands for pulse width modulation.

PWM basically turns on and off the pumps electric motor multiple times per second, minute, hour (if needed) or on full time to achieve a particular pump “on” time to ensure a particular effect. In this case it might mean that the pump is always running a little (pulsed on several times a second) to ensure there is movement to minimize hotspots. Then as the temp rises, the processor will step up the “on” time of the pump based on the coolant temp coming from the motor. By that same token with control of the fans you would measure the in temp of the radiator to the out temp of the radiator and adjust the radiator fan on time (if needed) and could even bring in a second fan to ensure an appropriate temperature differential of the in to the out.

Having some coolant movement all the time is valuable even if it is only a minor amount.

I don’t know what the algorithm is for the DaviesCraig controller but I would hope it offers a bit more adjustability to allow some constant flow with a rising flow rate to support a given temperature target.

Karl
 
How to make the water pump push the water around in both modes (stat open and stat closed) might be tricky...
It might not be too difficult because the X's T-stat has the bypass built in (2 way system). So circulation is always happening; partially to the whole cooling system and partially to the recirculate/bypass portion, depending on the relative position of the T-stat. The e-pump could be plumbed the same as the stock pump.

I believe the electric water pumps are variable speed? If so, the flow rate can be adjusted that way also.
 
This is the control scheme of the pump:

EWP.JPG


It is designed to keep the temp at the desired (adjustable) target. When it does, the pump runs continuously, but the speed is varied (portion of graph labeled LINEAR RAMP, Area C).
Problem is that I chose too big a pump (115 L/min because it is aluminum and not plastic like the smaller 80 model), so that the pump never get out of the two warm-up modes where the pumps cycles (Area A and B). At least in the current temperatures.
This may be different come summer.
 
Thanks, I wasn’t aware that they were so “crude” with their control parameters. Hmmm.

So the lower capacity pump would likely have more consistent run times getting into the far end of B and then C as the temps rose spending more time in the variable speed constant on of the C regime.

Hmm, perhaps I am being too tight with my control scheme which would mean my error worries should be less real for my 850’s controller. For my 850 I have been playing with a pump intended for intercoolers (similar to BBrown’s pump he used on the trailer queen) and have been worried about capacity so I have been thinking about adding a second pump which would be run as needed to boost flow but it may not be needed.

In thinking about this furtherI need to add some instrumentation to my X to look at factory radiator inflow temp versus outflow temps to better understand what I should do as I move forward.

Food for thought. Thank you very much for bringing this subject up and the issues you are currently having.
 
For my 850 I have been playing with a pump intended for intercoolers (similar to BBrown’s pump he used on the trailer queen) and have been worried about capacity so I have been thinking about adding a second pump which would be run as needed to boost flow but it may not be needed.

There was some discussion in the BBrown thread, but you need to be careful when running pumps in parallel that you don't end up just recirculating the coolant in a loop. I think a properly set up electric pump as a replacement (rather than a supplement) to the mechanical pump is a far more elegant solution as it allows more precise control of the temperature with the ability to increase the flow rater under specific conditions (e.g. when the car is stuck in traffic). This is the biggest drawback of the mechanical pump in that the speed is proportional to the engine speed, which is not always desirable.

I'll be interested to see how Ulix goes with his setup.

Best regards,
Dominic.
 
Ulix, interesting that the pump might be too big. Although it seems that should not matter considering it's speed can be controlled; reduce the maximum speed limit on it and it will act like a smaller pump. Can the controller be reprogrammed to work in various modes?
Also, what would happen if the T-stat was used in conjunction with the E-pump...especially since the X's system is a 2-way bypass design? One concern with not running a T-stat is it will 'leave the back door open while the AC is running' (guess you might have to live in an extremely hot climate to get that). What I mean is, the by-pass circuit will be open all the time, diverting some of the coolant circulation away from the main cooling system. Not very efficient control of coolant flow over the range of temps.

I have been playing with a pump intended for intercoolers
Karl, those pumps have a VERY low volume/flow rate. Check their specs and compare it to the type of pump discussed by Ulix...HUGE difference. I do not believe one of the smaller pumps you are referring to will be enough. Even two of them is not a lot of volume, but if you run multiples they would need to be run in parallel and not series.

you need to be careful when running pumps in parallel that you don't end up just recirculating the coolant in a loop
Dom, that was my thought earlier that two different pumps might be competing (so to speak) with one another. I agree, it would be better to have a replacement rather than a add-on pump.

The X's front rad / rear engine layout, with those long tubes in between, is a good candidate for a E-pump design.
 
Last edited:
The small pump is for my 850 whose radiator is 3” from the side of the block and its a whopping 817cc so it doesn’t need as much flow as an X :) We have some members here that use one successfully on a healthier 850 so I am not too worried.

I have a large DC pump I bought from a member here slated for an X or my 124 or ?

I am especially interested in the control process as the pump is just a thing to manage, its capacity is however very important.

A non thermostat system needs to be dealt with quite differently. You no longer need a bypass, you need to bring all of the coolant to temperature but the process one uses to do this is quite different and needs some replumbing. This means you need to know the temp leaving the engine, if it isn’t at a preferred temp then you need to reduce the flow. This is where a PWM approach to pump control works well as you can throttle the pump down to deliver the desired flow.

One can continue to use a thermostat with an electric pump with a low flow for a given time period to deal with the reduced need while recirculating and then after the thermostat is open a higher flow rate. One will still want to have a temp sensor at the outlet to cue the start of the higher flow to move the greater volume of coolant.

Being someone who likes thinking through systems, creating a multiple feedback system to control the pump, fan and temperature system is fascinating. I know, I am a boring nerd...
 
its a whopping 817cc
Excellent point, much like the cooling system for a small motorcycle. The smaller pump may be enough. I looked into these for a possible turbo inter-cooler set-up and was quite surprised at the limited capacity, so do the math and determine your needs.

You no longer need a bypass
I guess if you are controlling the temp 100% via water pump flow (i.e. no T-stat), then you would need to modify the X's head/T-stat housing assembly to eliminate the whole by-pass/recirculate system?

Very interesting to see what develops on this...but no pressure to get back to us right away with results ;) :D [Note: I haven't accomplished shat on my projects lately]
 
Yep. I have modified the T-stat housing with a blanking plate.
There is no more shortcut in the system, only full flow through the radiator.
The temp sensor controlling the pump is also in the T-stat housing, so it monitors the water temp right where it exits the engine.
Eliminating the shortcut is what Davies Craig specifies in their manual, so that's what I did.
I have not thought it through if it would work with a T-stat.

Karl, why do you feel the need to monitor inlet and outlet temps on the radiator?
If it is because of the notion that water needs time in the radiator to cool off, that is an internet myth.
The faster you pump, the more efficent you heat transfer will be and the more you are cooling the engine (unless it is too fast for the system and it cavitates etc.)

I think the control strategy of the DC system is fine, if I quickly make it into "Area C" where pump speed is controlled by water temp exiting the head and then stay in that zone.
This may be the case when summer gets here. :)
 
On your system Ulix, I guess my only heartburn is that the control system doesn’t maintain a better target temperature, 160° isn’t ideal, 180 would be much better and something closer to 190 would likely be better still from a combustion stand point assuming it can hold it. It should always hold the target temperature whatever that is, by flow control and the radiator doing its job. But that is just me with my non engineer novice knowledge :), as we all know opinions are like a**holes, everyone has one but one test is worth more than all the opinions...

About the 850 solution
The two sensors (in the 850’s case) are a sensor in the thermostat housing, just a few inches from the radiator entrance and in the exit hose from the radiator. The reason for that is to look at the temp differential and modify the fan speed accordingly to reduce electrical use when not needed and ensure that I am getting effective cooling.

This output sensor is effectively being the installed temp switch in an X radiator, the 850 comes with a belt driven pump and fan on one shaft so their action is based entirely on engine speed which tends to give a temp spike when idling at a light after coming off the highway for example. I am trying to eliminate that behavior and don’t want to modify the radiator. The 850 does have a bypass but it is very small and easily isolated.


I would agree with your points about the radiator on an X, I would use the temp sensor in the thermostat housing to give the ‘inlet’ temp of the radiator and use an inline sensor in the exit hose rather than the standard radiator sensor position if I went that way. I am not concerned about dwell time in the radiator as ultimately I don’t see the flow on my system exceeding the standard flow rate or if it does it would be a minor difference.

You are right, monitoring the output temp is something one really doesn’t need to do on an X as one can easily use a two speed fan switch in the normal position to control the fan(s) quite satisfactorily. Simple direct and reliable, follows the KISS principle which I generally prefer anyway. Unless I modify the 850 radiator I can’t go this way and an inline sensor in the hose is very simple.

In any case thank you for posting about your system, it is a good learning opportunity and writing about what I am doing helps me as explaining it to others helps to crystallize what it is I am trying to do and gives me valuable feedback from others which I don’t often receive.
 
opinions are like a**holes, everyone has one but one test is worth more than all the opinions.
So how do you test a a**hole?

I agree with Karl, use the temp sensor in the T-stat housing to control the E-pump, and the temp sensor in the rad to control the E-fan(s)....which is what you are doing I believe. As for the temp control, it sounds like maybe just a matter of getting the E-pump's program dialed in?

I also agree with Ulix about the coolant not needing more time through the rad to cool off. If your rad doesn't cool the fluid to a desired temp level, then it is undersized for your application. I've seen this comment many times, people thinking the system needs to run slowly in order to allow the rad time to cool the fluid more. You do not want to slow down the flow rate. That will certainly lead to overheating...which would likely cause such logic to want to slow it down even more? Look up the laws of physics and do the math.
 
Last edited:
Missed this. I'd like to see pics of the install.

Don't understand why the pump can't be set for continuous run, though. Erratic coolant temps seem like a very bad thing.
 
Back
Top