Cooling System Build Thread

[Dr.Jeff]

While at a trade show this week I was talking to a company that supplies OEM replacement electric water pumps. I mentioned about the idea of converting a vintage car engine's belt driven pump to a electric one. Unfortunately I haven't had the time to keep current with this thread, but I recalled something about a BMW pump being used. So I said that to the guy at the show. He reached over and picked up a sample unit from his display, and said "it is probably one of these". From my vague memory it did appear to be similar to what's been discussed here. Being the last day of the show he told me to keep it and see if I could use it for a project. After getting home and searching the part number on the tag I see it is the CWA400 pump. Considering the pump was free makes me want to do this mod. However I am concerned about the amperage draw (35.5) for use on a X. Not to mention I already have way too much going on and really need to get some vehicle builds completed soon. So it likely won't happen. But for the sake of conversation, given these pumps have an internal control is there a way to lower the amperage draw by how it is configured? I apologize if this has already been discussed here...as I said I have not been able to keep up with the whole thread.

 

[Jonohhh]

While at a trade show this week I was talking to a company that supplies OEM replacement electric water pumps. I mentioned about the idea of converting a vintage car engine's belt driven pump to a electric one. Unfortunately I haven't had the time to keep current with this thread, but I recalled something about a BMW pump being used. So I said that to the guy at the show. He reached over and picked up a sample unit from his display, and said "it is probably one of these". From my vague memory it did appear to be similar to what's been discussed here. Being the last day of the show he told me to keep it and see if I could use it for a project. After getting home and searching the part number on the tag I see it is the CWA400 pump. Considering the pump was free makes me want to do this mod. However I am concerned about the amperage draw (35.5) for use on a X. Not to mention I already have way too much going on and really need to get some vehicle builds completed soon. So it likely won't happen. But for the sake of conversation, given these pumps have an internal control is there a way to lower the amperage draw by how it is configured? I apologize if this has already been discussed here...as I said I have not been able to keep up with the whole thread.

I don't blame ya for not being able to keep up with the thread. Far too many words for the amount of info conveyed, admittedly, but whatever.

The CWA400 is fine. It's absurdly powerful, but can be easily tamed by running it at a slower speed. The tecomotive controller has a "virtual smaller pump" function that caps the pumps maximum speed at some defined value, so you could easily do that.

The only real downside (other than cost which is irrelevant in this case) is decreased pump efficiency at an equivalent flow rate to a CWA200. Also, I think the opposite direction housing may be a little harder to plumb up, but that's solvable.

I sized the wiring for my project to be suitable for a CWA400 despite using a 200. I have these pierburg pumps coming out of my ears these days so I left the option open. I can provide info on the wiring if wanted, but the B+ source is the starter terminal, trigger source for time delay off relay is the fuel pump feed. Works excellently.

 

[Dr.Jeff]

The CWA400 is fine. It's absurdly powerful, but can be easily tamed by running it at a slower speed. The tecomotive controller has a "virtual smaller pump" function that caps the pumps maximum speed at some defined value, so you could easily do that.

Thanks, that was exactly my thought....throttle it down a bit by capping the max motor speed to reduce the load draw. But I haven't reviewed the Tecomotive controller yet so I did not realize it offered that feature. If I use this pump for a X-engine then I really wouldn't be too concerned with a little less efficiency, because it is so overkill to begin with.

This would also make a great choice of pumps if I ever build the next project car on my list; a 450-500 HP front engine-rear drive tire shredder (just for fun). That might even be a Fiat (perhaps a 600 or 126 ?) with the body only being a silhouette over a custom frame/drivetrain.

 

[Jonohhh]

Thanks, that was exactly my thought....throttle it down a bit by capping the max motor speed to reduce the load draw. But I haven't reviewed the Tecomotive controller yet so I did not realize it offered that feature. If I use this pump for a X-engine then I really wouldn't be too concerned with a little less efficiency, because it is so overkill to begin with.

This would also make a great choice of pumps if I ever build the next project car on my list; a 450-500 HP front engine-rear drive tire shredder (just for fun). That might even be a Fiat (perhaps a 600 or 126 ?) with the body only being a silhouette over a custom frame/drivetrain.

That sounds like quite a bit of fun. I hope for my 65' Chevy to fit that role one day, but it currently has about half of its body panels and approximately 0hp.

One day...the fiats are the priority lately.

 

[Jonohhh]

Update time!

Good news, the car is finally drivable again - and better than ever I must add. Cleaning up the main electrical connections and upgrading the alternator-starter wire have made the electrical system, even with the water pump at 100% speed (~15 amp load), maintain a higher voltage than before. Another advantage of the CWA that I don't think I covered previously, is that voltage fluctuations down to as low as 8v or as high as 16v have no effect on pumping performance.

I had trouble getting the pump to prime, and as a result, accidentally ran the pump dry for a very long time. However, now that it's primed and working, the performance is exceptional, and it's near silent.

The build is far from complete, there's still a lot of cleaning up to do and features missing, but in about 2 weeks, I'll have time to do just that for a few weeks.

Belt:
With the WP pulley delete, I found that the configuration mentioned previously works as I had hoped. The belt is a little tighter than I'd like, but the next size up was far too loose. The alternators tension adjustment now works backwards to how it did before, backwards is looser, forward is tighter. The adjustment range is also smaller- requiring the movement of the AC compressor for large adjustments, but that should rarely be needed, and still worlds better than dealing with the split pulley and almost impossible to remove alternator belt nonsense.

My alternator bracket was modified by a previous owner, so your mileage may vary, but I settled on a Gates Powerated 6840. The outer section of the split pulley has been removed, as all that is currently used is the inner V.


Additional Temperature Sensor:
I plan on repurposing the factory temp gauge for oil temperature once the controller (which shows coolant temp) is repositioned to the dash, but I'm not quite ready to do that yet.

For now, I have a few options. A spacer plate on the thermo housing, a T fitting on the thermo housing, or a T fitting on the expansion tank feed line. Given that my future plan is to remove the original gauge sender, I'm tempted to just tee into the expansion tank line for now. If accuracy/temp loss is a concern I can put the FI temp sensor on the hose instead since accuracy matters far less for it.

 

[Jonohhh]

Heater routing:

This is a tricky one. Lots of options, hard to choose. Currently the heater is routed from the port near the snail mount to the second inlet on the expansion tank. It works well at high pump speed, but not so well otherwise. I am also unhappy with how much coolant bypasses the cooling system through the expansion tank. While those are currently separate issues, I recently saw something here on Xweb that may resolve both of them in one- a four port, electronic heater control valve. But we'll save that for later.

To improve performance, I am going to try to get the inlet from the coolant feed pipe rather than the head. Hopefully the coolant will still be plenty hot enough, we will see.

 

[Dr.Jeff]

Additional Temperature Sensor:
I plan on repurposing the factory temp gauge for oil temperature once the controller (which shows coolant temp) is repositioned to the dash, but I'm not quite ready to do that yet.

For now, I have a few options. A spacer plate on the thermo housing, a T fitting on the thermo housing, or a T fitting on the expansion tank feed line. Given that my future plan is to remove the original gauge sender, I'm tempted to just tee into the expansion tank line for now. If accuracy/temp loss is a concern I can put the FI temp sensor on the hose instead since accuracy matters far less for it.

The stock temp gauge will need to match the oil temp sender resistance requirements to work correctly.

Just off the top of my head, the idea of adding a coolant temp sensor in the expansion tank line does not sound like a good solution. I don't think it will be anywhere near accurate. Maybe add another bung onto the T-stat housing for the extra sender? Or place it in the larger coolant hose feeding from the engine with a inline adaptor like this?

You can get them very inexpensively on eBay, Amazon, AliExpress, etc.

 

[Dr.Jeff]

To improve performance, I am going to try to get the inlet from the coolant feed pipe rather than the head.

Various years of X's had the heater supply coming from a few different places: You can get thermostat housings with the extra nipple, coolant pipes (from T-stat to water pump) with it, off the 'front' end of the head, and even directly off the water pump (and possibly others that I don't recall). I've never heard anyone say any location was better or worse than the others. So just look for the most convenient spot to fit your design. Then you could get the needed (stock) part with that nipple, or make your own attachment point.

 

[Jonohhh]

Hi Doc, hope you're well!

The stock temp gauge will need to match the oil temp sender resistance requirements to work correctly.

I haven't settled on how yet, but I may just end up moving the factory coolant temp sender to the oil pan. It won't have any issue being submersed in oil rather than coolant, and the temp range is very similar. I'll get to that one day, and will probably implement it using something similar to this:

Just off the top of my head, the idea of adding a coolant temp sensor in the expansion tank line does not sound like a good solution. I don't think it will be anywhere near accurate. Maybe add another bung onto the T-stat housing for the extra sender? Or place it in the larger coolant hose feeding from the engine with a inline adaptor like this?
View attachment 68420

I do like those in-line bungs, at least for convenience sake. Of all the in/out ports on the thermostat housing, the expansion tank one is the only one which has both constant flow, and is upstream from any cold coolant which is introduced by the thermostat. I'd put it as close as possible to the housing, and additionally, if accuracy is a concern I can put the water pump sensor in the factory FI temp sensor location, and put the FI temp sensor in the expansion tank hose.

Accuracy doesn't matter much for the FI temp sensor. Once it reads above about 60c, it does very little as far as I am aware. It's primarily used for warm up enrichment.

It'd be inline with a hose in position 1 on the housing, which should be pure outlet temp minus a couple degrees potentially due to heat loss.

 

[Jonohhh]

Most of the headache is from me not yet wanting to move the factory temp sender to oil temp for a while. It'd be hard to compare before/after if I change where and how I'm reading coolant temp, and I am curious to see the difference, all else equal.

 

[Ulix]

If you want to use the factory gauge, I think you may have to use the factory sender.
Veglia is completely different from VDO for instance.
At least for oil pressure, this is the case.
A VDO oil pressure sender will be of different resistance and make you gauge read backwards (it shows high when pressure is low and vice versa).
I haven‘t tried temperature.

 

[Jonohhh]

If you want to use the factory gauge, I think you may have to use the factory sender.
Veglia is completely different from VDO for instance.
At least for oil pressure, this is the case.
A VDO oil pressure sender will be of different resistance and make you gauge read backwards (it shows high when pressure is low and vice versa).
I haven‘t tried temperature.

I'm not very hopeful either, the sender is definitely not Bosch (didn't realize Veglia also did the sensing side of the Instrumentation),

The coolant temp sender is NTC (decreasing resistance with increasing temperature), like the Bosch, of course, but the chances of it being equivalent are pretty slim like you say, so it'll probably end up being the factory gauge sender stuck inside the oil system somewhere

 

[Russe11]

Most of the headache is from me not yet wanting to move the factory temp sender to oil temp for a while. It'd be hard to compare before/after if I change where and how I'm reading coolant temp, and I am curious to see the difference, all else equal.

To get a before/after comparison, install them in both locations, and run them to a double throw switch. You’d see any difference on the gage by toggling back and forth.
Brian

 

[Jonohhh]

To get a before/after comparison, install them in both locations, and run them to a double throw switch. You’d see any difference on the gage by toggling back and forth.
Brian

I have absolutely no idea why I didn't think of this, good call. If I end up doing something involved for the temp sensor I may even consider leaving it that way.

Thanks!

 

[kmead]

Personally if I wanted to know the temp coming from the head I would use a spacer between the head and the thermostat housing and put the sensor between the thermostat housing and the head. This will give you unadulterated coolant temperature at the source.

In regard to the heater system, I would T into the suction side of the hose leading to the pump, this would have the system function in the standard way (supply at the end of the head, return to the pump). This is a fundamentally simple system, why add further complexity if it isn’t absolutely needed?

 

[Jonohhh]

Personally if I wanted to know the temp coming from the head I would use a spacer between the head and the thermostat housing and put the sensor between the thermostat housing and the head. This will give you unadulterated coolant temperature at the source.

Thanks for the input as always, I hope you're doing well.

This and the simple hose tee are both on the table at the moment. I will likely do the latter until I get a water pump blockoff plate finalized in CAD, and have a machine shop do both of them at the same time.

In regard to the heater system, I would T into the suction side of the hose leading to the pump, this would have the system function in the standard way (supply at the end of the head, return to the pump). This is a fundamentally simple system, why add further complexity if it isn’t absolutely needed?

The water pump inlet is back to being a hose only, the geometry needed to get everything to work there has actually become a major pain in the a**. So, I'm pretty limited on the suction side, hence routing it back through the expansion tanks extra inlet barb. If I can get a better inlet configuration that enables a barb there without needing a hose splice w/ barb, I will absolutely do that.

And why add complexity? That's what I do! But really, this car is my outlet for little things like this- things you could do but likely shouldn't- so I don't mind. The electric heater valve possibility I mentioned will likely end up being the "learn to code for an embedded application" project that I initially hoped this would be...something along the lines of temperature targeting HVAC. That's very far down the road though, not a high priority.

 

[Jonohhh]

After another semester of school, I'm back home, and making progress again.

The pump has been relocated to where the original coolant elbow went, which works flawlessly now. Previously, the pump would get air locked, not only during initial filling, but also after the car sits for a while.

Now, it just works immediately. Super happy about that.


That leaves a few more small electronic things, but otherwise the water part of this project is done. Next comes air.

 

[stayros32001]

May I ask, why you kept the half of the water pump on the engine block, instead of making a plate that will connect the pipe to the engine block directly?
What are the advantages by doing this?
I am asking because I am thinking to make a conversion like yours.

 

[Dr.Jeff]

May I ask, why you kept the half of the water pump on the engine block, instead of making a plate that will connect the pipe to the engine block directly?
What are the advantages by doing this?
I am asking because I am thinking to make a conversion like yours.

Just my opinion, but I think some of it might depend on whether or not you intend to retain the stock thermostat function vs utilizing a electronically controlled T-stat to control the electric pump. If the stock T-stat arrangement is to be retained, then keeping the stock water pump housing as well as the "water return pipe" that connects them is a far easier approach. Why would you want to retain the stock T-stat? At least for the X1/9, given the overall coolant system layout, with two large hoses connecting to the two metal pipes running to the front mounted radiator, and the internal "bypass" T-stat design between them, it avoids having to redesign those components. In other words it's a much easier task to keep it. Kind of a 'all or none' situation I suppose; change a whole bunch of stuff, or keep it simple and just change the water pump. Personally I would change everything, but that's just me.

I think someone else took a slightly different approach to converting the water pump to a electric one. They used a different style E-pump and they replaced the entire stock one with a block-off plate as you suggest. So it is one viable option.

 

[Alex(Tenerife)]

I think someone else took a slightly different approach to converting the water pump to a electric one. They used a different style E-pump and they replaced the entire stock one with a block-off plate as you suggest. So it is one viable option.

This is exactly what I did. I use the craig davies water pump, which is smaller and a bit cheaper than the BMW model. I still have the stock thermostat, and I have the pump between the thermostat housing and the new pipe that I had made that replaced the mechanical water pump.

I run my pump permanently, while the ignition is on. I did try running it at different speeds using a pulse width modulator, to be honest slowing down the pump only maintained or increased the temperature. So in the end I removed it.

Great thing with this set up, once I threw a fan belt and the car continued to run without overheating. Also I find that on a hot day when I stop in traffic the engine temp goes down and not up like it used to.