Advice on my watercooling setup

Hey guys, I'm looking for advice or suggestions about watercooling my Delta printer. I'm using it to experimentwith very, very high speed printing. It's going to have a heated chamber with a max most likely around 90C.

At these temps, my fan on E3DV6 is insufficient, so I need to either to resort to watercooling or having cool air funneled in. I've decided to go with watercooling.


I'm not sure how much watercooling I'll need. I'm going to drive my stepper drivers at higher currents (2A, can I push it further? Or diminishing returns?) to get the most out of my NEMA17s to get as high as accelerations as I can so I will be watercooling an E3DV6 hotend, 3 NEMA17s, the smoothieboard stepper drivers, and at least 2 Extruders inside of the chamber (Floating Extruder setup).




I haven't convinced myself on what kind of water pump i'll use. I have thought about the Kraken hotend pump


Filastruder Waterpump



,but it's only 240L/hr and was not designed with a heated chamber in mind with multiple components.


The rest of components will draw from PC watercooling components, a waterpump/reservoir combo, and generic waterblocks (aluminum) and an RC waterjacket for the E3DV6.


I'm trying to keep all the parts in the loop aluminum due to galvanic corrosion if I introduce other metals. I'll be using a biocide to prevent algae growth which I haven't seen anyone who has done watercooling for 3d printing mention, distilled water, and most likely an anti-corrosive mix just in case since I may not be able to find an aluminum radiator block for cheap.


Now about the radiator block, this is the most important of the loop because it controls how much heat transfer is possible. In the PC world, it is recommended to have 120mm for each component (CPU, GPU, etc). Now, our 3D printer motors don't usually get to that temperature, but with such a high temperature chamber on the extruders, it may be necessary to use a 240mm radiator (I don't mind, the delta printer will be quite large, relatively). I just want to avoid so much cooling that the hotend can't do its job properly (which is my largest worry).




Any tips or critiques are very welcomed. Thanks and Happy Holidays.

I think that guessing at components isn't going to work for this one. You need to do some arithmetic with Watts, Joules, heat dissipation and work out how much heat you want/can get shut of then buy the appropriate components.

I can't work out what you want to water cool or why you want to water cool it to begin with so I can't offer you any advice on it.

Makes a lot more sense to me to put either Nema23 or Nema 34 steppers on it and power them with some reasonable current using TB6560 or similar than to bake the Nema 17. It isn't just heat that kills things, they are designed to work within certain constraints and once you start going outside that then mechanical/electrical failure will occur.

There is a huge difference between quality steppers and the ones that these machines tend to have. I good set of German/US/UK made steppers will work so much better than the ones form China but they will easily cost you ten times as much.

Horses for courses as they say.

I want to watercool my hot end and my extruders and the NEMA17s are operating within spec, they are rated for 2A, but they are going to be subjected to a high heat chamber (90C).

A fan on my hotend blowing ambient air (90C) at the heatsink isn't going to cool it down. Mine are 86oz/in steppers.

most 3x120mm radiators are rated to dissipate over 600 Watts. Thats beyond what I need, but I do have questions on how much cooling ill need to cool a stepper running at 2A which will naturally run hot + a 90C chamber.

Do you understand why i want to watercool and what I want to watercool now?

My Wanhao works just fine with a heated chamber. I normally use it around 70 degrees as that seems to be hot enough to prevent warping without overheating anything inside. I can't see why you would ever need 90 degrees in there really.

My Stratasys also runs the heated chamber at 70 degrees.

Neither of those have anything other than forced air cooling.

Stratasys recommends a chamber temp of 70-90, I put 90 as an upper limit to what I'd be hitting. Does your Wanhao use an E3DV6? I don't believe that my stepper motors can handle that kind of heat with the kind of aggressive acceleration settings that I will be using on top of that. I'm honestly surprised to hear that you can run your printer with a 70C heated chamber and not have to cool your stepper motors at all. You don't get skipped steps? http://forum.e3d-online.com/viewtopi...7adadcedc8181d

This guy's steppers shoot up to 120C without watercooling in his heated chamber so something isn't adding up. Also doesn't Stratasys use a massive air duct to force air in and out?

The Stratasys chamber is sealed. It used one four inch fan to blow air over the back of the extrusion head, 4X4 inch fans to circulate air and one four inch to blow air over the printed material on the SST, the BST doesn't have a fan to blow the print,. The temperature is 70c and is fixed under normal use. It is possible to change it but only by "hacking" the machine and I have never found a need to, 70c is optimum for ABS, I have tried lots of temperatures above and below but 70 performs best.

My Wanhao will print for ever without overheating or skipping steps. I say for ever but the longest print I have done on the Wanhao is around 22 hours.

My extruder is just the bog standard Replicator dual.

One thing to say, a delta printer lends itself to keep most of the electronics and motors out of the heated chamber. If your not using a bowden setup your most likely moving too much weight around to reach any reasonable speed, even if you have large motors, the larger mass your swinging around will either wear components to slack or induce flexing on the arms. The whole point of delta is to have a light mass and move it quick.

Keep the nema17's out of the chamber and attach heat sink to them. Keep the extruder out and do the same, insulate the rest of the parts and construct the heated chamber. I am working on something similar myself.

One more thing I wanted to say, I assume your going to be printing in ABS/Nylon, I currently print with no cooling on my printer in ABS. It simply does not need it, when going heated chamber I will be evaluating the need to cool the hot end, however I would steer away from water cooling. If you fatigue a tube to the point where it reaches breakage you will seriously cause issues. Most silicone tubing that would be used to water cool will be porous and will become a maintenance item, also these cheapo pumps are not designed to longevity most 3d printers have 100's of hours of work time, do you trust a ~15usd pump to last more than 100 hours? I used a pump that had a floating impeller with a jewel on the end when I was watercooling pc's. Those get expensive quick. Water cooling anything properly requires good quality components, and those are not cheap nor light.

If you insist on water cooling get a mcp355 pump and a single 120mm rad and matching fan. It will be plenty to cool the hot end as well as anything and everything you put on your printer. A single 120mm rad can dissipate 1600W of heat easily before it becomes heat soaked, not 600w, in pc's the fight is to get the parts closer to room temp, in your case the parts do not need to get there and the hotter the rads run the more efficient they get.

I am using a bowden setup, but the extruder is "floating" above it with a counterweight to keep all the weight off of the effector, but with a very small bowden tube to negate as much of the disadvantage as possible. The effector is still very light. The water jacket on the E3D is only 14g. It's made for RC boats so they are designed to be incredibly light and used for many hours.

I'm using an E3DV6, running without cooling across the heatsink is not recommended by the E3D guys. That water pump is the water pump that is a paired to watercool the Kraken Hotend. I was going to use something similar to the mcp355 or a D5 pump if watercooling.


I'll just go ahead and use fans, if they end up being inefficient I'll switch over to watercooling.

I should mention that the 3 motors to move the belts are not subjected to that kind of intense heat of the chamber, but I wanted to account for overtemp protection due to high acceleration settings and max current.

I thought 600 Watts is plenty accurate for a pc watercooling system, if you average 1 component per 120mm. With decent cooling and fans per inch. I'm trying to run them with low temps, I don't want my NEMA17s over 50C, but a single rad will be more than adequate.

When water cooling a pc users generally want the cpu/gpu/northbridge as close as possible to room temperature. Watercooling loops get more efficient with higher temperature differentials.

Let me put it this way, when watercooling a cpu we want to dissipate 100-200w of waste heat as fast as possible to bring said chip close to room temperature. The reason for this is that the chip itself once overclocked has a much lower tolerance for heat. My last watercooled pc was in the prescott days. The chip was rated at ~90w with a thermal shut off of around 75c. The chip got unstable when running at 160% of its rated speed much sooner than the thermal shut down. Therefore for me to keep the chip happy I had to be in the 30-40c range. For me to do this I had to have a very large surface area to dissipate the heat from said chip and other chips that were operating out of the design spec GPU and northbridge in my case. I used a 3x120mm radiator for it, the radiator had a temperature difference of roughly 1c idling and around 2c load from in to out verified with thermal probes. This allowed me to run my prescott under full load for weeks on end at roughly 35c with a room temp of around 27c....

What you are talking about is a much larger differential temp to meet, which a single 120mm radiator is more than happy to provide. If your hot end is in a 90c chamber you add the thermal emissions from the heater cartridge to the mix, you realistically are not dissipating more than 30w of heat on it, however even if you were your not trying to cool the hot end to room temp ~25c, your trying to cool it to workable temp ~ 40-50c, a radiator is designed to increase surface area, however that surface area becomes way more efficient if the thermal differential is in the 40c range than it does in the 5-10c range. Your steppers are designed to acomodate up to 80c, therefore you realistically only need to bring down the temperature by 20-30c for them to be happy.

I remember in the old days, we used heater cores for our water cooling systems, a typical heater core can put out more than 8000BTU of heat ~2500w of disipation. However for it to reach 2500w its designed to operate with 90c water.... You get my point? The hotter the source is, the larger the delta T the more efficient it becomes, bringing down to nearly room temps is the issue.

In either case, I would do a flex hose with a squirrel cage fan for both of them. Its simpler and less problematic, the last thing you want is water to hit a 100c bed. The only hoses designed for high flex are made of silicone, unfortunately they are porous and will loose water and add vapor to the chamber, not only that, they do tend to burst if you have a powerful enough pump and enough fatigue.

Here is a graph that shows this, notice the delta T and the watts dissipated. This is a 480mm rad, but the principle applies. Modern radiators for water cooling pc's are designed for low back pressure so you don't have to use a vacuum cleaner to keep it cool.
Attachment 4028