What is the current view on water cooling ??

[dazz]

Hi
I am on a path to build a Trident. I have been considering what mods to do.
I have looked at water cooling a Revo nozzle. I have the tools to do that.
I have found very little on water cooling published in the last year. I did a search on this forum before writing this and nothing.

So I am wondering has water cooling gone out of favour because it is a waste of time/effort, or
is it because making a heat exchanger requires tools that most people don't have or
some other reason.

 

[120decibell]

I am not an expert on that but have found it less popular because these is typically little to no gain with enclosure temperatures below 80-90C. Cooler air is easily able to cool the hotend with just air movement. Once the chamber temperature get hot enough ( > 80C or so) the air is no longer cold enough to sufficiently cool the hot end so the heat has to be moved outside the enclosure. Combined with the cost and complication of water cooling I believe that explains the lack of it.

 

[dazz]

OK I am not interested in heating the enclosure. I don't use any filament type that would need it. I am interested in surface finish of prints. I see vibration artifacts from my Prusa and printers I see on-line (YT). My background in electronics engineering includes control engineering. So I am looking at the potential for reducing the effects of vibration.

There are a number of techniques that can be applied. One is to choose the right type of machine. I think the Trident has greater potential for improvement than the 2.4. Another technique is to reduce the moving weight. I can make a water jacket that is lighter than air cooling and with no spinning moving parts. PC water cooling is now common enough that the equipment is not to expensive and the tech is proven.

I may look at enclosure warming purely to achieve consistent conditions and finish from print to print. That would mean controlling the enclosure temperature at a set point of, say, 30 deg C. I think that should be possible without adding heating. 3D printing is a process, and like all processes, consistent output needs consistent input.

So in asking my question, I was looking for something I might not have thought of.

 

[claudermilk]

Start with making sure everything is tight on your printer. Since almost all Vorons run Klipper, you can get an ADXL345 and run input shaper. That should virtually eliminate vibration artifacts (ghosting). My 250 Trident and V0.2 are far better with that than my little Prusa Mini. It's a large reason the Mini mostly gathers dust now.

 

[dazz]

Start with making sure everything is tight on your printer. Since almost all Vorons run Klipper, you can get an ADXL345 and run input shaper. That should virtually eliminate vibration artifacts (ghosting). My 250 Trident and V0.2 are far better with that than my little Prusa Mini. It's a large reason the Mini mostly gathers dust now.

My plan is to start with a standard Voron build, including clipper and the ADXL345.
Next step will be to identify the sources of vibration. One of the simplest ways to do this will be to add small temporary test weights to suspect points. The added weights will shift the frequency of vibration so this will allow correlation of the vibration source to the input shaper plots.

I will also measure hysteresis at key points to plot hysteresis curves. This will allow me to measure the flex of the printer and figure out where rigidity should be improved and damping added.

I already suspect that improving the mechanical connection between the side panels and frame will improve rigidity. This especially applies to the front doors. Using cam-lock type fasteners to clamp the front door to the frame to make the doors part of the structure. This should reduce unwanted motion, but what I do will be guided by the measurements.

Using input shaping is treating the symptoms. So my strategy is to treat the root cause of the vibrations before applying input shaping. The less work that input shaping has to do, the better the potential outcome.

I should add, I like tinkering with stuff.

 

[claudermilk]

I should add, I like tinkering with stuff.

Clearly. You have found the right hobby and community.

 

[dazz]

Clearly. You have found the right hobby and community.

I won't really feel part of it until I build a Voron. That could be a few months away.
When I do start tinkering, I will post my results.

 

[NoGuru]

I would say skip the water cooling unless you have a really hot chamber and doing extremely long prints.

But on the other hand you have the "cool" or "bling" factor that water cooling bring. I just think keep the complexity out of the mix at first.

 

[claudermilk]

I won't really feel part of it until I build a Voron. That could be a few months away.
When I do start tinkering, I will post my results.

You're just in the research-and-prep phase. We all know--like quicksand--once you're started in there's no going back. You will eventually have a Voron.

 

[dazz]

My son has purchased a house and will be moving out FINALLY!! Unfortunately he will be taking the Prusa with him. Fortunately that will give me the excuse to build a Voron.

I have a number of reasons for investigating the water cooling option. It is my assessment that the Voron is a well designed and engineered machine tool, so any possible improvement is going to be incremental. I am chasing ants looking for improvements in motion control and process consistency (chamber temp control).

It is my observation that most fans have axial play on the rotating part. If acceleration of the head causes the fan to move, it would be like having a tiny hammer tapping the head and that would not be a good thing. Is this a significant issue? I won't know until I take measurements on a Voron. If the problem is real, it won't directly show up on the input shaper plots because an impulse (tapping) seen in the frequency domain appears as broadband noise. I doubt the input shaper could directly detect or compensate for this effect.

An impulse acts as a unity function and will excite all existing resonance modes. An input shaper plot would show an increase in amplitude across the entire spectrum. The changes in amplitude of each vibration mode will depend on the Q of each resonance mode. So although the impulse energy will be evenly spread across the frequency spectrum, the amplitude response of each mode will vary. I expect the input shaper would detect and compensate for this but impulse energy will degrade the signal to noise ratio seen by the input shaper.

Note that I haven't read any technical info on the input shaper, so I am making huge assumptions about how it works based on my knowledge of control system engineering.

A correctly engineered water heat exchanger and tubing should reduce moving weight compared to a fan. That can only be a good thing for motion control. For the same reason, I will look at installing a remote part cooling fan. Nothing that hasn't been done before.

I have a workshop full of machine tools just waiting to make the parts I need. I can make custom cooling components at low cost. This sort of project lets me justify their existence to SWMBO.

I could say that I am in the pursuit of engineering excellence, but that would just be pretentious ego polishing. Really it's just for the cool bling and bragging rights. I advise young engineers, if you spend too much time picking up ants, you will get trampled by the elephants. I am too old to care about accusations of hypocrisy.

 

[shiftingtech]

I have seen high frequency spikes in input shaper measurements that seem to corolate to fans running. Now, that's a little different than what you're talking about, I think, but it shows the adxl is at least capable of detecting stuff related to the fans. (Doing anything about it is a different question, of course)

 

[dazz]

Hi
There are a few things that can be done to improve motion control relevant to 3D printers.
1. Remove any backlash (belt slippage, loose pulleys, slop)
2. Reduce hysteresis. (flex, stretch, bending).
3. Increase rigidity while reducing weight.
4. Add damping to reduce resonant peaks (Tuned Mass Damping)
5. Use input shaper.
6. Apply Active Tuned Mass Damping.