Digital Mist Coolant Pump

[JimDawson]

I have been having inconsistent flow problems with my pressure fogless spray coolant system. The flow control valve seems to get gummed up as it’s running, and you have to tweak the valve a bit to get it to flow again. I am running it at 60 PSI, have an in-line filter, running WD-40 as a fluid. I suspect that WD-40 has something in it that builds up in a small orifice that needs to be periodically cleared. I have tried several different types of valves with the same result, so time for a change. I would like to be able to walk away from the machine on long runs, but right now its not really advisable given the coolant issues.

So to eliminate the orifice I decided that a controllable flow, positive displacement, low volume pump is the answer. I had a FloJet carpet cleaner pump on the shelf so I started there. These are a dual diaphragm,1.8 GPM pump. Way too much for mist coolant, so it needs to be slowed down…… A lot! So I removed the original motor and installed a NEMA 23 stepper motor. That didn’t work very well. A few problems: The Viton parts inside don’t like WD-40 very much, that particular pump has had some very strong oxidizers run through it for extended periods so the valves leak a bit, and it’s obsolete so I can’t get the petroleum compatible buna-N seals and valves for it. So back on the shelf it goes, probably for another 20 years. FloJet is still in business, and have lots of different pumps, but they are primarily geared towards water and liquid soaps. They have only one pump with buna-N parts in it, and I am not convinced that it will pump properly when running very slow. I needed something more positive.

Given that I didn’t have anything on the shelf that I could cobble together, I started really thinking about what I actually needed to make the system work. I guessed at the needed flow rate, so I settled on 0.05 to 5 mL/min. The viscosity near that of water, pump components compatible with petroleum or water based liquids, adjustable flow over a wide range, and positive displacement.

Next I evaluated a number of different pump types and settled on a peristaltic pump. A peristaltic pump is a positive displacement pump and the flow rate is a function of the tubing size and the rotor speed. Fluid compatibility is a function of the tubing used. Good peristaltic pumps are generally expensive, especially the reasonably accurate lab grade pumps with controllers. Junk peristaltic pumps are available on FleaBay cheap, but you get what you pay for.

I also have another application for a very tightly controlled pump that must be very consistent. The pumps I am currently using for that application are about $6000 each, and will dispense as little as 0.005 mL. They are a tiny progressing cavity pump, coupled to a tiny servo gear motor. One of the problems with existing peristaltic pumps is that the output pulses, OK for most applications, including coolant pumps for a machine tool, but not for my dispensing application. For some time I have been thinking about how to build a non-pulsing peristaltic pump that would meet the needs of my dispensing application.

So killing two birds with one stone and in the hobby machinist tradition of ‘’why buy it when you can build it’’, it’s time to build a peristaltic pump that will run my coolant system, and use as a test bed for my dispensing application. If I can get the pump to work for the dispensing application, I can do a nice cost reduction on the assembly.

The first order of business is to draw up a basic design, then scrounge up the parts to build it. I needed a small NEMA 23 size gear box, found a 15:1 gear box with NEMA 23 stepper motor for $61.00 on FleaBay. I designed the housing, rotor, and controls around materials I had in stock. Then order the small bits & pieces from McMaster and others.

I also had to order another stepper controller because when trying to test the FloJet, I miss-wired the power plug on one of my stepper drives and fried it. Watch what you are doing, it can get expensive in a hurry. That one was my 4th axis drive. I have also purchased a few cheap stepper drives for non-critical applications, like pumps, just in case I do something stupid again.

So what is a digital pump? It will be controlled from the computer over a RS232 line (serial port), so I guess that makes it digital. The stepper drives I am using will operate in 3 different control modes, Pulse & Direction, Analog Voltage, and Serial Communication. All I need to do is turn it on and off, and control the speed, so serial comms will be perfect. That way it doesn’t tie up an axis on my motion controller. It could also be externally controlled in a couple of different ways for use on a manual machine.

Here is a drawing of Rev1 (with the secret anti-pulse stuff removed)



Here is the material, a piece of 2 inch 6061 round stock for the rotor, and a 2x2.75x3 6061 bar, squared up.



Laying out the band saw cuts, this is where I screwed up in the orientation.



Drilling and reaming the 0.125 radius. I didn’t have a 0.250 ball end mill that would reach that deep



After the band saw cut, the drop will be used for the front cover.



Cleaning up the back.



Pocketing the lower housing



The cover bolted down and ready for machining



A little clean up



Pocketing the inner shape in the cover



Cutting the path where the tube will sit. Normally I would not have the tool stick out this far, but this tool has had the shank cut down for another project.



Getting ready to clean up the top. I stuck this piece of shafting in there vice jaw would clear the mounting foot on the bottom of the housing



And drilling the holes in the mounting foot.



The completed housing



Next the rotor

Turned and bored, ready for cutoff. I don’t own a cutoff tool for my lathe, so over to the band saw



I turned the hub to fit the pocket I already had in my vice soft jaws, so all I had to do was drop it in place. I drilled & tapped the 8-32 screw holes……



….then used it as a fixture to hold the cover plate



The rotor assembly after machining for bearing clearance.



This is REV2 of the rotor, I added a nose to support the cover plate and take some of the load off of the shoulder screws



The assembled rotor. 3/16 x 5/16 shoulder screws. I may use dowel pins in the next version.



The finished assembly mounted on the motor.



………..More to come!
.
.

 

[JimDawson]

Rev1 didn’t work out so well, it had several problems: The mounting foot was in the wrong location, couldn’t be mounted on a flat surface, would not allow the tubes to exit. I cut the material in the wrong orientation so had no room for the cover bolts to be installed vertically. And the biggest problem, I cut the internal tube path too deep, so the rollers did not pinch off the tube inside and the pump didn’t work at all. And with the cap screws installed in the face, there was no way to adjust the height of the cover. There were a few other minor issued that needed to be addressed also. Machinists don’t make mistakes; we make prototypes.


So back to the drawing board!



The machining operations are about the same as in REV1 so I’m not going to bore you with too much repetition. This time I did not trim the material on the band saw, I just left it to hang on to while machining since I was not intergating a mounting foot. I’ll make a plastic mount that grabs the gear box housing, that way I have a range of mounting options.

The finished housing. Changes: Radiused entrance and exit for the tube, and the depth of the entrance and exit aligns the tube with the internal tube path.



This time I made sure the cap screws would install in the correct orientation. I counter bored so the heads of the cap screws would be below the cut in subsequent machining operations.



Profiling the outside. That’s a lot of stickout on a 0.250 end mill, but I had to go 1.135 deep. Slow and careful!



Corner rounding the face. Harbor Freight carbide, corner rounding, router bit with the bearing removed.



Cutting the internal tube path



All of the machining is done on the front, so time to remove the excess material from the back.



The finished housing



And mounted on the motor



And another view



I didn’t get a chance to test it yet. I ran out of time and had to go enjoy Thanksgiving dinner with my son and family.

I know there are a couple of things I need to address before I can test it, I’ll get that done today. I’ll post the results and pictures of the complete mist coolant system later today.

I’m going to redesign this so it can be simply built on a manual machine and I’ll post the drawings in case anybody wants to build one. It could easily be powered by a small DC gear motor, maybe like a windshield wiper motor with a cheap PWM controller.
.
.

 

[JimDawson]

I got the finish machining done, assembled it, and fired it off. After some fitting of the cover and installing a 1/4 inch OD tube, up from 6mm, it pumps fine. I really think the bearings in the rotor are too narrow, not getting enough of the tube, going to order some 0.312 wide bearings, currently has 0.197 wide bearings. I still haven't had time to experiment with it too much, but I can turn it down to almost zero, and can pump almost a flood on the top end, with great control through the range.

But I had another couple of problems. First I ran out of WD-40 in the tank so in to town a pick up a couple of gallons. The next thing I found was the flow from the tank was almost nonexistent. ?????? I back blew the line out with shop air and tried again, it flowed great at 10 PSI on the tank.....for about 15 seconds then dwindled to nothing. What the He!!? I put in a new in-line filter about a week ago, like this one

The filter was clean, only a couple very small pieces of crud in it. It turns out that something in WD-40 plugs off the screen in these filters. Wax maybe? You blow it out and it works fine for a few seconds. Anyway, I took the screen out and it works fine, so I'll be replacing the screen with some 100 mesh screen I have on the shelf and see how that works. If that doesn't work, I'll try some screen door screen. There are no valves in the system now, so all I really need to to is keep the big stuff out.

The bottom line is that it works, but needs some fine tuning.

More later.....
.
.

 

[John Hasler]

An interesting trick with peristaltic pumps is to use wide rollers and route the tube in a helix. That way you can have a full turn rather than just half a turn or even multiple turns. I used to design controllers for peristaltic blood pumps for heart-lung machines and dialysis systems when I worked at Sarns Inc. The rollers in the dialysis pumps, which always used the same tubing, were spring-loaded. The rollers in the heart-lung machine pumps were adjustable.

I actually have some small precision peristaltic pumps on hand left over from an unsuccessful product development that might be able to reach that output level. They have low voltage DC motors on them but bulding a controller is trivial. A source of similar pumps is fuel transfer pumps for model aircraft.

 

[wrmiller]

Hey Jim,

I noticed some flakes of 'wax' or something floating in the WD-40 I put in my FogBuster. I've been trying this A-9 aluminum cutting fluid of late and it works well enough that I don't know if I'll go back to the WD-40.

 

[coolidge]

Its CNC threads like this that get me drooling over the Haas mini mill.

 

[JimDawson]

An interesting trick with peristaltic pumps is to use wide rollers and route the tube in a helix. That way you can have a full turn rather than just half a turn or even multiple turns. I used to design controllers for peristaltic blood pumps for heart-lung machines and dialysis systems when I worked at Sarns Inc. The rollers in the dialysis pumps, which always used the same tubing, were spring-loaded. The rollers in the heart-lung machine pumps were adjustable.

I actually have some small precision peristaltic pumps on hand left over from an unsuccessful product development that might be able to reach that output level. They have low voltage DC motors on them but bulding a controller is trivial. A source of similar pumps is fuel transfer pumps for model aircraft.

I like the helix idea, would be much more positive than a half turn. Would be perfect for my dispensing application!

Hey Jim,

I noticed some flakes of 'wax' or something floating in the WD-40 I put in my FogBuster. I've been trying this A-9 aluminum cutting fluid of late and it works well enough that I don't know if I'll go back to the WD-40.

I might just give that a try, $42/gal for A-9 vs. $21/gal for WD-40, not that much price difference given I might be able to use less of it. I have used A-9 for tapping, but have never tried it for cutting. Does it run OK in a mist system?

Its CNC threads like this that get me drooling over the Haas mini mill.

Go for it! You need a CNC. Besides you can still walk through you shop in a straight line, that's just not right.
.
.

 

[wrmiller]

Can't speak to using A-9 in a mist system as the FogBuster spits droplets, but I doubt there would be a difference. I noticed that I need less of the A-9 during cutting than the WD-40.

 

[RJSakowski]

I'm not sure what the helix design gets you other than orienting the exit tubing in the same direction as the entrance tubing. There will be a sliding friction component as the tubing slithers from one side to the other which kind of defeats the purpose of the rollers. The flow rate delivered depends upon the internal diameter of the the tubing , the rotational speed, and to a certain extent the spacing between the rollers. Carrying the packets around for an extra half revolution should not improve anything. Peristaltic pumps being constant displacement beasts, more or less are pressure limited by the tubing. Some of the pumps we used were capable of pressures exceeding 100 psi but this was largely a function of the tubing used.

Bob

 

[JimDawson]

Can't speak to using A-9 in a mist system as the FogBuster spits droplets, but I doubt there would be a difference. I noticed that I need less of the A-9 during cutting than the WD-40.

Thanks Bill. My coolant system also spits droplets much like a FogBuster, so it should work well. Later today I'll post some pictures and a complete description of how it works.