Rotary tank - how to build?

Oops - I mean hemisphere. Where did the "s" go...

I built a setup like this years ago, we used a large custom built rectangular tank, and there were various types of paddles run by a Bodine variable speed motor, which oscillated vertically up and down.

The fish-flake solution (that what the visualization fluid is actually made of, BTW) was used to watch the flow dynamics for various paddle shapes and speeds.

Later they fitted a laser doppler velocimeter to the setup to get three-axis velocity profiles.

Jim

================================================== please reply to: JRR(zero) at yktvmv (dot) vnet (dot) ibm (dot) com ==================================================

have gotten a

tank could be

can be mounted

could work

more expensive.

and tank shaft

without much

but I need

particular fluid

Sounds like another application for the ubiquitous DC treadmill motor and controller. With the one I've got, I'd definitely gear it down or use a jackshaft to reduce the speed, top speed is around 6000 rpm. See

Stan

How would a metal wok do for this? They are relatively cheap, readily available, and the right shape. Tack weld a shaft to the bottom, and spin it to your heart's content.

Michael

could have gotten a

The tank could be

that can be mounted

leftover I could work

would be the ~

with an additional

bottom that would

the more expensive.

water. If it's

speed is around 200

available with

shaft and tank shaft

about 220 rpm or

actuators, and if it

velocity without much

ideal but I need

can take the tank

particular fluid

Must be a hemisphere with height = dia/2. A stainless steel mixing bowl has the correct shape (except small flat on the bottom), but trying to spin it by hand while floating in water makes it difficult to fine tune. If the spin is off center, especially when trying to reverse rotation, the bowl wobbles and there's no control. Only one attempt was successful in producing an especially interesting pattern in the bowl consisting of water and a measured amount of baking soda . So I'm trying to figure a way to gain some control in the process, and it needs to be precise. If it were possible to obtain a direct drive DC servo with a bearing for direct mounting of the tank, capable of fine resolution over a

Make yourself a wooden mould for the bowl shape, a wooden ring and make the bowl in fibre-glass. The wooden ring would form the basis for the shaft attachment. Keep application of the fibre-glass and resin even around the shape. A rotating flat table (with suitable clamps) and a sander held at a fixed point will assist in making the outside smooth (the inner would be smooth from the mould form).

On the drive side, a stepper motor (at least 100 steps per rev) with worm drive gearing to the base gear wheel would provide a quite smooth rotation which you have total control over. The drive electronics would be quite simple.

There are also salad bowls which are round bottomed. I would not have tried to do precise control with the wok by just spinning it by hand. Tack weld a rod to the center of the bottom. Put some bearings on the rod so it spins around the rod axis.

while floating in

consisting of water

gain some control

resolution over a

accuracy isn't

controllers.

Perhaps you could use a servo motor from an old 51/4" floppy drive. These are build on an aluminium frame and have good bearing. There is also a hub to mount the bowl to. Some old turn tables used servo motors also.

could have gotten a

The tank could be

that can be mounted

leftover I could work

more expensive.

shaft and tank shaft

about 220 rpm or

velocity without much

ideal but I need

particular fluid

Your idea sounds like a good one except I'm not sure I'll make it with little runout. I made some spherical speaker enclosures using a rounter mounted on a turntable setup, but don't have that anymore. Though it may cost some, a metal spinner could make the tank. They'll make it more precisely in terms of TIR than anything I could do.

The stepping motor may work but I'm concerned with torque ripple transmitted to the tank contents, although less thru a gear reduction. It's unfortunate that servodisc motors are so expensive since they have no cogging and are brushed dc motors. Well, it's just like anything - it'll take some experimenting to see what effects things like cogging have on the contents. Perhaps it won't be a problem due to its mass, just like platters used for record turntables. I'm just a bit paranoid about building this contraption, because it's then a lot of trial and error for a certain pattern. The only other possibility is using say, three small motors with rubber tires around the circumfrence to turn a mixing bowl floating in a larger bowl of water. Just thinking out loud....

smooth

Now that's a thought. I do have a 5-1/4 floppy drive. Will it have enough torque to handle what I described? There's a lot of reversing going on.

The fibreglass bowl was an idea to keep it light but strong. Naturally some thought needs to go into the design details. However, if you know a metal spinner who could do your task to your satisfaction then that is an option to explore.

Having done quite a bit of work with stepper motors over the past couple of years (very special application where we were concerned with the potential for forming striations within an extrusion process) we found that having a large number of steps per revolution of the stepper motor means that you have already minimised the ripple factor. Our motor drives a linear translator which advances 5mm for every rev. Our current motor is 400 steps per rev and the mechanical output in the extrusion process is very smooth (couldn't see the steps in the extrusion force measurement at close magnification).

You can only try. I noticed that someone mentioned disc drives as a source of stepper motors. Old hard drives are aslo quite good if the use stepper motors in them for head movement or as the main drive. I know I have obtained a couple of quite good motors that way which had 160 steps per rev at reasonable torque.

Synchronisation might be your problem with this last idea.

You might be able to obtain an assembly with allowable run out by welding an over sized shaft as accuratey as possible to the bottom of your bowl. The end of the shaft to be welded to the bowl should be milled square, so the bowl will have best chance of being perpendicular to axis of the shaft. To eliminate any offset between shaft and bowl axes, place the weldment in a 4 jaw chuck on a lathe. Adjust 4 jaw chuck till you reduce runout at bowl edge to an acceptable level. You can now turn the shaft down to your bearing size and the axis of the bowl should coincide with the new shaft axis.

As an alternative you might design a coupling element to use between bowl and shaft that can adjust offset and tilt. It could be as simple as a long oversized collar with 8 set screws. Kind of like a Chrismas tree stand.

could have gotten a

Don't sweat the small stuff.. For example, you had mentioned: i) The small flat in the bottom of a mixing bowl. This would have very little impact on your results. Get on with your stuff and don't worry about perfection until everything else is working perfectly. ii) Back lash in the speed reducer. That will have no effect at all except to introduce a very tiny time delay in your direction reversals. Again, first make it work; then perfect it.

I did some very primitive experiments along these line using a phonographic turntable. It worked fine. Try borrowing an AC signal generator and driving a turntable with one of these. The variable frequency together with the 78/33/45 gearing would give a pretty good range. 16 RPM versions also exist. I would not recommend welding to the bottom of the bowl. I suspect you will waste a lot of time trying to adjust it and it will never be quite good enough.

I'd really love to see some of the pictures.

Walter.

I may try a stepper, but already ordered a small surplus Servodisc motor. It's shaft will take a 7.5 lb axial load, and the tank (decided on 5" dia) will weigh under 2 lbs, so it won't wear out any time soon. The Servodisc has zero cogging, and through the point of reversal the torque would be smooth.

That would be a problem.

Ben

I'm hoping, that at a reasonable cost, a metal spinner could make the tank with a 3/8" shaft at the bottom. Then have an adapter made to fit the motor shaft. Since I already ordered a motor, the adapter will have to fit an 8mm dia. shaft. Just in case the motor doesn't work out, I'll just have to try another. Appreciate your suggestions. I'm trying to get this thing together as simply as possible. I was going to use an additional bearing and flexible coupler between the tank and motor in fear the motor bearings wouldn't support it, however I later learned that the particular motor I ordered will take up to a 7.5 lb axial load, and a 5" dia hemi filled with water will be well under that.

Ben

You don't know how much time I've wasted thinking about results before getting them. :-) The motor won't need a speed reducer since it's a Kollmorgen Servodisc (found surplus recently), which has zero cogging and can turn as slow as 1 rpm, perhaps even slower. The speeds in the experiment would range between 0 to 150 rpm or a little faster, but certainly slower than centrifugated water all over the place.

it > will never be quite good enough.

No welding. A metal spinner should be able to provide a shaft at the bottom of the hemisphere for mounting. It will be direct coupled to the motor shaft with an adapter.

Well, if anything interesting happens there'll be movies.

Ben