Original poster: "Scott Hanson" <huil888@surfside.net>

Rob -

The ultimate solution is to balance all the rotating components as a set, with the angular position ("indexing") of the RSG disk locked to the motor armature (or rotor) via a hub with a keyway, etc. Almost all motor armatures are balanced at the factory to a certain tolerance; i.e. there is always some remaining uncorrected imbalance. Likewise, an RSG rotor assembly that has been "balanced" by itself will have some residual uncorrected imbalance.

When you assemble the RSG rotor to the motor armature, the remaining imbalance values can either be additive, resulting in greater vibration, or subtractive, with less vibration.

A very good static balancing support can be assembled in an hour or so using a set of four small instrument ball bearings. For best results, the bearings should be of the flanged type (flanged outer races), as this will minimize the area in contact between the motor shaft and the bearings. Build a simple structure to hold the bearings in pairs to create a "rolling vee-block", and far enough apart to ride on the armature shaft just outboard of the ball bearings (or the portion of the shaft where the bearings seat). Make sure to meticulously wash the bearings in solvent to remove any grease or oil that would damp free rotation.

Next, you will need to decide if you want to correct imbalance by ADDING mass, or by SUBTRACTING mass. Both methods are used extensively in commercial balancing. Removing mass can be done by drilling, filing, grinding, etc. However, it is permanent, and you better be sure you know exactly where you need to remove mass before you attack your RSG rotor with a cutting tool. Adding mass is a bit more forgiving, especially if you have designed-in a series of small radial tapped holes all around the OD of the rotor to allow the addition of correction mass by using small set screws.

Assuming that you have done all of the above, remove the motor armature from the motor housing, set the motor armature (with RSG rotor attached) in the rolling vee blocks, give it a slight nudge, and allow the assembly to coast to a stop. If you are going to correct imbalance by adding mass, mark the exact top dead center of the rotor ("light spot") with a pencil when it comes to a rest. Repeat this several times to ensure you have accurately located the light spot of the assembly. If you are going to remove mass, mark the "heavy spot" at the bottom of the rotor.

Now that you have found the light (or heavy) spot, you will need to correct the imbalance. If you are adding mass, install a set screw in the hole nearest your "light" mark. Repeat the rotate & coast-to-a-stop process. If the light spot still ends up on top, use a longer (e.g. heavier) set screw.

If you have installed the heaviest set screw you have and its still too light, add another set screw in the nearest adjacent hole, or better add two more set screws, one at each of the holes on either side of your first correction mass. To provide some range of coarse adjustment you will need to have a selection of set screws of masses (different lengths). Also, adjusting the depth that the set screw is installed into the rotor gives you the ability to "fine-tune" the balance. The closer the mass is to the OD of the rotor, the more effect it will have.

Repeat this process until the rotor stops at completely random positions each time it comes to a stop. Permanently mark the relationship of the RSG rotor to the hub to ensure that it can be reassembled correctly if it is ever disassembled.

Without an electronic balancing system, this is a pure trial-and-error process. As you approach (near) perfect balance, you may also have to make subtle adjustments to the angular position of the correction masses to achieve the best possible balance.

Obviously, if you are correcting imbalance by subtracting mass, you will perform this exact same process, but remove material from the heavy spot.

For very small (lightweight) rotor assemblies, you can also throw together a "knife-edge" support system using a pair of brand new box-knife blades hot-melt or cyanoacrylate-bonded to a piece of plate glass or other dead-flat, hard smooth surface. Some additional support on one side of the

blades is necessary so they don't just fall over under the weight of the rotor. You can use hot-melt glue or cyanoacrylate instant glue to kludge together small pieces of 1/2" aluminum square stock, etc to keep the blades absolutely vertical and the edges parallel to the glass and so they can support the weight of the rotor. Use a level and paper shims to get the plate glass completely level on the work surface so the rotor won't roll off. Use extreme care to VERY GENTLY set the rotor shaft onto the knife edges. The slightest impact will create a dent in the sharp edge of the blade and the shaft will no longer roll freely.

Now, this entire process yields a rotor assembly that is STATICALLY balanced. This assumes that nothing deforms significantly or changes location at the actual RSG operating speed. This assumption is probably valid for rotors made from 1/2" thick epoxy-glass laminate ("G-10", etc) or 1/2" thick linen-reinforced phenolic. Any rotor made from 1/4" or 3/8" thick polycarbonate or similar unreinforced polymer material WILL distort and change its balance characteristics (and is not a sound design from an engineering standpoint anyway).

Also, remember that as the electrodes erode, or are reground or replaced, the balance will change. This is best controlled by very carefully matching the weight of the electrodes initially, and keeping the set matched in weight as they are reground or replaced. What you really want to do is perfectly balance the rotor, and then keep the electrodes matched in weight, even if the absolute weight of the electrodes changes over time.

Commercial balancing systems, for car tires, peanut-sized 100,000 RPM dental turbines, or 100,000 lb powerplant generator rotors all work on the same principal. Basically, the rotor is supported on a "table" that is suspended so it can move very slightly in response to the imbalance of the rotating mass. The table is equipped with two very sensitive displacement transducers that convert the table motion into an electrical signal. Finally, there is some sort of angular position transducer that provides an electrical signal for rotor angular position.

The item to be balanced is rotated (it can be at very low speed, just enough so the imbalance causes some table motion). The output from the three transducers is processed by a computer to yield an imbalance force magnitude, plus an imbalance force angular position. The balancing system will tell you the angular position at which the balance correction(s) must be made, and the magnitude of the correction mass to be added or subtracted.

Regards,

Scott Hanson

----- Original Message -----

From: "Tesla list" <tesla@pupman.com>

To: <tesla@pupman.com>

Sent: Thursday, December 16, 2004 4:05 PM

Subject: RSG unbalance

> Original poster: Rob Maas <robm@nikhef.nl>

> Hi All,

> I am in the process of building an RSG. The disk spins at

> 2950 rpm (the motor is asynchronous, mains frequency is 50 Hz).

> The first trial run indicates a slight unbalance. It is

> probably OK, but if possible I like to improve the balance of

> the disk. Is there a systematic method to do this 'on line',

> i.e. with the disk mounted on the motor shaft?

>

> In a garage they use equipment to balance wheels: after a short

> spin the apparatus spits out what balance weight should be

> placed where on the rim to balance the wheel: anybody knows what

> the principle of operation is here?

>

> thanks in advance, Rob