View Full Version : Si3N4 ball bearings

09-18-2003, 03:25 AM
My balls have flat spots - on my Flexbar Ball Bearing V-Blocks. This is normal - it happens over time. Replacing them is becoming annoying - they don't make balls like they used to!

I ordered some Cerbec (Saint-Gobain Advanced Ceramics) Silicon Nitride balls (Si3N4) to replace the Chrome Steel balls. They are Grade 5 balls and exhibit better properties than the steel balls. One thing I have noticed is they are extremely slippery compared to Chrome Steel. You can feel the friction differences by rolling the 15/16" test ball for the blocks on the regular BB's vs. the Cerbec BB's.

I am sold on these things.

I am working on a few projects with them - I want to try them on a set of ball centers. I just have to figure out how to join then to the centers. So far, my best guess is a blind hole drilled in the Cerbec with a diamond burr and then bonded to the steel center with a high temperature adhesive. I don't think they can be brazed on.

Any suggestions?

09-18-2003, 08:23 AM
hello again Dave.

might be out of my league here, but if you are talking about lathe centers, would a 60* countersink not work. center drill tailstock center and part. ball between them.

can't take credit for this as an original idea. saw something similar in a taper article i was reading.

how big are your balls? [loaded question there] i sure would like ot be able to afford some of those ball bearing v-blocks.


09-18-2003, 10:37 AM
How much of the ball are you going to use for a center? Couldn't you just bore a socket and then roll the lip over the ball. Sort of like a ball point pen. Maybe drill a little hole in back for lubrication.

09-18-2003, 08:17 PM
Just a thought, carbide teeth are brazed onto steel sawblades. Maybe those balls are too expensive to try brazing unless you know it works.

09-18-2003, 11:12 PM
I can tell you this much,I wouldn't let anybody braze my balls http://bbs.homeshopmachinist.net//biggrin.gif

Maybe would it work to insert the ball into the center rather than vice versa,you could bore a blind hole with a radius bottom slightly deeper than half the diameter and maybe a half a thou smaller then shrink and roll the outside egde of the cup to hold it in.

09-18-2003, 11:19 PM
wierd, see rustybolt's earlier post.

I don't think your suggestion is an optimal solution, there may not be enough of the ball protruding to give good bearing contact with the mating part.

But it's my opinion - circumstances are not general - it might work just fine, depending on the situation.

09-18-2003, 11:41 PM
Well,I was thinking of a say 3/4" ball with a 7/16"deep hole,maybe even roll the lip in while its hot?Coarse if hes making something like a tooling ball then no it won't work.At that point I would say get some trimetal and flux and go at it.Heck use HF to make it neat.

[This message has been edited by wierdscience (edited 09-18-2003).]

Bill Neufeld
09-19-2003, 01:28 AM
How about using an EDM to burn the hole to the right size and depth. Fit the rod and braze in an inert atmosphere using a nickel bearing braze? Induction heated perhaps.
Bill. http://bbs.homeshopmachinist.net//cool.gif

09-19-2003, 02:00 AM
Bill N
I thought of that, the problem is I do not know if you can even braze Silicon Nitride at all. There is a company that has a foil that welds ceramics/metals together - but at $150 for a 1" square my thoughts at the moment are a high temperature adhesive.

I can turn the end off a hardened center with little effort. I had planning on using some Jacobs mt#2 -jt33 shanks I have laying about, boring a depression and attaching the Cerbec ball to it.

Rusty, Randy

I thought of that, but for turning tapers the crimp would get in the way (sooner or later). I need 5/8ths. of the sphere free. The pen nib concept will work for my "other" project - thanks for the idea

WS, Bill R
I thought about a shallow cup to hold the ball by pressure against the tail end of the shaft. My concern with that is cutting pressures could pop the ball out under under power on steep tapers.

[This message has been edited by Thrud (edited 09-19-2003).]

09-19-2003, 02:26 AM
Could you use the end of a valve pushrod for this purpose? Seems like it would work, and they're easy to come by.

09-19-2003, 11:15 AM

Hi-temp adhesive sounds like the right choice. I don't think that silicon nitride can be brazed without first metalizing the surface to be joined.

Since the ball will be subjected to temperature excursions, the choice of the adhesive is interesting. The compound must be able to withstand temperature deviations and maintain the bond without fatiguing.

But the joint also must remain "flexible" enough to accommodate the relative movement between the silicon nitride and the metal to which it is attached because of differences in thermal expansion.

This is a common problem in the electronic industry and is the reason that "kovar" was developed. Kovar is a pretty good match for glass and for many ceramics (linear Tc).

I've also used 440 stainless for high temperature fixturing of metalized ceramics to be brazed. This would be an admirable choice (if the Tc fits silicon nitride) since it can be hardened.

09-19-2003, 11:30 AM
Brazing is out. Here are a few of the charateristics of Silicon Nitride. The last is the one that matters.

Good thermal shock resistance
Good high temperature strength
Creep resistance
Low density
High fracture toughness
High hardness and wear resistance
Electrical resistivity
Good chemical resistance
Good oxidation resistance and not wetted by molten metals

09-19-2003, 03:29 PM
Another thought, is there some acid or whatnot that can etch silicon nitride, mask an area and rough it up before adhesive bonding. Or arc the area to be bonded to rough it up.

09-19-2003, 03:53 PM

I did some poking around on Loctite's web site and found this: Hysol® E-214HP Epoxy Adhesive

The description:

LOCTITE® Hysol® Product E-214HP is a light paste, industrial grade epoxy adhesive. This one-component, no-mix, heat activated formulation develops tough, strong, structural bonds which provide excellent peel resistance and impact strength. When fully cured, the product offers superior thermal shock resistance, excellent mechanical and electrical resistant
properties, and withstands exposure to a wide variety of solvents and chemicals.

Bonds to a wide variety of materials, including metals, glass,
ceramics and plastics.

Try this link: http://www2.loctite.com/int_henkel/loctite_us/index.cfm?layout=6&productline=E214HP&disp_language=en

Bruce Griffing
09-19-2003, 04:47 PM
How about a variant on the ball point pen thing. Make a thread on cover that captivates the ball in a blind hole. That way you can remove it easily.

09-19-2003, 11:29 PM
Take a diamond wheel and grind a flat spot on that sucker and braze as normal with carbide.Works on carbide burrs?

[This message has been edited by wierdscience (edited 09-19-2003).]

Bruce Griffing
09-20-2003, 12:42 AM
As Evan points out, Silicon Nitride cannot be brazed. It is a wide bandgap insulator - similar in some ways to quartz. It is very hard and more or less impervious to diffusion. It is used as a diffusion barrier in semiconductor processing. People have experimented with ceramic turbines made with this stuff. I have heard stories of an experimental 200hp unit not much bigger than your hand. It operated at very high temperature.

09-20-2003, 12:56 AM
The ball may be roughed with a diamond wheel for gluing, it could be acid etched with flouric acid (dangerous stuff, deep flesh burns possible), or it could be blasted with CBN or diamond grit (expensive). It could also be roughed with an excimer laser. Roughing with a diamond wheel or burr would work fine and is cheap. The epoxy from Loctite sounds like it should work well. According to the spec sheet it actually becomes stronger with exposure to organic solvents such as alcohol and petroleum based solvents. Like, say, some cutting oils?

[This message has been edited by Evan (edited 09-20-2003).]

09-20-2003, 01:29 AM
I was reading the data sheet on Hysol and apparently it is used primarily for gluing the plastic to stainless steel needles (the tan colored goop). I will look in to it. Thanks for the idea.

I do know that I can get a special foil to weld them together (designed to bond heatsinks directly to the silcon chip (like the Power PC, AMD, or exposed Intel) but to get a sample is mondo expensive. You clamp the two pieces together and light the foil with a torch - instant bonding.

These balls are awesome - unbelievably low friction. It is one of the neat properties of them that makes them great bearing material. The coefficient of friction against steel is extremely low - ideal for high speed spindles and turbo chargers. The Cerbec spindle bearings use grease instead of spindle oil - even at extreme high speed. They have a surface finish of .15 Ra - far better than steel Jo Blocks!

BTW, these are made by Norton (Saint-Gobain Advanced Ceramics). I wonder if they will ever make Diamond BB's?

09-20-2003, 01:29 AM

It seems to me that it is silicon nitride that is used for blue and UV light LEDs. It's the high bandgap that makes it possible but it was a really tough item to manufacture. It was a rogue Japanese scientist that figured it out.

09-20-2003, 01:35 AM

The one I recommended is grey. They have about ten different Hysol types. The 214HP is very different from some of the others, being monocomponent and heat curing. You must be thinking of a different one as the 214HP cures at about 250 degrees. Not good for most plastics.

I looked at the website for that foil stuff a while ago. Verry interesting. Strange material.

[This message has been edited by Evan (edited 09-20-2003).]

Bruce Griffing
09-20-2003, 07:13 PM
As to blue LED's I think most of them are Silicon Carbide. Silicon Nitride has a wider bandgap and thus could in principle produce blue or even uv light, but I don't think it can be doped to produce conductive n and p type material. Could be wrong tho - maybe someone has figured it out.

09-20-2003, 07:41 PM
Now that you mention it Bruce I think you are right. Maybe it was blue diode lasers I read about?

09-21-2003, 10:56 PM
The blue diodes and the UV Lasers are both SiC as far as I understand it.

You see those new triple gate SOI 50nm transistors AMD came up with - nickle gates, 2 picosecond rise time.

God I feel old, I remember when tubes where the HOT thing! (bad, BAD pun! http://bbs.homeshopmachinist.net//biggrin.gif)

Bruce Griffing
09-21-2003, 11:20 PM
If you haven't designed/built/fixed tube based electronics you haven't achieved manhood in the field of electronics. IMHO

09-22-2003, 01:43 AM
When I was in high school we built a five tube superhet radio as a project. The next year I was teaching assistant for the first year class and got to "bug" the projects for troubleshooting tests. I just loved breaking a wire inside the insulation. Little did I know that this is a common failure mode on equipment that has moving wire harnesses, such as copy machines.

Oh yeah, I still have a Hallicrafters S40A short wave receiver, all tubes, still works.

[This message has been edited by Evan (edited 09-22-2003).]

09-23-2003, 04:35 AM
I have been considering bulding a triode amp for my electrostatic speakers - I only need 5w to blow the glass out of the house. Then if I could just get two 2-100Hz subwoofers dirt cheap I could hook them up to the 400W/C monster amp - NIRVANA!

Tubes should couple well with the reactive load of the electrostatic panels transformers.

[This message has been edited by Thrud (edited 09-23-2003).]

09-23-2003, 10:59 AM
Thrud, others:

I have five VT amplifiers, ranging from 15 to 85 watts and a Hammarlund receiver, all of which require tube replacements from time to time.

You may know this already: a quick source of tubes and associated parts is partsexpress.com

09-24-2003, 04:44 AM
I thought the best audio tubes came from Russia (out of mig29's)? http://bbs.homeshopmachinist.net//biggrin.gif

09-24-2003, 12:04 PM

Entirely correct and from other former soviet-bloc countries, also China. Some of the brand names: "Sovtek" and "Svetlana". The "Tesla" brand is made in Czechoslovakia, oops I mean somewhere in that part of the world that used to be called Czechoslovakia.

Quality of these tubes is spotty, I usually buy several and expect to get one good one. I save the others for the occasion when I need a matched pair or matched quad.

BTW, there are lots of mom/pop operations on the web that sell used tubes. From the price lists, must be a lucrative business !