spindle design resources

[dagray3]

Hey, this is my first post.

I am involved in a project that requires the design and construction of a spindle for a friction stir welding tool. As we are designing the tool and building it ourselves, our specifications are sort of vague right now. I think the requirements of the project will evolve with our understanding of the materials, components and process.

That being said, I am trying to find some good resources for spindle design. Are there websites that are better than others? Are there books that any of you recommend?

If anyone is curious, as it stands right now, we are looking for a spindle that will handle speeds between 1000-3000 rpms. I do not know the loads that this will require yet (not had I known even part of the questions I needed to consider until I found this forum). We are trying to maximize the diameter of our shaft for several reasons, and unfortuantely I don't think I am at liberty to discuss all of them. But, we are looking at 6inches or more outer diameter of our shaft. Also, due to the nature of FSW, it is likely that our entire tool will be subject to a good amount of heating (external -- not from bearings). So, we might need to look at some active cooling.

Right now, I am trying to model my tinking along the lines of the floating axle of a pick up truck -- i.e. oil bath, external shaft housing, axle seals with tapered roller or angular contact bearings.

So, if anyone has any input, I would appreciate it. I am a Materials Engineer, so I am mainly interested in the product we are making, rather than the tool. However, no one ever gets to work with what they are 'mainly concerned with', so i will need to start doing my homework to follow the lead mechanical engineer on the project.

So, I am not really asking anyone to answer questions for me on my design, but I am rather asking, "What questions do I need to consider in the design?" and "What are some places places to go to answer those questions?"

Thanks for any help,

Dave

[RICHARD ZASTROW]

Dave, You may have noticed many questions pertaining to bearings / spindles will eventually involve the resident bearing guy NC Cams. Pay attention, he knows of what he speaks though sometimes a bit "disciplining".

Is the process you refer to similar to simple friction welding where parts are forced against each other while they are counter rotating? If so, that axial force will have to be quantified and compensated for.

Ref. heat: how much will reach the bearings and soak into the spindle?

What about torque required?

I'm sure NC Cams will inquire of more required info, but hopefully this will start the discussion and helpful suggestions.

Dick Z

[Al_The_Man]

If you are looking at designing the whole machine and including the control of the spindle, you are going to have to calculate the HP required, If under 3000 rpm, this is in the range of standard 3phase 2 pole motor or 4 pole run at 120Hz, for the speed control, a VFD could be used, but I suggest you use a model with encoder feedback. This will give you better control than a normal Sensorless Vector model.
Al.

[dagray3]

We are looking at a 5HP Baldor motor. 208, 3 phase. For control, we are planning on an ABB ACS 350 series sensorless controller. I know that the open loop configuration will give less control but it cuts our cost by 50%.

I have read some of NC Cam's replys, and of course, I am a little frightened of his response. I know that he does know his stuff, or at least it comes across that way. That's part of the reason I am not trying to get someone to recommend the bearings or motor or lube technique, but rather to point out some of the issues that I will need to address.

The Friction Stir Welding is similar to the friction weld, except that in stir welding, a joint of two materials (normally aluminum I think) are butt joined and welded together by a steel tool. The tool is forced into the materials to be joined while spinning. The heat from the plastic deformation of the material is enough to weld the two materials, but not enough to affect the microstructure of the material far outside the weld. Niether is the heat enough to actually melt the material.

As far as heat reaching the bearings, the answer to that will depend on the bearings we can find. I think our current goal is to find bearings that are large enough, can handle the speeds that we are looking for and are rated for the highest temperature that we can find. We will then taylor our cooling scheme and/or process parameters (time, speed, duty cycle, etc) in order to ensure that our bearings are happy.

As far as torque is concerned, I am not certain yet. My supervisor had a sort of intial FSW setup with which he was not completely satisfied. I personally have not yet done any welding. I think the old system they had consisted of a DC 3 horse motor turning a spindle via a belt on a pulley. But, the heat generated was enough such that the CNC stage was heated, and the 3 horses weren't enough for the process. So, that's why he wants to kick up the size of the spindle I think. I think he is hoping that the increase in mass of the spindle will increase the momentum and the thermal mass of the unit. I am sort of under the impression that increasing the mass of the system will increase the heat storage and decrease the effectiveness of the cooling mechanism.

So, right now we are planning a 5 HP, Voltz/Hertz sensorless system. We are then considering something like a 7324 angular contact bearing (two really) or some kind of tapered roller bearing. I have seen that tapered rollers seem to be lower speed, higher load than the angular contact ball bearings. I have also heard that using Krytox as a lube could increase our operating temperature a little.

Since our application is research more than produciton, we have the luxury of having a very low duty cycle -- we can operate far below several hours per day. So, we can keep our heat down on that front. Also, we are not overly concerned with lifetime of the bearings or the system, as the total lifetime will be measured in less than hundreds of hours I would imagine, as opposed to hundreds of days. We are mainly looking for proof of concept and characterization of the final product rather than setting up a welding shop.

I do appreciate the help guys. I suppose I have found at least a partial answer to the question of "where do I go for answers".

[RICHARD ZASTROW]

Dave, Can / must your spindle shaft be solid? You could cool your spindle bearings with air or liquid cooling thru a hollow spindle unless you must "swallow" the workpiece. As far as bearing type, angular contact and tapered roller bearings can absorb axial and radial forces to varying degrees. Don't forget seperate radial and axial bearings, they function well also.

You appear to be starting with blank paper, don't rule anything out.

[brunog]

Dave,
in order to choose the right bearing you need to know the following basic information:

Usage: hours per day usage

RPM the shaft will be running at

Dynamic load

Static load: whats the weight involved when the machine is not working this can create as too much weight can cause strin hardening in different spots and increase wear

Shaft diameter

How many years are you expecting this to last.

Basically you will be choosing a bearing that will need to be changes after x million revolutions.

then other factors come into the puzzle which are mainly environmental: heat, soil, dust etc.

Most Bearing catalogs and or websites will explain all of the above

Best regards

Bruno

[dagray3]

Richard,

Unfortunately, our spindle will be hollow, but our process will involve an added heat source through the core. I cannot go into specifics due to company policies, but suffice it to say that we will be adding heat through the core of our spindle. So, any cooling will have to be outside of the spindle. So, we are going to have to counteract heat from the bearings, heat from the spindle-to-material friction, and heat from the core of the spindle. Sounds fun, right?

Also, I had not considered separate radial and axial bearings. That is a good point, and I will check into that as well.

Bruno,

Those questions are very helpful. I will have to look into them. The dynamic and static loads are issues that are more difficult for me to consider, but I can see that they are critical. I will research those and try to find a good way to answer these questions.

All,

When temperature ratings of bearings are given in data sheets, are these limitaitons more because of lubricant or from a change in the material properties of the bearings themselves? I wouldn't imaigne much change in hardness of steel bearings at temperature, but I could see lubricants going crazy above 300F. What are osme of the contributions to the temperature ratings of bearings?

Again, thank you all so much for the advice. This forum is a great resource, and I hope to learn enough over the next few years to be able to contribute to someone else's projects.

[NC Cams]

You also can NOT, NOT NOT use the rated capacity of the bearing for design purposes. For "infinite life", max start up load should not exceed 10-15% of the rated capacity. The lower the static capacity you use, the less chance of static damage you'll incur.

Heat affects grease life more so than bearing life. Cook the grease, fry the bearing. If you get enough heat into the spindle, you CAN encounter grain growth (size change) of the bearing over time. This requires "heat stabilized" bearings and, surely, metallic cages - probably machined bronze for this as opposed to plastic/nylon.

If you can deal with the friction, go with tapers - cheaper and a bit more robust than A/C's.

If you KNOW the actual loading, the task of doing a load life calc with varied load/durability ratings can be readily calc'd.

My "disciplining" comes when guys do SWAGging half-fast kluging of their bearing applications work. The formulae for doing bearing calcs are as viable as E-IR for the electronics guys - yet, guys would not trust SWAG for doing their electronics designing (some would) so why do they SWAG the science of bearings??? Go figure.

A continuous temp rating of 300F won't allow the bearings or grease to live. Grease life PLUMMETS as temp gets about 200F. Figure life cut of 50% for every 10deg F above 212. Do the math, life is short at 300 deg continuous.

Forget silicone grease - these work at high temp but they have LOW film strength - they survive temp but dont necesssarily lube well under high loads. Keep the temp at or below 250F. IF you can't, find a way how.


Re: temp affect on bearing hardness: it all depends on time and temp. Relatively speaking, 300 F is a viable tempering temp for bearing steels. A special temper would be needed and/or special steel for continued high temp operation. Forget finding/getting such speical H/T'd parts easily or cheaply. If you can get them, buy a bunch and prepare to WAIT and pay dearly for them....

Essentially you're making a variation of an intertia/friction welder. This technology is quite well known and/or developed. From the ones I've seen, they find ways to isolate their bearings from the hot end of the work zone.....

[307startup]

It seems to me that your main concerns will be from the plunge into the material as well as the side load from forcing the tool through the materials being welded. Your spindle should be able to be kept cool if you can consider a toolholder that will insulate your spindle.

I have made a FSW tool using a Bridgeport milling machine. 7hp with a CAT40 taper. The biggest problem I faced was the amount of side load on the spindle bearings from feeding the tool along the weld seam. The second biggest problem was isolating and delaying heat transfer from the tool itself to the holder and then to the spindle.

[dagray3]

wyld, I agree. We are concerned about the plunge portion of the weld. We are thinking about preheating for the plunge, maybe with IR heat lamps?

And, yes, the heat transfered to the spindle is going to be a problem. I had thought about cooling, but I also feel that maintaining the proper temperature of the workpiece is critical as well. So, I don't want to take too much heat out the bottom of the piece, as I feel that could degrade the quality of the weld. So, I imagine that I am going to be forced to sink the heat generated into the spindle and then actively cool the spindle. Isolation of the tool would be nice for short term, but I don't think there would be enough thermal mass for a long weld.

Does anyone have any recommendations of vendors for pulleys and synchronous belts? I think we are planning on driving the spindle with a set of synchronous pulleys and a poly belt. Any input? Our spindle is sitting at 3 inches in diameter right now. We are looking at Gates pulley right now. They have a nice online design aid that seems to be giving us some good starting points. Anyone ever use Gates products? The local distributor seems very helpful, so that's always a plus.

Thanks.

[RICHARD ZASTROW]

Dave, I regularly specify Gates belts and pulleys. Check into their carbon fiber GT2 belts. Should be good stuff.

Dick Z

[307startup]

All the setups that I have seen or read about regarding FSW don't mention preheating. In fact, I think the idea is bad altogether, as the heat from the process is a byproduct of friction. In fact, as I'm sure you already know, it's not really welding, it's the "stirring" of the parent metal to a plastic state and redistributing the mass in the wake of the tool. The goal of FSW is to be able to weld materials that commonly can't be welded by any other means, to weld dissimilar materials, and to keep the HAZ from being larger than absolutely necessary.

I keep going back to the idea of a commonly available spindle (Tormach sells a BT30 upgrade for $595 that is a industry standard 80mm cartridge type) and building a water jacket for it...run the coolant through a closed-loop TIG torch cooler. I think a hydraulic tool holder would be a pretty good idea for this kind of setup.

It would probably be pretty expensive, but an idea I had for insulating the tool from the holder and thus preventing heat transfer to the spindle...
AlMMC. Aluminum Metal Matrix Composites are stronger than cast iron (tensile strength and compression...approaching many steels, in fact), hard, light (up to 25% lighter than "regular" 6061), ungodly abrasive to machine...and work wonderfully well for heatsinks. Better in fact than 6061 or 6063 extrusions. Typically AlMMCs are made from 6061 with SiC3 (I think that's right...silicon carbide, by any rate), and are available for extrusion and in billet form. I suggest AlMMC rather than ceramics due to fracture toughness and thermal shock/cycling. MMCs are the best of both worlds...metals that are lighter and stronger than possible with current alloying methods (minus exotics like Aluminum/lithium or aluminum/beryllium) while exhibiting amazing thermal resistance to excessive growth or creep.

What material are you building the FSW tool from? I used H13, it works well, but degrades over time due to heat cycling.