Machining a DEEP square hole

[T-Gold]

A friend of mine has a job to do... not due any time remotely soon, thank goodness. The piece is, if memory serves (I will correct later if need be), a large piece of 304 stainless. It needs to have a 1" square hole bored in it, 10" deep, not blind, parallel and true to +/- 0.0001". Yes, one ten-thousandth. Precision shaper is what I was thinking... or pre-drilling to 15/16" and pressing a multi-stage carbide broach through. I'm looking for out-of-the-box solutions to this, because frankly, I'm pretty well stumped. (For the curious, it's for some sort of laser interferometer.)

Things which I've thought of and discarded:
Cutting the part apart and machining the hole, then bolting it back together (Not allowed)
Waterjet (Inadequate precision, esp. at depth)
Laser (Inadequate depth)

Possible stuff:
Electron beam cutting??
Magic
Voodoo


Just something to strain y'all's brains over. :)

[irnsrgn]

Wire EDM

[T-Gold]

Precise, but not enough depth, at least not the folks I've found who do it... they could do a 60mm deep hole of that precision. I will search further.

[Ed Thomas]

T-Gold: Take this question over to the Practical Machinist board. They thrive on this sort of puzzle.

http://www.practicalmachinist.com/

I was thinking "Big Deal... use a shaper..." until you started getting all precise on me. :mrgreen:

[T-Gold]

I'll bug them too. :)

[bruce wilcock]

i think you should steer well clear of friends that get them selves into jobs like that, your entire life will be bashing your brains, get a friend that wants to ,say make needles out of rail iron using a candle to get it hot and a snowball for a anvil, then we will all be able to help you .
but as to the sq hole you are on your own. :)

[metalmaster1766]

only way your gonna get a 1 inch square hole 10 inches deep is with a broach, parallel and true within .0002" good luck on that, you may get it within .002" if your lucky, sounds like its overkill on the engineering part, but then again i dont know what you want to use this part for, that always helps

Ron

[T-Gold]

Goldangit, I said it here and I said it on the machinist board, and in both places, y'all seem to have missed it...

(For the curious, it's for some sort of laser interferometer.)

It's in there.

LOL, Bruce. :)

[metalmaster1766]

maybe you could broach it a little smaller than the finished size, then lapp it to size, gonna take a while, and its still gonna be tough to hold that kind of tolorance

Ron

[HWooldridge]

There are of course, different size machines, but wire EDM is the best choice for this project. Broaching would be something for production - not a one-off job. Wire EDM is infinite in what it will cut so long as the piece will fit between the guides. However, the first thing the shop must do is get a small hole all the way thru so the wire can be threaded. A big "hole popper" EDM could do it but it would probably be better/faster to drill a 3/8 or so hole all the way thru the block and wire the rest out. Your buddy might be able to do that ahead of time. The 60mm response may either be the distance between their wire guides - or - how deep their hole machine will go.

[frogvalley]

A google search for wire EDM yields two shops with a 16 inch depth. 10 minus 16 = 6 inches to spare. Just teasing. You'll have to ship or drive, there ain't many shops with deep EDM machines.

Wire EDM would be my first choice. Broaching would be #2. I might be tempted to try a shaper just because I like them.

I used to do prototype work for mil spec and specialty electronics firms, some of the work was real trick stuff. I can't talk about it or I'd have to kill you all. When I took a one off, hi accuracy, tight tolerance job , I simply bid it to have a failure rate that was usually two in three. That is two were probably going to be off spec and fail. If I didn't get the tolerance right in three shots that was my bad. But usually it was right first or second try. If not then there was time and materials to spare.

I do hope your friend bid this job right.

IMHO.

[Strine]

1 ten thou accuracy is the same accuracy as marking out a kilometre to the nearest 100mm...easiest thing in the world to achieve with the right tools but it costs money...big money. You can't do it with a ten dollar plastic tape measure. Tell the bloke that to achieve such accuracy is going mean his you-beaut interferometer will be so expensive it won't have a market and he's wasting your time and his.

Allow me to tell a yarn about an engineer I did a job for. We were building a roof over a reservoir, maybe five acres in area. The roof was to be held up with vertical "I" beams at around five metre centres. The I beams were to be supported on concrete pads, caste in situ with four holding down bolts. My job was to set out the location of each holding down bolt. The engineer required half millimetre accuracy. After much negotiating and telling him he was insane we managed to get it out to 1mm. Had it got to 5mm the cost of setout would have reduced by 80%. The accusations of insanity remained but they were kept in house.
Do you know a concreter that can work to 1mm? No, and the engineer didn't either because as it turned out, after all the expensive survey work the most useful tool on site when the I beams arrived was a 14 pound sledge hammer :wink:

So as the half millimetre was just a fanciful number to the engineer so too might the ten thou be to your mate. Baffle him with science. Ask him to explain to the nth degree why it has to be so accurate. Whether you understand it or not, pounce on any hestitation with suggestions of less accuracy. You might strike it lucky and get to the 0.002" MM suggests

Another solution to the problem is to point out that he is jeopardising his relationship with you as a mate with such difficult conundrums and that he should only consult you for answers to the easy things like what is the meaning of life or what existed before the big bang if there was a big bang.

Can you build/fabricate the hole around an accurately machined core then take out the core.

I think I'll move on to the next problem solving thread

Good luck T

[Ed Thomas]

When working with light and other wavelengths, this isn't ridiculous precision... it is just a little silly to ask blacksmiths.

You can see some comments that professional machinists made about this:

http://www.practicalmachinist.com/ubb/ultimatebb.php?ubb=get_topic;f=1;t=015122

[irnsrgn]

got curious did a search for laser interferometer, I don't see any reason for me to measure that close LOL

[metalmaster1766]

yeah after years of machining, iv found engineers over engineer everything, i proved my point a few times to then with thier unrealitic tolerances, everything looks good on paper, and can be achieved on paper, i can pour 10 gallons of water in a 5 gallon bucket on paper and not spill a drop, but lets try this in for real with water and a bucket, your gonna spill 5 gallons, id suggest opening up the tolerance, i dont know what this part is being used for (i know nothing bout lasers and dont wanna know anything)

Ron

[bruce wilcock]

making needles on a snowball defenitly is the sort of problem we should pooling our gray matter on, somthing of use to all of us :)

[ThomasPowers]

look into laser and maser wave guide work and you may find out that the tolerances *were* loose!

Not everything works "to the thickness of a worn shilling" tolerance.

Thomas

[Ed Thomas]

T-Gold: You are being paged on the Practical Machinist board. They want to know what ever became of this square hole problem. :)

[T-Gold]

So I am aware. :)

[Ironscot]

I've been thinking this over for the last several days. I have no idea what a laser interferometer is, but the part description is hard to let go of.

This is my idea, so throw rocks as you see fit. How about making the part out of stainless sintered metal, forge to density on an under sized mandril then lap the bore to tolerance? If density isn't an issue, then copper flash fill and bright hard chrome the bore.

Like I said, I don't know how the machine works, but I keep thinking about it while going about my normal chores.

My other idea was metal injection molding with a lapped and honed bore finish.

What are the part quantities and lead times looking like?

[T-Gold]

Wow, talk about a throw money at it solution You work at Scot Forge or something? I like it, but geez. This is a one- or two-off part. Being jobbed out, and probably won't be made to tolerance... this is what happens when you let the scientists make the decisions... yuk. Anyway. No real call to use sintered stainless and forge and lap when we could use heavy plate and bore and probably use a pressed-through broach, maybe followed by lapping if necessary.

Don't know much about MIM -- but heard of it before. Didn't know you could do it with steel, does that require graphite molds or something? Just curious.

[frogvalley]

Ok, here goes a bit of a rant. Sorry in advance...

First the basics.

An interferometer is a precision device for measuring using two or more input sources. It works on the principle that two waves(of laser light in this case) that coincide with the same phase will amplify each other while two waves that have opposite phases will cancel each other out. It is used to measure things with precision. And I mean extreme precision. Not like some dial micrometer measuring in thousanths. Starrett don't make anything that accurate. Thats amateur stuff comparatively.

There are millionths and billionths of an inch. And sometimes that matters.

The ones used in astronomy measure the distance between stars. Here on earth a good interferometer can measure to within a fraction of an inch, the distance to the moon. Or they can measure a wave length of light. Or a machine set up that makes other precision parts. Can't do that with abench mark based on your boot size. ( I use that one for a 12inch/one foot mark myself).

Even a slight error or out of tolerance component can exaggerate itself into miles of error in your final calculations. Or light years if your doing astronomy. Or microns if your thinking small.

When measuring small, say something .001 of an inch, and you miss that measurement by .0005, you would be 50 percent wrong. Plus or minus. And that is smaller by a long shot than a piece of dust.

If you send a probe to another planet and your interferometer is off by 1/1000 of 1 percent over 100 million miles, thats 1000 miles off, you could miss your target.

As a blacksmith, I rarely need ANY kind of accuracy that my eye can't see. As a machinist, I have needed .0005 accuracy many times.

Accuracy is not ALWAYS needed, but in precision machining for interferometry it is.

Second the solutions.

Wire EDM Sounded good and would have been my first choice. Ironscot may also be onto something. Especially about lapping to tolerance and/or fill treatments. Sintered stainless would need a surface treatment to be smooth enough I think. Correct me if I'm wrong but even when forged to density it is still, under microscopy, a bit grainy.


IMHO

[HWooldridge]

Frogvalley,

That wasn't a rant but it was informative - I know nada about interferometers, thank you for the knowledge. For the machining part, I stick by my original suggestion of wire EDM although I am not sure of the practical tolerances over that distance. It may need to be lapped and air gauged to check for the proper dimensions/tolerances.

I worked in plastic injection molding for 18 years and know a small amount about MIM. Any of the normal production methods will not really be suitable for a one-off job.

Plunge EDM is another option but would be more cumbersome than wire. Wire is tailor-made for cutting pockets in plates, which is essentially what is needed in this app. The hole could be cut undersize by 3-5 thou and lapped to size.

[Ironscot]

I have to second HWooldridge on the explanation of an interferometer.

I read an article in a science journal a few years ago about the bearings in an experiment. They were going to try and justify another part of the theory of relativity by doing a picture framing experiment in earths orbit. The parts were manufactured to fifty electrons. I myself always quit at two tenths.

For one or two parts, I don't think sintered metal would justify the cost. Forging to density aside, the copper flash and hard chrome would overcome surface issues. At least from my limited experience.

The last time I went to IMTS, metal injection molding was doing some amazing things. Basically one to one reversal of medium with roughly ninety-nine percent accuracy.

I'm going to go ahead and get my ofishial biz license again just to go back to IMTS. For those who make or do anything with anything that has to do with manufacturing, fabricating or automating, that place is Mecca.
Be sure to block out at least four days. That's if you're going to try and see everything at a dead run. Be forewarned, the robotics and lights out tech will keep you enthralled like a youngin at the lion pen at the zoo. :shock: :shock:

For the parts in question, I tend to agree again with HWooldridge, plunge EDM. A month or two lead time could definitely yield the desired parts.

[metalmaster1766]

i agree with HWooldridge, informative, i think the best way to get a tolerance like that is from lapping it, iv never measured in millionths of an inch, but i have measured 50 millionths .00005 , could only have .0001 runout on an O.D. grinder, could only do it when nobody else was working in the shop vibrations from other machines made it hard to get a good reading, like to get my hands on something that accurate, always liked doing that kind of work, precision machining that is ill let the others measure stars LOL

Ron