Announcement

Collapse
No announcement yet.

Tool and cutter grinders, - a new slant

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Tool and cutter grinders, - a new slant

    OK the dogs got the bone and it's not letting go

    Some thing I have been thinking about for a while, even got some bits but as I have more pressing matters I'll put this forward in case anyone else want's to run with it.

    Take a say 6" wheel and present a tool directly to it horizontally and you get a flat face [ well nearly as there is a 3" radius to take into account.
    Now do the same but say 1/2" above centre and you will get an angle of about 10 degrees. Many T&C grinder use this method to get clearance angles, some tilt the head and some raise or lower the wheel.
    No right or wrong way just different ways to achieve the same thing.

    Now take this horizontal head, fit a collet chuck to it and power it by a stepper driven indexer drive to get the 2,3, 4, divisions etc needed, that takes care of the indexing.
    It also needs to be able to swivel at this point but can be manual.

    The horizontal head needs to be on a ballscrew driven linear rail, again stepper driven and the wheel head needs to be able to move up and down, again by stepper.

    Operation for a drill:
    Spin the head to 135 degrees or what ever, fit drill to collet, line cutting edge up horizontal and present to the wheel, [ not running ] , on centre line.
    Zero controller, back off say 0.200" and enter the drill program for that size from a library.
    Wheel starts, head drops and presents drill to the wheel and grinds one secondary angle, backs off, raises head, feeds in drill to form primary angle, backs off rotates the collet 180 degrees and repeats.
    Job done.

    End mills can he handles in roughly the same manner.

    Asbestos suit on, let the flames commence...............................

    .
    .

    Sir John , Earl of Bligeport & Sudspumpwater. MBE [ Motor Bike Engineer ] Nottingham England.




  • #2
    It would be tempting to have a go along those lines, just as a project, but I too have enough on already. If I had the space to keep it, putting steppers and a CNC workhead on my Ward grinder would be a good excercise, & I've got another idea for getting motorised feed onto the head. I *definitely* don't have room, though, it has to go.
    Maybe once I've got the shop somewhere near sorted, some sort of semi-CNC control on the Cinci grinder might be worth a try, just a workhead and rise & fall on the wheelhead. I've got some of the bits to do it.
    Trouble is, my Bridgeport control has been in bits for months as the day job has got in the way, and there have been quite a few times when I really could have done with it
    Shame about having to feed the family, there would be much more time for projects otherwise!

    Tim

    Comment


    • #3
      Don't know, John, if I'd trust the steppers to index that precisely. I've heard good and bad things about them, but never actually built anything with them. Maybe I'm just too used to mechanical indexing with positive stops. Hey, did I ever tell you how I hate CNCs? I thought so.

      Comment


      • #4
        Originally posted by Millman
        Don't know, John, if I'd trust the steppers to index that precisely. I've heard good and bad things about them, but never actually built anything with them. Maybe I'm just too used to mechanical indexing with positive stops. Hey, did I ever tell you how I hate CNCs? I thought so.
        Good enough for me. see

        http://web.onetel.com/~duttondock/In...eelgear-5b.jpg

        188 teeth using a stepper-driven rotary table. The flywheel is about 150Kg.
        Cutting a '189th tooth' as a check didn't remove a whisker.

        Tim

        Comment


        • #5
          OK, I would be impressed with that set-up.

          Comment


          • #6
            I see three problems,
            1. How do you do the flutes?
            2. The drill will have to turn as it is ground back to keep the lip horizontal due to the twist.
            3. The nearest library is six miles away.

            I was at an auction yesterday, a B&S T&C grinder sold for $75.00 & a Cincinatti for $50.00. Both were in operating condition if a bit shabby. I can't handle the large machines myself, but a couple of people got good deals. No grinder people were there.
            Jim H.

            Comment


            • #7
              Originally posted by JCHannum
              I was at an auction yesterday, a B&S T&C grinder sold for $75.00 & a Cincinatti for $50.00. Both were in operating condition if a bit shabby. I can't handle the large machines myself, but a couple of people got good deals. No grinder people were there.
              Grinders can be bought cheaply enough, but as John said the other day you need a ton of tooling to go with them. That's where some sort of CNC head might score, may be able to take the place of a number of special items. If you wanted to take it a step further by including table movement in the control system, then you could even do spiral flutes without the need for fingers or fancy controlled spiral heads.

              Tim

              Comment


              • #8
                Originally posted by JCHannum
                I see three problems,
                1. How do you do the flutes?
                2. The drill will have to turn as it is ground back to keep the lip horizontal due to the twist.
                3. The nearest library is six miles away.

                I was at an auction yesterday, a B&S T&C grinder sold for $75.00 & a Cincinatti for $50.00. Both were in operating condition if a bit shabby. I can't handle the large machines myself, but a couple of people got good deals. No grinder people were there.
                JC,
                Just a concept idea.
                [1] Flutes can be done by moving the head forward along the side of the wheel whilst rotating.
                I didn't want to bring this into the equation at the moment due to complexity and confusing the natives, Norman stop scratch your nose, it's big enough

                Have you noticed on newer end mills that they have the lead angle already etched on ?

                Personally I'm not too interested in doing flutes as once you do the flute you then have special offsets just for that cutter to compensate for the amount you have removed.

                [2] True if you have to remove shed loads, touching up will work on the one setting just like a drill grinder.
                If the drill is that bad then hand grind back to a good point and start at that point.

                [3] Move closer to the library.

                T&C grinding is very dear to people hence the number of designs out there and the following it attracts on these forums,
                It won't suit many but will also appeal to many others as the answer they have been looking for or as Tim siad an attachment for an existing T&C grinder.

                I have a nice all singing, all dancing grinder but to do the odd cutter it's a pig to setup as regards time.
                Unless I have a load of all the same cutters I can't afford to switch it on, it's cheaper to buy new or even pay to have it sharpened.

                The local T&C guys to me have these 5 and 6 axis grinders that once the cutter is in the collet it does the whole thing automatically then spits it out.
                A 20mm 3 or 4 flute cutter is in the machine for less than a minute for an end grind, two or three for a end and flute and it spits out the new diameter so you can reset the offsets.

                Seeing these in action prompted me to think of a DIY CNC grinder.

                .
                .

                Sir John , Earl of Bligeport & Sudspumpwater. MBE [ Motor Bike Engineer ] Nottingham England.



                Comment


                • #9
                  [[once the cutter is in the collet it does the whole thing automatically then spits it out.]] Kinda' makes you sick, doesn't it?

                  Comment


                  • #10
                    Originally posted by John Stevenson

                    Seeing these in action prompted me to think of a DIY CNC grinder.

                    .
                    Me too!

                    The Walter site is fascinating, and got me interested looking at their 5-axis machines. They use a species of gang tooling to put multiple wheels or even EDM on the spindle. But they're made for churning out new cutters from blanks by the thousands, and they cost $250K.

                    My question, is what would be the modern day CNC equivalent of something the relative size and cost (in the day) of the Deckel T&C's? I would not think you need huge capacity to make a small shop happy. It doesn't need to churn out thousands of cutters in other words.

                    As to the machine, 3 normal mill axes, one of your CNC rotabs set in normal 4tth axis mode with a collet to hold the cutter, and a second CNC rotab put where the spindle is so your grinder can rotate about Z and that's the basic geometry needed.

                    The working envelope need not be large. Seems to me a max cutter diameter of 2" (50mm?) and length of 8-10" (250mm?) ought to be plenty. The thing would then not be that much larger than a Deckel, and could sit on top of a bench or it's own stand. You've probably seen the little millers various folks are making these days, it would be similar in size:



                    I think the real challenge is this application would require some no messing around accouracy (+/- 0.0001") and then there is the job of programing the silly thing, which is no mean feat. Probably needs servos, ground ballscrews, and linear slides, but at least it is all to a small scale, so costs won't be completely out of sight. I would think you could grind the slide mounting beds on a surface grinder and have no need to scrape anything in. Likewise, given the scale of the machine, you could give it a really beefy frame. I'd think about something like 2" OD 1" ID DOM filled with polymer concrete.

                    FWIW, this is all figured out for you on a much larger scale. Google "Principles of Rapid Machine Design". It's an MIT student's PhD MechE thesis. Lots of fascinating innovations for machine design discussed and he wound up building a 5-axis T&C grinder that's being sold today. I did a sketch over the weekend of a 5-Axis T&C frame over the weekend based on his principles and readily available materials:



                    I have a hard time believing you could adapt one of those Siegs or most any other mill and still come out with enough accuracy to do cutters well, but maybe I'm all wet. To program it, I envision a custom VB or Java program that takes all the parameters (axis angle, rake angle, helix angle, axial angle, cutting edge length, etc.) and spits out the g-code.

                    In the end, I decided this was way beyond my time, expertise, and attention span. Still, it is an interesting daydream.

                    Best,

                    BW
                    ---------------------------------------------------

                    http://www.cnccookbook.com/index.htm
                    Try G-Wizard Machinist's Calculator for free:
                    http://www.cnccookbook.com/CCGWizard.html

                    Comment


                    • #11
                      Originally posted by John Stevenson
                      JC,
                      Personally I'm not too interested in doing flutes as once you do the flute you then have special offsets just for that cutter to compensate for the amount you have removed.

                      .
                      I'll take opposite view, if its not going to sharpen the flutes, not interested. i rarely plunge with an end mill, its always cutting on the sides, that's where it needs to be sharpened. grinding 1/4" or 1/2" off the bottom to get to a 'fresh' part of the periphery seems like both a waste (how many resharpens will you get) and a misery (grinding away 1/2" of hss is tedious).

                      before I started grinding my own endmills, I used to think that loosing the dia dimension would be an issue, but have found that in the minority of times that i am cutting to a shoulder, I'm taking roughing cuts and a finish cut with some measurement to determine the final cut anyway - whether the end mill is .5000 or .4975 doesn't matter. when cutting slots I find i have to use a smaller cutter than the the finish dimension anyway, to get the accuracy and finish i want.

                      Any idea that helps put sharpening capacity in the home shop is a good one, and as you've pointed out linear ball bearing rails and ball screws bring some pretty accurate off-the-shelf motion control into the home shop. Cost will be reasonable, (but not insignificant), ie ball screw, linear bearings, angular contact mounting etc.

                      The two missing pieces (imo) are 1) and precision spindle and 2) an accurate way of tracking the helix while sharpening the periphery.

                      No easy fix on 1), imo you won't get a decent job on the flutes without a high quality spindle.

                      As far as tracking the helix goes, a stepper/servo is interesting. Personally I see the optimal solution as being a tooth rest and air bearing - the pita isn't the tooth rest, its moving the thing without friction in the cylinder. Now they problem with air bearings is they are expensive and difficult to make so the stepper/servo idea is used to create the helix motion.

                      then you've got a challenge of very accurate motion control between the x axis and rotary axis. very similar challenge to a cnc leadscrew - to accurately follow the helix of a thread lets say, the linear motion must be encoded or other wise present to a very small number of subdivisions per revolution. Then you get into challenges with the controller hardware keeping up. an analogy might be, you can knock together a sort of cnc lathe, but to really make a cnc lathe where the x motion can be accurately enough controlled independently of, but in sync with, the spindle so that it can cut any thread is not a trivial matter.

                      what about different approach; say use server or stepper to drive the indexing. x axis moves steadily along (could even be via capstan, well maybe that wouldn't work, but it could by hand). in place of the tooth rest is a sensor (yet to be designed) that controls the stepper/indexing function. avoids all the hassle of coordinating x/indexing and getting the helix angle right and also avoids tricky setup
                      Last edited by Mcgyver; 09-14-2006, 11:03 AM.
                      .

                      Comment


                      • #12
                        Quick reply to Bob.

                        Whilst 0.0001" is ideal i don't think we need to set our HSM sights that high.
                        After all can you hit a tenth on a manual one ?

                        To McG [ sorry can't see a name ]
                        Tracking a flute could be done by pushing the cutter forward a known amount against a finger and getting a shaft encoder to give you the rotation per known amount.
                        This can them be fed back in. That's how the professional ones do it but they do it automatically.

                        Sorry for the quick replies but trying to get two machines setup.

                        Bloody customers.

                        .
                        .

                        Sir John , Earl of Bligeport & Sudspumpwater. MBE [ Motor Bike Engineer ] Nottingham England.



                        Comment


                        • #13
                          Originally posted by John Stevenson
                          Quick reply to Bob.

                          Whilst 0.0001" is ideal i don't think we need to set our HSM sights that high.
                          After all can you hit a tenth on a manual one ?

                          .
                          I could with a surface grinder or toolpost grinder, but you are right, it might not be necessary.

                          Somehow, if I'm going to the trouble to CNC it, I hate tracking the existing helix either mechanically or with a finger and encoder. All the software you'd want to leverage is just not really set up to "think" that way. It wants g-codes and assumes the machine will accurately track the g-code. You might save some mechanical cost (i.e. a much less accurate or simpler machine) only to have a much bigger nightmare on the software and controller side.

                          Besides all that, once you get reading about CNC tool grinders, you quickly discover they do all sorts of neat stuff besides just sharpening cutters. For example, creating custom form tools, taking an existing endmill and putting a customer radius on it, converting it to do undercuts, etc., etc. It would be easy to write a g-code program to true the grinding wheel by sticking a diamond in the collet. You get the picture.

                          Once you have the hardware, a little bit of software makes it do all kinds of nifty things. Bye bye toolpost grinder for things that fit the machine, for example (never liked the mess and grit anyway).

                          I think the small relative size of the machine would let one deal with the high accuracy more easily than a conventionally sized machine as well. The ballscrews (ground, with mapping enabled in Mach 3) with linear slides are up to the task if you can get the axes square and true. Here is how I would go about doing that in a home shop, based on that PhD thesis I mentioned.

                          You need a separate structural chassis/frame from the precision chassis frame. Let's build the structural chassis/frame from 2" OD/1" ID DOM filled with polymer concrete. We'll Tig weld it together in a jig. Shape ala my sketch.

                          Now you've got a T-shaped plate, of relatively small size, that is welded to the structural frame. Let's take that assembly to the grinding shop and get it ground really true and flat. Let's also grind a mounting for the y-axis linear slide rails. Take a look at how nook shows those shoulders along the edge when mounting their rails, midway down this page:

                          http://www.nookindustries.com/profil...ilGlossary.cfm

                          That gives us a true, very accurate Y-axis. We need to attach the X-axis to the rail blocks and make sure that's true to the Y-axis travel. Should be possible to machine that plate with ground shoulders for the X-Axis and then get it in place on the Y-Blocks and accurate to a tenth for less than 6" of travel, no?

                          Two axes are down. Now let's build the gantry (which doesn't move, btw) for the Z-Axis. My sketch shows 3"x3" welded. Put that all together, and that's the structural piece for the Z. I'd use thick wall square tubing (or more DOM if you like) and fill with polymer concrete. You need a mounting flange on the bottoms to go into the T-plate that holds the Y-axis. To that structural piece, bolt down the steel plate you mount the Z-Axis slides to. I'd machine and grind it just like the other 2 plates. We want to bolt it to the structural portion of the gantry so it is square relative to all axes. Now this is the tricky part!

                          I would grind the flanges at the bottom of the gantry, and bolt it down to the base. I've now got my ground Z-Axis plate in hand, and I need to mount it so that it winds up square and true relative to the T-Plate, essentially. We won't worry about the gantry being anything more than close, because we're going to adjust the Z-Axis plate on the gantry in 2 dimensions to take out the remaining error. To do that, I'm going to borrow another trick from that guy's PhD thesis and do it with the help of Moglice.

                          Set up the Z-axis plate so it not only bolts to the gantry, but also has 4 set screws with fine threads. Bolt it losely, leaving a gap of perhaps 1mm. Apply tension with the set screws so the thing is rigid. Bolt the Z-plate onto the gantry, add the Z-rails and carrier so it can go up and down. Now mount your tenths indicator to that Z-axis and apply the usual fixturing so you can raise the Z-up and down, watch the indicator against a cylindrical square of some such, and see whether you are square to X and Y. Remember, travels are pretty short, so it ought to be straightforward.

                          When it is close, fill the gap with a bead of silicone cement and let it cure. Once cured, recheck the alignment, and tweak until it is close as you need. Now using the holes you drilled in the z-plate (oops, forgot to mention those!), fill the gap between the plate and the structural gantry flat side with Moglice. They have a formulation especially for the purpose, and the silicone will act as an o-ring to keep it from running out. Be sure you applied mold release to the mounting bolts, setscrews, and perhaps the Z-plate if you ever want to remove it.

                          Once the Moglice cures, you have a solid surface. Remove the setscrews and carefully torque down the bolts. Double check, but everything ought to be very square and true. No hand scraping was needed to get here, just a bit of work with a surface grinder that could be farmed out to the grinding shop if you desire. Your fleaBay bargain linear slides should function very happily in this arrangement. If you did your homework right mounting your ground ballscrews (which were cheap because they're small and nobody else is building a small machine like this), the machine should be very accurate. The frame should be very rigid relative to the size.

                          I leave it to the rest of you to fabricate the 4th and 5th axis and get them mounted true perhaps using the moglice + setscrew technique. You could drive a rotab, but I'd worry about backlash. Harmonic drives are available pretty cheaply on eBay, and are perfect for this application:

                          http://www.cnczone.com/forums/showthread.php?t=19448

                          Now as to the spindle, a machine like this does need a nice one. But there are nice toolpost grinders readily available, and everyone hates to use one on their lathe. Why not here? Get a Dumore or Themac and rock on.

                          Sorry for the length of thread. I guess I thought about it pretty hard this weekend! (he said sheepishly)

                          Best,

                          BW
                          Last edited by BobWarfield; 09-14-2006, 01:42 PM.
                          ---------------------------------------------------

                          http://www.cnccookbook.com/index.htm
                          Try G-Wizard Machinist's Calculator for free:
                          http://www.cnccookbook.com/CCGWizard.html

                          Comment


                          • #14
                            A little off topic BUT

                            Off the current topic . But still tool an cutter grinder. Question Thanking of building something . What do you thank SIR John Norman or any one else . For table movement Linier bearing on round rods Like Thompson shafting Or something like THK rails and linier bearings ?????
                            Every Mans Work Is A Portrait of Him Self
                            http://sites.google.com/site/machinistsite/TWO-BUDDIES
                            http://s178.photobucket.com/user/lan...?sort=3&page=1

                            Comment


                            • #15
                              like most machines though, the bigger the envelope the better, that big slab milling cutter JS posts pics of 'll need more than an inch, and different fixtures and attachment and tools all seem to present themselves at different places along the x axis.

                              the need for accuracy is in the spindle and a an X axis with minimal variance along a straight line. maybe better than a tenth. I don't really know but since a typical cut is a few tenths, if your x axis is wobbling back and forth, you're not going to get a very pleasing results along the flutes. most (well ok the few) grinders i've worked with accomplish this by the table sitting on rollers in a V channel (no clearance therefore needed) moved by a capstan. A linear slide, if the tolerances where tight enough, might accomplish this, although i'd guess they have to have some clearance or they'd jamb?? maybe the could be preloaded to one side, possibly via gravity at an angle?

                              I don't think though that it requires a ground ball screw. The accuracy you need in the motion, rotating wheel and X axis truly straight, you are not dimensioning as you would with a surface grinder. For that matter, a ball screw isn't any more accurate than an acme screw. since there are no climbing cuts, backlash isn't a factor, why not save the mucho's $$$ and use an acme screw if cnc and capstan if manual on the x and acme on the y?

                              if you're buying dumore spindles, high end linear bearings and ground ball screws, you're starting to get to where you can buy a new, good quality benchtop unit. But you'll buy them from ebay you say? the counter to that is buy a T&C used or from ebay. Therein lies the challenge; bloody tough to come up with something better and cheaper than all those designers that have come before thee, not impossible but tough.
                              Last edited by Mcgyver; 09-15-2006, 11:48 PM.
                              .

                              Comment

                              Working...
                              X