Buying first machine. Questions about Speedio vs. Okuma. - Page 5
Close
Login to Your Account
Page 5 of 14 FirstFirst ... 34567 ... LastLast
Results 81 to 100 of 267
  1. #81
    Join Date
    Jun 2006
    Location
    Near Seattle
    Posts
    4,980
    Post Thanks / Like
    Likes (Given)
    3445
    Likes (Received)
    1424

    Default

    It's great you are looking at models. More below...

    One of your next steps should be to CALL the local HFO on the phone (or appear in person.) Remember, they pay their rent, literally, by finding ways to make these machines work for people. As in, if there's a way to get it in, they should be quite motivated to find it. (If not, that's a bad omen.)

    OK, now, modeling and workholding for 5-axis machines.

    Think about chucks (as in flat back lathe chuck.) Or parts of chucks that you may have to make or have made yourself. (Go find a video about large VTLs and look at the jaws-on-table arrangement they use.) Model this. You may suddenly be "out of Z"...

    You've already found the "oh we do a 5-axis smart-pet-trick and we can machine a much larger part"....

    Next, you must get used to the idea that the part does NOT have to be lined up at any particular angle (one's natural tendency is to line something up along X - you can still do that, but open your mind.)

    And finally, you must learn to manipulate the CAM and machine, or more realistically, lie to them. Expect to spend lots of time with CAM and controller, and it's really best to do this with various test parts.

    [Side note - a Heidenhain iTNC530 has a couple of startling limitations in an otherwise awesome control, and so one resorts to loading part, probing it,
    feeding that skew data back into CAM, and then programming for the skew you actually ended up with. No production shop will go near this kind of stunt, but there's no reason you can't get really good at it. You want to find these issues in whatever control EARLY and learn to deal with them.]

  2. Likes empwoer, SVFeingold liked this post
  3. #82
    Join Date
    Aug 2015
    Country
    UNITED STATES
    State/Province
    California
    Posts
    140
    Post Thanks / Like
    Likes (Given)
    137
    Likes (Received)
    52

    Default

    Jaws on table...you mean something like this?

    2020-03-08-23_51_39-large-vtl-google-search.jpg

    I'm envisioning creating a larger subplate with radial slots. Then clamp the outside of the part. I suppose as long as the trunion can swing it and it doesn't bash into anything you're golden.

    Speaking of bashing into things, what's the cheapest reasonable verification tool? The UMC offers up the exciting opportunity to smack the spindle into the trunion. I'd like to avoid that...

    I've done a little digging into Vericut (lots of money), NCSimul (slightly less money), and I'm not really sure what else is out there that can show me precisely what the mill will do. Will these verification programs also show you when the mill decides to unwind a rotary axis in an unexpected (potentially crashy) way, or is that just up to the controller.

    EDIT: More questions! Sorry I just keep going on. Looking at that 20 arcsecond accuracy figure, say I needed to hold a very tight parallelism. Maybe half of what the rotary accuracy guarantees. Is there typically a way to fine tune the 0° position of the trunion without doing a whole recalibration? Maybe sweep an indicator around from the spindle, or use a spindle probe, to pick up several locations on the platter and then nudge the trunion and lock it down when it's as close to 0° as possible?

  4. #83
    Join Date
    Jun 2006
    Location
    Near Seattle
    Posts
    4,980
    Post Thanks / Like
    Likes (Given)
    3445
    Likes (Received)
    1424

    Default

    Bingo.

    Or, some little parallels, I call pucks. (I'll try to find the company name at the shop tomorrow) - think flat discs ground to match like parallels - with holes in them that allow a SHCS to bolt them down to a table slot. (Also some side reach arms I've never used) Those make a base - so you can dril through the part, etc.

    Then, around the edge, some variant of a true toe clamp. (Lots of people say toe clamp when they mean strap clamp, you want something like an actual toe clamp - which might be a ground flat slider that bolts to one or more t-slots, with a "riser" on the end with, say, a pitbull on it.)

    Spend some time on the mitee-bite, teco, etc. web sites. Sometimes you have to see something 3 or 4 times before it's apparent how to combine it with other things. Or that you need to make your own vaguely similar thing. (But use standard components whereever possible...)


    As for verification....
    1. The verifier is still a software entity (may have bugs of its own) and in any case can only be as accurate as the stock, workholding and machine models. Seems there's been more compettion for these in the last few years.
    2. Controllers are also software entities and can and so sometimes have bugs. Often exciting ones.
    => learn to step through programs, have optional stops at key places, etc.
    3. I've always only had the machine simluators that are available inside solidcam (which can also call out to vericut, ncsimul, etc.) - and that's been good enough.... But you are young and smarter than me, so investing in more may well be wise.

  5. #84
    Join Date
    Aug 2015
    Country
    UNITED STATES
    State/Province
    California
    Posts
    140
    Post Thanks / Like
    Likes (Given)
    137
    Likes (Received)
    52

    Default

    Quote Originally Posted by bryan_machine View Post
    Bingo.
    But you are young and smarter than me, so investing in more may well be wise.
    Yeah it's the young smart ones that always rack up the highest repair bills...

  6. #85
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    3,398
    Post Thanks / Like
    Likes (Given)
    3859
    Likes (Received)
    978

    Default

    Quote Originally Posted by SVFeingold View Post
    BTW, just to put a number on the accuracy. 20 arc-seconds at a 225mm radius (example part) is about a 20um chord length. Which I think is sufficient for my purposes. I haven't worked out what that would mean for, say, a 10mm bore axis on the end of that part that needs to be parallel to the top or bottom face to within X. As far as true position type tolerances within a particular orientation (when doing 3+2) it seems the rotary is out of the equation as it's clamped. Ignoring of course the aforementioned issue on a deep hole.

    I could see that accuracy maybe being an issue for surface finishing while doing simultaneous 5-axis. Again, what that means in practice I have no idea. Then again
    I doubt an Okuma 460-5AX has an accuracy better than like...5 arcseconds if that, and it can do this. Not sure how that rates for y'all but as far as I'm concerned that's a fantastic finish unless you need a mirror. So there's got to be more to the story than raw indexing accuracy when it comes to high quality finishing during 5-axis moves. I mean it's not like even a Hermle has a sub-arcsecond accuracy, right?
    Just a quick one over Monday morning coffee...

    Okuma Genos 460V -5ax … Depending on which documents you have repeatability is of the order of +/- 2 arc seconds or better. Interestingly the machine does not have any direct read scales. No linear scales nor direct read rotary scales / slightly indirect smaller high count encoders on shafts.

    Hermle surprisingly accurate for example the c axis rotary can repeat / index down to 1 arc second , that's their entry level model. However the brochures won't show you that, rotary plots are needed.

    What you are picking up on is 1. Bandwidth of the control and 2. servo tuning.

    So the Okuma and Hermle-(Heidenhain 640 control) have excellent TRUE motion control for sim 5 axis. It's hard to describe they just MOVE beautifully.

    Methods (Machine Tools) - They have rolled out their entry level Litz base Taiwanese machines to their specs and they have 5 arc second encoders and linear scales as standard. Machine looks like it moves very nicely …

    They have the MB- 650 U as well as new smaller Mb 450 U

    MB 650U - Methods Machine Tools

    MB 450U - Methods Machine Tools

    The HAAS control does not seem to have quite enough bandwidth for multiple simultaneous channels to read scales on all axes fully closed loop. However some demo footage of the NEW UMC 750 shows some decent swarf milling type moves.

    Empwoer is getting into the data optimization of point distribution for tight curves versus shallow ones.

    Machines like a Makino D200Z (small 5 axis machine that would fit in your garage but would set you back $375K ) position down to nearly 1 micron (most of the time) volumetrically.

    20 micron chord ~ that's 1 micron - 40 millionths - so 20 micron is 0.0008" two tenths shy of a thousandth. Depends on your application.

    So correctability of angular errors probe wise depends on backlash and minimum sensitivity of movement with the drive mechanisms of the table.

  7. Likes Jmaks, SVFeingold liked this post
  8. #86
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    3,398
    Post Thanks / Like
    Likes (Given)
    3859
    Likes (Received)
    978

    Default

    Also I wouldn't 100% dismiss DMG Mori , it's hit and miss as where you are and how things unfold and which models one may have invested in.

    So for example in Empwoer's case his experiences are REAL and I would never do anything to "Minimize" that.

    But locally For example Triad / DMG Mori there are plenty of "Peeps" that are making out WELL with the CMX 1100V (s) etc and the 4th axis units (that repeat to +/- 2 arc seconds / 5 arc second order ) DD units.

    Folks I know ARE doing well with the DMU 50 3rd gen in different continents and the USA.

    Empwoer's boss bought one of the lemon versions of the 4th axis horizontal and [email protected] Vice bought a later model that's really very good, and two models before that were also very good. [Seems Empwoer's boss and DMG Mori seem to have gotten up to their eyeballs in sh*t over the Defective machine and seems DMG Mori have been far from gracious about it. ~ As far as I have been able to sense given Empwoer's accounts - seems valid / not unheard of scenarios. ].

    I can't really validate or dismiss DMG Mori's "Treat most customer's like they are a Bug" thing (Gkoenig's observation / paraphrasing his statement) ~ I'm not saying that's a bad observation in some instances.

    In a lot of cases it is about making some sort of personal relationship with the "Peeps" you are dealing with (as gkoenig said and as [email protected] vice have said in previous tales by the fireside online.) . Locally my local sales people seem good and they try to steer me away from bad fits or potential lemons as they don't want the hassle either.

    Shocking as it may seem they want to make money too, and the second and third machine sales are more important than the first.

    Some folks are doing well locally with the DMU 65 (monoblock).

    Some folks have been left in the lurch by HAAS as well (in other regions- but locally HAAS seem very good), so unfortunately I can't think of a single MTB / vendor combo that doesn't have some sort of "Horror" story associated with it; YAMAZEN seem to have a very good track record for minimizing those types of 'Event". Hence popularity of the Brother 30 taper mafia.

    Nothing is perfect, bad things happen - Depends on what and how you want to roll the dice , but honestly it should not be like that.

  9. Likes SVFeingold liked this post
  10. #87
    Join Date
    Oct 2014
    Country
    UNITED STATES
    State/Province
    Maine
    Posts
    1,096
    Post Thanks / Like
    Likes (Given)
    166
    Likes (Received)
    519

    Default

    I'll throw out what a former co-worker of mine chose. He was in a similar boat as you. Engineer with a day job, wanted a machine to prototype his own stuff and do a little bit of production. His stuff was, at most, 3+2. He wasn't doing full 5-axis. He ended up with, IIRC, a VF2SS (or maybe it was a VF4SS) with a Haas 4th/5th axis sitting on it. He had the 4th/5th on a doweled subplate, so he could easily switch between 3-axis for big stuff, 5 axis, and 4 axis with a tombstone and a tailstock. It seemed to work well for him and gave him the most versatility. With an engine hoist, swapping "modes" was pretty quick.

  11. Likes SVFeingold, Jashley73 liked this post
  12. #88
    Join Date
    Sep 2018
    Country
    UNITED STATES
    State/Province
    California
    Posts
    656
    Post Thanks / Like
    Likes (Given)
    1243
    Likes (Received)
    317

    Default

    Quote Originally Posted by SVFeingold View Post
    Jaws on table...you mean something like this?

    2020-03-08-23_51_39-large-vtl-google-search.jpg

    I'm envisioning creating a larger subplate with radial slots. Then clamp the outside of the part. I suppose as long as the trunion can swing it and it doesn't bash into anything you're golden.

    Speaking of bashing into things, what's the cheapest reasonable verification tool? The UMC offers up the exciting opportunity to smack the spindle into the trunion. I'd like to avoid that...

    I've done a little digging into Vericut (lots of money), NCSimul (slightly less money), and I'm not really sure what else is out there that can show me precisely what the mill will do. Will these verification programs also show you when the mill decides to unwind a rotary axis in an unexpected (potentially crashy) way, or is that just up to the controller.

    EDIT: More questions! Sorry I just keep going on. Looking at that 20 arcsecond accuracy figure, say I needed to hold a very tight parallelism. Maybe half of what the rotary accuracy guarantees. Is there typically a way to fine tune the 0° position of the trunion without doing a whole recalibration? Maybe sweep an indicator around from the spindle, or use a spindle probe, to pick up several locations on the platter and then nudge the trunion and lock it down when it's as close to 0° as possible?
    for verification, you have Vericut, camplete and ncsimul (no personal experience with the last one) for 3rd party stuff, and then most cam systems will have their own machine simulation. we have both camplete and hypermill sim. i've yet to see either of them make a mistake. gotten to the point where i program a part, if it verifies good, i let it rip.

  13. Likes cameraman, bryan_machine, SVFeingold liked this post
  14. #89
    Join Date
    Aug 2015
    Country
    UNITED STATES
    State/Province
    California
    Posts
    140
    Post Thanks / Like
    Likes (Given)
    137
    Likes (Received)
    52

    Default

    Quote Originally Posted by cameraman View Post
    So correctability of angular errors probe wise depends on backlash and minimum sensitivity of movement with the drive mechanisms of the table.
    Interesting stuff for sure. So taking a more holistic system-level view, would you expect to get better accuracy/finish on say a 6" cube part in a dedicated 5ax machine w/ 20 arcsecond accuracy, or on a 5th axis rotary with ~5-10 arcsecond accuracy?

    For what it's worth none of the tables I've looked at - except for the DDR tables which seem much bulkier and more expensive - will promise that number. Hell even the Nikken tables spec a 60 arcsecond accuracy on the tilt, despite a post I read on here saying they measure more like 5 arcseconds.

    If anyone has real data on the Haas UMC that would be illuminating. Maybe the 20 arcseconds is a worst-case spec but they actually gauge much better in use?

    I'd further wager that for the best finishes repeatability is more important than accuracy. I haven't seen an official repeatability spec for the UMC but I probably haven't looked hard enough. In accuracy terms 20 arcseconds is plenty for the type of work I plan on doing. I certainly don't need to hold form tolerances that exceed that, but repeatability is what should play a bigger role in achieving good finishes. Is that true?

    EDIT: Unrelated thought: it seems to be that the way the UMC is configured, it should be better able to resist damage from abrasives getting into the bearings. The rotary axes are sealed, and rotary seals will be better at preventing egress of abrasive particles. The linear ways are all covered and way above the work so they won't have coolant circulating around them, plotting their demise by depositing whatever is floating in it. This will only matter for the 2 times a year I might consider doing some light grinding (say those aforementioned O-ring grooves). Filtering coolant is another story, but that aside am I wrong in thinking that the UMC would be better able to resist abrasive damage than a standard VMC?

    EDIT 2: Interesting notes from Haas about the UMC750:

    It is important to remember the Linear accuracy of the machine is 0.0004/10" and the angular accuracy is +/- 15 arc. sec. These tolerances can add up quickly.
    10um linear accuracy isn't fantastic, but it's enough. More accuracy notes here. Of particular note to me:

    When circular interpolating a hole at low feed rates, you can expect a circularity error of up to 0.001 inch (0.025 mm) when the program has a finish pass.
    On a UMC-750, when using a drill from two sides of a part you can get up to 0.002 inch (0.051 mm) error per side.

    This can add up to a total of 0.004 inch (0.102 mm) mismatch between the two drilled holes.

    Use a probe to get separate offsets for the B90 C0 and B90 C180 side. This can reduce this mismatch error.
    This is also not fantastic, but I guess livable. I wish they gave some bounds on what's possible with an additional probing cycle. There is no date on article so I'm not sure if it applies to the current generation. It's also not clear what size part this mismatch is on.

  15. #90
    Join Date
    May 2018
    Country
    UNITED STATES
    State/Province
    New York
    Posts
    138
    Post Thanks / Like
    Likes (Given)
    109
    Likes (Received)
    79

    Default

    Quote Originally Posted by SVFeingold View Post
    EDIT: Unrelated thought: it seems to be that the way the UMC is configured, it should be better able to resist damage from abrasives getting into the bearings. The rotary axes are sealed, and rotary seals will be better at preventing egress of abrasive particles. The linear ways are all covered and way above the work so they won't have coolant circulating around them, plotting their demise by depositing whatever is floating in it. This will only matter for the 2 times a year I might consider doing some light grinding (say those aforementioned O-ring grooves). Filtering coolant is another story, but that aside am I wrong in thinking that the UMC would be better able to resist abrasive damage than a standard VMC?
    How good is thermal comp/management on the umc, this might affect accuracy and repeatability.

  16. #91
    Join Date
    Aug 2015
    Country
    UNITED STATES
    State/Province
    California
    Posts
    140
    Post Thanks / Like
    Likes (Given)
    137
    Likes (Received)
    52

    Default

    Quote Originally Posted by Jmaks View Post
    How good is thermal comp/management on the umc, this might affect accuracy and repeatability.
    For what it's worth, the space will be climate controlled and I'd plan on eventually adding a coolant chiller to maintain the temperature as well as possible. I'm not sure if coolant is circulated around the spindle on the UMC500 since a spindle chiller isn't offered as an option, but a chiller is an easy thing to add aftermarket.

  17. #92
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    3,398
    Post Thanks / Like
    Likes (Given)
    3859
    Likes (Received)
    978

    Default

    Quote Originally Posted by SVFeingold View Post
    Interesting stuff for sure. So taking a more holistic system-level view, would you expect to get better accuracy/finish on say a 6" cube part in a dedicated 5ax machine w/ 20 arcsecond accuracy, or on a 5th axis rotary with ~5-10 arcsecond accuracy?

    For what it's worth none of the tables I've looked at - except for the DDR tables which seem much bulkier and more expensive - will promise that number. Hell even the Nikken tables spec a 60 arcsecond accuracy on the tilt, despite a post I read on here saying they measure more like 5 arcseconds.

    If anyone has real data on the Haas UMC that would be illuminating. Maybe the 20 arcseconds is a worst-case spec but they actually gauge much better in use?

    I'd further wager that for the best finishes repeatability is more important than accuracy. I haven't seen an official repeatability spec for the UMC but I probably haven't looked hard enough. In accuracy terms 20 arcseconds is plenty for the type of work I plan on doing. I certainly don't need to hold form tolerances that exceed that, but repeatability is what should play a bigger role in achieving good finishes. Is that true?
    Today's a busy day so I can't do justice to this within the time I have...

    But just quick pass off the top of my head and will maybe edit accordingly later.

    Depends on the specific machine.

    Personally I prefer a dedicated 5 axis machine to a trunnion mounted onto a 3 axis vertical for a many many reasons.

    IF memory serves me right the HAAS UMC may state +/- 20 arc seconds ~ 40 arc second spread (I'm guessing to two sigma). One Sigma 20 arc second ish possibly 10 arc seconds from move to move.

    These quoted values are static figures and I much prefer to see the whole rotational plots. Unidirectional positioning versus bi-directional positioning.

    Sometimes MTBs quote only single static value small angle unidirectional moves... To appear more "Accurate" and precise.

    Most Universals out there can position to +/- 8 arc second @ 1 sigma.

    So in spite of ALLLLLLL those numbers I would expect the HAAS UMC 500 SS (for example and even the NEW UMC 750 reboot) to contour better / much better than any of their current 5 axis trunnion offerings.

    Accuracy as such is referenced to the machine initial homing whereas repeatability is what I look at as your first reference plane that will BE produced as result of what you cut as your first reference plane (or features ) to the XYZ axes of the spindle. So with probing and Dynamic Work offsets (for 5 axis) (+ Work piece error setting compensation ) or equivalent, the second and third rotations will be referenced to that first cut set of reference features or plane (or should be) - Depending on your scheme. That's why high precision rotary scales can be really important as well as to map out in cut deviations from commanded position to actual position. There are physical torques in cut against the driven rotary axes dynamically. Your finishing cuts should be light and in-line with commanded positions and tool paths / tool wear compensation(s).

    Some of these ultra precise machines are such so they better enable automation where CpKs etc. are really critical and operators are not standing at the machine to make subtle tweaks all the time to keep things on track.

    Those errors really stack up in an automation scenario ~ More hands on prototype work you can compensate a lot fairly artfully as long as the machine can cut really straight along one axis (that's what I look for) . Some of the Matsuura machines can be quite sweet for that + very low spindle runout … But that's maybe not what you are looking for.

    "These are not the Matsuuras you are looking for... " (hand wave gesture)..

    "Move along, move along ."

  18. #93
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    3,398
    Post Thanks / Like
    Likes (Given)
    3859
    Likes (Received)
    978

    Default

    Quote Originally Posted by SVFeingold View Post
    Interesting stuff for sure. So taking a more holistic system-level view, would you expect to get better accuracy/finish on say a 6" cube part in a dedicated 5ax machine w/ 20 arcsecond accuracy, or on a 5th axis rotary with ~5-10 arcsecond accuracy?

    For what it's worth none of the tables I've looked at - except for the DDR tables which seem much bulkier and more expensive - will promise that number. Hell even the Nikken tables spec a 60 arcsecond accuracy on the tilt, despite a post I read on here saying they measure more like 5 arcseconds.

    If anyone has real data on the Haas UMC that would be illuminating. Maybe the 20 arcseconds is a worst-case spec but they actually gauge much better in use?

    I'd further wager that for the best finishes repeatability is more important than accuracy. I haven't seen an official repeatability spec for the UMC but I probably haven't looked hard enough. In accuracy terms 20 arcseconds is plenty for the type of work I plan on doing. I certainly don't need to hold form tolerances that exceed that, but repeatability is what should play a bigger role in achieving good finishes. Is that true?

    EDIT: Unrelated thought: it seems to be that the way the UMC is configured, it should be better able to resist damage from abrasives getting into the bearings. The rotary axes are sealed, and rotary seals will be better at preventing egress of abrasive particles. The linear ways are all covered and way above the work so they won't have coolant circulating around them, plotting their demise by depositing whatever is floating in it. This will only matter for the 2 times a year I might consider doing some light grinding (say those aforementioned O-ring grooves). Filtering coolant is another story, but that aside am I wrong in thinking that the UMC would be better able to resist abrasive damage than a standard VMC?
    As to your second part … I'll come back lol very interesting _ I really have sh*t to do for once ;-)

    Ohhh f*ck it... OK

    Grinding is very high on my list and I figured it would be fine to destroy HAAS's "economically"

    I think the matsuura's are well set up for grinding and so are some of the top flight DMG Mori offerings.

    Here's the reality IF you are doing high precision work then surface grinders , older jig grinders and even old jig boring machines + HAAS machines + a few custom build hydraulic spindle / air spindle machines can still work out cheaper than the MAKINO equivalent to achieve the same or similar result... ~ Comes back to automation.

    Hermle are not grinding friendly.

    Using the "Host spindle" for grinding need to check out spindle runout …

    Filtration is a very BIG deal and seems HAAS has these new filter bags... The DMG Mori machines (in some instances have some really awesome filtration equipment so grinding won't necessarily damage the spindle (thru spindle coolant vs. labyrinth seals ).

  19. Likes SVFeingold liked this post
  20. #94
    Join Date
    Aug 2015
    Country
    UNITED STATES
    State/Province
    California
    Posts
    140
    Post Thanks / Like
    Likes (Given)
    137
    Likes (Received)
    52

    Default

    Quote Originally Posted by cameraman View Post
    Some of the Matsuura machines can be quite sweet for that + very low spindle runout … But that's maybe not what you are looking for.

    "These are not the Matsuuras you are looking for... " (hand wave gesture)..

    "Move along, move along ."
    By all means...I would love an Okuma 460-5AX, or any of the Makino/Matsuura offerings. I just flat out can't afford them. My max budget for this including running power, workholding, compressor, phase converter, initial tooling, taxes, etc. is 175k. Maybe, maybe that can stretch to 200k if I am more confident I can get paid work or I can get someone else to go in on this with me. Anything north of that really isn't a viable option, and trying to go that route means an additional year or two of saving, which is another 1-2 years of my life where I'm not able to do the things I want to do. Hell ask my friends, I've been talking about getting a mill for years. I specifically only looked at houses where I could fit a mill. Found the tallest garage door I could. Time to pull the trigger, waiting is not worth it. Maybe the Haas can make enough money in those 2 years to buy one later anyway, who knows.

  21. #95
    Join Date
    Sep 2018
    Country
    UNITED STATES
    State/Province
    California
    Posts
    656
    Post Thanks / Like
    Likes (Given)
    1243
    Likes (Received)
    317

    Default

    Quote Originally Posted by cameraman View Post
    As to your second part … I'll come back lol very interesting _ I really have sh*t to do for once ;-)

    Ohhh f*ck it... OK

    Grinding is very high on my list and I figured it would be fine to destroy HAAS's "economically"

    I think the matsuura's are well set up for grinding and so are some of the top flight DMG Mori offerings.

    Here's the reality IF you are doing high precision work then surface grinders , older jig grinders and even old jig boring machines + HAAS machines + a few custom build hydraulic spindle / air spindle machines can still work out cheaper than the MAKINO equivalent to achieve the same or similar result... ~ Comes back to automation.

    Hermle are not grinding friendly.

    Using the "Host spindle" for grinding need to check out spindle runout …

    Filtration is a very BIG deal and seems HAAS has these new filter bags... The DMG Mori machines (in some instances have some really awesome filtration equipment so grinding won't necessarily damage the spindle (thru spindle coolant vs. labyrinth seals ).
    get your ass back to work! lol

  22. Likes SVFeingold, cameraman liked this post
  23. #96
    Join Date
    Aug 2010
    Location
    Medina OH
    Posts
    2,252
    Post Thanks / Like
    Likes (Given)
    281
    Likes (Received)
    1088

    Default

    Quote Originally Posted by SVFeingold View Post
    By all means...I would love an Okuma 460-5AX, or any of the Makino/Matsuura offerings. I just flat out can't afford them. My max budget for this including running power, workholding, compressor, phase converter, initial tooling, taxes, etc. is 175k. Maybe, maybe that can stretch to 200k if I am more confident I can get paid work or I can get someone else to go in on this with me. Anything north of that really isn't a viable option, and trying to go that route means an additional year or two of saving, which is another 1-2 years of my life where I'm not able to do the things I want to do. Hell ask my friends, I've been talking about getting a mill for years. I specifically only looked at houses where I could fit a mill. Found the tallest garage door I could. Time to pull the trigger, waiting is not worth it. Maybe the Haas can make enough money in those 2 years to buy one later anyway, who knows.
    Wait, I thought I read something about the west coast, but your post says Georgia? It has been a long time since I have been to Georgia, but like Ohio, there ought to be a pole building about every 1/4 mile that will fit just about anything you want?

    When looking at 5ax trunnions, I just never found anything worthwhile. I wasn't even worried about the accuracy levels you guys are talking. Price wise it just wasn't worth it.

    For example, I looked at a 5ax to put in my S1000. Nikken quoted something like $40k. For $150k ISH I can buy a Brother M140.

    Which might be my next machine, since I can get to 5 sides of a part.

    Have you really sat down and looked at the accuracy required by your parts? I realize you are doing development/prototyping, but have you REALLY considered what 20 arc seconds means for your parts?

    My point is, super nice, super accurate machines certainly are easier to make parts on. But EVERYTHING I have ever run, has required tweaking, adjusting, workarounds to some degree or another.

    If the parts you are potentially prototyping are going to be difficult to make with a machine that STATES .001" error (based on the article you referenced) then you need a better machine, if your budget does not allow for that machine... it is time to reconsider.

    I know that I cannot hit a .0001 tolerance on the first shot with my Brother. I program accordingly, and work into my final dimension. I RARELY need to hold a few tenths. My budget, and work, does not provide for a Maats or Hermle. So when the requirements get tight, I have to work with the machine. I've gotten some pretty DAMN close accurate work out of my Brother, the closest of which that I measured, was within .0002" flat, perpendicular, and ROUND for something like an 8" disk. Not super precision, but I was pretty happy with it.

    Hell, I grind everything that comes through my shop on a $300 surface grinder that I bought almost 10 years ago. I can hit a .0001" with it, and when I need to hold a total tolerance of .0002" it does take some work. But my work just doesn't justify a new/nicer grinder.

    If all I could have was one machine, and I had to get rid of my lathes and mills and grinders, I would cry, die a little inside, and then probably buy a Brother M140. I could turn, I could mill 3+2, and maybe, after talking to Brother a little, I could probably do a little form grinding on it. Would it be a perfect all around machine, NO. But for a VERY capable, versatile, and pretty ACCURATE, for what I think is a STUPID VALUE

  24. Likes empwoer, eaglemike, Pete Deal, Jmaks, SRT Mike and 1 others liked this post
  25. #97
    Join Date
    Jun 2006
    Location
    Near Seattle
    Posts
    4,980
    Post Thanks / Like
    Likes (Given)
    3445
    Likes (Received)
    1424

    Default

    If looking at a UMC (or any other modern machine) you want to search out at least the following 3 things.
    1. A first rate probe.
    2. Some kind of kinematic adjustment - what I've seen is a routine where you mount a tooling ball at known spot, this gets moved about and probed, and it's movement used to adjust machine kinematics. At least for a C-axis table you put the ball as far out on the edge as you can. I don't know what features like this haas, brother, or anybody other than DMG/Heidenhain has - this is a thing to look into.
    3. Probing routines that can be called from a program, and used to adjust for position. Figure you'll spend time learning to do this too.

  26. Likes cameraman liked this post
  27. #98
    Join Date
    Sep 2018
    Country
    UNITED STATES
    State/Province
    California
    Posts
    656
    Post Thanks / Like
    Likes (Given)
    1243
    Likes (Received)
    317

    Default

    Quote Originally Posted by bryan_machine View Post
    If looking at a UMC (or any other modern machine) you want to search out at least the following 3 things.
    1. A first rate probe.
    2. Some kind of kinematic adjustment - what I've seen is a routine where you mount a tooling ball at known spot, this gets moved about and probed, and it's movement used to adjust machine kinematics. At least for a C-axis table you put the ball as far out on the edge as you can. I don't know what features like this haas, brother, or anybody other than DMG/Heidenhain has - this is a thing to look into.
    3. Probing routines that can be called from a program, and used to adjust for position. Figure you'll spend time learning to do this too.
    all of those are included with UMC500/750/1000

  28. Likes gkoenig, bryan_machine liked this post
  29. #99
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    3,398
    Post Thanks / Like
    Likes (Given)
    3859
    Likes (Received)
    978

    Default

    Quote Originally Posted by Fal Grunt View Post
    <snip all really excellent advice>

    If all I could have was one machine, and I had to get rid of my lathes and mills and grinders, I would cry, die a little inside, and then probably buy a Brother M140. I could turn, I could mill 3+2, and maybe, after talking to Brother a little, I could probably do a little form grinding on it. Would it be a perfect all around machine, NO. But for a VERY capable, versatile, and pretty ACCURATE, for what I think is a STUPID VALUE
    I understand / hear you with a "Die a little" getting rid of the "therapy machines" .

    A One-machine "everything machine..." ~ Integrex I-100 or I-200 but with the ability to crane in a 3 axis table.

    In other words a mill turn machine with preferably a B axis tilting head , travelling column machine with a 12" in Y 50" to 60" in X table. OR something like spinner or HAAS UMC Duo 1500 that got shrunk in the wash ~ So I could do dedicated five axis grapefruit / soccer ball sized parts on the right side of the machine (knuckle trunnion embedded into casting "wall/bridge" ), but also mill longer "skinny" 4th axis and 3 axis type parts. + Some grinding capability in terms of low runout spindles high RPM / use air spindles and positive pressure seals and filtration. ~ Preferably synthetic granite... Don't want much lol. [B axis mill head would remove one trunnion movement or vice versa fixed head with two rotary axis knuckle-trunnion.]

  30. #100
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    3,398
    Post Thanks / Like
    Likes (Given)
    3859
    Likes (Received)
    978

    Default

    Seems like HAAS is really killing it these days for 5 axis entry level...

    @SVFeingold so what's the plan if the UMCs can't get through your garage door ?

    VF2, VF3 ?

    Speedio/ Brother ?


Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •