CAM vs Hand Coding - Page 7
Close
Login to Your Account
Page 7 of 8 FirstFirst ... 5678 LastLast
Results 121 to 140 of 153
  1. #121
    Join Date
    May 2017
    Country
    UNITED STATES
    State/Province
    Minnesota
    Posts
    2,006
    Post Thanks / Like
    Likes (Given)
    2739
    Likes (Received)
    1402

    Default

    Quote Originally Posted by SeymourDumore View Post
    OK, so tell me what CAM program can generate the following for say a pocket:
    In Mastercam, and likely many others, if you want your post setup to be able to do that, you can. You can make it behave exactly however you want; the post is a touring complete programming language. People write in probing routines that they can access as custom drill cycles.

  2. #122
    Join Date
    Jan 2014
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    5,701
    Post Thanks / Like
    Likes (Given)
    1009
    Likes (Received)
    3215

    Default

    I'm pretty sure the guys that believe their CAM system can generate code that will run as fast as hand written or hand edited CAM have never really tried it.

    In the case of the tier1 auto supplier that I did some work for they had very advanced CAD-CAM and folks who were not dumb to run it. In all production cases that I saw, they ran hand coded or hand edited CAM programs to eke out all possible speed. I'm quite certain that were they able to post CAM code that was just as fast to run, they would have done it.

    The particular project I helped on was a die-casting about 2.5" diameter about 3.5" long with about a 6" diameter flange on the end. Had to face the flange and face groove it for an o-ring. Then 2 different diameter bores with snap ring grooves and then turn a bearing fit diameter on a cast boss inside the bottom of the bore and snap ring groove on it. All in a sub 30 second cycle time.

    For this to work, there was an absolute bare minimum of code. Not one single extra character or line. No "safety lines". No safe restart points. If something happened and the cycle had to be stopped, the machine was homed, the part tossed and replaced and the cycle started from the top. Tools were all PCD. The bearing fit was periodically air gaged and results graphed. The operator was not allowed to change an offset until a dimension was within 20% of edge of the tolerance band.

  3. Likes SeymourDumore liked this post
  4. #123
    Join Date
    Apr 2018
    Country
    UNITED STATES MINOR OUTLYING ISLANDS
    Posts
    8,858
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    4293

    Default

    Quote Originally Posted by Vancbiker View Post
    The operator was not allowed to change an offset until a dimension was within 20% of edge of the tolerance band.
    I kinda learned that same thing the hard way. If you chase size too hard, you're going to be chasing it always, every single part. But if you let it squinch around a little, there's a lot less offset-chasing. And the parts are just as good.

  5. Likes Ox, Bobw, doug925 liked this post
  6. #124
    Join Date
    Jan 2015
    Country
    UNITED STATES
    State/Province
    Missouri
    Posts
    1,510
    Post Thanks / Like
    Likes (Given)
    229
    Likes (Received)
    958

    Default

    Quote Originally Posted by Vancbiker View Post
    I'm pretty sure the guys that believe their CAM system can generate code that will run as fast as hand written or hand edited CAM have never really tried it.

    In the case of the tier1 auto supplier that I did some work for they had very advanced CAD-CAM and folks who were not dumb to run it. In all production cases that I saw, they ran hand coded or hand edited CAM programs to eke out all possible speed. I'm quite certain that were they able to post CAM code that was just as fast to run, they would have done it.

    The particular project I helped on was a die-casting about 2.5" diameter about 3.5" long with about a 6" diameter flange on the end. Had to face the flange and face groove it for an o-ring. Then 2 different diameter bores with snap ring grooves and then turn a bearing fit diameter on a cast boss inside the bottom of the bore and snap ring groove on it. All in a sub 30 second cycle time.

    For this to work, there was an absolute bare minimum of code. Not one single extra character or line. No "safety lines". No safe restart points. If something happened and the cycle had to be stopped, the machine was homed, the part tossed and replaced and the cycle started from the top. Tools were all PCD. The bearing fit was periodically air gaged and results graphed. The operator was not allowed to change an offset until a dimension was within 20% of edge of the tolerance band.
    What you are talking about is a single part that runs it’s whole life. Let’s talk about every other job that comes into our shop. There is no way you can compete with a good program that use all high speed tool paths and dynamic paths. The tooling costs that it saves easily pays for a machine each month. thinking hand coding is the way to go is crazy. I have a machine running AR uppers for 4 months straight and have replace about 3 tools the whole time. These new tool paths have lots of code but very impressive what they can do
    Don


    Sent from my iPhone using Tapatalk Pro

  7. Likes Mike1974 liked this post
  8. #125
    Join Date
    Mar 2013
    Country
    UNITED STATES
    State/Province
    Oregon
    Posts
    2,242
    Post Thanks / Like
    Likes (Given)
    924
    Likes (Received)
    2898

    Default

    Quote Originally Posted by Vancbiker View Post
    In the case of the tier1 auto supplier that I did some work for they had very advanced CAD-CAM and folks who were not dumb to run it. In all production cases that I saw, they ran hand coded or hand edited CAM programs to eke out all possible speed. I'm quite certain that were they able to post CAM code that was just as fast to run, they would have done it.

    The particular project I helped on was a die-casting about 2.5" diameter about 3.5" long with about a 6" diameter flange on the end. Had to face the flange and face groove it for an o-ring. Then 2 different diameter bores with snap ring grooves and then turn a bearing fit diameter on a cast boss inside the bottom of the bore and snap ring groove on it. All in a sub 30 second cycle time.
    So, I think this absolutely depends on the complexity of a part and the environment.

    I've never done automotive work, by my impression is that cycle time is absolutely life in this world, and the reliance on very high quality net-shape methods (forging, die casting, molding) really leaves machining as very secondary processes for critical dimensions/surfaces. If you're just turning a simple profile, or PCD face-milling a mating surface, and drilling/tapping some holes... Sure, why not hand code? Your programs are probably ~100 lines. Nothing is ever easy at that level, but this is manageable and most of the job is in the rest of the process (fixtures, automation, QC, documentation, custom tooling development, etc etc).

    My experience is in consumer/outdoor products (read: knives, firearms, etc). Volumes are healthy, but not so big that these companies are willing to invest mid 6 figures in near-net tooling, and even if they did, consumer demand at the mid-high end of the market doesn't want metal products with as-cast surfaces (especially not to compete against the Chinese). These parts often have a lot of machining - 3+ minute cycle times, signifiant material removal (over 50%), complex profiles that cannot be easily hand coded, and we're starting to see trends towards simultaneous multi-axis surfaces. Finger CAM either does not make sense for these parts, or we're dealing with geometry that simply cannot be coded by hand.

    IDK how anyone else does work like this, but I like to crawl/walk/run. Crawl - figure out the workholding, cut strategy, and tooling. Walk - harden the fixtures, start pushing tools, start running off to capture CPK, tool life, etc etc. Run - turn off safeties, cut entry/exit motions to the bone (never zero though), etc. My value add isn't that I'm a fucking rocket scientist level machinist, but I own a production class machine and I am pretty nerdy with how to make it do things... so I am cutting parts rapidly and constantly through this process.

    Personally, I never ever hand code anything - I want to be able to open the NX file, hit Post Process and have 100% prduction code. Why? These markets are nowhere near as stable as automotive or Apple, where a print change requires a Fortune 500 Board of Directors meeting. Processes change, downstream stuff changes and the upstream parts need a tweak to accommodate, etc etc. Hand edits also add risky tribal knowledge; will I be around in 2 years to make them? Will anyone managing the process implement them? Probably not.

    Look, if you want to get anywhere as a professional machinist, you better know G-Code, but that knowledge is usually learned within deeper context, and often comes along with learning the languages behind common post processors (Java, XML, TCL, etc). G-code, knowing the guts of a major CAM system, and knowing how to bend it's post processor language to your will are absolutely critical skills...

    But they are irrelevant to Machinist 101 level folks, and I think focusing on these skills (often at the introduction of these courses) is totally gods damn stupid. Classroom stuff needs to be deep in CAD/CAM from Day 1, and everyone in the class should walk out with a (very) simple part they set-up and machined in their first shop day. Then the course should progressively add complexity across that workflow.

  9. Likes metalmadness, Winterfalke liked this post
  10. #126
    Join Date
    Jan 2014
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    5,701
    Post Thanks / Like
    Likes (Given)
    1009
    Likes (Received)
    3215

    Default

    Quote Originally Posted by D Nelson View Post
    What you are talking about is a single part that runs it’s whole life. Let’s talk about every other job that comes into our shop. There is no way you can compete with a good program that use all high speed tool paths and dynamic paths. The tooling costs that it saves easily pays for a machine each month. thinking hand coding is the way to go is crazy. I have a machine running AR uppers for 4 months straight and have replace about 3 tools the whole time. These new tool paths have lots of code but very impressive what they can do
    Don


    Sent from my iPhone using Tapatalk Pro
    Absolutely not saying that all jobs are best hand coded, just that in certain situations it has a place.

  11. #127
    Join Date
    Jan 2014
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    5,701
    Post Thanks / Like
    Likes (Given)
    1009
    Likes (Received)
    3215

    Default

    Quote Originally Posted by gkoenig View Post
    So, I think this absolutely depends on the complexity of a part and the environment.
    Absolutely correct.

    Quote Originally Posted by gkoenig View Post
    I've never done automotive work, by my impression is that cycle time is absolutely life in this world, and the reliance on very high quality net-shape methods (forging, die casting, molding) really leaves machining as very secondary processes for critical dimensions/surfaces. If you're just turning a simple profile, or PCD face-milling a mating surface, and drilling/tapping some holes... Sure, why not hand code? Your programs are probably ~100 lines. Nothing is ever easy at that level, but this is manageable and most of the job is in the rest of the process (fixtures, automation, QC, documentation, custom tooling development, etc etc).
    Again correct.

    Quote Originally Posted by gkoenig View Post
    My experience is in consumer/outdoor products (read: knives, firearms, etc). Volumes are healthy, but not so big that these companies are willing to invest mid 6 figures in near-net tooling,.....
    Quote Originally Posted by gkoenig View Post
    Personally, I never ever hand code anything - I want to be able to open the NX file, hit Post Process and have 100% prduction code. Why? These markets are nowhere near as stable as automotive or Apple, .....
    My experience with consumer/outdoor products is that the product lifespan is relatively short compared to say auto drivetrain components. The part I used in my example ran for at least 7 years unchanged. Last year's popular knife is this years old news. Consumer products have to change to keep consumer interest high. Nobody cares that the same wheel hub has been used on nearly every GM FWD car for the past 5 plus years. As long as the body, interior, and infotainment system gets a little update regularly, buyers stay interested.

  12. Likes gkoenig liked this post
  13. #128
    Join Date
    Oct 2009
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    1,527
    Post Thanks / Like
    Likes (Given)
    1052
    Likes (Received)
    849

    Default

    Quote Originally Posted by Nagol View Post
    From the conversational guy I don't see contours I can't do in the part pictured above. I can't contour it with a ballnose and 15K lines of code but I sure as hell can index a horizontal mill and cut it from the side. My view on some of that stuff is guys getting more out of their machine with possibly a easier set-up. Not saying everything can be done this way but I see a lot of stuff stitched that could be done faster/ more accurate with the proper machine/ fixture. Can't beat facing a part at an angle or finishing with the side of an endmill with .01-.02 step overs. Guys used to make some pretty complex shit before we had access to all this stuff. I think engineers have gotten a little out of hand with some of it. Probably could do away with a lot of the features that require the software and design it to be made with reasonable means. Attachment 329036 Attachment 329037 Attachment 329035 First 2 pics were designed and made on a Mazak horizontal in conversational. Last is a customer part. Kind of off the main topic but I sometime think these guys couldn't machine themselves out of a paper bag without the almighty CAM. Anyone I've spoken to that is familiar with code says it's necessary while guys who don't seem to know think they are getting by but they are likely leaving a lot on the table.
    The kinds of things that need to be programmed with CAM that puts out lots of lines of code are things with surfaces that aren't flat or basic non-flat shapes. 15,000 lines of code is pretty small in some instances. I've written programs for wakeboard and water ski molds that ranged between 30 and 40 megabytes once posted. As mentioned previously, there's no one-size-fits-all solution to this topic.

  14. Likes Ox, metalmadness, mhajicek, Winterfalke liked this post
  15. #129
    Join Date
    Nov 2015
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    481
    Post Thanks / Like
    Likes (Given)
    94
    Likes (Received)
    270

    Default

    Quote Originally Posted by gkoenig View Post
    So, I think this absolutely depends on the complexity of a part and the environment.

    I've never done automotive work, by my impression is that cycle time is absolutely life in this world, and the reliance on very high quality net-shape methods (forging, die casting, molding) really leaves machining as very secondary processes for critical dimensions/surfaces. If you're just turning a simple profile, or PCD face-milling a mating surface, and drilling/tapping some holes... Sure, why not hand code? Your programs are probably ~100 lines. Nothing is ever easy at that level, but this is manageable and most of the job is in the rest of the process (fixtures, automation, QC, documentation, custom tooling development, etc etc).

    My experience is in consumer/outdoor products (read: knives, firearms, etc). Volumes are healthy, but not so big that these companies are willing to invest mid 6 figures in near-net tooling, and even if they did, consumer demand at the mid-high end of the market doesn't want metal products with as-cast surfaces (especially not to compete against the Chinese). These parts often have a lot of machining - 3+ minute cycle times, signifiant material removal (over 50%), complex profiles that cannot be easily hand coded, and we're starting to see trends towards simultaneous multi-axis surfaces. Finger CAM either does not make sense for these parts, or we're dealing with geometry that simply cannot be coded by hand.

    IDK how anyone else does work like this, but I like to crawl/walk/run. Crawl - figure out the workholding, cut strategy, and tooling. Walk - harden the fixtures, start pushing tools, start running off to capture CPK, tool life, etc etc. Run - turn off safeties, cut entry/exit motions to the bone (never zero though), etc. My value add isn't that I'm a fucking rocket scientist level machinist, but I own a production class machine and I am pretty nerdy with how to make it do things... so I am cutting parts rapidly and constantly through this process.

    Personally, I never ever hand code anything - I want to be able to open the NX file, hit Post Process and have 100% prduction code. Why? These markets are nowhere near as stable as automotive or Apple, where a print change requires a Fortune 500 Board of Directors meeting. Processes change, downstream stuff changes and the upstream parts need a tweak to accommodate, etc etc. Hand edits also add risky tribal knowledge; will I be around in 2 years to make them? Will anyone managing the process implement them? Probably not.

    Look, if you want to get anywhere as a professional machinist, you better know G-Code, but that knowledge is usually learned within deeper context, and often comes along with learning the languages behind common post processors (Java, XML, TCL, etc). G-code, knowing the guts of a major CAM system, and knowing how to bend it's post processor language to your will are absolutely critical skills...

    But they are irrelevant to Machinist 101 level folks, and I think focusing on these skills (often at the introduction of these courses) is totally gods damn stupid. Classroom stuff needs to be deep in CAD/CAM from Day 1, and everyone in the class should walk out with a (very) simple part they set-up and machined in their first shop day. Then the course should progressively add complexity across that workflow.
    Very good post. One thing that the hardline hand coders are missing is that most shops aren't running 10K production. There are some large batch volume machining companies that specialize in that type of automation and there are certainly some smaller job-shop type outfits that have a few contracts for tens of thousands of parts as well. My first company was one of those, they were basically a job shop but had some DoD contracts where they needed 10s-100s of thousands. Fine, but those aren't going to be "complex" parts. TO be sure, there are some shops that run "production" on complex parts - a company I visited did just that. Extremely complex aerospace parts in the hundreds and thousands. They still ran that in CAM though.

    Quote Originally Posted by Nagol View Post
    From the conversational guy I don't see contours I can't do in the part pictured above. I can't contour it with a ballnose and 15K lines of code but I sure as hell can index a horizontal mill and cut it from the side. My view on some of that stuff is guys getting more out of their machine with possibly a easier set-up. Not saying everything can be done this way but I see a lot of stuff stitched that could be done faster/ more accurate with the proper machine/ fixture. Can't beat facing a part at an angle or finishing with the side of an endmill with .01-.02 step overs. Guys used to make some pretty complex shit before we had access to all this stuff. I think engineers have gotten a little out of hand with some of it. Probably could do away with a lot of the features that require the software and design it to be made with reasonable means. Attachment 329036 Attachment 329037 Attachment 329035 First 2 pics were designed and made on a Mazak horizontal in conversational. Last is a customer part. Kind of off the main topic but I sometime think these guys couldn't machine themselves out of a paper bag without the almighty CAM. Anyone I've spoken to that is familiar with code says it's necessary while guys who don't seem to know think they are getting by but they are likely leaving a lot on the table.
    This is true - for guys without a solid CAM background or setup, conversational is a very solid tool. Programming on Mazatrol or Hurco Winmax is a great alternative to CAM, even doing surfacing. But what that lacks is ability to do 5-axis contouring. Just can't do it. 15K lines was used as an example of a single roughing toolpath - my programs tip past 1,000,000 lines of motion all the time and even that isn't that big compared to what I have seen some people doing. When you start having Mastercam files that are 100MB+ then you know your working with some complicated shit. All that data doesn't sound like something I want to conversationally program.

    Yes you can hit sloped walls by hand coding a subroutine - Titan had a video talk about doing this exact thing and it was very slick and a great solution. Compound swept surfaces simply cannot be hand coded PERIOD. I don't care if you are a G-Code savant, you can't program this finishing op by hand:
    swept-surface.jpg

    It is very likely that the engineering of some of these complex surfaces has been exaggerated by engineers in the past decade and that is a problem, but the machines are plenty capable of machining them. "Just because you can doesn't mean you should" kinda things happening with part design these days. IMO the best way to harness a modern controller is with CAM software and no one on this forum is going to change my mind that hand coding is better outside of some select circumstances, which we have pretty well covered here in-depth.

    Regardless of where you land on the topic, it is pretty clear that
    A) understanding basic Gcode is an absolute necessity
    B) being able to write gcode programs is a huge boon to a skilled programmer
    C)Editing code can sometimes be necessary and knowing how to do it is very important.
    D) Post processors can do all (most) those pesky hand edits on the back end so you don't have to think about it.
    Attached Thumbnails Attached Thumbnails swept-surface.jpg  

  16. Likes mhajicek, charlie gary liked this post
  17. #130
    Join Date
    Apr 2018
    Country
    UNITED STATES MINOR OUTLYING ISLANDS
    Posts
    8,858
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    4293

    Default

    Quote Originally Posted by charlie gary View Post
    The kinds of things that need to be programmed with CAM that puts out lots of lines of code are things with surfaces that aren't flat or basic non-flat shapes. 15,000 lines of code is pretty small in some instances. I've written programs for wakeboard and water ski molds that ranged between 30 and 40 megabytes once posted.
    And if controls had progressed to more advanced interpolatin methods, like they were doing before the Japs took over, we would now have something better than 400 million dots to connect. TERA was right up by you ...

  18. #131
    Join Date
    Sep 2002
    Location
    People's Republic
    Posts
    6,324
    Post Thanks / Like
    Likes (Given)
    748
    Likes (Received)
    3808

    Default

    Some of my commercial work involves drilling lots and lots of 6-32s in aluminum. I have often wondered if there was time to be saved if rather than running the old 2 1/2 axis toolpaths if I ran it all angles and radius' in z so it never really stops at the top of the hole

    never tried it, too lazy

  19. Likes Bobw liked this post
  20. #132
    Join Date
    Mar 2013
    Country
    UNITED STATES
    State/Province
    Oregon
    Posts
    2,242
    Post Thanks / Like
    Likes (Given)
    924
    Likes (Received)
    2898

    Default

    Quote Originally Posted by gustafson View Post
    Some of my commercial work involves drilling lots and lots of 6-32s in aluminum. I have often wondered if there was time to be saved if rather than running the old 2 1/2 axis toolpaths if I ran it all angles and radius' in z so it never really stops at the top of the hole
    Hate to tell you, but there is *significant* time to be saved. Lots of variables involved, but we had a project to drill about 200 0.030" holes, about .01" deep to add a grip texture to a component. Very fast times per individual hole, and very tiny transitions - switching from G00 rapids between holes to max feed rate transitions using arcs shaved 30% off the cycle time. It helps that in a Speedio, I can turn off all the look-ahed, acceleration control interventions and let it rip; those numbers are nowhere near as impressive when the machine is working to control the accuracy of those transition arcs.

    Lasers won out in the end, but changed the nature of the project dramatically as they took anodizing off the table as a finishing option (they could never get the lasered sections to match the rest of the part for some reason that wasn't my puzzle to solve).

  21. Likes gustafson, mhajicek, Bobw, Winterfalke liked this post
  22. #133
    Join Date
    Sep 2002
    Location
    People's Republic
    Posts
    6,324
    Post Thanks / Like
    Likes (Given)
    748
    Likes (Received)
    3808

    Default

    Quote Originally Posted by gkoenig View Post
    Hate to tell you, but there is *significant* time to be saved. Lots of variables involved, but we had a project to drill about 200 0.030" holes, about .01" deep to add a grip texture to a component. Very fast times per individual hole, and very tiny transitions - switching from G00 rapids between holes to max feed rate transitions using arcs shaved 30% off the cycle time. It helps that in a Speedio, I can turn off all the look-ahed, acceleration control interventions and let it rip; those numbers are nowhere near as impressive when the machine is working to control the accuracy of those transition arcs.

    Lasers won out in the end, but changed the nature of the project dramatically as they took anodizing off the table as a finishing option (they could never get the lasered sections to match the rest of the part for some reason that wasn't my puzzle to solve).
    Well I may have to try it sometime.

    My DMG is pretty quick, not speedio quick but quick.

    Recently realized how slow the rigid tap is, bummed me out how much of they cycle time is just tapping

  23. #134
    Join Date
    Aug 2005
    Location
    CT
    Posts
    8,444
    Post Thanks / Like
    Likes (Given)
    464
    Likes (Received)
    2266

    Default

    Quote Originally Posted by gkoenig View Post
    Very fast times per individual hole, and very tiny transitions - switching from G00 rapids between holes to max feed rate transitions using arcs shaved 30% off the cycle time. .
    Not seeing what you mean by arcs ...


    '

  24. #135
    Join Date
    Oct 2009
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    1,527
    Post Thanks / Like
    Likes (Given)
    1052
    Likes (Received)
    849

    Default

    Quote Originally Posted by EmanuelGoldstein View Post
    And if controls had progressed to more advanced interpolatin methods, like they were doing before the Japs took over, we would now have something better than 400 million dots to connect. TERA was right up by you ...
    What is/was TERA? Never heard of them, and all the internet tells me is it's some kind of game.

  25. #136
    Join Date
    Oct 2009
    Location
    Oregon
    Posts
    6,311
    Post Thanks / Like
    Likes (Given)
    6743
    Likes (Received)
    3614

    Default

    Quote Originally Posted by SeymourDumore View Post
    Not seeing what you mean by arcs ...


    '
    Instead of hitting the brakes in Z, accelerating X/Y to a stop above the next hole and then accelerating Z- to drill you do it arc moves.

    Makes sense to me.

  26. Likes Bobw, Winterfalke, gustafson liked this post
  27. #137
    Join Date
    Nov 2015
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    481
    Post Thanks / Like
    Likes (Given)
    94
    Likes (Received)
    270

    Default

    Quote Originally Posted by Garwood View Post
    Instead of hitting the brakes in Z, accelerating X/Y to a stop above the next hole and then accelerating Z- to drill you do it arc moves.

    Makes sense to me.
    Definitely makes sense.

    Mastercam added this functionality in 2021 as well, and I have seen definite decrease in cycle time.

    This is a small move so it can only fit a small arc on there, but you get the idea.

    arc-fit.jpg

  28. #138
    Join Date
    Jan 2014
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    5,701
    Post Thanks / Like
    Likes (Given)
    1009
    Likes (Received)
    3215

    Default

    Mitsubishi has offered an option called “High Speed Drilling and Tapping” which does the arc motion between holes since at least as far back as 1995. If one has the option on their machine, you just add addresses ,I and ,J (note the commas, they are needed) to your G81, G82, G83, G84 line. The main caveat is that the value in I has to be less than the minimum distance between holes and the value in J has to be less than the distance from your initial level to your rapid level.

  29. Likes Garwood liked this post
  30. #139
    Join Date
    Aug 2005
    Location
    CT
    Posts
    8,444
    Post Thanks / Like
    Likes (Given)
    464
    Likes (Received)
    2266

    Default

    OK, so accelerating X/Y before Z is at rest makes sense, but ...
    Where does the radius start/end?
    At the R level, or the Z-start level?

    So let's say the drill cycle is:
    G00 X0 Y0 Z1.
    G83 G99 R.05 Z-1. Q1000 F2.
    X1. Y0
    X2. Y0
    G80

    What Z-level is the move from X0 to X1. and X1 to X2. ?

  31. #140
    Join Date
    Jan 2014
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    5,701
    Post Thanks / Like
    Likes (Given)
    1009
    Likes (Received)
    3215

    Default

    Quote Originally Posted by SeymourDumore View Post
    OK, so accelerating X/Y before Z is at rest makes sense, but ...
    Where does the radius start/end?
    At the R level, or the Z-start level?

    So let's say the drill cycle is:
    G00 X0 Y0 Z1.
    G83 G99 R.05 Z-1. Q1000 F2.
    X1. Y0
    X2. Y0
    G80

    What Z-level is the move from X0 to X1. and X1 to X2. ?
    I don’t know how other control manufacturers do it, but on Mitsubishi the I and J addresses roughly control the arc. Note that the arc is not true circular interpolation, but more like a controlled accel/decel of each axis. I sets the value for X and Y moves and J sets the value for Z. In the canned cycle operation, if one set an initial plane at 1” and the rapid plane at .1” and set ,I to .2 and set ,J to .3 then the X and Y moves would allow Z to start moving to the rapid plane when the X and or Y were .2” from the programmed position and when they reach the programmed position, Z would be approximately .8”. Z would continue at rapid to the .1” position then execute the drilling operation. When the drilling operation is complete, the Z would rapid up toward the rapid plane. When it reached .7” then X and or Y would start to move toward the next programmed position and Z would continue up until it reached 1”.

    Hope that makes some sense.


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
  •