We do a lot of small dia long shafts (think needles) on our swiss machines. We always have to account for stretching if we remove a lot of OD

With small cutters, you must minimize runout. A small bit of runout can double (or more) your radial depth of cut on a flute or two. .06 is a fairly large cutter for us. SFM is material dependent. But with 12k, chances are you will be low anyway. I like Harvey's guidelines, although they are a bit conservative on the SFM. But hey, I like conservative.

Nice job!

I know it's not an add-on, but do a search for probe posts. Ron and I made one that works extremely well on Horizontals. I use it for our Makinos. Pretty post and play, including error traps and DPRINT

Good to know. Thanks, Mike

That happened on our mam too. But in normal surfacing and HSM, that did the trick. Sorry, I got nuttin else.

John, set your Rapids to about 390 in the control def. That's what brought it closest for our mam. I can get you a pic when I get in. Oops, double post instead of edit

John, set your Rapids to about 390 in the control def. That's what brought it closest for our mam.

You're just crazy!

We found a lot of Muda. 1260 ft of travel, among others. We were able to reduce Muda of all sorts over 60%. We now have kitted setups, 3 more shadow boards, radius end the bar feeder blades (#1 in our Kaisen due to a injury), and great info for better setup books. We had 14 people in a combined Kaisen for setup reduction and autonomous maintenance.

We have ground masters for our unilocks we check them on a scheduled basis, or if there is a questionable location. That was the verify them statement above. For our fixturing, all fixtures are on a maintenance schedule as lean policies dictate. The schedule varies, based on quantities run, material abrasion and hardness, etc.. We try to control all aspects of a process. Obviously we miss on occasion, and that is where Kaisen events and or continuous improvement kicks in. We just had a week long Kaisen for Swiss setup reduction and Muda.

Our repeat work runs as intended. There is no change. It goes on the machine and runs. G10's do not change. We verify on schedule, and if for some reason a unilock is not on location, we make it so. We do not change our G10's. That would invalidate other programs. We would have to re run capability and validation checks if we did change. So changes, while they are made in the spirit of continuous improvement, must be justified. We have development cycles for all parts. We have engineers that design, inspectors that check, and programmers that handle the machine side. When we go production, it is a rock solid process. Like I said earlier, we don't operate under job shop parameters. We can't.

See, that's where we differ...we don't do it in process. The accuracy is built into the process. It is done Before it hits machine, off line. We don't stop and move parts. Because we already knew they were off before we put them on the machine. Mistakes are caught off line. Before they hit the machine. That's how we can make such a big deal about no operator involvement.

That was the third time I said the same thing. You see what you want for argument sake.

You assume wrong. I don't probe. I know where everything lies from the check station (much more accurately than a probe). I don't rotate (sine errors). I simply open the file, translate the offending geometry and repost the correct (to the tenth) program. Just as fast as changing any transform pattern. Faster than breaking up transforms. This also works after our scheduled fixture maintenance that may alter a part location due to a cleanup. There are some instances or machines we do probe. But not on our horizontals. Even then, I don't transform. If, for no other reason than to make you wonder why.

I'm done hijacking JP's thread. Sorry John.

Ok... Let me put this to bed for you. I have a fixture. Said fixture is locating 26 pcs. This particular fixture is located on a unilock. The centerline of the unilock and face of the unilock is my sole offset for said fixture containing 26 different pieces. Now, this fixture goes through a checking station. Four random locations do not fit the original grid by .0003 or so... Now, my entire rectangular transformations are off for these positions. Can I break into the transformation and relocate those four without creating separate transformations breaking my grid of 26 pieces into smaller, more localized transforms? Now, because we do families, planning asks "can you put a pivot pin in 9 pieces for a shortage we have on the 128-169-4-3a parts?" I can continue, but why?

I never said transforms don't work. I said they don't work for us. I'm not quite sure what your question is, other than why i don't use transforms. My answer is the same as why I don't use certain softwares... They don't fit the way we do things. I like to have control of each and every station, on each and every fixture, on each and every pallet individually. I don't think I have that control using transforms. It really is that simple.

Our parts do not generally meet the normal job shop operating parameters. We do short run, high variety frequently repeating parts as we need to Kanban them. We can go from running a thousand piece run to a one off, then back (or to a different part run) with no intervention other than loading a program and tombstones. We kit dozens of parts per machine that we can select from on the fly. When we make parts, we don't just get a FA and run. We are in the medical world. EVERY part we make goes thru capability studies and run validations. We MUST have fully capable processes running to cpk 1.0 or better. So we tend to front load in programming for the process, not the programmer time. Transforms are ultimately not accurate or flexible enough

No. That is the last thing we want. Operators don't set offsets, they don't edit programs. They load, unload, and check parts. The machine tells them when to change a tool if it is dull. We still have work to do, but that is our goal shop wide.

Yes. A lot of our parts are laser blanks that are form fixtured. Due to inconsistencies in fixtures and or blanks, we have to have ultimate flexibility. Transforms can work sometimes, but I've been bit too many times to rely on a set number.John, I understand... I was giving our perspective. Not saying it's right or wrong for any given situation.

Because we have a policy of zero hand edits, we rarely transform. We have dedicated unlock locations on our tombstones, that all offsets come from. If I need to fudge one part, I do it, without affecting any other parts that would be included in a transform. A little more up front, but we don't need genius setup guys at each machine this way. Everything is controlled in programming, not at the machine.

It's been a bit since I've programmed the Siemens, so I don't have any templates handy. The main is a *.mpf file, the sub is a *.spf file. That's how the control likes them. As for the Z value, some machines have a tool change height of Z-1. And maybe why that's why the post was set up that way??? Your MTB would be the place to get that info, along with a couple samples of main and sub calls. From there, I could help a bit with the post.

The command G00 SUPA Z0 D0 positions the spindle head "up" to its machine zero position. The Siemens' SUPA (suppress all [offsets]) has been used instead of G53 because with Siemens, G53 does not cancel the base offset. D0 is tool zero, (no tool) I have always used SUPA Z0

Not having ever seen a Roders, I'm kinda curious what makes it a super duper high level machine, as opposed to "just" a high level machine ala Makino. Serious question. Are there super functions? Accuracies beyond our micron accuracies on our Makinos? Or other reasons why it would be superior? Also, what makes their control so complicated I couldn't efficiently use Mastercam?