I have found that the filtered toolpath always has little gouges (sometimes big), but the unfiltered point-to point is fine if you reduce your tolerance to .0001". Makes huge program, but very smooth motion, as long as the surfaces are smooth. Also may take verify a day or two to get through it, and will probably crash before it finishes. Had some luck using an stl of the model as stock and running verify to check for gouges-. Good luck!

Jay thats a great time for a laptop- I wanted to try what you are doing before I bought my laptop, to see if a Mac was faster- but the Apple "genius" (or whatever they are) didn't want to set up that much software for a test. So, its good to see how that turned out... I might have saved some $$ with the Mac, and not given up much speed.. FWIW, I have a Xeon3370 q4 at 3.0 ghz, 4g ram, 80G SSD, FX3700 graphics. Time 3:40 in MU1. I hope to try out X4 soon.. Steve

James, you are right- you can modify a post for just about any situation very easily- we have posts formatted for 3x, 5x, 5x with pallet management, and several other oddbal things we run across from time-to-time, each having some specific additions or deletions depending on the situation. Very nice and easy to do. If your cost estimate is right, Camplete is cheaper for 5x than ICAM...not to rain on the ICAm parade. I'll shut up now...

+1 on Phil's post- we have run tons of 316 proj 70 at 800-1000 sfm...Sandvik inserts. The premium for the Carpenter material was more more than made up by the increased machinability and consistent results. Cheap, off-the-shelf material would knock us back to 300-400 sfm and who knows what for tool life.

OK, I see my mistake...lol! Thanks John, at least I will quite wasting time. On that, anyway... +1000 on your sig. Steve

I have some left-and right, mirror image parts and am trying to use the Mirror Toolpath c-hook. When I select mirror geometry and create new ops, it brings up the transform-mirror menu, I select the axis I want to mirror, and then it creates new ops but not new geometry. Then it hangs and doesn't finish. Can't close, escape, or find a way to make it complete the process. I have to use task mgr to end mcam and restart it. This is in X3mu1. Anyone know what I am doing wrong? Thanks for any replies...

Hmm, looks very similar to Camplete's post and simulate software... I have Camplete, and it is really nice for 5x stuff when you are trying to fit a part inside the machine limits and not crash or over-travel. I wonder how the price compares?

Newbieguy, you have more mail...thanks for the opportunity

FWIW, we just roughed 6al4v using Dataflute 5-flute medium-helix AlTin coated 3/4 dia endmills in shrinkfit holders. Cutting at 400sfm, .75 DOC, .040"radial engagment, .006"per tooth (62 ipm). Used coolant AND air-blast to blow off chips and cool the cutter. Cutters lasted 120 minutes at this rate! Programmed using Cimco's Adaptive Clearing for a constant .04" radial engagement. This beat a 2" Sandvik insert cutter by 15% for metal removal, and inserts only would run maybe 10-15 minutes. Run time for this process was almost 3 hrs roughing, so the cutter life and cost was much better with the endmill vs indexable inserts. I was really surprised at this result..

I have a Hermle C40 after being a 23-year Mazak and Kitamura user, and I will never go back. The Hermle is much more accurate and the Heidenhain 530 control is incredible. I chose the Hermle over the DMG because of the service issues with DMG, as well as people reporting accuracy problems. The Hermle service has been excellent, even though I am a small job shop. I can do parts on this machine that used to go to a jig-grinder, and the 5x accuracy has never been an issue. Its built on a synthetic granite base, has .1 micron scales, and thermal comp built-in. Tables are torque-drives, no gears. I will buy more Hermles as soon as I can.. Mazak service is pretty bad if you don't have 40 machines, and they really don't want to hear about it if something is not right with a new install. And I KNOW the accuracy is not even close... Hermle all the way for me...I wish they made mill-turns as well. Just my experience so far.

Matt, I have the quality down to less than half- any less and its no better than blind...

Mmm, I have had a lot of problems with verify- always, not just with X3. The last four 5 axis parts I did would crash verify before completion, and usually crash X3. CNC logged a bug, but that was months ago. I still haven't heard anything back. I can't seem verify anything I would call a medium-complexity part without a crash, so I have been flying blind for some time. It also still takes forever, and then gets even slower if you do a compare. I can't zoom or rotate in less than 1 or 10 minutes...if it doesn't crash doing the compare... However, MU1 is faster calculating, (if it doesn't crash)and most other things seem to be OK. And it acts the same on an XP64 and Vista64 system.

MAtt, nice system! Its good to see what a new i7 will do without any tweaks. Definitely go for the HASP upgrade, its really no big deal. Gcode, I agree withyou about not switching to software licensing- the hasp allows many of us to run mc on different systems, and in my case, has kept me alive when the main computer crashed and burned...more than once! As for X4, I know its in beta, I just was wondering if multi-threading was something available to us now, or if only X4 can use it. SOunds like X3 doesn't use it. I'm just not computer-savvy enough to know.

Hah! phooey...

Definitely set it up to use that chip blaster- it'll be worth it!

What brand of machine are running? I've had good luck getting q's answered by our Hermle guys. Neil, sort of off- topic, but what do you think of nx6?

Watcher, if you have the Heidenhain control, there should be a geometry file for checking 5x positioning yourself. We have a Hermle, and the file is called "Zyklus 19"- it checks the postioning at A + and _ 90, with C0,90,180,270 and prints a file. You can then change params to get it straight. Ours is a trunnion-table type, but I have seen DMG guys use a similar cycle for doing the nutating-head machines..

Gene, X3 doesn't have multi-threading, does it?

Gcode thanks for checking that out- its interesting that it works with an X_B file but not a sldprt file...at least it appears to be an X3-in-Vista problem..

Gcode, its MU1, and I don't have Solidworks, it just seems to be what my customers use the most...I will ask one of them to do X_B file and see if it works. I have tried all the fixes from other threads here, and none have worked, so it looks like some kind of Vista thing. Thanks Steve

Gcode, I have had the same results regarding Vista64 vs XP64. I have both systems on the same laptop set up as dual-boot, and the Vista install runs the benchmark in 3m40 sec, while the XP64 runs it in 4m30s. Both running out of the same SSD hard disc on the same computer... I thought something was wrong with the XP64 side, so I was going to re-install XP and X3 to see if it got better. I may still do that just to be certain, but it appears that Vista is faster than XP on the same machine right now. Only issue I have with Vista is that X3 won't read parasolid file from Solidworks on the Vista side. Same software on the XP side works fine, so right now, I have to open .sldprt files in X3 on the XP side, then save as mcx, then reboot into Vista if I want the extra speed. I am sending a file to tech support to see if they know what is wrong..

Bernie I think you want the OS and Mcam both on the ssd for best performance. I have a 500G secondary drive that has all data and extraneous crap on it. My next desktop will definitely have one or two of these ssd's in a raid to make it as quick as possible...

A couple people here have expressed interest in SSD's and how they may affect performance in Mcam, so I will relate my experience. I just got a laptop with Intel's X25M 80gig solid state drive, a Xeon 3370 3.0ghz quad cpu, 4g Hyper-X ram, Vista64. It runs the benchmark in X3mu1 in 3m49s.. For comparison, my desktop is an E6600 oc'd to 3.2ghz, 10k w/d raptor drives in raid 0, 4g Hyper-x ram, XP64. Benchmark is 4m40s... Since X3 still only uses ONE core (#@!$$$!!), the 6600 at 3.2ghz I think should be about as fast as the quad cpu with no overclock. Seems like it makes a notable difference- I noticed a bigger difference on 5axis processes- I just had one take 9m56s on the desktop, same one ran in 6m23s on the laptop! Also, everythhing starts and runs quicker on this drive- X3 starts in about 2 seconds! It boots into Vista in 9 sec.. I did some research before getting the Intel drive because it is more expensive than a Samsung-based ssd. Anandtech did really good test with these drives, showing that Intel's first shot at ssd technology blows away the competition. Do a search on the X25m on anadtech's site if you are interested. All in all, if you are doing a lot of toolpaths that take forever to calculate, it may be worthwhile to look into it, it really does make difference, and if your time is worth money, you will see payback pretty quickly.

MU1 on the new laptop No overclock, out-of-the-box. 3:49... Yeehaw! I'm pretty excited to get under 4 on a lappy! In Vista 64. Are there tweaks for a Vista system? This is the first experience I have with Vista, and I tried it because the Passmark scores were higher than on XP64...Bogus, any Advice?

I use it also, for everything but 5x stuff. It is a LOT faster, and the adaptive clearing is very good for roughing without ever burying the tool. And it supports multi-core now, so it is even faster than it was before... I agree that its very annoying to have to use add-ons to get what should be in the mcam you already spent money on.