X5 tool paths with arc filter twice the size in x5 as x4

[gcode]

Yeah,

We've got a couple of older VMC's that are good machines, but have 128K of memory.

They run these high speed paths very well, but can't hold enough code to do much.

One of them runs well with DNC drip feed, the other can't read the code fast enough.

We've got 3 Misubishi HBM's with 256K memory

We run a bunch of high speed rough programs on them with DNC and its PIA

They quoted me $7K to upgrade them to 1 meg of memeory. The ROI would be about a month, but I can't talk the boss into it.

[gcode]

You could get a quote on a memory upgrade.. Its not cheap but it is an alternative

[neurosis]

Yeah,

We've got a couple of older VMC's that are good machines, but have 128K of memory.

They run these high speed paths very well, but can't hold enough code to do much.

One of them runs well with DNC drip feed, the other can't read the code fast enough.

We've got 3 Misubishi HBM's with 256K memory

We run a bunch of high speed rough programs on them with DNC and its PIA

They quoted me $7K to upgrade them to 1 meg of memeory. The ROI would be about a month, but I can't talk the boss into it.

 

 

I am in this exact same boat. The owner feels that the older machines are not worth adding memory to due to the costs. He also feels (old school) that we got along for years without these new fancy tool paths and we can get along without them now.

 

We have tried drip feeding in the past and it is a pain in the xxxx. Unless we had a better setup and software I wouldnt even consider it. As it stands, we use Cimco to transfer files to every machine in our shop except for one. Our mori came with its own software.

 

These machines are old but it is what it is. They are what we have to do the job. Until the owner decides to buy new machinery (could be never) then we are stuck with their limitations. This frustrates me somewhat because one version of software makes the difference of whether I get to use this tool or not when the path seemed to be working just fine before. I guess that we will just have to see what happens in the near future.

[Guest]

One of them runs well with DNC drip feed, the other can't read the code fast enough.

 

Yup, if the older control has a slower baud rate, then lots of small move code is going to just cause issues.

 

It's all about competitiveness today.

 

Time = Money

[neurosis]

Neurosis,

 

I downloaded your files and I have noticed that in a few places X5 breaks straight line moves into smaller segments than in X4 (seen in Backplot). Even with the filtering switched off the NC code in X5 is 2000 lines longer than the NC from X4 (32000 lines vs 30000 lines)

 

 

Thanks for taking the time to look at the files. I wanted to make sure that it wasnt just me..

[Mark H]

As for the difference in files size from X4 to X5...

In the control def compare the Tolerance settings there - particuarly check;

-NC Precision (minimum step value)

-Choral deviation (used in post)

-Deviation of 'vector' endpoints for planar detection(used in post)

-General math function tolerance (used in post)

For the most part, if you have tighter tolerances in here then your posted code will be larger.

[Tyler Robertson]

I feel the same pain with robots

some of them having 14,000 line maximum memory ... memory upgrades are like $1k for like, 14,000 additional lines. Frightening. DNC isn't close to an option

*sigh*

[Mike Whitten CAD/CAM Contractors]

Hopefully I am not missing something, but everything looks 100% normal to me. Sure, the posted out NCI file and NC file are slightly different in size, but percentage wise, not much considering we have two versions of Mastercam and most likely slightly different toolpath algorithms. My guess is that if you studied it in detail, the X5 toolpath would be a bit more optimized. This would account for the extra size. This is just a guess though. I am not sure why you have the cut tolerance and filter tolerance the same? Maybe this was just for the test or I missed it in the above posts. I changed the filter to .002 total / 2:1 I then posted out both files with a known good mpmaster based post, I get 81kb for X4 and 84kb for X5. I would adjust my settings as required and make chips.

 

Mike

[Mike Whitten CAD/CAM Contractors]

By the way, this file is a good example of the NC precision setting(located in CD) issue that I posted up a while back. If I use a standard post with the NC precision setting set to .0001, it will have the linearization issue that is mentioned several times above. This has been an issue for many versions now in Mastercam. I just use an NC precision setting of .00001 and don't have this issue anymore. I have had this problem in X2, X4 and X5. I feel that CNC knows about it already as there have been a few posts on it. It would be a good idea to send it in again though.

 

Thanks

[neurosis]

Hopefully I am not missing something, but everything looks 100% normal to me. Sure, the posted out NCI file and NC file are slightly different in size, but percentage wise, not much considering we have two versions of Mastercam and most likely slightly different toolpath algorithms. My guess is that if you studied it in detail, the X5 toolpath would be a bit more optimized. This would account for the extra size. This is just a guess though. I am not sure why you have the cut tolerance and filter tolerance the same? Maybe this was just for the test or I missed it in the above posts. I changed the filter to .002 total / 2:1 I then posted out both files with a known good mpmaster based post, I get 81kb for X4 and 84kb for X5. I would adjust my settings as required and make chips.

 

Mike

 

Thats pretty amazing. All of my posts are based on MPMaster. I changed the tolerance setting to .007 and 3:1 and I still get a NC file that is over 208K. I even changed my NC Tolerance to match what you have shown in your other post.

 

This is slightly smaller than the generic post that I was using as a test.

 

No, I am not using sequence numbers just in case you are wondering.

[Mike Whitten CAD/CAM Contractors]

If I post out using the file as is, my nc file sizes are the same as what gcode got.

the X4 NC file was 101K

the X5 NC file was 108K

[neurosis]

OOPS! I had been playing with the path last night (apparently in my sleep). I agree. I get 108K in X5 and 85k in X4. If I use a non MPmaster CD in X5 posts @ 95K. I wonder why the huge difference from one control def to the other.

[Mike Whitten CAD/CAM Contractors]

It's the NC Precision setting. I always run .00001 instead of .0001. This way you won't get all of the unexpected linearization. This is an issue that goes way back. You can do a few searchs if you want.

 

Mike

[neurosis]

I did a file compare on the two files and they were quite different. Some of the arcs are broken in to two segments and quite a few of the numbers do not match by as much as .001. I found that during import of the Mpmaster, it changed the break arc in to quadrants. At least that explains the differences in the two X5 differences.

 

 

If I remember correctly, I have had the linearization problem in the past. It was breaking only some 2d arcs in to segments. I couldnt remember how I had fixed it but this must have been the solution.