PDA

View Full Version : A axis speed



Gregore
06-21-2013, 11:01 AM
I am wondering how to get madcam to control the A axis speed not linked to the speed chosen for the tool, in the tool cutter box.

attached is a gcode for a very large lift cam / wing section sort of part and the A axis keeps the same speed for the whole rotary section , it should go slower on the tips of the cut like it does in the sim.

Am I forgetting to do something in madcam?

Dan B
06-21-2013, 11:45 AM
See if this helps:

http://www.cnczone.com/forums/haas_mills/69433-inverse_time_feedrate_4th_axis.html

Don't forget to read the quote from Haas.

Dan

Gregore
06-22-2013, 01:27 PM
Dan,

Thanks for the response....

Your link to the Haas response illustrates that for 4th/5th axis work specifying feed-rates using Inverse Time Feedmode (G93) is the best practice. I would like to know how to get madCAM to do this.

Do you know if this is possible?

Dan B
07-01-2013, 05:22 PM
Sorry, I've been on vacation for a while. I do have experience with inverse time feedrates when we used to use a Tritech 5-axis head. I'm back to work tomorrow, and I will try and dig up some old examples of code that uses inverse time feedrates. Then we can figure out if it's possible in madCAM 5.

Dan

Gregore
07-01-2013, 06:30 PM
thanks , that would be helpful as I would like to not have to use only controllers that can handle not having it in the post.

subnoize
07-02-2013, 11:03 AM
Gregore,

I think we are looking for the same thing. I use LinuxCNC which reguires G93 Inverse Timing Mode for 4 or higher axis cutting. What post processor do you use? Try the 4 axis EMC post processor and see if it outputs the G93 code.

I would but I have the demo version which will not generate code.

I have wasted so much money on CAM software where they claim it supports G93 but then you get using it and it doesn't that I was to see this BEFORE I buy!

I've tried emailing twice now the maker of MadCAM and we'll see if he responds. I strongly suspect that MadCAM does NOT support G93 inverse timing. Trust me, I'm hoping! I have a wonderful machine but can't use it because of this and a Rhino solution would be great...

Dan B
07-02-2013, 11:59 AM
I've been going through the documentation on the post-processor, and I don't see anything that would support inverse time feedrates. However, that doesn't mean it's not possible. And if it is not currently possible, that doesn't mean that it will remain that way. Like the saying goes: "Necessity is the mother of invention". :)

Dan

subnoize
07-02-2013, 12:34 PM
I've been going through the documentation on the post-processor, and I don't see anything that would support inverse time feedrates. However, that doesn't mean it's not possible. And if it is not currently possible, that doesn't mean that it will remain that way. Like the saying goes: "Necessity is the mother of invention". :)

Dan

Thanks, Dan!

I can almost say with 100% confidence that MadCAM does NOT support NIST RS274/NGC G93 inverse timing mode. The reason is the European market doesn't even recognize the United States' National Institute for Standards and Technology as a standards body. I bet you they probably have never even heard of RS274/NGC at MadCAM.

If MadCAM doesn't support RS274, I put a minimum 2 year time frame on it. Not many CNC machines are made in the US and LinuxCNC is mostly 3 axis which is why everyone claims they support it even though they know they don't!

Keep your fingers crossed! I really wish I am wrong on this one! I like what I see in the demo version of MadCAM and I like the control it gives me. Give me G93 and I will give you my money!!!!

Gregore
07-02-2013, 01:34 PM
Your demo should output gcode . I know our did .

emc 4 axis is what we are using right now for a post and it is not g93

I am hoping that joakim will spend some time to get the g93 out put for madcam as it was my understanding when i spent the money that he would make post code for any machine i made or bought and that includes from my point of view a controller that needs g93.

He has been on things very quickly in general so I assume that when he returns he will get to our questions about this.






How complicated is it to make madcam out put g93 ? Any one know?

subnoize
07-02-2013, 02:05 PM
How complicated is it to make madcam out put g93 ? Any one know?

Very complex if done correctly. I'm actually trying to write a converter for G94 to G93 and it hurts just trying to get my head around the problem.

Here is the back story on "Inverse Timing" as I have heard it;

Linear and angular axis were not in sync with each other so the CNC manufactures started building in code to compensate for this anomaly. Eventually these algorithms became defining features of the machines themselves. That led to patents being filed and the world suddenly became dependent on the machines! So, back in the early 90's a group of US engineers decided to fix the problem with g-code and came up with a solution. That solution became know as RS274/NGC and was eventually published in 2003.

EMC (LinuxCNC) was the "reference implementation" of RS274/NGC. So, today, LinuxCNC fixes the problems with mixing linear and angular movements but it pushes the calculations off to where it should be, the CAM software.

So, European and Asian CNC makers simply do not support the newer standard because it relegates their "secret sauce" to the scrap heap! US companies like Haas use RS274 and love it! We LinuxCNC users find that our only available CAM options in a price range we can afford is either from Asia or Europe, where they don't support inverse timing.

As to how easy? Very hard for me personally. You have to calculate rate of any axis not by units/minutes but by the surface speed determined by the diameter of the model. If you have an irregular model of say, Venus De Milo's bust then you have to vary the axis by the distance from the center of the angular motion. Worse yet is like on her nose, you will need to vary the speed slower as the Z axis rises and then increase the speed as the Z falls to her cheek to maintain the calculated surface speed...

We haven't even touched on applying Pythagoras yet and it already complex!

So, if it aint supported it aint going to be support for a very long time. It isn't something to just plop out and be done with. Luckily, for our MadCAM the calculations are already being done in 3D space so we may see it sooner that some of the other CAM packages. But only IF it makes our MadCAM guy big money!


PS. Here is to wishing I am terribly wrong! *Gulp* *Gulp* *Gulp*

PSS. And no, my demo version does not generate code. Throws a bunch of errors but no code...

Dan B
07-02-2013, 03:01 PM
Can you post a screen grab of those errors?

subnoize
07-02-2013, 03:21 PM
I will as soon as I get home! I am about to go do a few laps around the pattern to keep current. Probably be around 9ish!

subnoize
07-02-2013, 09:10 PM
Ok, here is the popup. It says I can't use the post processor so I can't finish enough to actually do a cut.

Also, simulation doesn't do anything like the demo videos do. it only draws the path, doesn't show what it looks like as it machines the model. I'd like to see it from the removal of the material perspective.

Dan B
07-03-2013, 05:45 AM
Okay, that must be new for madCAM 5. I wasn't aware of that limitation. That doesn't seem like a good idea to me. You need to be able to check the quality of the tool paths when evaluating a CAM software. As for the simulation, material removal is currently only visible in 3 axis. Material removal in 4 and 5 axis is coming in the future. Joakim has mentioned the idea of a completely new simulator with more advanced features.

Dan

subnoize
07-03-2013, 08:09 AM
I wasn't aware of that limitation. That doesn't seem like a good idea to me. You need to be able to check the quality of the tool paths when evaluating a CAM software.

That was why I installed the demo! I'm just not buying unless I know the software performs what I want. Simply having a post processor doesn't mean the CAM package supports your controller. It just means they created a text file with the same name as your controller and included it in the install.

As to the simulation of 4 or more axis, that figures. Its a difficult concept to calculate in the first place. Doing repeated booleans on a 3D model gets expensive.

Dan B
07-03-2013, 08:24 AM
Simply having a post processor doesn't mean the CAM package supports your controller. It just means they created a text file with the same name as your controller and included it in the install.

Very true. We fought with a post for Esprit for the longest time. It took months before we could be productive with the software. When I hear from Joakim I will ask about changing the evaluation to a fully functional demo, limited in other ways (number of saves or time limit). I think the current limitation will hurt potential sales.

Just hang in there a bit longer...

Dan

subnoize
07-03-2013, 08:52 AM
Hanging describes what us 4 axis guys are doing! Blowing in the wind is another good description.

all we can do is wait I guess...

subnoize
07-12-2013, 08:10 AM
Hey, does anyone know when madCam people come back from vacation? Being from the US I had to look up the definition of what a vacation is... it sounds terribly boring. Maybe they will get bored and come back early?

Dan B
07-12-2013, 08:56 AM
I'm just assuming Joakim is on vacation. I base that off of working with him for the last 4 or 5 years. Usually he takes the summers off and enjoys the nice weather with his family. Considering how far north he lives, summer is not to be squandered!

Dan

subnoize
07-12-2013, 10:04 AM
Ah, ok. So really he could have been hit by a tourist from Cali and we're SOL and don't know it... ;)

On a serious note, I'm trying to do a roughing op on Venus De Milo's bust using a 1/8" ball nose and MadCam has been working on it now for two hours and is only half way through. Is this normal? I have no real baseline to know if this is slow and I did something wrong or this is blazing fast. The pretend 4 axis packages are really fast but that's probably because they aren't real 4 axis, right?

Dan B
07-12-2013, 11:51 AM
Roughing with a 1/8" tool? Wow, what different worlds we come from!!

How big is your project? What step-overs, and step-downs are you using?

I'm guessing it's finished by now. How long did it end up taking?

Dan

svenakela
07-12-2013, 12:50 PM
Joakim is way off any communication, right as Dan assumes.
Summer may be short here. But the sun is not setting until 11 PM and it never gets dark, 3 AM the sun is back again. Further up north it never goes down.

This pic is taken 0:30 PM a cloudy "morning". Beat that if you can. :)
191840

subnoize
07-15-2013, 10:11 AM
Joakim is way off any communication, right as Dan assumes.
Summer may be short here. But the sun is not setting until 11 PM and it never gets dark, 3 AM the sun is back again. Further up north it never goes down.

This pic is taken 0:30 PM a cloudy "morning". Beat that if you can. :)
191840

We forget that the sun stays up a while longer during summer! I'm would go nuts without my sleep!

Joakim contacted me and sent me a key to generate code, so I'm happy! I've been so impressed with just what the software does on screen that if these cuts come out as good I will buy the software even though it only support 3 axis on my machine. Hopefully Joakim will find it in his heart to support G93 and I will be a happy camper!

As to the process taking too long last Friday, I goofed and entered too many zeros after the decimal. The model is 3" x 2.4" x 2.4" so not very large so an 1/8" end mill is a monster! My test is if I can see the pupils of Venus and the her nostrils and lips (which are far smaller than the 1/8" end mill) then I know the program is worthy of my money.

Gregore
07-15-2013, 01:32 PM
I have been talking with my friend who is making my mill (his understanding is way beyond mine and anything I say here is just how I understand it) he mentioned that for Joakim to go in and change the gcode to output g93 is not all that difficult as the calculation are already being done to some extent by madcam when it is tool pathing the 4 and 5 axis g code. . His first thought with out looking behind the scene of madcam was that it was about 30 to 40 hrs for him to do it. So I would assume that for joakim ,who built the software it should be way less.
It seems to me that it would open up whole new markets for madcam to be sold to operators who do not have machines that work without g93. My industry (jewellery) is one of them that really need 5axis and 99 percent of the machines in use right now need g93 to do it easily . Add to that the drive surface ease of tool pathing for 5 axis in madcam and joakim might have one of the top choices for any smaller company that is needing a 5 axis option.

If I was able to figure out how to do drive surface tool pathing in one evening then you know he is one to something good.
it was actually way more work learning how to do the axis tool pathing in madcam.

subnoize
07-15-2013, 02:23 PM
Gregore,

I'm not wanting to put words in Joakim's mouth (or for that matter assign hours to his tasks) but he told me he has many of the calculations done for angular movement. That said, I would put a few more hours on it than 40, my background is software and I will be the first to ask for it done right, not done fast!

So, I am pretty confident that Joakim will have inverse timing for us sooner than later. It is to his advantage being the only CAM package under $9K that supports G93 (to my knowledge).

Gregore
07-18-2013, 12:47 AM
Exactly , imagine the extra sales he will get with the program under 9k and g93 plus really easy to do tool pathing for 5 axis. So many jewelry 4 axis machine sellers have tried to scare me away from 5 axis because they claimed it would take hours to tool path one part.
I sure am glad I found out about madcam.

subnoize
07-18-2013, 08:26 AM
...imagine the extra sales he will get with the program under 9k and g93 plus really easy to do tool pathing for 5 axis.

I like your enthusiasm but 4 axis machines are less than 1% of the CNC market. 5 axis machines are even fewer. The simple numbers means the CAM software guys have two approaches; sell for really expensive or sell really cheap in hopes of selling to hobbyist.

I know a guy who does the business plan research for startups (which is where I got my numbers, he is also a DIY CNC guy). He says the CNC/CAM market looks hot. The wow factor is great but the sad truth is it is a very small market with a huge buzz, but only in 3D printers. You got a 3D printer, doesn't even matter if it actually works, they are flying off the shelves. CNC routers like what I built requires too much knowledge to run for the average joe to use. They get very little buzz despite the fact that even a complete crap one is amazingly usable. The 3D printers are mostly unusable crap. Especially the ones that print with water soluble plastic! You can't even use the vast majority of what they print in production ready product, much less to make molds with (which was the original reasoning for using water soluble plastic).

What surprised me is how many CAM software providers are single person shops. I can name eight (8) off the top of my head. Two of those eight aren't even programmers. They outsource their development 100% to either India or China.

I guess what I'm getting at is this is a really narrow, niche market. That is why we as consumers have to protect these guys (aka. not allow their software to get pirated) and do word of mouth advertising as much as we can. Our new found friend, Joakim, is a special person providing very narrow resource stream (aka, MadCAM). Maybe supporting RS274/NGC wins him more business, but that actually depends on us users talking about the product in other places.