Ideas On home made HSM mill

[yoopertool]

As of late the company I work for hs been growing more and more interested in getting a High Speed Machining Center. I think, why not build one? I understand the limitations of a home made machines and realize that high speed machining of steel is probably out of the question, but I think that aluminum is doable. It of course would have to be of steel construction or a E/G-steel combo machine. This is slightly out of the normal DIY catagory, but for those of us with access to metal working equipment available, I think that it can be done. What I am looking for is ideas and tips on the following topics:

1) machine type. gantry vs knee mill or what ever other style there is out there.

2) The spindle. I would like to aim for the 25,000 rpm range. design, balancing, and spindle motor ideas are needed.

3) control. Will mach. 3 handle high speed machining? or would I have to make my own?

I am not asking just for the heck of it. there Is actually a good chance this is going to be made. I am a full time machinist and a full time engineering student. The company I work for is interested in sponsoring my senior project for school and having us build them a small (20" x 14" travel) high speed machining center. I just need to give them some information and proof that their money would be well spent and not waisted. All of the work on the machine would be done by engineering students and we can get alot of companies to give us good deals on conponents, so I think that I will be able to stick with a 8-10k usd budget. Please let me know your input! Thanks everyone!

[walter]

Here's fairly simple but rigid design.

Can't comment on spindle or Mach software, not enough experience.
Good idea, btw. I will be watching this thread with interest.
_

[Zumba]

If this is a serious project with a deadline, skip the E/G for now. The stuff has great potential but you'll be taking classes at the same time. Stick with something simple and worry about the other stuff.

I'd use continuous cast iron bars, milled and ground to spec. You can buy solid gray iron bars in various sizes up to 2'x2' cross section, manufactured by companies such as "Durabar" and "Versa-bar".

1.) Neither knee nor gantry. Stick with a bed type mill. Rather than machining T-slots, drill and tap/threadmill holes.

2.) Use a commercial electrospindle for sure. Consider a high frequency router spindle, 5-10hp that uses ER32 or ER40 collets. The main consideration will be that the spindle can handle the splash of the flood coolant.

3.) Mach 3 will work fine, esp with a GREX.

BTW, what companies will give you deals on components? I'm curious, and admittedly, dubious.

The best piece of advice I can give is to draw out your machine completely in Solidworks, down to the last bolt, asap, and expect many revisions. You can get solidworks part files from Nook for all the linear motion components. Drawing your design on a napkin will likely result in a very disappointed sponsor .

[andy55]

Originally Posted by Zumba 2.) Use a commercial electrospindle for sure. Consider a high frequency router spindle, 5-10hp that uses ER32 or ER40 collets. The main consideration will be that the spindle can handle the splash of the flood coolant.

I would think that the idea with HSM is to get parts done faster? If so you really need a tool changer too! So the spindle needs to have an ISO or HSK or similar taper. And you need to design a rapid tool changer.

3.) Mach 3 will work fine, esp with a GREX.

There won't be problems with simple G-code that has 'long' segments(lines/arcs). But if you machine 3D parts and your CAM program spits out G-code with lots and lots of short segments then the demands on the controller are tougher.

With EMC people were reporting problems when the block rate (lines or arcs per second) was approaching the traj_period (that's the delay between successive points on the trajectory). Normally people run with traj_period = 1 ms, so you could process 100s of G-code blocks per second. On most systems traj_period could be lowered, but I don't know anyone who has done testing...

Another thing you need for HSM if you have lots of short segments is lookahead. I don't know how good Mach is here but the lookahead in EMC is not super great. It looks forward one or two segments. If you have motivated engineering students willing to work on a HSM/Lookahead trajectory planner then I am sure the EMC community would more than welcome that! (I'll do what little I can to help you get going...)

Finally, I don't think Mach, even with a grex, is still a truly closed-loop system. With EMC you'll get that, which means that the controller is fully aware of the machine position, following error, and commanded motor speed/torque at all times.

Oh, while we're day-dreaming, there are not a lot of affordable or open-source CAM programs out there that can produce HSM paths. The important thing here is that while most CAM algorithms work based on the part geometry only, a HSM toolpath needs to take into account the cutting load at all times. So you need something simulation-based to generate a toolpath that will not push your endmill at 20mm depth-of-cut at full speed into the material but take many light passes instead. That's another opportunity for someone to really make an impact in the open-source cad/cam/cnc world.

AW

[yoopertool]

Thanks for the input and keep it coming. I had been contemplating a tool changer, but I wasnt sure if any of the pc based softwares supported tool changes. once you add tool changing into the mix, you have to worry about spindle orientation, release mechanisms, and a few other things which I dont have any experience with. Right now most of my machine building experience is what I have learned while building my CNC router (50% done) and my classes. There will be computer, elctrical, and mechanical engineering students on this team if it works out, so I guess that modifying an existing program to meet our tool changing needs would be possible for a few of the other people. Andy55, you mentioned a program called EMC, do you have a website for that?

Walter, I like the design that you showed, but when I was thinking of E/G construction, I wasn't really thinking of 100% construction with it, more of a way to knock the vibrations down. A few more questions now.

1) Zumba, are the cast-iron bars you mentioned already aged so that they wont move around after I machine them? because I think that I read that it takes a few years for them to actually finish stress-relieving.

2) Does Mach. 3 or EMC support tool changing?

3) Would I be able to orient an electrospindle? if so how would I go about that? Do these things have encoders on them?

4) Could anyone point me toward some free CAM programs like andy55 mentioned?

I have to get to work, but thanks for all of your comments everyone! I think that this will be an interesting project!

[RICHARD ZASTROW]

yoopertool, I spent more than 2 weeks at deercamp, eh!!

Back to your fantasy. How big (size & weight) are the workpieces with set-up tooling?

If your HSM includes high feed rates as well as high spindle speed, you have to accelerate-decelerate either the machine components or the workpiece etc. That can be a serious consideration with any rapid motion or fast feedrate machining.

Your control will need adequate look-ahead capability, particularly at axis reversals etc.

[andy55]

Originally Posted by yoopertool Thanks for the input and keep it coming. I had been contemplating a tool changer, but I wasnt sure if any of the pc based softwares supported tool changes. once you add tool changing into the mix, you have to worry about spindle orientation, release mechanisms, and a few other things which I dont have any experience with.

If you want rigid tapping you will need an encoder on the spindle anyway. With a modern VFD spindle orientation for toolchanging should not be that hard to implement

Andy55, you mentioned a program called EMC, do you have a website for that?

EMC or the Enhanced Machine Controller is an open-source real-time machine controller that runs under linux. The big advantage I see over Mach is that EMC can run in true closed-loop mode, i.e. the servo PID calculations run on the PC in real-time.
Homepage http://www.linuxcnc.org/
Wiki http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl
IRC: #emc on irc.freenode.net (very knowledgeable people here, come discuss your machine sometime!)

2) Does Mach. 3 or EMC support tool changing?

EMC supports tool changing, but you need to write some simple HAL (=hardware abstraction layer) code that interfaces to your particular hardware.

4) Could anyone point me toward some free CAM programs like andy55 mentioned?

There are no good(Mastercam/Surfcam level functionality) free general-purpose CAM programs out there.
I'm playing around with some basic toolpath algorithms myself - when they are ready I will publish them and we'll see if there is interest among other developers to build a GUI and everything else that is required...

AW

[yoopertool]

I only have a few minutes during my lunch break, but i will try to fit a quick post in.

The workpieces I will be working on will be small. probably in the 5lb or less range. The company I work for is in the process of purchasing WorkNC, does that support hsm? I have only worked with mastercam in the past so I am unaware of what other packages offer. Does anyone know of any elctrospindle manufacturers that make fairly sturdy spindles? I would like to find some rough pricing on them and maybe even just look to see how they are built so I can look into designing one. I have full access to swiss lathes, gundrilling and precision CNC grinders, so I think that I could make one. anyway, I have to get back to work. Thanks again for the info everyone, eh!

Joel

[Zumba]

A high speed spindle with toolchanging will blow your entire budget. $7K for a 5hp with an HSK spindle. Doesn't even include the tool carousel, which you'll have to build yourself.

[Caprirs]

I hate being a cynic because I like seeing new ideas but you'll find the machine tool builders of the world have come to some common attributes regarding construction of metalworking machinery. Learning from them can save a lot of headaches.

The number one feature of a machine tool is mass. The more, the better. Mass increases rigidity and absorbs vibrations. Despite how strong carbon fiber is, no one builds machine tools out if it. Huge lumps of iron have been the material of choice for centuries even when cost is no object. The most accruate encoders are wasted if the machine deflects under load. By contrast, the machines that stuff circuit boards are stupid fast because they encounter no load other than their own mass, so lighter is better in that application.

The second feature is big motors. If the mass needs to move faster (i.e. HSM), then bigger servo motors are installed with high current/voltage amps to drive them. The mass is never decreased so smaller motors will be "faster". The servos just get bigger.

You probably need to define what constitutes HSM to this project and establish some parameters regarding accuracy and material removal rates. HSM seems to mean different things to different people often. Do you have some goals that you're trying to meet/exceed?

[yoopertool]

Thanks again for the input. Caprirs, I agree 100% that there would have to be alot of mass. I like the idea of using the pre-cast iron bar for construction as Zumba mentioned earlier. I guess that I should list my hopes for the machine so that we all know. I would like a small mill roughly 20" x 14" travel in XY and 16" in Z. These numbers are just estimates though. I would like to have a spindle which goes up to 25k rpm and be able to feed accurately up to 400ipm. The company it would be going to is really concerned about saving floor space in their plant, so the smaller the footprint the better, but functionality is first priority. I would like a tool changer, but not a carousel, probably just a line of tools in a holder off to the side which the spindle would move to. The more research I do the more I see that 10k USD is a rather ambitious budget even for the most resourceful of us. I am still in the stage of getting a proposal together, and the 10k budget was just a number I picked. I just need to make sure that the price isnt getting anywhere close to the range of a used hsm mill. In the past the university I attend has gotten great deals from companies like fanuc, baldor, and a few metal suppliers, so I think that I will be able to get nearly at cost products from most of them since this is an educational project. Please feel free to give your opinion and share your thoughts as to what would make a better machine. I am also going to need to start gathering prices for components, so if you know of a place that might supply what I need please let me know so that I can contact them. Sorry for the long post. Thanks again everyone!

[RICHARD ZASTROW]

yoopertool, For that spindle you might consider Precise. We used one on some 4-axis Bostomatics 20 years ago. 40,000 rpm max. but we only used 26k. These machined D-2 & D-5 @ 58-62 Rc rotary dies. Speed was controled by variable frequency drive. No toolchanger as this was an auxiliary spindle.

Keep it UP eh!