Big router small brain

[spoiledbrat]

I am about to finalize my designs, and though I would tap into this wealth of knowledge before I fabricate a paperweight.

12 foot X axis:
I intend to use a 1.25" ballscrew with 0.5" lead. I will use servos to drive this thing. My question is this: Can I direct drive this thing with a servo, i.e. hook the motor directly to the ballscrew with a connector?

880 OZ/IN servo motor:
I purchased a set of three servo motors from ebay. The seller stated 880 in/oz, and I am now having troudle tracking down the specs. It was recomended to me that to drive my 150 lb gantry (670 Newtons?) along its 10 feet of track, about 700 oz/in would be more than sufficient. Does this mean 700 peak, or 700 continuous? I imagine I can use the servos somewhere else, and just pick up two new servos to drive the gantry.

Controller Card: I see the benefits of using steppers and geckos, but what of using the 8 axis pci cards available from camsoft, or other places? I intend to use visual mill to operate this machine, as we will create an AB head (5th axis) for it soon. Has anyone here tried these PCI or ethernet controller cards? I know they cost around $3000 U.S., but I am no electrician/ PC board creator and this machine is intended for commercial use.

Any answers would be greatly appreciated, I need to get this ball-screw rolling. I will post as many pictures as the admin will allow.
Rob.

[ViperTX]

Well, you can direct drive. It's a design related issue...how fast do you want to traverse down each axis....what is the torque of the servo....what is the rpm of the servo.....do I want to increase torque, or speed....they are all inter-related. There is also the issue of space......can I directly mount the servo to the ballscrew without having the ballscrew requiring a large mount (5inches in length or more) for the axis....

[ger21]

You'll want to gear down the servos on the ballscrews. Servos need to spin fast to use their power. They won't work well direct drive.

Can't help you with your second question.

Visual Mill won't run a machine. It will just give you g-code. The downside to the PCI cards is cost. In addition to the price of the card, you'll also need to use their software to run your machine. This will probably cost a minimum of $4-$5000. And you'll need to pay an annual maintenance/support fee, minimum $500/year.

The cost of 3 Geckos and a copy of Mach2 will only set you back $600. And the absolute best, as well as FREE, support you'll ever find. And it will probably do at least 90% of what the Camsoft package will do.

It sound like you should be doing quite a bit more research before you start spending any money, or you might end up spending a lot more than you need to.

[ViperTX]

Servos have a torque constant......so much torque for each amp....maximum torque occurs at maximum rated current.

[jeffs555]

You may want to reconsider your ball screw size. I'm no expert, but 1.25" sounds awfully small for a 12 foot axis. This site has a lot of design info about screw loadings and speeds, and also has online calculators for critical speed and column strength. http://www.nookindustries.com/ball/BallCharts.cfm

[spoiledbrat]

Originally Posted by ger21 Visual Mill won't run a machine. It will just give you g-code. The downside to the PCI cards is cost. In addition to the price of the card, you'll also need to use their software to run your machine. This will probably cost a minimum of $4-$5000. And you'll need to pay an annual maintenance/support fee, minimum $500/year.

Visualmill's website stated that the software provided "Direct numerical control", I guess I may have been mislead.

I understand that the servo's need to be running at their nominal speed to produce max torque; I would like to keep the machine simple (no belts) and run direct drive. The servos I am looking for for the long X-axis are to be rated at 1750 rpm. If I were to run these at 1200 rpm, using 0.5" lead ball screws, this would produce 600 IPM jog rate. I would indeed be buying a larger motor and running at less than optimum rpm.

The cost of the software is not the driving issue for this endeavor. Are there any other benefits, besides cost, to using cheaper setup?
I thought that when using geckos, the signal from the motor's encoder is sent only to the drive, and not back to the controller. This seems to negate the advantage of using servos (at least the almost failsafe encoder part) as opposed to steppers.

As you mentioned, Visual Mill will not run the machine. I did not know that. If I were to use a Gallil card, I imagine they have some kind of proprietary software they will want me to buy.

The reason I cannot use the software you recommended is that I need 5 axis software. Most programs seem to be able to do 3 axis milling with 4rth axis indexing. I can settle for 5th axis indexing.

When selecting a servo for an application, when can I rely on "peak", and when should rely on "continuous" torque? For instance, to make the machine "jog" up the length of the table, it seems safe to demand a lot from the motor for a brief period. If I were doing continuous cutting for 30 minutes, it seems I should remain below the continuous torque.

I am trying to find machines similar in weight and application to mine, and base my design off of them. It seems they do not advertise servo torque or ballscrew lead/size combinations. I still have much more to learn, and at least a few months to do it.
Thank you for your help.

[turmite]

Mr Brat the Mach2 controller will do your 5 axis plus 1 if you want. You can have std xyz use a for a rotary 4th and b and c for your articulated head for the fifth axis. Mach2 cost $150 US and come with unbelievable support and free lifetime upgrades. They alnost have Mach3 ready to go also.

What do you intend to cut with this machine? If it is light weight materal i.e. balsa or foam you will get better speeds on your rapids usisng timing belts or rack and pinion.

Mike

[HuFlungDung]

Originally Posted by ger21 You'll want to gear down the servos on the ballscrews. Servos need to spin fast to use their power. They won't work well direct drive.

Sorry, Gerry, but I don't agree with this at all. As you know, I'm sure, that power is the rate at which work is performed. But the servo motor torque is proportional to the servo loop position error, so a very large torque can be output if needed. That doesn't always mean that a lot of work is being done (thus low horsepower requirement), because the movement may be slow against high friction or mechanical loading.

[ger21]

My mistake. It just seems that all the servo systems I've seen are usually geared down. Is that just for more torque?

[HuFlungDung]

I suspect so.

My old knee mill retrofits (with box ways) use 2400 rpm 660in/oz DC brush type servos in direct drive in X and Y on the table, and Z on the spindle. All screws are .200" pitch. When set up to be fairly "snappy" in accel/decel, these motors will move the table at up to 400ipm, but I think thats about their practical limit. Otherwise, I need to use quite a spongy accel/decel curve.

For the Z axis, I use a 2:1 reduction, in case I should want to drill with a large ordinary twist drill, which is probably one of the most brute force operations undertaken: this can require a lot of thrust pressure. I consider the performance of this Z axis to be equivalent to that of a 1.25" pillar drill press.

[ger21]

Our router uses very high lead ballscrews (~1" or so), and all the axis are geared. (Not sure how much). The X-axis has a spinning ballnut on about a 15ft screw. It moves the ~1000lb gantry at about 500ipm, with very quick acceleration.

I guess if you direct drive with a low lead screw, or gear down with a high lead screw, you'd get about the same thing, right?

[spoiledbrat]

15 foot screw... wow.
What is the diameter of this 15' ballscrew?
I did not want to gear down because of backlash or belt stretching. Probably just being superstitious.
Rob