[Runner4404spd]

you have to create a free body diagram to figure this out, on one side you have the mill head and gravity going down, on the other side you have your weight and gravity going down. these cancle each other. also on the first side you have static friction going down or up and this will counter the motor force depending on which way your trying to move. ideally you need to add to it, the bearin friction and the ball screw friction, etc, but really once you add the weight to counter balance the head your really only fighting friction and not the mass of the head. at this point you should also be able to release some tension on the gibbs as your now only using them to hold the head against the ways and not using them to hold the head up in place.

[Cruiser]

Pneumatics would be to slow to change directions, the volume of air for the force needed has to pass through lines and fittings etc. Hydraulics on the other hand could change direction instantly and controlled for holding, accelerating, and stopping. It could be run off of an accumulator, that could be kept charged by a constant pump on a pressure switch. A few good solenoid control valves would be the way to control it electrically. It would take some math to compute the pump flow required, per cyl size, per force needed @ flow velocity to choose pump. And pi x (radius squared of piston) times line pressure to get force available from cyl. Then it could all be adjusted for the weight of the assembly to make assembly movement more neutral at the servo. Most of the big Lathes that I worked on used the hyd system to lift the tool carousel, and they can move FAST !
With air, think about how long it takes to let the air out of a tire !

[ninefinger]

I seriously doubt that it would be too slow. 1st off the link I posted showed someone who made it work. 2nd lets do some math:

Speed of a pneumatic cylinder (inches/second) = 28.8 * CFM / piston area in inches^2

so lets use a very modest 1 cfm through our air line and assume a 3" bore and a 1" rod

pi X D^2 / 4 - pi X d^2 /4 = working surface area = 6.28 in^2
swap into our equation and we get speed = 28.8 * 1 / 6.28 = 4.585 in /sec
or 275 inches per minute.

I think even a 1/8" air line can deliver 1 CFM and a cylinder this size would likely have a 1/4 or 3/8" connection.

Next - it doesn't "consume" air - it displaces it back and forth from a reservoir. The only air required is to replenish that which leaks out past the seals. Not very much as far as a compressor is concerned.

Linearity - with a good size tank the pressure variation is small - so the force is quite linear over the stroke:
displacement of piston = area x stroke = 6.28 x 21 = 131 in^3
Volume of tank = 9 gallon x 231 in^3 per gallon = 2079in^3
change in overall air volume over piston stroke = 131/2079 = 6.3% to reduce this number use a larger tank or smaller diameter piston and higher pressure.

The mass will always matter in terms of acceleration. Even in an imaginary ideal world of no friction and no gravity the force to accelrate something depends on its mass. Double the mass (ie counterweigtht) and you double the force to accelerate it.

The free body diagram is usefull for a static situation, this is kinematics which I am rusty with but its coming back to me

Mike

[thebodger@roger]

No industrial CNC mill that I have seen in 35 years has ever used pneumatics for counter balancing. Why? Because air is very compressable. With temprature changes so does the pressure. Also the cylinder movement and force is very different at stand still and at maximum speed. Accurate speed and force with pnematics is very difficult to do. But this all seems like over engineering a simple problem. Let me ask you this. The cost of air cylinders, valves, linkages, hydraulic pumps, ect. ect. and what ever else you may run up against, how much cost and trouble are you willing to go through? Why not a simple lump of lead or steel, a few pulleys and an eyebolt on the head and on the weight. Job done. By the way, you don't need to fully counter-weight the head. A bit of extra down force can work to your advantage. Think about it when drilling or plunging in Z axis. And when do you ever cut with the Z axis going up?

[letmefixit]

Well I think we have covered all the basics. It really looks like either will work. They both have there PRO's and CON's. The counter weight would look kind of medieval unless you could hide it somehow. Hydraulics is expensive, air is ever changing. I think im going to try the air cylinder using a tank as an accumulator. Ill use nitrogen gas instead of air (tends to be a little more stable, and will make my grass green when it leaks out) hey eco friendly. My mill is supposed to be here Tuesday... am totaly stoked. Looking forward to further posts.
Thanks for all the info.

dj

[altaic]

Sweet! Hey, if you have those plastic handwheels do yourself a favor and get some new ones. Jergens cast iron are wonderful; part number 22201 (MSC #82403247) for the quill feed and 22402 (MSC #82403064) for each of the table axes. Enco has cast aluminum ones by Jergens for a bit less, which might be nice (haven't tried them, myself). Regardless it's not a pricey upgrade, and it's well worth it since it's the interface between you and your workpiece. The cast iron ones are polished for an excellent feel, and they have smooth and precise turning aluminum handles.

However, you'll have to bore them out to fit the shafts and drill and tap for a set screw. The table axes also require three teeth to be cut, which is a bit of a pain if you don't have a rotary table or indexer/spacer. I have a DRO, so I milled six small holes in the area to be cut away to use as alignment guides, which I stuck (two at a time) upended drill bits in to butt a parallel against (and clamp it to the handwheel) to indicate on so I could align it with my x-axis (powerfeed). I then indicated the shaft hole (Blake co-axial center finder) and offset to make the cuts-- note that the teeth on the shaft collar (on mine) are cut 0.1-0.2mm too small, so make sure to offset a bit extra to leave more material for a good fit.

BTW, I recently figured out a trick for the blake co-axial center finders-- as you are nearing getting the needle still, lightly press (momentarily) on various points around the main collar to determine the direction to go. Actually, you can press hard, but you may deflect it too much and the needle will swing more; what you want is to press to make the needle swing less and note the direction you are pressing.

Cheers,
Will

P.S. Sorry for going a bit off-topic. More on-topic, the reason industrial CNC machines don't use pneumatic counterbalances is that the pneumatic seals don't last as long as pulleys and/or spring steel; for them it's all about maintenance schedules. Huge machines would require huge pneumatic cylinders which are more expensive and will wear faster, and simply don't have the duty cycle to deal with that kind of use (or abuse).

Pneumatic cylinders are totally fine for our machines-- with a decent sized reservoir, the force will be near linear. And, they move damn fast-- I seriously doubt that people are seeing some kind of lag with them. IIRC, the parts can be had for around $100 from McMaster Carr.

I wouldn't bother with nitrogen; assuming it's CNC, surely the motors are beefy enough to deal with a slight change in load with temperature, let alone cutting forces! I mean, it's not like you're ditching your motors and encoders for pneumatic actuation. I would be much more concerned with the configuration of the counterbalance WRT the direction it is pulling, and how the z way will deal with it friction-wise.

[BobWarfield]

A better place to spend the money is one shot oiling for the ways. Sticktion is a bigger problem for the Z-axis than the weight in my experience. You can tell when really short moves fault the servos that it is stiction.

But, yes, lots of peeps have done counterbalances with both lead and gas springs.

Some pix on these pages:

http://www.cnccookbook.com/CCBlogSep2006.htm

Cheers,

BW