Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: Katoh on September 20, 2013, 11:53:19 AM
-
G'Day Friends
I am searching for suggestions on my next build/conversion. I have a Bridgeport knee mill that I bought solely for this purpose that has being sitting on ice. I am now slowly getting thoughts together to purchase parts for the conversion.
I'm not really a newby, I built a router from the ground up and converted a lathe, both are Mach controlled and stepper run with gecko's as the drives, actually the router runs twin steppers on the X and Y, but its a big machine.
Now back to the mill.
I'm tossing up ideas, thoughts and parts. I definitely will be changing all lead screws to ball screws, preferable ground If I can get them at the right price. Motors and drives I'm still not sure.
On a mill of this size would you go Steppers or Servo,s?
Has anyone had any experience with the AC Stepper drivers advertised?
http://www.ebay.com.au/itm/CNC-Mill-High-Speed-Torque-Stepper-Motor-Driver-Controller-AC80-250V-8A-400KHz-/321170854171?pt=UK_BOI_Industrial_Automation_Control_ET&hash=item4ac746511b
If we go steppers would it be better to go one bigger motor than two smaller on the same axis?
I have read a bit about stepper encoders, or is this just a waste of time?
Thanks for reading looking forward to your thoughts.
Cheers
Katoh
-
Personally I would use Servos but they are the more expensive option.
Steppers will work well just you will not get the same Velocity or Acceleration that you could expect from servos.
I did have steppers on my Bridgeport for a long time and it worked well, just a bit slow. They were 916 ozin and geared 2:1.
Regarding steppers with encoders, I think that is a waste of time.
Hood
-
Thanks Hood
I was thinking servo's, but I have never used them before. This could be a new experience.
With servo's you need the motor and controller as unit or do you look for motors and use something like a Gecko G340?
Looks like a lot of study coming up!
Cheers
Katoh
-
There are two (well actually 3) types of servos normally used with CNC machines.
Ac
DC Brushless
DC Brushed.
Of the 3 the DC Brushed is the easiest to mix and match drives and motors.
However AC will in my opinion get you the best performance.
It all depends on your budget though as DC will be cheaper, probably not much more expensive than steppers.
There are some nice second hand AC servos about however which can make them closer to DC servo prices, still not cheap however.
Servos need tuning which can be a bit daunting for people but a lot of the modern AC Servo drives do an excellent job of Auto Tuning and some even do constant tuning which means you never have to tune them and the drives automatically tune them constantly as they are running.
Hood
-
I've being looking at servo's on ebay, must say its getting a bit daunting.
I was under the impression AC servos are the best but you need to have matching drives and the drives need to be able to accept step and direction.
Looking at the DC motors and drives I can't seem to work out how much power I need. I know with steppers on the mill I would have started with 1200oz/inch motors as a min and worked up from there. but I can not make a comparison to servo's. What I thought would be large enough only seem to have 1/4 the torque and the prices are just way high.
Maybe I'm not reading this right.
Thanks
Katoh
-
Servos and steppers are two very different animals. Steppers produce their max torque at stanstill (Holding Torque) and it drops away with speed, quite rapidly after a few hundred RPM usually. A servo on the other hand normally is rated for constant torque and peak torque. The Constant torque is what the motor can produce constantly from zero RPM right up to Max rated RPM. In addition Servos have a big overhead for short periods, usually 2 to 3 times the constant torque.
For a Bridgeport sized machine I would think 750Watt AC servos would be more than adequate at 2:1 reduction, some people reckon even 400W would suffice although I am not so sure about that, suppose it depends what you will be doing and cutting and what accel and Vel you are wanting.
There are currently Samsung 750W drives, motors and cables on eBay, they are second hand obviously and they are asking, if I recall correctly, about $1800 for the three sets and 400W ones for about $900.
Hood
-
Thanks Hood
Again you have enlightened me. I do like the Ac servo's, and the Samsung ones do look mighty attractive, but with a price tag to suite. This definitely gives me a starting point to know what to look for.
I was also looking at those whopping big steppers, size 42's 4000 odd Oz/Inch beasts that you run with the big AC stepper drives, not as expensive as servos but still when you put 3 of them together and the drives and freight the servos really aren't that much more.
Well Ill have to start putty my penny's away, and keep an eye out for a bargain.
Many Thanks
Katoh
-
Servos are expensive compared to steppers but once you have used them, especially AC ones, then you will know the extra cost was worth it :)
Hood
-
In addition Servos have a big overhead for short periods, usually 2 to 3 times the constant torque.
For a Bridgeport sized machine I would think 750Watt AC servos would be more than adequate at 2:1 reduction, some people reckon even 400W would suffice although I am not so sure about that, suppose it depends what you will be doing and cutting and what accel and Vel you are wanting.
Hood
These days it is difficult to compare machines by 'size' so I suggest focusing on the weight (mass) of the load and not the size. For example, I just finished a servo conversion on the largest asian bench mill, namely the Industrial Hobbies mill with the 'oversize' table. While the table is close in length and width to the a bridgeport table, it is only half the weight. Another difference betewwn servos and steppers is that you can oversize a stepper without negative consequences. Oversizing a servo can result in all sorts of problems including making the system un-tunable, defeating built in auto tuining and/or having the motors break into uncontrolled self exciting harmonics (given an elastic connection to the load).
The IH conversion has 400 watt motors moving the 150lb table. An actual Bridgeport table is over 300lbs and, as you suggested, would require the 750watt motor to achieve similar performance.
On the IH conversion, the head was scratch built and weighed in at over 300lbs so the Z axis in fact has a 750 watt motor.
Mitsubishi publishes a software specifically for sizing their AC servo motors and that is what I use to spec the motor size for a particular application. It is a bit complicated and requires a lot of technical details about the application including things like the coeficient of friction of the sliding members and stuff like that, but one could also use a 'rule of thumb' approach and probably not get into too much trouble. If one accepts that 400 watt is good for a 150lb table (with dovetails) and 750 watt is good for a 300 lb load (on ball slides), then it might be reasonable to extrapolate a rule of thumb like 1.25 to 1.35 watts per lb. of total mass to be moved (for a machine tool application).
Note that for calculating the horizontal movements of the table, the weight of the table (but not the mass) can be discounted. A 'live load' however shoudl be included since this is the actual typical operating condition. You might consider 1x the table weight as a reasonable WAG. So in our examples, the IH table would carry a 150lb load for a total mass of 300lbs to be moved. 400 watts / 300lbs = 1.33 watts per lb. Similarly the 300lbs bridgeport table carries a 300 lbs live load (mill vice, 4th axis, fixtures, workpiece, etc) for a toal weight of 600lbs. Therefor 750watts /600lbs = 1.25 watts per lb.
The Z axis is a bit different in that ther is no live load . . other than the head itself . . which has 'weight' to be lifted in addition to the mass to be accellerated. So if the head is fully counterbalanced (air springs, aka struts, or equiv) Then just use the all up weight of the head as the load to be moved. If the head is not cojnterbalanced, then add the 'weight' of the head to the mass of the head and use that.
DISCLAIMER; ::) the proper calculation method for motor sizing are complicated and the 'rule of thunb' described above is not presented as the correct nor the preferred, nor accurate method to calculate servo motor size. It is a theoretical context without warranty express or implied . . . use at your own risk . . . batteries not included.
-
Oversizing a servo can result in all sorts of problems including making the system un-tunable, defeating built in auto tuining and/or having the motors break into uncontrolled self exciting harmonics (given an elastic connection to the load).
Interesting, I found the exact opposite on my Bridgeport when I swapped it over to servos. I had some 1.5Kw motors and they were extremely easy to tune, in fact they were the first ones I have had any real success with the Auto Tuning in the Allen Bradley drives. Other setups (Computurn Lathe and Beaver Mill) the auto tuning has been only partially successful and needed a lot of tweaking to get to my liking.
Hood
-
Oversizing a servo can result in all sorts of problems including making the system un-tunable, defeating built in auto tuining and/or having the motors break into uncontrolled self exciting harmonics (given an elastic connection to the load).
Interesting, I found the exact opposite on my Bridgeport when I swapped it over to servos. I had some 1.5Kw motors and they were extremely easy to tune, in fact they were the first ones I have had any real success with the Auto Tuning in the Allen Bradley drives.
Would I be correct in guessing that you have a big bridgeport with dovetail slides and the orignal leadscrews directly coupled to big low RPM servo motors with nice heavy armatures?
-
Steve, it was a Bridgeport Boss 6 series 1 CNC , so dovetails on short table, ground ballscrews and the motors I put on it were Allen Bradley MPL (Light inertia) motors with a rated RPM of 5000 and I kept the original gearing of 2.5 :1
Hood
-
I kept the original gearing of 2.5 :1
Gears or rubber belts?
-
Timing belts, original Imperial sized ones.
Hood
-
Mr. Hood, sometimes you just get lucky, I guess. Probably your proximity to Ireland . . ;)
-
This thread seems pretty much dead, but before I sign off, I want to mention a few caveats and clarify some terms in case there is confusion about the 'rule of thumb'.
*Stepper motors ratings and Servo motor ratings cannot be compared directly as they are different measurments made in different ways. For the same reason, industrial AC servo ratings cannot be directly comapred to DC servo ratings, particularly the typical hobby level DC servos and drives.
-
In hindsight, I think a 'rule of thumb' is a bad idea for this topic, but I already stuck my foot in it, so this post is a final attempt to clarify some of it before I sign off the thread and hopefully it will prove genrally useful in selecting and purchasing Servos for new builds.
UNDERSIZED:
*Stepper motors ratings and Servo motor ratings cannot be compared directly as they are different measurments made in different ways. For the same reason, industrial AC servo ratings cannot necessarily be directly compared to DC servo ratings, particularly the typical hobby level DC servos and drives.
for example a Keling 90V 40A motor would theoretically be 90 x 40 = 3,600 watts peak. But paired with a Gecko drive which only puts out 20A max, that number gets cut in half. A typical linear power supply drops voltage as more power is drawn, so you will never see full voltage at full amps. Therefor take away for various losses and you might have 1,200 watts peak. Industrial AC servos are rated very conservatively and typically have several times their rated torque as a useable reserve.
Bottom line is that a 400 watt industrial AC servo motos is roughly equivalent to the 3,600 watt hobby DC servo/Gecko combination . . . so direct comparisons are not meaningful.
* Given the vast difference in 'rated' vs real world performance, it should be clear that a 'rule of thumb' must be taken in context and not applied across the board. In the case of my rablings, they apply to Industrial AC servos.
OVERSIZED:
*Not all 'Auto Tuning' is created equal. Newer models can be vastly improved over previous drives and if you are buying 'whatever' brand used on Ebay for example, do not assume you will be getting the same level of capability as the setup you just read about on a foruim (unless you buy the same parts).
for example: Mitsubishi J2S series servos are very susceptable to the harmonics issue that I described earlier (elastic connection of an oversized motor to a large load). The newer J3 series drives can measure the machine's resonance and has input filters and smoothing and can detect a variety of harmonic issues and compensate automatically.
Bottom line is that the J3 is nearly immune to a common issue that will absolutely confound an earlier model drive. Don't assume that buying a 'Mistubishi' will erase a bunch of problems. Make sure you are getting the same model as are reading about.
*'Oversize' in the context of my 'rule of thumb' means that the motor falls outside the drives inertia ratio as published in the drive manual. This varies quite a bit between brands and between models and also with the types of motor.
*'Inability to tune' in conrtext means that the tuning can only be accomplished with the gain set very low, which is a typical bandaid for a host of problems. Prozac for Servo Drives. In particular this is one of the solutions to the harmonics problem I described. With the gain (responsiveness, frequency, etc) set low to overcome vibration issues, the overall performance envelope suffers and the super responsive 'stop-on-a-dime' behavior without overrun or underrun is not achievable.
*'defeating AutoTune' in context means that if the overall system responsiveness (gain, stiffness, etc) has to be set so far from the default (to counteract instability) that the auto-tune cannot achieve the performance as described above, then the Auto-Tuning has been defeated, in my opinion.
Bottom line; tuning in general for ALL servo drives and Auto Tuning (which comes in m any flavors) are able to work withing a certain range of parameters. This 'window' is pretty much a guessing game with the hobby level drives, and you find out it won't work when it doesn't work . . usually after many frustrating hours of trial and error. However, on the commercial/industrial side, the envelope is well defined and published in the documentation. Ultra low inertia motors are targeted at super-response applications like pick-and-place machines and similar ultra-resonsive, light-load robotic applications. Given a simiar ineria ratio, 'oversize' for these motors (moving the same load) would differ from the same size, same power 'standard' motor by a factor of 2.
That's a wrap and I'm outta here.
Either way, good luck with all projects and endeavors. That's a wrap. I'm outa here.
-
Only times I had trouble with tuning a servo were when the servo was undersized.
Dan
-
Only times I had trouble with tuning a servo were when the servo was undersized.
'trouble' is not a useful description. Anyone who knows how to tune servos can certainly be more specific than that.