Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: trek_fan on April 26, 2008, 08:57:01 PM
-
Hi all,
I would like to say my noobishness knows no bounds when it comes to electronics. Electronics are a bit out of my realm. Anyways , here goes: I have had this project in the works for quite sometime now, and am finially moving forward. I have tried to research this for several weeks on the net and ended up confusing myself even farther. I want to use servo motors instead of steppers, due to the fact of lost steps. I would be willing to use steppers if a encoder could be mounted to close the loop. I am planning on using 1/3 HP 90V Permanent Magnet DC Motors. These should operate nicely with the off the shelf Gecko Drives. I got the idea from this source:
http://www.truetex.com/servomod.htm
I plan to purchase (3) motors from Grainger (X,Y,Z) http://www.grainger.com/Grainger/items/2M509
I already have a 1/3 HP 7,000 RPM Spindle Servo Motor with a cat 40 taper in hand. What I am now overly confused about is this:
1. I was planning to use this http://www.geckodrive.com/product.cfm?pid=13 Gecko 320 drice. Will I need a 80V DC power supply to feed the motors somehow?
2. Does Mach3 feed this unit directly?
3. Does Mach3 Feed the Gecko G100 which in turn then drives the 320's?
I guess I am only asking what hardware would I need? I will then try to go from there.
I already have started construction on the machine, and most of the pieces are already in hand , Linear Rails, Ballscrews, Etc. I have built a lot of the components where I work using our CNC laser to cut 3/8 plate steel for the machine parts. Nothing like laser cut holes to line up the linear rails and such. Also using our CNC lathe and mill to make the mounts, etc. As far as building the machine, I feel pretty confident. I just am lost on the electronic parts of the connections and equipment required for the servos.
Basics on the machine are thus : 36" X 36" x 12" machine area. All running on THK linear bearings and 1 1/2" ballscrews. The gantry shows in my 3D models to weigh in at around 125 pounds, so that is why I want such beefy motors.
I really apprictae any help I can get on this, and hope the flames for my noobishness will be kept to a low level ;)
Thanks,
Rob
-
You can probably find some decent dc brush type servo motors with encoders already fitted on ebay for a little less. There is a fellow that pretty regularly has three new in the box reliant servos for aprox $500 .
In addition to the gecko drives you will need a dc power supply that can provide the needed amperage and voltage.
You will also need a breakout board of some sort that will go between the computer and the gecko drives. Computer hooks to the breakout board , break out board attaches to geckos , limit switches , solid state relays etc. The power supply will attach to the drives. There are a number of breakout boards avail , I like this one http://www.cnc4pc.com/Store/osc/product_info.php?cPath=33&products_id=161
-
You might know a lot of this is you have been reading up on it.
Mach 3 puts out signals on the LPT1 - the old computer printer port. There are only 12 output lines and 5 input lines ( It is the input lines you might run short of) - so run through what you might need.
Each axis requires 2 wires (plus a return wire) to drive it, one wire drives the step, one wire drives the direction. If you have four axis you need eight wires etc. Don't just allow for three axis - most people eventually want a rotary table and other goodies.
Outputs must also drive things like the spindle control (normally three wires) coolant (one or two wires) axis enable wires (if required) one for each axis. etc.
Inputs are for limit switches, home switches, spindle speed, and any other switch you might want to put on your lathe like an "emergency stop" button or other safety device.
You can see that it is very easy to run out of wires, when your system gets a bit advanced - and you end up having to add other ports to your computer.
Many firms offer driver boards and breakout boards etc. A breakout board merely provides an easy way ( usually screw terminals) of connecting your bits and pieces to your computer. Many breakout boards have electronics on board to isolate signals and provide a 5 volt power supply for switches etc.
The disadvantage with a lot of breakout boards is that they are pre- wired and fixed in their applications - rather like painting by numbers - fine to get started with, but in the end, very limiting.
The Mach 3 admin boys are going ******** at the moment about a new interface board called a Smooth Stepper. This is a board that provides inputs and outputs (more than you need), it provides on board pulse engines to drive your axis, which means that your computer does not have to generate the pulses and can get on with calculating the job. I understand that it is opto-isolated , etc etc . The beauty about it is that it connects to your computer via a USB lead. I would recommend that you look at one of these ( if you are starting from scratch).
Whichever type of board you start with - you then need to interface this with your machine.
Geckos are good drives and I am pleased with mine - so I cannot crticise your choice. Use seperate drives - Do not get a single board that has breakout board and three drives all in one - if one goes faulty it all goes faulty. If you have seperate units you can change anything at will.
The output from your breakout/smooth stepper board will drive the Ghecko board directly. Mine require a 5 volt connection, a "step" connection and a "dir" connection to each board. The inputs to the Gecko are isolated from the rest of the board to protect the low voltage side of the system.
I have looked at the Grainger website and your choice of axis motor. I am at a bit of a loss. These appear to be a straight DC brushed motor, and there is no mention in the blurb of feedback for positioning, There is mention of a tacho - is this for positioning.
Mach 3 gives out step and direction signals. It is preprogrammed with the number of steps per inch of axis movement, so if you want to move one inch - it puts out say 60,000 pulses and says - "there you are - one inch". With steppers you hope the motor got all the pulses, but with servos (as I understand it) there is a measurement system either inside the motor, or external to, which checks that your motor has gone the relevant distance. I do not see that system mentioned in the Grainger.
If you read the PDF file on the Gecko 320 it tells you how to wire them, and also about the feedback circuits. You can get external feedback circuits, but I thought that servo motors ( marketed as such) came with them ready installed integral to the motor. I would check the motors.
If you need any more - come back again.
-
Thanks for the fast responses. In answer to your concerns on the axis motors, I referenced an articale that demonstrated how to mount a digital encoder to the back of the motor to turn a regular dc motor into a servo. My gantry alone shows a weight of 125# on my 3D model, not counting the Z axis assembly, Rails, Drives, etc. which is why I wanted to go as large as I can on the axis motors.
http://www.truetex.com/servomod.htm
I was going to use that feed back to make the "closed loop" on the motors. I had planned on using those motors, and have machined a plumbers block to support the ballscrew and attached a Renco digital encoder to the ballscrew, not the motor. I am using Lovejoy couplings and they will flex and compress a few thousands as the machine runs, so in order to maintain true positioning, I am enconding the drive shaft, not the motor shaft. After reveiwing this website I am loosely basing my machine on his:
http://www.oneoceankayaks.com/madvac/madvac_index.htm
I would be willing to go with stepper motors if I could find a driver that would accept encoder feedback and drive steppers. I just don't like the thought of lost steps.
I was also looking the usb driver board you mentioned, but I thought it said mach3 didn't support the USB at the moment.
After reading your post, I think I may understand a bit more. You can use a breakout board / smooth stepper to "run" seperate drives, OR you can use a all in one wonder like the gecko g100.
-
The SmoothStepper works great with Mach, it still has a few things to get worked on such as the soft limits and threading sync for turning but I know Greg is working on that now and with the excellent work he, Brian and Art did getting the plugin issues sorted during the Beta testing I have no doubts whatsoever that they will be finished shortly.
As for steppers you can get the encoder Interface board from Ron Rogers www.rogersmachine.net This allows you to use encoders with steppers and if the steppers get out of position by the user set following error it will halt Mach and thus parts wont be ruined. In reality a well matched and setup stepper system will never lose steps but it is always an added safety feature.
Servos are the way to go and if you can afford them AC are the best but if you can compromise on speed then steppers will suffice.
Hood
-
Oh just another thing regarding mounting the encoders on the ballscrew rather than the motor, if you have play anywhere, and you have said you will in the couplings, then you may get issues with the motors constantly hunting back and forth trying to keep position.
Hood
-
Ahhhhhhh the Good Old Lovejoy Coupling! I love 'em but I think you're going to see a lot of backlash. I guess you could weld the 2 couplings together to form a rigid coupling. But then, why not just get a rigid coupling? Better still, check out http://www.sureservo.com/couplingconsiderations.htm there is a list of servo coupling manufacturers there. One Company, Ruland has a really good pdf explaining the differences between the diff types.
(Quick Hijack) Hey Hood, have you ever tried the Rogers encoder interface? I was speaking with Ron the other day about how it works and it seems like it would be great for steppies & Servies?
Regards,
Sid
-
Sid
Yes I borrowed one way back in the very early days of the encoder interface. It worked well as far as when I deliberately made the following error so small or deliberately stalled the steppers Mach would get stopped. There were quite a few niggly problems with it back then, cant really remember but I think homing was one of them. Since then however Brian has done wonders with it , think its now a plugin that controls things rather than the macropump. I dont have one now so I cant test it out but I have only heard good and all issues are reported to be in the distant past.
Hood
-
Cool! Thanks for the NON-"Silly Answer" ;) BTW- What happened to your "Bad-A$$" Avatar- I kinda liked it!
Regards,
Sid
-
it was Benny that gave it too me, I let him have a few days of fun then removed it, well you know what these Ozzies are like, they go in the huff easily if you dont humour them :D
Hood
-
Hope this doesn't over flame your noobness - but.... can I ask a simple question? what is your rig designed to do? I ask because I'm thinking 1.5inch ballscrews 36inches long - 125 lb gantry - this is for serious high speed deep milling of carbon steel - yes? :)
-
I will be doing some high speed machineing on this machine center. I am not sure how much though. I plan to do mostly aluminums and woods, but am designing to be a sturdy workhorse capable of handling steel and stainless as well. That is why I am going with a 7k rpm servo as a spindle motor to give me the option for rigid tapping and other options. I will be using a 2" wide timing belt to run the drive shaft on a 2:1 ratio. I have a air release cat 40 taper to automatically release and grab tools. Also I am hopeing to get 10k to 12k rpm comfortably at the tool. This machine center is in the "from scratch" stages. I have all the machine frame designed, and most of the parts fabricated. I have only ran into the problem of the electronic parts (what all pieces needed to drive servos), but thanks to some good responses here, I will be moving forward shortly. I had to stop fabrication due to motor mounts, and the electrical enclosure needed.
The gantry weight of 125# is just the gantry FRAME, not the drives rails, motors, etc. I expect it to come in around 160 - 180 pounds when complete. That is why I was looking at 1/3 hp dc motors for the axis drives and retrofitting encoders to the output shaft or drive shafts, to turn them to servos. I would prefer to ac motors and encode them to work as servos, but I have not found much information about ac drives. There are a ton of them out there, but no good explanation of how they work, hookups, etc. I prefer ac over dc for a hp and economincal reason, more power for less juice, and motors are far cheaper 1/3 the price.
I have the machine designed in SolidWorks, and this weekend I can generate some 3D models and post them if you like. You would be able to open them in Internet Explorer and rotate in 3D measure, etc. As a quick background note, I am the Engineering Manager at a medium sized fabrication plant. We have just about every cnc fabrication tool you can think of, from plasma's, laser's, shear's, turret's, brake's, mill's, lathe's, and on , and on, and on. I am good friends with the owner, so I have the benefit of using anything I need to build my machine (mainly cause he wants one too :) , which is a project for later)
My plan is thus: I want to finish this machine and use the experience (and hopefully revenue) to build a larger 5'x10' machine mainly for wood routing, and extremely light machining on soft metals. I would like to incorporate a 4th axis and tool changer onto my bigger machine, but at the moment it is just wishful thinking.
-
I'll make a couple of observations that may or may not help. You say you're happy to use steppers if you can use feedback/encoders because you're concerned about lost steps. Steppers only lose steps if they are under-spec'ed for the job or they are incorrectly driven. Adding encoders does absolutely nothing to change this. If motor x loses steps without encoders - motor x will lose steps with encoders.
But guess what - an under-spec'ed servo will also lose steps/position regardless of the fact it has encoders. Conclusion - encoders will do nothing whatsoever to stop either motor from losing position if it isn't powerful enough for the application.
Your spindle motor.... 1/3HP - 7000 revs - I think you need to reconsider this. Considering the massive construction of your machine this seems to be waaaaaaay too puny - 1/3HP is ~250Watts - that's about dremel power!!! you'll be lucky to cut plastic. 7000rpm - not for wood at any sort of feedrate - you'll need way more.
Just my 2c
-
Please forgive me if I am wrong, I thought that in the closed loop system, ie. an encoded motor, if it were to lose steps, then encoder would provide the feedback, and the system could do error correction to get back into position.
In regards to your concern about the spindle servo, you are correct, I had not thought about that. I was sizing everything to run off the standard off the shelf gecko drive, but I guess I will need to seriously upgrade the spindle servo. I will have to do some research to see what a good size spindle servo would be.
-
Please forgive me if I am wrong, I thought that in the closed loop system, ie. an encoded motor, if it were to lose steps, then encoder would provide the feedback, and the system could do error correction to get back into position.
OK a fair thought - and a very common one. Let's try to answer by asking a couple of questions...
Question: Why does a motor (stepper or servo) lose steps/position?
Answer: Because it doesn't have sufficient power to overcome the resistance it's being subjected to.
Question: So how many steps does it lose?
Answer: all of them for ever and a day. (or maybe it vibrates the obstruction sufficiently to eventually get going again - this is typical for example of a "sticky" leadscrew)
Here's a pretty good read: http://www.torchmate.com/motors/electronics_selection.htm
-
Very, very interesting article. I was under the impression that steppers were notorious for losing steps on a milling application (from an article I read somewhere). That is why I was heavily leaning towards a servo application. 1200 - 1500 oz steppers would be a more economical situation. I was not being smart or crass in my last response, I am working hard to try to understand the electronics part of the motors and their drives.
-
http://www.anaheimautomation.com/manuals/L010165%20-%2034Y%20Series%20Spec%20Sheet.pdf
That motor for instance is only approx $175 new. (1700 oz one)
-
Hey - no problem. If I had a quid for everyone that thinks steppers lose the occasional step here and there I'd be a rich man. The question to ask is why would they do that? Maybe they loose concentration and forget what they're doing or something ;D
Seriously though - I'm not dissing servos - far from it. For a start I don't know an awful lot about them - but I do know steppers, and I know there's lots of cr*p talked about them. Steppers generally don't live up to expectations because they're either not sized or driven properly or both. As a rule of thumb (FWIW) I use this: if the application demands more power than you can get from a (properly driven and properly sized) stepper then use a servo. But first - know the power you're going to need. So far I've never had the need for a servo.
Cheers
Ian
-
1700 oz would be Kong size to me :o . My steppers are only around 700-800 oz-in, and the first thing to wimp out is my weak porter cable router. I weigh in at 200 lbs, and I can't stop my gantry pushing against it with all my weight, and thats at a speed thats faster than my cutting speed. Steppers have more power the slower the speed, so If I were to really slow it down and run in reverse, I think the earth would start to slow down. As Stirling stated, a properly driven system is where all your power potential comes in to play. For instance, I can grab the ball screws or motor shaft and make the motor loose steps with little effort. It would be impossible for me to make the motor loose steps by hand from putting pressure against the ball nut on that same screw (or whatever the ball nut is attached to. Spindle, gantry, etc...) Same goes for my belt driven axis. By the time the power is transfered through the different pulleys, lost steps are of little concern. Most likely you won't need your machine to go 1000 inches/minute, so a properly built machine matched to do what you NEED it to do is fairly easy, even with small steppers. However, if somebody was to say, "hey, I'll set your machine up with this nice servo setup in trade for your steppers" I would gladly swap them. 1700 oz/inch?? Yeah, thats pretty big looking through my eyes. Wish mine were that powerful, just for the bragging rights. :P
-
I would rather spend the extra money on the front end and get more than I think I would ever need, than to reflect back, and wish I had got more. I have never heard anyone complain their car has too much horsepower :) !! I want to only by at my upper limit using 75% of available power. I have to seriously upgrade my spindle. I hadn't given it much thought, until Stirling mentioned how puny it was (my manhood was threatened :) ) Anyways, my project goal is to be able to comfortably run a 1" (25.4 mm) 4 flute endmill through carbon steel at a reasonable clip.
I am having serious fun with this project, and really can't wait to complete it!
-
You may not get the speed you want with steppers, depends what you want I suppose :D
I have steppers on my Bridgeport and they are great but on my Lathe I have AC Servos, mainly because its a big brute and steppers wouldnt be big enough but also because I want the speed, especially the acceleration you can get with servos. I have just been testing out a 5.5KW AC Servo that I am putting on the Mills spindle, had it going 4000RPM CW and put it 4000RPM CCW, had the acceleration set to 160,000RPM/Sec and it did the reversal in 0.05seconds with no positional error reported in the software :). The mill I am doing at the moment also has AC Servos on the axis and dont think I would go back to slow steppers again after seeing what servos can do. Saying that however steppers can do a great job if you are not in a hurry and as has been said if they are sized correctly you wont have problems.
Hood
-
comfortably run a 1" (25.4 mm) 4 flute endmill through carbon steel at a reasonable clip.
Yeah, that would definitely go beyond my description of a hobby machine, if theres any depth to that cut. The price for a good capable spindle alone would cost more than my entire setup. :) Lets see those models you speak of! 8)
-
I would like to qualify that and say I WOULD LIKE TO BE ABLE :) I don't expect I can get there, but it is my target, I will be more than happy to see the machine milling aluminum with a 1/2 inch single flute, but I want to maintain my target of the 1" in steel. I have to set a high goal, and if I fall short it will still do everything I want it to.
I will get my models into a some decent eDrawings and post them this weekend.
-
trekky - don't be too easilly seduced by size - big is generally better - but not allways. Nothing comes for nothing - everything's a compromise. Large stepper motors can introduce problems such as difficult tuning due to large detent torque. etc. etc.
Question: why do we need lots of power in our drive motors? - answer: to push the tool through the stock? - WRONG! - We need power in the spindle for that. A good spindle will go through stock like a knife through butter - a bad spindle will never go through it no matter how hard we push. We need torque in our drive motors to accelerate the mass of the gantry and we need power to maintain that torque accross our speed range.
But what speed? - well - the right speed for the job, our spindle and cutter (with maybe some to spare - for luck!). OK so you have a rig that will travel at a Km/sec - but your spindle and tool will only cut at 1mm/hour - that was money well spent then!!!!
That said, Hood is absolutely right - accel is also hugely important - you can never have too much, because what we'd actually like is instant speed, i.e. infinite accel. I'm no expert on servos as I've said - but I think I'm right in saying that at lower speeds, steppers will out accelerate servos by a gazillion miles (per sec per sec ;D) because they have more low speed torque. To get torque out of a servo - they have to be spinning fast. - so what happens at low speeds? - in order to work - you have to gear them down. That makes me think I'd want servos for either really high speed applications or really heavy rigs - course - that's just my opinion.
So..... read up - then do the maths - it may turn out that you'll have way more power than you need from 360 inoz Nema 23's. It may turn out you need 800 inoz Nema 34's OR it may turn out you need to buy Hoods servos ;D
Tip for the day - learn about inductance and what it means to your motors - it's far more important than a seductive holding torque. Remember holding torque only tells you how good the motor is at standing still ;)
-
That said, Hood is absolutely right - accel is also hugely important - you can never have too much, because what we'd actually like is instant speed, i.e. infinite accel. I'm no expert on servos as I've said - but I think I'm right in saying that at lower speeds, steppers will out accelerate servos by a gazillion miles (per sec per sec ;D) because they have more low speed torque. To get torque out of a servo - they have to be spinning fast. - so what happens at low speeds? - in order to work - you have to gear them down. That makes me think I'd want servos for either really high speed applications or really heavy rigs - course - that's just my opinion.
I have to disagree with that I am afraid and to back up my case I have attached a screenshot of the torque curve of the AC Servo that I am putting on as my spindle (the other AC Servos I have on the axis are exactly the same with regards to their torque). As you will see the continuous rated torque is constant from zero RPM up to the rated speed(about 2/3 of the max speed) and even after that the drop off is nothing like a stepper.
I dont know whether DC servos are the same but I think they are but wont say they are definitely. As I said I have steppers on my Bridgeport series 1 CNC and the acceleration is noticeable, on the Lathe you dont notice it at all, it seems instantaneous.
I am not sure why people gear their servos, might be that they have to do that to get the torque they require as their motors are smaller than they actually need, my lathe has big motors and is geared 1:1 and the latest mill will also be that way.
Hood
-
Hi Hood - well I did say "I'm no expert on servos as I've said - but I think I'm right in saying..." - that'll teach me to stick to what I know ::) ;D
.....That said.... ;D ;D ;D
Here's why I said what I said....
http://www.allbusiness.com/sector-92-public-administration/justice-order/1177040-1.html
-
Yes but from a quick squint at it they are not comparing like for like. I may be wrong as I am in the middle of working on a boats engine at the moment and dont have time to thoroughly read it. What I am seeing is they are comparing the wattage of these motors and not the torque. Also I dont know how they get the servo as having its peak torque at 3000RPM as I cant see any diagrams or charts, maybe they just arent loading because I use Opera. As can be seen by the above chart I posted my motor actually drops off slightly at higher revs and below is another to support my case, this is a motor I have on the lathes X Axis. Also I have attached a quote from the Rockwell Automation manual for the motor
What I can say for sure is you need to size the motor to your application and no matter whether its a stepper or a servo if you undersize you will get bad results. If you get the size right you will have a good system, just with different characteristics. Oh and just to show I am not Anti Stepper I have just made a coil winding machine for a customer. Money was not a problem with this build in as far as hardware was concerned but there was no way I was going to spec servos for it. Plain and simple reason was they were not needed so would have been a waste of money, the steppers are more than fast enough and as has been said if sized properly then accuracy is not an issue either.
Hood
-
Hey Hood - No argument from me - you the servo man ;D - I was just letting you know what I'd read.
Cheers
Ian
-
I am certainly not an expert on servo motors and I am just hoping the torque curves are correct. However I very much doubt that Rockwell would dare put out false representations of what their motors can do.
Hood