Servo motor drives in order of quality and features?

[Apples]

Hi,

Lets talk about servo motors, and drives.

How does one know what servo drive to go with. My understanding is bottome end scale might be Geckodrives like the G320 and G340.

Then you might go with Galil.

Then maybe higher up again with manufacturer specific drivers etc.

.......
So what I really want to know is what are the limitations of say the geckodrive G320 to say a Galil etc. I beleive that geckodrive is step/direction only and is to be used with mach series pulse engines.

Galil from my limited research seems to have to be used with Camsoft, and are of the +-10volt type.

Geckos don' have closed loop where as Galil do.

With all the different manufacturers around, how can one tell who has the better servo drives to use. for example. gecko g320 will sure enough drive servos quite fine if the machine is not overloaded and will probably position the machine accuratly all day. But so will a Galil setup, if it is used carefully and not pushing the machine hard.

So why would one need to go "higher up the ladder" so to speak tand use a Galil servo driver over a gecko servo driver if you are going to run the machine at non destructive positioning speeds.

Surely a geckodrive would be fine wouldn't it?

Then how can you tell the features and the real life benifits and differences between gecko drives, teknic drives, rutex drives and larken servo drives etc. These all seem to be the most commonly talked about drives on cnczone that I have heard of. Is this because they are lower "grade" drives that are not used on bigger industrial machines, and that they are just the most common as this is what people are using for DIY and smaller conversions and setups etc?

I'm just trying to figure out how to select a good reliable servo drive system that would be good enought to offer as an OEM solution for some machines I one day hope to make. As I would be building machines to sell to customers, I would want to make sure that the servo drives were of highest reliabilty.

Having said that, the market that I would be targeting for these smaller machines does not want to pay tens of thousands of dollars just for a servo drive etc.

So where do you draw the line?

Cheers,
Peter

[Al_The_Man]

Apples, This is a little bit apples and oranges
First Galil use AMC drives, but it can be anybody's, which as you say are ±10vdc analogue drives, these are relatively non-intelligent drives as they rely on the servo loop being closed back to the controller, which would be the Galil card or whoever's.
Due to the fact that through-the-parallel port servo systems do not close the loop back to the PC, they rely on semi-inteligent drives to close the loop, so the encoder has to return to the drive, rather than the controller.
Supposedly, if it gets off the ground, the Brain that is being touted as a closed loop control, there a quite a few benefits from true closed loop, electronic gearing being one, If they get the ±10vdc analogue drive feature working it will allow a larger selection of drives, for e.g.
Systems such as Mach, give up a few possible features for the sake of economy.
Al.

[samco]

sick of me yet?

There are a few printer port interfaces that close the loop in the control to run servos. (these work with emc2 <- free - open source machine control http://www.linuxcnc.org/ )

pico systems
http://jelinux.pico-systems.com/univpwm.html <-pwm control
http://jelinux.pico-systems.com/PPMC.html <- +/-10v

Mesa Electronics 7i43
http://www.mesanet.com/parallelcardinfo.html

pluto
http://linuxcnc.org/docs/2.1/html/hal/pluto_servo.html

other hardware for closed loop servo control can be found here.
http://wiki.linuxcnc.org/cgi-bin/emc...orted_Hardware

sam

[antichip]

Really the only need for the controller card in EMC is for the +-10 signal? So why is there not a Servo drive made or being made with a 5v pwm signal input?

[Al_The_Man]

You mean like this http://www.a-m-c.com/download/datasheet/azbdc6a8.pdf ?
Al.

[snoopy27]

Originally Posted by Apples Hi, Lets talk about servo motors, and drives. How does one know what servo drive to go with. My understanding is bottome end scale might be Geckodrives like the G320 and G340. Then you might go with Galil. Then maybe higher up again with manufacturer specific drivers etc. ....... So what I really want to know is what are the limitations of say the geckodrive G320 to say a Galil etc. I beleive that geckodrive is step/direction only and is to be used with mach series pulse engines. Galil from my limited research seems to have to be used with Camsoft, and are of the +-10volt type. Geckos don' have closed loop where as Galil do. With all the different manufacturers around, how can one tell who has the better servo drives to use. for example. gecko g320 will sure enough drive servos quite fine if the machine is not overloaded and will probably position the machine accuratly all day. But so will a Galil setup, if it is used carefully and not pushing the machine hard. So why would one need to go "higher up the ladder" so to speak tand use a Galil servo driver over a gecko servo driver if you are going to run the machine at non destructive positioning speeds. Surely a geckodrive would be fine wouldn't it? Then how can you tell the features and the real life benifits and differences between gecko drives, teknic drives, rutex drives and larken servo drives etc. These all seem to be the most commonly talked about drives on cnczone that I have heard of. Is this because they are lower "grade" drives that are not used on bigger industrial machines, and that they are just the most common as this is what people are using for DIY and smaller conversions and setups etc? I'm just trying to figure out how to select a good reliable servo drive system that would be good enought to offer as an OEM solution for some machines I one day hope to make. As I would be building machines to sell to customers, I would want to make sure that the servo drives were of highest reliabilty. Having said that, the market that I would be targeting for these smaller machines does not want to pay tens of thousands of dollars just for a servo drive etc. So where do you draw the line? Cheers, Peter

My first thought to you is this: begin with the end in view.

That will make all your decisions easier, since you will either know or find out what is needed to reach your objective for your machine designs.

Galil is a manufacturer of motor control cards that work well in a PC, by taking over much of the control of motors usually providing closed loop feedback for each axis. The advantage is speed and control. Certain software like Mach and Camsoft might have drivers supporting a particular PC based motion control card like a Galil.

Selecting motors and drives can be easier if you have an engineer from a representative company working with you and knowing what size, speed and type of feedback you will need is essential.

Cost and performance are important, but there is no short-cut to basic understanding of servo motors and drives. Who would try to build a house without first counting the cost?

Snoopy27

[antichip]

Serious al, you are the man, any idea on price of those units?
Ok how about a hobby level unit that does this?

[Al_The_Man]

Originally Posted by antichip Serious al, you are the man, any idea on price of those units?

I wouldn't swear to it, but I seem to remember that these or models close were $200~$300, but AMC will give you price by email.
Al.

[samco]

the main reason is you want a card is it will count encoders into the mhz range. The computer will count encoders but only to around 50khz (using emc2 and a printer port). Plus the pwm output will be at a higher frequency and less taxing on the computer.

Originally Posted by antichip Really the only need for the controller card in EMC is for the +-10 signal? So why is there not a Servo drive made or being made with a 5v pwm signal input?

[CarbideBob]

I use Galil and Delta-Tau controller cards, AMC and Copley amps, and Baldor, MTS, and Torque Systems (CMC) motors. I've also used Sanyo Denki drives and motors.

Many of my machines have been running for over 15 years, 2-3 shifts, 6 days a week and the only field failures I've had are CPU power supplies, CPU fans, and one hard drive.

Given the low cost of counter chips I think encoder counting in software is pure folly. Eventually you are going to miss a count when the CPU isn't looking.

Bob

[Apples]

I've been thinking about it last last few days. Snoopy27, has hit the nail on the head. I really need to find out what the end reslut has to be.

For example.
If you were going on a holiday and you were jumping backwards and forwards trying to workout how to get there by either car, boat or plane you will end up doing that all day. Why, well like Snoopy27 said, you need to know where you are going. If I say I want to go to an international location well straight of the bat you can kiss the car goodbye. And then you can either boat it or fly it.

Same with my delima.

if I want to make and sell plasma and oxy cutting machines, yes I can use high end controls etc etc, and I can use low end controlls. But unless I find where I want to end up at, which in this cse is what my market is, I can't really make my mind up.

So really I need to find the market first, call it the holiday destination. Once I have found where I want to go, only then can I work out how to get there.

Hmm, make sense.

Try to sell a "freezer" in Antartica (we will call it a multi head plasma and oxy cutting machine, with self loading table, 8 oxy cutting heads etc) you will have bucklys, absolute waste of time trying to sell it. Even if it was sold at cost price, no one is gunna want to buy a freezer in Antarctica.

But maybe if I was to make and sell "open fire places" (we will call this a small basic plasma and/or oxy cutting machine, single torch in 4x4 or 4x8 size, running mach3 and gecko g320 servos) then maybe just maybe I'd have a better chance of selling an open fire place in antarctica.

You would agree wouldn't you?

So whilst mach3 and gecko servo drives may not be the best possible controller option, the customer might not need a thermostatically controlled central ducted heating system in his igloo, when a open fire place will more than suffice. Yeah sure he might get a few sparks and ashes and smoke in his igloo, but for him that open fireplace is all he needs....




Just my thoughts. I am starting to learn and understand that no matter what product you have, you need to have the market for it first.

What are your thoughts?

Apples

[CarbideBob]

You're right, you need to understand who you are going to target and what they expect from the equipment.

The DIY market is ultra-sensitive to purchase price. Much of this market consists of hobbyist who place no value on their time as they are doing it for fun. They will tolerate an occasional failure or lower performance to save money.

The industrial market cares much more about downtime and machine speed.
If you are making $120 per hour with a machine and have to ship parts tomorrow it's a real problem if something hickups.

You also need to consider what kind of warranty cost you will have to cover.
Two of my customers demand a million cycle warranty. Anything goes wrong during this period and I have to eat the cost for a service call and parts. One phone call and there goes 500 to a 1000 bucks out the window. Support and customer service time can get expensive fast.

You're not going to sell a Mach powered mill to GM and the guy with a mill in his garage isn't going to buy a 5 axis laser scale feedback Fanuc retrofit.
Bob