Yaskawa SGDH and SGDA servo motor/drive wiring questions

[mactec54]

Hi jasonmbeamer

You can get a PDF from the Yaskawa site for your correct wiring of the encoder motor & drive I would not get it from anywere else the encoder wiring is not the same on all the motors you have to make sure you have the wiring for the SGM motor & for incremental encoder because that is what you have on your motor it is also a 2048 encoder

I have said in other post that Position control is not so good for CNC but because you have it that is what you have to use

This is quite easy to set up because of control software like Mach & others that have step/dir output & because of the yaskawa drive you have

Your drives have electronic gearing so this is how it works When your electronic gear ratio is set 1/1(in drive) the motor will move (1) reference unit per pulse input so the electronic gear ratio represents the number of encoder pulses per reference units from your Control

0.01mm,0,1deg or 0.01inches=1-pulse input moves axes by (1) reference unit

Reference unit=0,1um so if 50,000 pulses are input the axes moves 5mm
50,000x0.1=5000um=5mm

This will never be as good as using Velocity (speed)/Torque control but will work quite well for a home built machine if set set up correctly

[jasonmbeamer]

Thanks for all your help. I would have bought the S versions if I would have known. I didn´t know about that pixe100 that converts step/dir to anolog. I didn´t have enough time to research. I was lucky just to get a drive that will work with mach3 and be compatible with the servos. I took advantage of a trip to the states to buy this D-Tran and three drives.

This robot moves rather far per rev. Just guessing real quick I would say about 1 inch. I will be using it as a cnc router not a mill so I would think this would be great for woodworking. This is so much better quality than I could have ever have built. the only thing is that it uses greese. But I had already figuered out to put a positive filtered air flow through the three ballscrew channels preventing 99% of the sawdust from getting in there. Not to mention having a high powered vacuum system.

[HillBilly]

This will never be as good as using Velocity (speed)/Torque control but will work quite well for a home built machine if set set up correctly

The only difference in a velocity, torque or position CNC is where these loops are closed. This block diagram will be the same with all systems.

Darek

[mactec54]

Hi Hillbilly

I don't think you have used the 2 types of control or you would not be saying that they are the same the end result is night & day different in practice they are almost the same at the input stage but the output is very different as to how the machine will perform

I don't think every CNC machine manufacture would be wrong they are aways looking for better ways to do things but position control with step/dir is not one of them

I think it is good for any one that is wanting to build a system that want something that is not a stepper motor these AC drives & motors fill that spot also the software is good as well with Mach & others it makes it easy to get something going

[HillBilly]

Hi Hillbilly

Hello friend

I don't think you have used the 2 types of control or you would not be saying that they are the same the end result is night & day different in practice they are almost the same at the input stage but the output is very different as to how the machine will perform

Well I have used, setup and maintained both types of systems. They can definitly make the same parts.

I don't think every CNC machine manufacture would be wrong they are aways looking for better ways to do things but position control with step/dir is not one of them

The 2 main players in the CNC arena ( Fanuc and Mitsubishi ) have switched to a servo bus years ago ( FSSB = Fanuc Serial Servo Bus ). If Yasnac (Yaskawa's CNC) is still making CNC's I would imagine they can control these drives though the com port also. I am not saying they are sending step and direction commands over this bus, but I can assure you they are not velocity commands. Velocity drives existed before the CNC control. The first CNC's were designed to control these drives and this became the standard.

Darek

[mactec54]

Hi Hillbilly

Fiber-optic cable use does not make any difference as to how the motors are controlled
I have been using fiber-optic links between control & drives for cnc control for many years & before any of the big players had this

You can take a look at it here at ( Softservo Systems ) this control is made in the USA as well it can cut at 60m/ 2,400 IPM & rapid at 300m pm

Here is some information for you about GE Fanuc this is one of there newer drives which is used mostly for position control

The versamotion AC motors & amplifiers a family of motion controllers are for low-end position control applications they can be used as standalone single axes position controllers for Simple Indexing or electronic gear follower applications

Please note GE Fanuc say it Low-end position control Simple Indexing

GE Fanuc From Control Engineering
GE Fanuc motion controls ability to use powerful motors in small spaces enables building machines that have encoders with higher resolution offer better ( VELOCITY CONTROL ) of servo motors says Paul Webster, product mananger in GE Fanuc's CNC Division

As I have said before Velocity/Torque is the method used by all the main CNC manufacture's

[HillBilly]

The FSSB can be fiber or cable depending on application.

All CNC systems will have a trajectory planner that outputs step and direction. If it is a open loop stepper system this signal goes straight to the drives. If the system is a close loop servo system there will be a position loop followed by a velocity loop finished with a torque loop ( There is really no way around that.).

Darek

[jasonmbeamer]

Got the six conectors from newark the other day. I have the robot in a concrete base. and i have a cabinete mounted for the cpu and drivers. Now comes the fun part (wiring it all up).

I just have an important question beforte I start. I´m in mexico where the power is not very stable. The manual refers to a line filter. I thought of a regulator but 220v regulators are next to none down here, and i dont feel up to importing a regulator from the states right at this moment. even finding just a simple line filter has been impossible. do any of you guys have a simple setup for the line filter with a capacitor or somthing. Thanks for all your guys help.......

[mactec54]

Hi jasonmbeamer

You don't need a line filter as long as the grounding is good & you have shielded cables all power feed wires want to be twisted together

[jasonmbeamer]

well i got the X axis going from mach3. I don´t have the encoders going to mach3 yet. I believe it is necassary in order to keep all 3 axis inline right, or is it ??? I would have to get another lpt port in order to feed the encoder through to mach3.

this is very acurrate (for a router) and extremly fast, as fast as i could possibly want it. I am having a problem with the optical limit switch as the driver or motor causes interference. it works fine untill i turn on the driver. I´m a bit confused becuase thats how the robot was wired from the factory. the controller is only about 2 feet from the hookups to the robot. When you said to twist all power did you mean the wires to the motors ? I can do that from the controller to the robot but up into the robot would be a nightmare.

[HillBilly]

Hooking the encoders back to Mach is optional. What you do want to do is tie the error lines of the amps back to the mach E-Stop. You will set an allowable following error in the drives and if exceeded it will halt Mach. Set your drive display to show this error and move the axis at full speed. Using a value slightly higher than the error you see at full speed would be a good place to start.

Darek

[jasonmbeamer]

Ok my second axis is a Omron R88D-UP12V (yaskawa SGDAxxP 400w) I have it up and running with svmon just fine. I have autotuned the servo, and i can jog it around just fine. But in Mach3 it reacts real weird. jumps and is real jerky. it acts like the signal levels are wrong. I have swapped cables from my first axis just incase it was the cable but it did the same thing and the first axis ran fine with the cable that i made for the second axis. So i know the cable is ok at least compatible with the yaskawa SGDAxxP 750w version (axis 1). this is a 400w version (axis 2). I have configured the cn1 and cn2 switches the same on both drives. I am afraid the port on the drive might be bad. Or I am hopeing that this drive does need the buffered inputs.

If it does just need buffered inputs than great. I will be making a circuit board to make all these cables nice and neet anyways and can just add some optical isolators with buffered outputs. But the SGDAxxP 750W works just fine with 180ohm resisitors. Any thoughts ????

[HillBilly]

I would definitly use buffers or line drivers. The SigmaII drives call for 7-15mA of drive current on the position inputs. You will be lucky to get that out of most P ports. I have found that output pins 1, 14, 16 and 17 usually do not have as much drive as pins 2-9.

Darek

[dominus164]

Hi all
First please forgive my English.
Second i want to ask what controller do you guys use to control AC servos?
I m in the processes to convert my optimum BF20L mill on CNC and i have bought yaskawa SGMAH 01 motors and SGDH drivers and i don't know what controller to use except the yaskawa mp series (2200, 2300, 2300s).
If you have any suggestion please help.
regards
Kyriakos
Atheens

[CPK_Jim]

well ive got some old yaskawa sgda _s drives and am looking at wiring them up using mach3, and step direction but they discontinued the pixie p100. anyone know of a different board to change step dir signal into v reference?

Also i am confused, can i run the drives throught the operator rs232? or is this only for testing and tuning?

Jim

[mactec54]

Hi CPK Jim

I am working with one right now it is made by Max-Mod he is a Zone member you can find the information there The boards are very well made & I think it may be better than the pixie I'm not sure untill all the testing is done when I get more time for it I will post the results

[macona]

well ive got some old yaskawa sgda _s drives and am looking at wiring them up using mach3, and step direction but they discontinued the pixie p100. anyone know of a different board to change step dir signal into v reference?

Also i am confused, can i run the drives throught the operator rs232? or is this only for testing and tuning?

Jim

Heres the one mactec54 is referring to, the YAPSC 10V:

http://www.cnczone.com/forums/showthread.php?t=74532

Open source so you can build them yourself. That will prevent what happened with the pixies from happening here. I have three of the pixies in my mill and if one goes bad...

As for mactec54 earlier statement about pulse input drives being no good for CNC he is dead wrong. There is a fourth type of servo drive that he mistook for pulse following (CW/CCW, Step/Dir, Quadrature in) drives. Those are position indexing drives. Those are totally useless for CNC use. They are basically a servo drive with a little PLC built in. You can tell them go to this position for this input and another for this input, wait, and go here. Stuff like that.

The pulse following input drives are basically a speed type drive with a pixie or yapsc built in. The reason the big boys dont use these in their machines is simple. First they have closed loop back to the control. So its just as easy to send the quadrature encoder signals back to the control and feed an analog signal back to the drive. Second, tuning is handled in the control and you can set everything up from the control panel. With position drives you need the drive software and a PC or the HMI. Third, drives without position controls are significantly cheaper. Why cut into your bottom line?

[jasonmbeamer]

Its been awhile (been buissy). But i wanted to let everyone know this thing has been up and running for awhile now with not one prob. this is as sweet as it gets. very smooth, extremely fast 1400 inch per minute. extremely reliable. thanx for all your guys help out there....

[mactec54]

Hi jasonmbeamer

Pleased that you got it all going, & have done some nice work with it, you just need to build a better spindle & you will have a very good machine

Did you make the board between the drives & computer or get a ready made board

[jasonmbeamer]

Ya I do agree, A real spindle would be nice. I have to change bearings about once every two days (48 hours of use). I did the board myself. Not too complicated.