Add an encoder to do the job?

[LURBY 1]

Greetings all - I have an application that is confusing me. I am retrofitting a machine with a Baldor flux vector drive and motor. This motor is coupled to a ballscrew that moves a large slide back and forth. My customer has thrown me a curve ball mid-stream. I am setting the machine up very simply using an analog out on a ML1500 to get my speed reference for the vector drive. The machine was so simple that I was going to use limit switches on either end of the slide to stop my motion. Here is the curveball - the customer now wants the limit switches to go. He would like me to program the distances using a touchpad. Not a problem (so far). My thought was to connect the encoder on the back of the motor to my ML1500 as well as the flux vector drive. Not sure if this will throw up red flags or not. The other problem is the encoder voltage. I am not positive what it is. Either 5v or 12v. I KNOW that it is not 24v (what I need to tie it in to the Micrologix). Is there a way around this? Is there a way to bump up the encoder voltage (novice question, I apologize). Should I add a second encoder? I was looking at the AB 844D which has a hollow shaft. I could possibly slide it right on the ballscrew. It just kills me because there is an encoder right there!! Is all of this foolishness? Should I scrap everything and go with a servo? (please say no!). Thanks for your help in advance everyone! Chuck Liedke

[JRoss 84]

By your description, I'm assuming you have some sort of homing routine, and only need bi-directional pulse inputs? I suppose you could also use a couple of optocouplers to change the voltage, though that will introduce delay into the feedback. Also, make sure the encoder outputs can drive both the drive and the optocouplers. If the delay isn't a problem and the encoder output has enough power, then do that. Otherwise you should probably just use a second encoder. No need to go to servo unless you need the better response or an absolute system. Or if a second encoder tips the price scale!

[LURBY 1]

I did not have a homing routine before because I did not need one. The limit switches worked fine. Now I will need a procedure to home. How would I accomplish that with the PLC? How do I get the homing mark on the encoder to show up in the logic? Pardon my naivete. I have never set up an encoder on a PLC before.

[BobLfoot 302]

Depending on the Baldor Drive you have many of them have a pulse train output which can be run at 24 VDC and sent to the MLGX. Some basics as I understand them about encoders. A simple encoders giving out three pulse streams A,B & X would be homed as follows. 1. Pulse A proceeds B when moving CCW and Pulse B proceeds A when moving CW. 2. Pulse X is on for once per revolution and usually the Leading edge of A when moving CW is home. 3. Homing routine starts out that you move in a given direction CCW for Example. 4. You continue moving CCW until the Home Input is True. 5. You then reverse Direction {CW} slowly and wait for home input is false. 6. You then creep CCW until Z pulse is on and off. 7. Now You Reverse direction CW and wait for Z to go hi. 8. You then creep CW until A goes hi. This is marked as Zero {0}. 9. Home can now be either 0 or some offset from 0. Couple of other observations. Leave the proxes and move them into position as Overtravel Limit Switches if you don't already ahve them. Stopping at a limit beats replacing the motor, gearbox and machine frame when the darn thing runs away for no reason.

[P Daniil 0]

BobLfoot is absolutely, 100% right on the use of limit switches. Apart from addressing the real overtravel safety issues, they may also be required by your local rules/codes (they are on this part of the world). Edited 21 Nov 2006 by P Daniil

[JRoss 84]

If your drive has the necessary pulse train output, use that. 'Nuff said. Encoders come in all sorts of shapes, sizes, and outputs. The most common in motion control is the A/B/Z-phase encoder, because it allows a system to know which direction it's going. I've also heard these called differential encoders, and quadrature encoders. A and B phases are 90 degrees apart (see attached image), which gives you speed and direction. Z phase tells you each time you make a full rotation. [NOTE: BobLfoot mentions both X and Z phases. They are the same, and I've always heard them called Z-phase] There are several methods for homing a motion system. However, be aware that unless you are using an absolute system, you will have to home every time you cycle power. The method BobLfoot describes is the typical way to home a closed loop system, like a servo. It requires a "home position" prox or limit switch. Another much simpler way to home a system is to physically move it to a mechanical stop (usually called a "dog"), and call that home. In either case, I definitely agree with BobLfoot. You need to leave proxes, limit switches, or something in place at your extreme ends of travel if you don't want to be replacing parts the first time the operator cycles power and forgets to home the system. I have helped set up servo systems without this safety precaution, but they were absolute. We homed it once in the factory and set software limits on the travel.

[panic mode 246]

just a note that differential indicates type of output, not number of channels. quadrature encoders can have the outputs with open collector (NPN for example), TTL, differential (line driver) etc. line driver means that each channel has two outputs so ABZ encoder from previous post would really have six outputs (A, !A, B, !B, Z, !Z). reason for having two always oposite signals for each channel is noise immunity (specially at high speeds and long cables) but also error detection (cut, shorted or unplugged cable etc.).

[panic mode 246]

i'm confused here... if you already have encoder (wired to vfd? ...or plc?) then you already have closed loop system and switching to servo is likely throwing money out of window. do you have reading from current encoder and where (at the drive or plc)? if you want it at plc what is stopping you? isn't there communication option (modbus, devicenet, rs232, anything...)? how are you controling position (where is the loop closed, at the drive or through plc)? i see you have way of controlling speed via analog output but what is the stopping/positioning accuracy? doesn't seam to be critical (unless the drive and the existing encoder are tied together and you just trigger drive to move to a preset positions). what is the total travel? if you need to make stops at different positions, and need to know where to stop, you can add additional feedback (encoder, laser, lvdt, whatever). note that some of them will be absolute (know position at any time without homing). for short travels like 60" or less it's hard to beat temposonic EP series sensors for example (simple analog output).

[BobLfoot 302]

I've used Tempsonics LDVT in several applications and with good repeatable movement.

[LURBY 1]

In response to panic modes's reply let me say....yes. I do have an encoder. I believe that it is a 5 volt encoder on the end of my drive motor. It is connected to the flux vector drive. I would try to connect it to the PLC, but the PLC requires a 24 volt encoder. The encoder (i believe) is being used for speed reference by the flux vector drive. I did check for the "pulse train output" that someone else was talking about, but did not see them as outputs on the drive. The only outputs appear to be drive status outputs. As far as the position goes, I am controlling it with limit switches. If I was to get rid of the switches for position and go to some other type of feedback, I would definitely leave the switches for safety stops. My travel is less than 48" total, so maybe I will look into the temposonic units. As far as homing goes, They will only cycle the power once a day. Even if they had to home the nachine once a day it would not be THAT bad. Chuck

[JRoss 84]

Panic, thanks for the clarification. I know how to use encoders but I'm still shaky on some of the terminology. The distance sensor option makes a lot of sense, because it is absolute. The SICK line that we carry also has a number of sensors that would fit the bill.