Ultra 3000 in follower mode

[LURBY 1]

Anyone have any experience setting up an Allen Bradley Ultra 3000 in follower mode? I want to synchronize the rotation of a servo motor to the movement of a slide. It looks like all you need to do is connect an encoder to the aux encoder input on the drive. How accurate is something like this? Also, is there any way to change the presets with out the software? In other words can I make one of the preset gear ratios a variable and change it "on the fly" This would allow me to use any gear ratio that I wanted and not be limited to the 8 presets available. Thanks for your help. If I have neglected to mention something please let me know! Chuck

[paulengr 44]

It is extremely accurate since there is no "loss" of information like it would with an analog connection. I'm not sure if you can change parameters on the fly since this drive is really meant for "standalone" mode of operation unlike some of the others. You just sort of configure everything in Ultraware in the pre-built application. It's kind of like an Ultra 6000 except the program is built-in and can't be altered. The Powerflex 755 is less expensive and can be configured to run in follower mode the same way if you add an encoder input card to it. And you can change the parameters on the fly if you add a communication card. I believe it's also a bit less expensive. For the money and time I'd get the PF 755. I didn't have a pleasant Ultra 3000 experience. If you already have a PLC involved, you could just as easily drive the Ultra 3000 in positioning mode off a 4-20mA. If you want variable gearing ratio, feed the slide input (LVDT, encoder, or resolver card available from AMCI for instance), do the scaling inside the PLC, and output to the Ultra 3000 analog input.

[LURBY 1]

If I control the rotation of the screw using a 4-20 output how will I be synchronized to the moving slide? I guess you lost me! If I use a Powerflex that would be an AC motor instead of a servo. Will I still have the accuracy, torque, and response time that I would get with a servo?

[paulengr 44]

The Powerflex 755 can operate among other things, AC motors and DC permanent magnet motors. The Powerflex DC drive can operate pretty much any kind of DC motor. The other Powerflex series (4, 40, 70, 700) are designed for AC motors. It is far more expensive, but the S in "700S" stands for SERVO. Servo motors these days tend to come in one of 5 families: steppers, permanent magnet DC motors, shunt wound DC motors, brushless motors (which can operate on AC or DC), and AC "servo class" motors. I'll ignore gear motors since we are speaking to motors themselves and saying "gearmotor" does not identify the motor itself. The key design differences with servo motors is that they are usually built with bearings intended for high speeds (3000-10,000 RPM), they are designed to be very long and narrow to minimize inertia of the motor itself (which decreases response time), and they are designed to operate at virtually any speed for long periods of time to avoid cooling issues compared with general purpose motors. At one time the speed/torque requirements could only be met with DC shunt wound or permanent magnet motors. The variety of power devices now available, especially the IGBT, has wiped out this limitation. Because of the IGBT, electronic commutation and especially field vector control became practical. This opened up the world of servo motors to the brushless designs, and later to AC motors designed specifically for servo use. Crack open any motor catalog now and you'll find all of these options available. Of the types available, the brushless designs (PMDC, AC, DC, or AC induction) are much more rugged and require little maintenance. In addition, the AC induction servo motor is king of the high torque world since you get all the advantages of a conventional motor (rugged, almost unlimited torque) operating in a servo package. "Accuracy" is NOT a function of the motor. This is a function of the encoder. The only time that motor "accuracy" becomes an issue for servos is if there is some sort of cogging-type effect happening while inching where the motor physical design becomes apparent. Not all motors are subject to cogging effects. It is of course possible to implement extremely large DC servo systems too. You just have to gang up the motors. I just upgraded a large servo system. It has 4 motion axes. Three of the axes consist of 4 conventional DC motors operating in parallel with a common gearbox. Each motor is 1300 HP, for a total of 5200 HP per axis, operating through a 50:1 gear ratio. The 4th axis is "only" a few hundred HP in size. The armatures are driven via a large MG set since the system was designed and constructed in the 1970's. I just upgraded it from an op amp (analog computer) system to an all-digital system. We kept the MG sets and the DC motors since it was cost prohibitive (no return on investment) to switch it all over to AC motors, although that has been done. This is for an 8200 dragline. Here is a writeup on one of the more current ones that has been built: http://www.bucyrus.com/media/24595/bni%208...cess%200605.pdf For the curious, the motors on Marion (well, really any) dragline are operated in sets of 2 or 3 motors wired in series with their corresponding generators in a Ward-Leonard loop arrangement (except more recent shovels that actually use drives outright). There can also be up to 3 parallel loops present. The load sharing is done via analog current follower arrangements. The AC versions use torque follower mode. On draglines when additional AC motors are added to the axes, the extra slave motors use analog feedback. On the AC shovel (P&H 4100 AC BOSS), they went all out and use fiber communication between the drives (ABB drives) and even the drives themselves operate in parallel (up to 4 drives per motor) to achieve the necessary output power. We also commissioned a P&H 4100 AC BOSS rope shovel this year. The main (hoist) motion has two 1600 HP AC motors operated in parallel. The motors are very custom with 0.5% slip due to the fact that they made the motors fit in the footprint of the original DC motors that are available on the 4100C machines in case the AC experiment went bad. Fortunately the AC experiment has been so successful that P&H probably won't even be building DC machines after this year except if a customer special orders it. Oh, and that little detail about making sure that your stiffness is as high as possible to stay out of the mechanical poles and zeroes? Hahaha, not a chance of ever seeing that on a dragline or a shovel. You just have to live with the fact that it is an electromechanical system and that there are resonances in the mechanical system that can and will impact the electrical system.

[Peter Nachtwey 15]

http://www.deltamotion.com/peter/Videos/In...tive%20mode.wmv The person in the video has another Tempo rod and the dart is geared to that. You can see him move his arm moving the slider which is out of view behind the table. The ballon is moved such that the dart can touch it but not push into enough to pop it. This is the demo mode http://www.deltamotion.com/peter/Videos/Demo%20mode.wmv It gets tiring doing the manual mode all day at a trade show. The demo mode is not really what I want to show because it doesn't use the slider but it does show the control and how programmable the controller is since there is no PLC involved. This is what the demo finally looked like at a trade show. http://www.mtssensors.com/fileadmin/media/...alloon_Demo.flv Simply following with a PID is not enough. PID response lags and works off of error and any error will cause the ballon to pop in this demo. To accurately gear to the slider the position must be differentiated to get a speed and acceleration. This requires fine feedback which is the point of the demo but they, Temposonic, does not get into the geek speak like I do. The slave units use the position, velocity and acceleration so the ballon can see how the dart is accelerating and compute trajectories that will touch without popping. The position, velocity and acceleration for the slider is used to generate feed forward values for the ballon axis so it always has the correct control output based on the motion of the slide. As for changing on the fly. This example shows that it easy. This demo goes into and out of gear mode and it looks seamless. Changing set points, gear ratios or just about anything except for the transducer scale factor and offset is easy to do on-the-fly. BTW, one of our engineers programmed that demo in the shop in Carry, NC from Vancouver, WA over the internet. There haven't been any specification mentioned as to how well the motor must gear to the slider. I think this is an extreme example that shows what can be done.

[LURBY 1]

Thanks for the help. I have used both the Ultra 3000 and the Powerflex (40P). I have had success with both products. At least they have done the job that they were asked to do! I have a spare Ultra 3000 and matching motor. I think I will pick up an encoder and try to play around with it. I am looking to cut a spiral inside a tube. If I can make this work I would like to offer it to customers with the ability to choose whatever spiral you want. That is why I wanted to do more than just the eight presets. Actually now that I think about it I will have even less than 8 because I need to assign one as a drive enable. I am also trying to do this as simple as possible. The more complex it gets the more expensive it gets. I will still have a PLC, but I was thinking of a simple ML1000 just to control some push buttons and the output selection for the drive. My idea was to also use the PLC with a small HIM that the operator would use to enter a spiral pitch. The PLC would manipulate this number and spit out a gear ratio (integer) that I could then input into the drive as one of the presets. Crazy?

[alan_505 24]

The presets work on a binary input so you only need 3 inputs for 8 speed presets.

[paulengr 44]

...for a total of 64 possibilites if you use all 6 potential bits. You can use the "sequencer" system to preload up to 64 different settings.

[LURBY 1]

You are right. I can get all eight gear ratios. I guess it is too much to ask for more than that?

[OkiePC 14]

I have not used follower mode with an Ultra drive, but I have dealt with applications that required commuincations to allow more flexibility than offered by the built in presets in an indexing drive with preset index positions. It is possible to use the serial port to write to parameters in the Ultra drive, but it can get complicated, especially in a PLC. The commands require that you calculate a checksum to include with them which makes for lotsa ladder logic. I'd recommend using a gateway device that "speaks" ultra for you, and you just have to fill in the blanks. Red Lion G3s have an ultra 3000 driver, and although I've not used that driver, I love the ease of programming offered by Crimson 3.0 software. The last device I used was many years ago for the Ultra 100 and 200 and was made by Quartech. It would monitor PLC-5 addresses via DH+ and when they changed, write the values to the Ultra drive. It has been so long ago, I don't recall the details, but it was not terribly hard to program the Quartech module and it was quite reliable, although there were at least two or three times (in a period of two years) that it "locked up" and required a power cycle. Looks like Quartech is still in business and now supports products for the Ultra 3000 drives. http://www.quartechcorp.com/products.htm Then, you can write a new value to one of the presets easily. Be careful though, the older line of drives used EEPROM to store parameters which was only good for about 100,000 write cycles, so don't write to it every PLC scan, and look for information about that in the Ultra 3000 literature. I know at one point they added some NVRAM to solve this issue for some parameters, but I wouldn't be surprised if the EEPROM write cycle limitation still exists for certain (if not all) parameters. Edited 16 Jul 2010 by OkiePC

[Peter Nachtwey 15]

You guys are way off track. You can't synchronize two axes over a serial port alone. If you must write serial commands to two different drives over a serial line there will be a huge delay between the two messages being sent and acted on. The whole concept of following shouldn't be used. When making threads on pipe or doing similar the axes must be synchronized. Otherwise there will be a lag or phase delay between the master and the follower UNLESS one is very clever in deriving the master velocity and acceleration so that the slave can use that has a feed forward. When one controller directly synchronizes two axes the proper feed forwards can be generated for both axes so they will follow the computer generated motion profile.

[LURBY 1]

Are you confusing the serial communication for the preset change with the follower? I was under the impression that there would be very little to no lag if I set the drive up in follower mode and wired the encoder directly to the drive. The PLC would only be for changing the preset gear ratio. That shouldn't have anything to do with the accuracy of the master/slave, correct? I know that I could do this with a two axis CNC control or even possibly with a CompactLogix but I am trying to keep the cost as low as possible.

[Peter Nachtwey 15]

That would probably be a better option.

[Peter Nachtwey 15]

That would probably be a better option. A lot depends on the precision you need but if you are cutting threads I would go for the two axis controller. Then the two axis controller can generate ideal target position, velocities and accelerations for the feed forwards and both axes are synchronous to the same controller clock.