Kinetix Motion Pros & Cons

[BobLfoot 301]

I am about to attend a class on ControlLogix and Kinetix Motion put on as a one day {8 hour} seminar by my local AB Rep. I am a neophyte to Kinetix and am curious about others experiences good and bad. Please post away.

[BobLfoot 301]

A bunch of reads but no comments so everyone is looking, but no one is jumping just yet. Let me say that if you can attend the Kinetix on Tour Seminar when it comes to your local AB Distributor - take the day and do so. The class is awesome for introducing new hardware {Compact Logix, Safety PLC and Kinetix} in a learn at your pace atmosphere. Not sure how much I like sercos and Kinetix yet, but it does have promise.

[TimWilborne 108]

I read it but have no experience with Kinetix. Perhaps I am in the Con category because at least a few years ago it would not work with our application. We do not use the encoder on the back of the motor in our servo loop. Because of our application there is continuous slippage. As of revision 12 (might have been 13) you could not feed an external (axillary) encoder as your primary feedback into your servo loop on a 1756-MO8SE. So we use 2098 drives, 1756-MO2AEs, and feed the external encoder into the motion card. After we worked through some stability issues it works great. Got 20 or so of them setup like this and can't find any advantages to changing them to SERCOS. So yes, put me in as a Con because of application.

[mgvol 0]

I have not used it yet, but since I work for an AB distributor, I get to see all the demos and try to soak up all the knowledge I can from our motion specialist. As someone who has used other motion platforms previously, these are some of the obvious benefits (at least as I see them): 1) A single programming environment 2) Greatly reduced wiring 3) Common communications architecture between controller and servos 4) Fewer pice/parts, since the motion control resides in the PLC chassis Coming motion enhancements (later this year) with RSLogix 5000 V16: 1) Added support for 2 and 3-axis articulated dependent and independent geometires 2) 2 and 3-axis SCARA independent geometries added 3) Will support packaging top-loader and material-handling pick-and-place applications 4) Enhanced coordinate config for handling non-Cartesian systems 5) Will allow dynamic path profile tranlation and rotation 6) Ability to modify existing CAM points online 7) Will incorporate 2 new Homing methods - Home to torque level and Home to torque marker 8) Will support the new Kinetix 2000 (low-power) servos, and Kinetix 7000 (high-power AC) servos and new linear actuators 9) Not necessarily motion only related, but V16 will give L6x controllers the ability to manage all system firmware. Firmware for all modules AC drives/servos connected to the Logix can be stored on the controller's ControlFlash card. When a component is replaced, the controller will automatically flash the new component to the coreect F/W level.

[PLCMaster 0]

Been playing with our 21axis machine ( Ultra3000s and Kinetix 6000 ) for the last 18months using Motion Axis control over a twin sercos card closed loop fibre ring. I love AB Motion control but it has problems. Any voltage fluctuations of any sort and the drives go into fault and they seem weak in regards to Estopping. 2 fried drive units and counting .The feedback capabilities are awesome but the cam profiling can be a headache with 3 or more axis' working together. The great thing is the intergration into RSLogix software.

[jstolaruk 9]

Ditto on the voltage fluctuations finding. A couple of Ultra jobs ago we started using the 2090-UXLF line filters desined for the Kinetix/Ultras and haven't had the problems since, http://literature.rockwellautomation.com/i...in013_-en-p.pdf

[gravitar 3]

Drives Headquarters is coming to Detroit next week, I will share my thoughts after I take the Kinetix class!

[PLCMaster 0]

Thankyou very much . These are exactly what we need. Cheers

[BobLfoot 301]

Guess I owe this post an update. I attended the class and we've started to use CLGX Motion. I love the fact that it integrates into logix so cleanly. Yes we've had a few "noise" or voltage problems, but no fried drives. I also recently learned that you can use imaginary axes and cams without actual motion modules present. Kindu neat to sequence things using a cam profile.

[Peter Nachtwey 15]

I am a big believer in simulating the motion profile before putting it into operation. Those that don't often damage things.

[jstolaruk 9]

Bob, is that documented anywhere at the AB website?

[BobLfoot 301]

I don't know, I had the Rockwell Motion Rep out of Nashville in for a couple of problems and hee mentioned it during our follow up training class aftrer things were fixed.

[Peter Nachtwey 15]

I have noticed some negative comments about Sercos without much to back them up. I am not a fan of Sercos, yet. Back around 1995 I participated in a motion control forum where one of the topics was Sercos. Sercos was just becoming popular then. We I got a chance to speak I asked the other panel members how they can generate motion profiles by just knowing the point 2 millisecond ahead? No one said anything meaningful. Perhaps they didn't want to give away secrets to the competition that was there. One think I have realized about secrets is that they are used just as often to cover up weakness as well as strengths. The problem I see with Sercos is that one can't generate a smooth motion profile because the limited data. The controller needs to know pos(n+1), pos(n) and pos(n-1) to calculate a second order equation and interpolate between pos(n) to pos(n+1). A second order position equation and its derivatives look like this: pos(t)= pos0+Vel0+(Acc0/2)*t^2 vel(t)= Velo0+Acc0*t acc(t)=acc0 This equation provides no more than a linear ramp between the points and the velocity and acceleration are not continuous from one 2 millisecond period to another. The acceleration is constant during each 2 millisecond period. One can't even generate a jerk for the jerk feed forward term. This is important because good motion control relies on the target velocity, acceleration and jerk to provide the feed forward terms. When these 2nd order velocities and target are multiplied by the feed forward gains the output looks like there is a lot of noise on it. I couldn't believe it. The emperor had no clothes and no one said anything! In the past few years Sercos has got faster and it allows one to download the next velocity as well as position. This makes the target generation acceptable. One can easily generate a cubic equation which would be far superior to the quadratic ( second order ) equations. So what does your Sercos use?

[gravitar 3]

I did one of the Kinetix labs today at the R-A Detroit HOT training. I must admit I have VERY little knowledge of motion control so I'm noone to objectively judge the product, but it sure did seem like a great idea to be able to control the drive directly from ladder logic instructions. It felt like the motion controller was just a card in the backplane, with all the pre-defined tags at our disposal. They sung the virtues of the greatly simplified "wiring" and noise-immunity of fiber in the SERCOS system, and I was hard-pressed not to buy into it. I'm nowhere near the level of experience necessary to be able to criticize it! the use of a "virtual axis" was part of the lab, and it did work. Seems like it might be a good means of simulation and testing at early stages in the design process. Again, the R-A guys were really geeked about it, but I must admit I didn't see what all the fuss was about :)

[Gerry 13]

One thing that sercos provides (at least in the A-B version) is the ability to select arbitrary feedback resolution. This is useful on rotary axes that might otherwise have a fractional unwind count and so avoids having to change pulleys,etc. The resolution can be up to 2^32 - 1 counts per motor rev. High resolution is good, isn't it, Peter? Virtual axes have more uses than just simulation - most often in gearing and camming applications.

[Peter Nachtwey 15]

High resolution is good. If an encoder has 1000 lines and 4000 counts per revolution then how does one get more counts from it? I know you can multiply the counts internally but then the counts don't increment by one anymore. How does one get more than 4000 counts per rev from a 4000 count/rev encoder. Btw, I have heard of people doing something like this with the M02AS so the internal motion generator thinks it has more resolution to work with when generating coarse update points. This doesn't really provide more feedback resolution. This can be done by treating a 13 bit encoder like a 16 bit encoder. Since the counts are left justified the 13 bit encoder looks like it increments the counts by 8 for each real count.

[Gerry 13]

I really don't know. However, I suspect that the drive does some interpolation based on what it knows about motor speed. Motor feedback to the drive is fixed by the catalogue number selected and expressed as cycles/rev plus an interpolation factor. I did one application where I set the resolution to 16002 counts/rev in order to get an integer unwind count. For this app. the motor resolver feedback was 2 cycles/rev with interpolation factor of 2048. Other motor drive combinations (with encoders) give 4000 cycles/rev and interpolation factor of 4. I don't think it's simple multiplication or bit-shifting.

[TimWilborne 108]

I would like to know more about how they are increasing the resolution. I understand using the detection of rising and falling edges to do the x4 but anything else is in between pulses completely. I don't see how the motor could anticipate it. Too many external factors. Unless I find more I'm going to have to guess a multiplier too

[Gary Burton 0]

I just completed a project that used kinetix 6000 drives and sercos...MO3SE card. It was very easy to setup and configure and I have experienced zero problems with drive faults other than program induced, by me I'm impressed although I do not have a lot of motion experience. I used camming and have achieved .0005" tolerance on a system travelling at around 10"/second, this was sufficient for my application.