To ACC or not to ACC

Filed under: FANUC Robot Happiness TP Programming

That is the question. And the answer is “probably not.”

Once a programmer learns that the ACC instruction exists (and figures out how to make it go over 100) this is often the first place they go to make the robot faster. It’s such an easy change to make with immediate results, but there are serious consequences. I’d argue that going over ACC100 should only be done as a last resort.

Here are some alternatives with a brief discussion of why higher ACCs might be a bad idea.

Going Faster

Here’s my method for increasing throughput:

  1. Make sure your Payload is correct
  2. Don’t wait if you don’t have to
  3. Don’t go out of your way
  4. Move smoothly
  5. Increase ACC


Accurate payload settings are very important. Not only do they protect the robot by allowing Collision Guard to make its calculations accurately, but they also alter the path and speed of the robot. An R-2000iB/165F will move much faster with a payload of 40kg than the full capacity of 165kg, and a M-3iA/6S will move much faster with a payload setting of 0.5kg than the default 6kg.

Don’t wait

I’ve come into projects where the customer is complaining about cycle time to find random waits for 1.0s here, 0.8s there, all completely arbitrary. If your gripper only needs 0.1s to close, why have a wait of 0.2s? Or better yet: if you have reliable gripper inputs, why have a wait for time at all? Just wait for the correct input state.

Don’t go out of your way

If you’re driving from Chicago to Detroit, you’re probably not going to detour up to Traverse City. In fact, the fastest way to get there is to fly in a straight line.

It’s very easy to be too conservative with your approach and retreat heights or clear something by 12” when you really only need 1”.

Smooth is Good

Robot motion should be smooth. Say it with me now, “smooooooth.” Very good. Who’s the best at creating smooth motion? The FANUC motion planner, not you.

Don’t use 5 points when 4 will do. Better yet: don’t use 4 points where 2 will do. Chances are those extra points are causing the slower axes of the robot to slow your cycle down. If you just move from point A to point B, the slowest axis has that entire segment to get to where it needs to be.

Termination Types

Use CNT100 wherever possible. Avoid FINE moves if you can.

On your pick position, see if you can use CNT0. If CNT0 works, try CNT5. As you change the termination type you might have to tweak the position itself to achieve the same result, but rounding that corner is going to be much faster than stopping on a dime. (Note: different CNT values are going to yield vastly different results depending on segment speeds and ACCs when you get there… remember that.)

Ok, can I change ACCs now?

If you have the correct payload settings, you’re not waiting on anything, your path is optimized and it looks like a smooth dance, I hope you’re making your cycle time guarantee. If not, your layout is probably bad or you made some bad assumptions when bidding the project! If you’re still not there yet, you may consider increasing ACCs, but here are a few things to watch out for:

Increasing ACC changes your Path

“But what about Constant Path?!” You

Constant Path makes the robot follow the same path regardless of OVERRIDE, not segment speed. Your pick position with Z=-5.0 at CNT5 may have worked great before, but the same move at ACC150 might drive your robot into the conveyor.

High ACCs can Hurt the Robot

FANUC has really smart people to tune the standard accelerations for speed and reliability. As soon as you start increasing these accelerations, you’re putting greater strain on the servos, reducers, brakes and the arms themselves. Tread lightly.

Remember what we said about being smooth? It’s fairly likely that your buttery motion ends up looking violent with higher ACCs. If it looks violent, it’s probably not good for the robot.

Driving the robot harder also increases your risk of over-current (OVC) and overheat (OVH) faults. Your Accord might drive around the oval until the gas runs out at 65mph, but how long do you think it can drive 100? 115?

If you’re using a genkotsu robot (M-1iA, M-2iA, M-3iA), did you know that there’s always a small chance the force of your motion may cause an arm to pop off?

These robots constantly monitor this likelihood and try to prevent a “dislocation” by stopping themselves if a dislocation is likely. If you’re lucky, the software works great and keeps the arms intact. If you’re unlucky, you’re going to be spending the next hour or so putting your fancy robot back together.

Want posts like this delivered right to your inbox?

If you liked this post, please sign up for my mailing list!