Knife intelligence

I have a material we run 24 layers, with some smaller patterns being cut smaller than digitized on bottom layers. The top 12 layers are to print and good, the bottom 12 layers are smaller by up to 1/8". They are not distorted, the pattern is correct but to small on the bottom layers. I have 14 other various patterns on marker that are good, just the smaller ones seem to be problematic. Does anyone know enough about knife intelligence to know if this would be a factor ? It appears like the parts are cut on a angle causing the bottom ones to be smaller.


  • Pertaining to Knife Intelligence (KI), When parts are cut smaller on the bottom plies then the top, typically the KI Gain and/or Deflection Angle are set too high.

    The default gains are high, (the gain is the sensitivity) I never set the gain higher then 10, if set higher, you'll see the c-axis meandering when cutting a straight line.

    I also never set the angle higher then 8 degree (the amount of overall correction).

    Try a gain of 10 and angle of 6 and reply.....
  • DougWDougW
    edited February 14
    Thank you Doug Hopkins. My KI Gain was set at 15, I lowered it to 10. The Deflection angle was set at 6.016, I left it at that setting. Looking at my records the only other settings that were implemented before this issue were the Max Angular Velocity- set to 240 and the Max angular acceleration was set to 20000. Would these have a effect on my issue ?   I will respond with the outcome. Your input is greatly appreciated.
  • Nick D'AlfonsoNick D'Alfonso Gerber Staff 💛
    Hi DougW, Some other things to consider.
    24 plies of a very hard material might be beyond the ability of the machine to cut constantly with equal sizes top to bottom. You may need to reduce the plies considerably.
    A new knife is much stiffer than a well used knife. Does changing to a new knife improve top to bottom?

  • Nick, the knife is usually new or fairly new. The material is a shoe liner and not really a hard material. The issue may result in going to less layers, but I wanted to eliminate any other possible causes before I went that route. I appreciate your thoughts.
  • I like the Angular velocity reduced to 240, as well as the Ang. Accel to 20000, those are good, I wold also reduce the Max cut acceleration (just above the params) by 25%.

    Have you checked your Yaw calibration (c-home to system zero) depending on the cut directions, parts can cut smaller if yaw is present.

    Or the Heel to heel setting, but using the white foam (ply height simiulator) to see if the cutter tube is "cocked".

    Reply if you need detailed instructions.... 
  • Doug Hopkins, I do not see a setting called Max cut acceleration, is it possible its called something else ?
  • Sorry,  It's called Maximum Acceleration, it is directly above Angular Acceleration.

    If running in imperial the default is 144, I reduce to 100, somewhere around 25%
  • What is the main difference between  Hardware KI Gain & Software KI Gain ,,, and Hardware KI Enable & Software KI Enable ?  And when making adjustments to these to correct parts cut smaller on the bottom layers, do BOTH settings need to be adjusted ?

  • Background:...Gerber has had Hardware KI for decades.  A transducer reads the flex of the blade, that information is fed to the Controller and a corrective angle is applied to the C-axis to straighten the blade (make it perpendicular to the bristle) much like when cutting a circle in plywood with a reciprocating jig saw, if you go outside the line, you twist your wrist (c-axis), in turn correcting top to bottom accuracy. 

    There are many variables that affect the amount of defection that is sensed and the amount of needed corrective angle to apply to straightened the blade, ie...material hardness and density, contours being cut, cut speeds etc.... to achieve acceptable results we adjust the Hardware Gain (amount of pressure needed to sense any deflection) and an amount to correct (deflection angle).

    Gain guidelines:  Light and delicate materials won't deflect a blade much, but if spread over 1" high, it may require a higher gain to sense the light amount of deflection.  While cutting hard and dense material spread over .75" high would need a low gain, because the tough material will deflect the blade easier, so it does not need to be so sensitive.

    And as mentioned way above, our defaults are set too high, when cutting higher ply of light weight material I probably would not set the gain above 12 and an angle around 6, if cutting tough material like denim I'd use a gain 6-8, and an angle of 10...

    All that said,....we have learned a lot of about how much to adjust (rotate) the blade when plunging into tight contours to maintain the perpendicularity, so in conjunction with using the Hardware KI, which makes corrections based on live transducer feedback, we can also predetermine an amount of c-axis corrective rotation PRIOR to plunging.  So if we see that we are plunging in to a small circular cut out, software KI will predetermine the best angle to plunge, and it does cut out a much nicer small circle with Software KI enable, with a reduced gain.

    So now,......LONG story short,...If cutting higher ply of material use hardware KI, but I recommend with reduced gain and deflection angle.

    If your application (automotive) often includes smaller sized circular contours (less then 3" dia), and require optimal appearance, use Software, with a gain around 20.  

    However using both Hardware and Software KI together does result in overall higher amount of "Knife straightening", so if you do not have a lot of small circular contours, Disable Software KI, or reduce the gain drastically to 5.

    Try and Reply.......Will work on getting a comprehensive C-home to system zero and Mechanical Knife offset test and procedure posted....
  • DougWDougW
    edited February 28
    Thank you Doug for the details and explanation, this helps a lot. The 2 materials I am having cut issues with are 24 layers, fairly soft and simple material ( shoe lining ) with a material height of just over 2 1/4" .
Sign In or Register to comment.