584,861 active members*
4,741 visitors online*
Register for free
Login
IndustryArena Forum > Machine Controllers Software and Solutions > Controller & Computer Solutions > Strange Acceleration problem with 990 MDb controller
Results 1 to 4 of 4
  1. #1
    Join Date
    Sep 2018
    Posts
    83

    Question Strange Acceleration problem with 990 MDb controller

    Hi all.

    I've set up a 990MDb controller to run a 1325 router. Generally, I've ironed most of the kinks out quite well, and I'm happy enough with most of the results. Because what we're routing is foam, it needs to be run at a high speed, which is generally fine. However, I'm noticing a weird pattern forming - there's an aggressive deceleration when transitioning from lines to arcs, but it's not on every arc, though. I've not determined what the specific pattern it follows is yet.

    The manuals for this controller are, as I've expressed before, subject to dubious Chinese to English translation.
    Does anyone understand the acceleration settings on these controllers well enough to help?

    The relevant parameter settings are:
    Code:
    4.7.2 Speed parameter
    1,X-axis's G00 speed(mm/min) 10000
    2.Y-axis's G00 speed(mm/min) 10000
    3.Z-axis's G00 speed(mm/min) 4000
    5.Manual maximum feed speed(mm/min) 12500
    6.Auto Maximum feed speed(mm/min) 12500
    7,G01/G02/G03 default speed(mm/min) 4000
    8,Null run speed(mm/min) 10000
    9,Feeding axis`s manual speed(mm/min) 10000
    10,Spindle`s manual speed(rpm) 12000
    11,Beginning feed speed(mm/min) 50
    12,The maximum mutation of feeding axis speed(mm/min) 50
    13,Limit G1G2G3 axis speed [1meanYes,0meanNo] 0
    14,XG1G2G3 max speed (mm/min) 12500
    15,YG1G2G3 max speed(mm/min) 12500
    16.ZG1G2G3 max speed(mm/min) 12500
    18,X acceleration [1?99999] 10000
    19,Y acceleration [1?99999] 10000
    20.Z acceleration [1?99999] 5000
    22,Auto run acceleration [1-500] 1
    23,Handwheel acceleration [500--30000] 1200
    24,Run program Handwheel acceleration [>500] 60000
    25,Run program Handwheel G00 speed(mm/min) [>10] 1500
    >There’s another parameter here that’s not listed in the manual: 25-1 Run Program Handwheel g00 start speed 100
    26,Handwheel X limit speed(mm/min) 4000
    27,Handwheel Y limit speed(mm/min) 4000
    28.Handwheel Z limit speed(mm/min) 4000
    30,Acceleration type [0meanline,8meancurve] 0
    31,Curve initial acceleration [>=10] 100
    32,Curve acceleration with second-order [>=10] 100
    33,Curve max acceleration [>=500] 1000
    34,X go home positive speed(mm/min) 3000
    35,X go home negative speed(mm/min) 100
    36,Y go home positive speed(mm/min) 3000
    37,Y go home negative speed(mm/min) 100
    38.Z go home positive speed(mm/min) 3000
    39.Z go home negative speed(mm/min) 100
    42,Spindle max speed in the first gear(rpm) 18000
    43,Spindle max speed in the second gear(rpm) 2500
    44.Spindle max speed in the third gear(rpm) 2500
    45,Spindle max speed in the fourth gear(rpm) 2500
    46,Second Spindle max speed(rpm) 2500
    47,G02/G03 reverse compensation mode (0meanA;8meanB) 8
    48,mode B reverse compensation speed (mm/min) 3000
    48-1,mode B reverse compensation Beginning feed speed(mm/min)[>10] 500
    48-2,mode B reverse compensation acceleration(mm/min)/s)[>10] 60000
    49,Speed mode[1meansyes,0meansno] 1
    50,Handwheel stop speed(mm/min)[>100] 100 
    There are four extra paramenters not listed in the manual here:
    	51 Follow the tapping knife when the spindle speed 1
    	52 When tapping spindle backlash compensation 1
    	53 Follow the tapping cutter withdrawal after reversal 1
    	54 Tapping back speed 1
    58,Forcedly limit drop speed critical(mm/min) 1
    > The following are not listed in the SZGH manual, but are in the NewKye manual
    60, smooth running function 28
    61, smooth running time 50
    62, handwheel smooth running time 50
    63, G01/G02/G03 smooth running time 50
    101, X-beginning feed speed 500
    102, Y-beginning feed speed 500
    103, Z-beginning feed speed 500
    111, X-jump speed at continuous track 500
    112, Y Jump speed at continuous track 500
    113, Z jump speed at continuous track 500
    200, G00 continuate run is valid 1
    > The following are not listed in either manual
    230 SP G00 speed 1000
    231 SP G01 mode 4
    232 SP Direction 0
    233 SP back to zero control mode 0
    234 SP stop angle 0
    235 SP back to zero speed 800
    Oddly enough, the SZGH manual says I should use linear acceleration for stepper motors (which this machine has), while the NewKye manual in front of me says " 30: The parameter is about the driver configuration, in generally, set as curve when match the stop, set as straight line when match servo"
    I tried running curve-based acceleration using the default settings, and it didn't really achieve much.

    Of the two manuals, the SZGH is more coherently translated, although neither it nor the NewKye manual covers every parameter the controller has. For what it's worth, even though the 990MDb is a unit produced under many company names, the one I bought is NewKye branded.

    For what it's worth, once I've gotten this nailed down and sorted, I'm happy to post my configuraton and wiring schema. I've even written a Vectric post processor for this controller that seems to work pretty well, barring this weird intermittent curve issue.

  2. #2
    Join Date
    Sep 2018
    Posts
    83

    Re: Strange Acceleration problem with 990 MDb controller

    I suspect that this may possibly have something to do with parameters 11 through 113, having trawled google for other manuals that were written by people with a better command of the language. However, this seems to govern feedrate changes, which is weird, given that there's no actual feedrate change implemented in the gcode I was running as an example.
    I've appended the gcode in question, along with a photo of the surface fnish to show the difference, The second photo highlights the points where it slows down significantly.

    Code:
    ( Curve Issue )
    ( File created: Thursday June 25 2020 - 09:19 AM)
    ( for MDb990 from Vectric )
    ( Material Size)
    ( X= 1000.000, Y= 500.000 ,Z= 50.000)
    ()
    (Toolpaths used in this file:)
    (Profile 1)
    (Tool used in this file: )
    (1 = End Mill (12.7 mm 10Krpm))
    N100 G91 G28 X0 Y0 Z0
    N110 G40 G17 G80 G49
    N120 T1 M6
    N130 G90 G54
    N140 G43 Z-20.000 H1
    N150 S12000 M3
    N160 G4 P6000
    N170 (Toolpath:- Profile 1)
    N180 ()
    N190 G0 X0.000 Y0.000
    N200 G00 X64.203 Y99.928 Z-10.000
    N210 G00   Z-5.000
    N220 G01   Z50.000 F2000.0
    N230 G01 X99.330   F10000.0
    N240 G01 X101.285 Y92.634 
    N250 G03 X107.418 Y87.928 I6.134 J1.644 
    N260 G01 X368.160   
    N270 G03 X373.231 Y90.456 I0.000 J6.350 
    N280 G01 X380.369 Y99.928  
    N290 G01 X414.037  
    N300 G01 X421.174 Y90.456 
    N310 G03 X426.245 Y87.928 I5.071 J3.822 
    N320 G01 X576.672   
    N330 G03 X583.022 Y94.278 I0.000 J6.350 
    N340 G01  Y189.278  
    N350 G03 X576.672 Y195.628 I-6.350 J0.000 
    N360 G01 X86.443   
    N370 G03 X80.309 Y190.921 I0.000 J-6.350 
    N380 G01 X58.069 Y107.921  
    N390 G03 X64.203 Y99.928 I6.134 J-1.644 
    N400 G00   Z-10.000
    N410 G28 G91 Z0
    N420 G49 H0
    N430 G28 X0 Y0
    N440 M30

  3. #3
    Join Date
    May 2015
    Posts
    684

    Re: Strange Acceleration problem with 990 MDb controller

    As you have found out at high speed a simple trapezoidal acceleration is not conducive to accurate machining. You end up with position errors at transitions. The solution is complex and that is to use jerk limited acceleration which results in an S curve, if velocity was plotted over time.

    I don't think you can use the S curve feature of your hardware for CNC. That is more for industrial applications where the distance of the motion segment is known. In this case you might have a 30% motion profile on a movement that spans 10 seconds. In this case, the motor would accelerate for 3 seconds, travel at desired speed for 4 seconds and decelerate for 3 seconds. In this instance, you might ask the motion controller to do its own S curve planning. But that is not possible in CNC becasue we never know how long a segment might be before the motion planner does its job.
    Rod Webster
    www.vehiclemods.net.au

  4. #4
    Join Date
    Sep 2018
    Posts
    83

    Re: Strange Acceleration problem with 990 MDb controller

    Interesting. The parameter settings do have the option for defining acceleration curves and I do remember there being a setting in there somwhere that inferred some kind of readahead functionality - I'll continue to trawl, and see what I stumble across.

    I don't suppose ayou'd be able to point me in the right direction for understanding the S-curve parameters, at least? While I don't doubt your word when you say that it won't be all that useful in this scenario, I'd like to be able to experiment with it at least in a manner that has comprehensible results, as opposed to just banging in a range of numbers and seeing how it behaves.

    I've just had a play with the parameters, and I've gotten the curve acceleratio behaviour to at least match linear acceleration in terms of total run time. It was initially significantly slower. I've changed the following parameters:
    Code:
    30 Curve initial acceleration 2500
    31 Curve acceleration 5000
    32 Curve maximum acceleration 1000
    I'm a little nervous about messing around with these values too wildly (this was about as drastic as I was willing to go, given how low they were initially) purely because I don't know what parts of the S-curve they specifically pertain to. You can't describe rates of change with a single number like that, so I don't what specific variabes they're altering. Have you or anyone else done much work with S-curve manipulation before?

    I'm also looking at parameters 111-113 as well, although if it's what I think it is, it should be kicking in a lot more aggressively on every other turn as well, I'd have thought.

    For what it's worth, the first small turn/step in the piece doesn't slow down anywhere near as hard as the paired one does.

Similar Threads

  1. DIY CNC Controller Strange Sound Problem
    By Jetpack_Was_Yes in forum Stepper Motors / Drives
    Replies: 2
    Last Post: 01-16-2016, 10:29 PM
  2. dx-32 Spindle Acceleration/Deceleration Problem
    By cybermachineman in forum Bridgeport / Hardinge Mills
    Replies: 4
    Last Post: 05-22-2011, 07:01 AM
  3. Strange X4+ problem
    By Mud in forum Syil Products
    Replies: 15
    Last Post: 09-24-2009, 03:08 AM
  4. Strange Problem
    By mrcosmos in forum Mach Software (ArtSoft software)
    Replies: 4
    Last Post: 08-17-2008, 01:55 PM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •