---In dynomotion@yahoogroups.com, <tk@...> wrote:

Hi Cristian,

(I hope you saw my earlier reply)

Open collector mode works well for drives that have opto couplers with the LED anodes connected in together to +5V.

Your drives are probably different and expecting TTL signals.

Regards

TK

Group: DynoMotion	Message: 8401	From: cristian_atzori	Date: 9/26/2013
Subject: Re: Possibly fault Kflop board

Yes, they expect TTL waves. Now I made it work with output 28 (LVTTL quadrature) and succesfully moved the motor. Maybe you could give me the clue to configure the closed loop, because when moving a linear encoder (related to input 0, for channel 0) accross the 0 point I only can see a very short quadrature train, just when the position value crosses the Destination one. Short signals only when passing that zero point, and then nothing. When using the DAC output mode I can see, instead, a constant DC output until Position and Destination are convergent, and that is what I expected to see in step-dir output mode. Could you please tell me what are the values to modify to have a constant train of quadrature pulses when I move le linear encoder far away from the cross point? Here is the code I was using: #include "KMotionDef.h" // Defines axis 0, 1, 2 as simple step dir outputs // enables them // sets them as an xyz coordinate system for GCode int main() {             ch0->InputMode=ENCODER_MODE;     ch0->OutputMode=CL_STEP_DIR_MODE;     ch0->Vel=10000;     ch0->Accel=60;     ch0->Jerk=600;     ch0->P=0.3;     ch0->I=0.005;     ch0->D=0.1;     ch0->FFAccel=0.0005;     ch0->FFVel=0.0005;     ch0->MaxI=200;     ch0->MaxErr=10000;     ch0->MaxOutput=20;     ch0->DeadBandGain=0.5;     ch0->DeadBandRange=0;     ch0->InputChan0=0;     ch0->InputChan1=0;     ch0->OutputChan0=28;     ch0->OutputChan1=0;     ch0->MasterAxis=-1;     ch0->LimitSwitchOptions=0x0;     ch0->InputGain0=1;     ch0->InputGain1=1;     ch0->InputOffset0=0;     ch0->InputOffset1=0;     ch0->OutputGain=1;     ch0->OutputOffset=0;     ch0->SlaveGain=1;     ch0->BacklashMode=BACKLASH_OFF;     ch0->BacklashAmount=0;     ch0->BacklashRate=0;     ch0->invDistPerCycle=1;     ch0->Lead=0;     ch0->MaxFollowingError=100000;     ch0->StepperAmplitude=20;     ch0->iir[0].B0=1;     ch0->iir[0].B1=0;     ch0->iir[0].B2=0;     ch0->iir[0].A1=0;     ch0->iir[0].A2=0;     ch0->iir[1].B0=1;     ch0->iir[1].B1=0;     ch0->iir[1].B2=0;     ch0->iir[1].A1=0;     ch0->iir[1].A2=0;     ch0->iir[2].B0=0.000769;     ch0->iir[2].B1=0.001538;     ch0->iir[2].B2=0.000769;     ch0->iir[2].A1=1.92076;     ch0->iir[2].A2=-0.923833;     EnableAxisDest(0,0);     EnableAxisDest(1,0);     EnableAxisDest(2,0);     DefineCoordSystem(0,1,2,-1);     return 0; } Thank you ---In dynomotion@yahoogroups.com, <tk@...> wrote: Hi Cristian, (I hope you saw my earlier reply) Open collector mode works well for drives that have opto couplers with the LED anodes connected in together to +5V. Your drives are probably different and expecting TTL signals. Regards TK Group: DynoMotion Message: 8402 From: cristian_atzori Date: 9/26/2013 Subject: Re: Possibly fault Kflop board Ok, I think I got it. In all your manuals and examples P is set to 0 or very small values, but now I found the system working as it should with P set at least to 50. The linear encoder reads 1000 pulses/mm, and the motor will have a rotary encoder set to 1000 (read as 4000 from the drive). Now I can finally see my dual closed loop working, it'll be just a matter of time to calculate the correct PID values using Ziegler-Nichols tables, before final optimization. Thank you Tom for all the support! ---In dynomotion@yahoogroups.com, <cristian_atzori@...> wrote: Yes, they expect TTL waves. Now I made it work with output 28 (LVTTL quadrature) and succesfully moved the motor. Maybe you could give me the clue to configure the closed loop, because when moving a linear encoder (related to input 0, for channel 0) accross the 0 point I only can see a very short quadrature train, just when the position value crosses the Destination one. Short signals only when passing that zero point, and then nothing. When using the DAC output mode I can see, instead, a constant DC output until Position and Destination are convergent, and that is what I expected to see in step-dir output mode. Could you please tell me what are the values to modify to have a constant train of quadrature pulses when I move le linear encoder far away from the cross point? Here is the code I was using: #include "KMotionDef.h" // Defines axis 0, 1, 2 as simple step dir outputs // enables them // sets them as an xyz coordinate system for GCode int main() {             ch0->InputMode=ENCODER_MODE;     ch0->OutputMode=CL_STEP_DIR_MODE;     ch0->Vel=10000;     ch0->Accel=60;     ch0->Jerk=600;     ch0->P=0.3;     ch0->I=0.005;     ch0->D=0.1;     ch0->FFAccel=0.0005;     ch0->FFVel=0.0005;     ch0->MaxI=200;     ch0->MaxErr=10000;     ch0->MaxOutput=20;     ch0->DeadBandGain=0.5;     ch0->DeadBandRange=0;     ch0->InputChan0=0;     ch0->InputChan1=0;     ch0->OutputChan0=28;     ch0->OutputChan1=0;     ch0->MasterAxis=-1;     ch0->LimitSwitchOptions=0x0;     ch0->InputGain0=1;     ch0->InputGain1=1;     ch0->InputOffset0=0;     ch0->InputOffset1=0;     ch0->OutputGain=1;     ch0->OutputOffset=0;     ch0->SlaveGain=1;     ch0->BacklashMode=BACKLASH_OFF;     ch0->BacklashAmount=0;     ch0->BacklashRate=0;     ch0->invDistPerCycle=1;     ch0->Lead=0;     ch0->MaxFollowingError=100000;     ch0->StepperAmplitude=20;     ch0->iir[0].B0=1;     ch0->iir[0].B1=0;     ch0->iir[0].B2=0;     ch0->iir[0].A1=0;     ch0->iir[0].A2=0;     ch0->iir[1].B0=1;     ch0->iir[1].B1=0;     ch0->iir[1].B2=0;     ch0->iir[1].A1=0;     ch0->iir[1].A2=0;     ch0->iir[2].B0=0.000769;     ch0->iir[2].B1=0.001538;     ch0->iir[2].B2=0.000769;     ch0->iir[2].A1=1.92076;     ch0->iir[2].A2=-0.923833;     EnableAxisDest(0,0);     EnableAxisDest(1,0);     EnableAxisDest(2,0);     DefineCoordSystem(0,1,2,-1);     return 0; } Thank you ---In dynomotion@yahoogroups.com, <tk@...> wrote: Hi Cristian, (I hope you saw my earlier reply) Open collector mode works well for drives that have opto couplers with the LED anodes connected in together to +5V. Your drives are probably different and expecting TTL signals. Regards TK Group: DynoMotion Message: 8403 From: Tom Kerekes Date: 9/26/2013 Subject: Re: Possibly fault Kflop board Hi Cristian, I don't really understand that.  You should start off with P=0.  With CL Step/Direction mode the "Output" is a Position Offset (not a velocity or torque/acceleration like in a typical servo).  Normally I Gain is used to ramp the Output Offset at a rate proportional to error. You should match the linear encoder resolution to the Step Pulse resolution by changing the InputGain0 setting.  This way CL Step/Direction Feed Forward will cause the output to follow the Input Trajectory without any feedback (PID=0,0,0). The Settings in your previous email had a MaxOutput=20.  This limits the feedback correction offset to only +/-20 Steps.  That is why you only saw a few output pulses when the error switch direction. You should set FFAccel and FFVel to zero. Hope this helps. Regards TK Group: DynoMotion Message: 8404 From: cristian_atzori Date: 9/27/2013 Subject: Re: Possibly fault Kflop board Yes  Tom, I realized what MaxOutput is related to. The fact is that I was testing the linear encoder by hand, not connected in any mechanical way to the motor. The rotary encoder has variable resolution (internal dip switches), I'll then try to find the best way to make the system close the loop even with PID 0,0,0. Isn't it better to change rotary resolution in order to increase Output Gain instead of Input? ---In dynomotion@yahoogroups.com, <tk@...> wrote: Hi Cristian, I don't really understand that.  You should start off with P=0.  With CL Step/Direction mode the "Output" is a Position Offset (not a velocity or torque/acceleration like in a typical servo).  Normally I Gain is used to ramp the Output Offset at a rate proportional to error. You should match the linear encoder resolution to the Step Pulse resolution by changing the InputGain0 setting.  This way CL Step/Direction Feed Forward will cause the output to follow the Input Trajectory without any feedback (PID=0,0,0). The Settings in your previous email had a MaxOutput=20.  This limits the feedback correction offset to only +/-20 Steps.  That is why you only saw a few output pulses when the error switch direction. You should set FFAccel and FFVel to zero. Hope this helps. Regards TK