Hi Dean,

PWM amplifiers are a more complex configuration and more complex to tune so hopefully you have a good background for this. One programming mistake and you are likely to apply full voltage to the motor. Although that drive should provide peak current limiting.

I can't see where they specify the voltage input levels for the control signals (optos). But with 100 ohm series resistance I would expect them to work pretty well with common anode 3.3V. Of if necessary use the level shifters.

Common anode at 5V probably won't work. The optos would come on and there wouldn't be damage to KFLOP but it is likely the optos would never turn off. The PMW outputs on KFLOP are 3.3V and not open collector. So when high the hold ~3.3V

which would put ~1.7V on the opto (relative to 5V) and the opto may stay on or turn off too slowly.

KFLOP has eight 8-bit PWMs. With sign + magnitude you would have 9 bits of resolution. KFLOP doesn't support PWM output mode directly. You must add a few lines of code to do the sign/magnitude and write the servo output to the PWMs. See the example:

OutputToPWM3axis.c

Although that is written for anti-phase PWMs (50% duty cycle is no voltage) rather than sign + magnitude.

Those AMT encoders are not real optical encoders. I think they are capacitive with software interpolation. I would use real optical encoders. But others have had success using them.

The Amplifier PWMs at up to 20 KHz so you shouldn't hear and "singing".

HTH
Regards
TK

Group: DynoMotion	Message: 7046	From: deanw1a	Date: 3/19/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

Thanks.

I already have the amps and I would like to use them. I was hoping the DynoMotion products would help me put them to work.

>>hopefully you have a good background for this

I probably have bitten off more than I can chew. I have an EE degree but I am a computer Programmer... and I think I can do anything!

This is a hobby for me and I have more time for it than I do money.

Am I going to have to write a software PID loop for these drives?
Does DynoMotion already include PID code for the KFlop? Or, is that only available with the SnapAmp or KAnalog for +-10v control?

Does anyone have any links to information that can help me understand what is needed?. For instance, what is expected as PWM input to control these amps. 1% duty cycle with DIR toggling to hold the motors position at 1% current? I am clueless.

I do not want to spend more on extras to support these amps than I would on new discount drives. Also, I do not want to spend hundreds of hours coding and tuning if that time can be saved by using different drives.

Last question:

Will a SnapAmp drive 4 DC Brushed 100-120v 6-8A cont., 25-30A peak motors without stopping due to current overload/fault?. They are on a knee mill - 1 for the knee (2 or 3:1) and X,Y,Z axis (direct - 1:1).

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> Hi Dean,
>
> PWM amplifiers are a more complex configuration and more complex to tune so hopefully you have a good background for this.Â One programming mistake and you are likely to apply full voltage to the motor.Â Although that drive should provide peak current limiting.
....

Group: DynoMotion	Message: 7050	From: Tom Kerekes	Date: 3/19/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

Hi Dean,

It sounds like you have an appropriate background and it will take some time to sort out so that is good.

You can download our software, run KMotion.exe, goto the C Programs Screen, load OutputToPWM3axis.c, to get an idea of what is involved and if you are comfortable with it.

No you won't have to write PID equations. KFLOP has all that. All you need to do is to take the final Servo Output value and write it to the PWMs. You will need to convert it from a signed value to sign and magnitude.

The PWM value actually represents a Voltage not a current (like a percentage of the power supply). This is a somewhat difficult servo type to tune. A position error generates an output voltage, which ramps up current, which generates acceleration, which ramps up velocity, which finally provides a position change. This type of feedback is inherently unstable unless some phase lead is added somewhere (ie D Gain).

Regarding SnapAmp: if you limit acceleration such that high currents are no needed then you should not trip any current faults.

HTH
Regards
TK

Group: DynoMotion	Message: 7054	From: deanw1a	Date: 3/19/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

Thanks Tom.

The code looks like it sets up the PWM outputs, and then runs continuously - setting the PWM registers to a value based on the ch?->Output value every time slice.

So... All I need to do is write some code to monitor the commanded direction and speed (or is it offset?) and keep the PWM registers up to date?
That seems too simple. Easy as cake... piece of pie... Uh Oh.. I got the sign wrong and crashed the machine! :-)

Can someone point me to the programming reference where the ch0,1,2 structs/ranges/settings are documented?

How about the FPGA(IO_PWMS_PRESCALE), etc. registers?

Some references to what documentation to read to get up to speed would be helpful.

Thus far, I see that I might need
FPGA(IO_PWMS_PRESCALE) = 2; // set pwm period to 21.7 Khz
To get below the 25Khz the amplifiers allow.

Please: Any help finding docs explaining the signals I am trying to generate for the amps PWM- and DIR- pins.. and their purpose / result would be highly appreciated. Or just a quick simple explanation may be all I need to understand?

Just a side note about SpindleMach3PWM.c

>> FPGA(IO_PWMS_PRESCALE) = 65; // divide clock by 65 (1 KHz)

This looks like divide by 66 to get 986Hz?
I think this code could have been
FPGA(IO_PWMS_PRESCALE) = 64; // divide clock by 65 (1 KHz)
or
FPGA(IO_PWMS_PRESCALE) = 65; // divide clock by 66 (986 Hz)

Am I missing something?

Thanks again... I think I am hooked... your products are amazing! :-)

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> Hi Dean,
>
> It sounds like you have an appropriate background and it will take some time to sort out so that is good.
>
> You can download our software, run KMotion.exe, goto the C Programs Screen, load OutputToPWM3axis.c, to get an idea of what is involved and if you are comfortable with it.
>
> No you won't have to write PID equations.Â KFLOP has all that.Â All you need to do is to take the final Servo Output value and write it to the PWMs.Â You will need to convert it from a signed value to sign and magnitude.
>
> The PWM value actually represents a Voltage not a current (like a percentage of the power supply).Â This is a somewhat difficult servo type to tune.Â A position error generates an output voltage, which ramps up current, which generates acceleration, which ramps up velocity, which finally provides a position change.Â This type of feedback is inherently unstable unless some phase lead is added somewhere (ie D Gain).
>
>
> Regarding SnapAmp: if you limit acceleration such that high currents are no needed then you should not trip any current faults.
>
> HTH
> Regards
> TK
>
>
>
>
> ________________________________
> From: deanw1a <deanwyant@...>
> To: DynoMotion@yahoogroups.com
> Sent: Tuesday, March 19, 2013 9:44 AM
> Subject: [DynoMotion] Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps
>
>
> Â
> Thanks.
>
> I already have the amps and I would like to use them. I was hoping the DynoMotion products would help me put them to work.
>
> >>hopefully you have a good background for this
> I probably have bitten off more than I can chew. I have an EE degree but I am a computer Programmer... and I think I can do anything!
>
> This is a hobby for me and I have more time for it than I do money.
>
> Am I going to have to write a software PID loop for these drives?
> Does DynoMotion already include PID code for the KFlop? Or, is that only available with the SnapAmp or KAnalog for +-10v control?
>
> Does anyone have any links to information that can help me understand what is needed?. For instance, what is expected as PWM input to control these amps. 1% duty cycle with DIR toggling to hold the motors position at 1% current? I am clueless.
>
> I do not want to spend more on extras to support these amps than I would on new discount drives. Also, I do not want to spend hundreds of hours coding and tuning if that time can be saved by using different drives.
>
> Last question:
>
> Will a SnapAmp drive 4 DC Brushed 100-120v 6-8A cont., 25-30A peak motors without stopping due to current overload/fault?. They are on a knee mill - 1 for the knee (2 or 3:1) and X,Y,Z axis (direct - 1:1).
>
> --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> >
> > Hi Dean,
> >
> > PWM amplifiers are a more complex configuration and more complex to tune so hopefully you have a good background for this.ÃÂ One programming mistake and you are likely to apply full voltage to the motor.ÃÂ Although that drive should provide peak current limiting.
> ....
>

Group: DynoMotion	Message: 7058	From: Tom Kerekes	Date: 3/19/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

That is correct. You would just update the PWM duty cycle every time slice.

Most all the C functions and Variables mirror what can be set and done from the Console Script commands or from the KMotion Executive Program Screens. So look at those places for detail descriptions.

http://www.dynomotion.com/Help/CmdsCategory.htm

http://www.dynomotion.com/Help/ConfigurationScreen/ConfigurationScreen.htm

http://www.dynomotion.com/Help/StepScreen/StepScreen.htm

But the available C structures, functions, variables, register definitions and such are listed in the KMotionDef.h file

Regarding the PWM Pre-Scale: the exact PWM frequency is usually not very important. But you are correct that is a mistake, the 8-bit PWM counter runs at 16.6666MHz and so the frequency set by 64 would be:

16.66666MHz / 256 / (64+1) = 1001.6Hz

Regards

Group: DynoMotion	Message: 7067	From: deanw1a	Date: 3/19/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

I have looked at these references but I cannot determine how to set the KMotion settings A summary of what to do would be helpful.
Such as: On the Config screen, set the Axis Output Modes to DAC Servo. On the Step Screen, set the Max Output to 255. This will cause the CHAN->Output values to range from ? to ?.... In code, determine the direction by checking CHAN var???

I did not see how to determine direction from the CHAN struct or what determines the range of CHAN->Output. It looks like the max is set on the Step Screen and the mode set in Config determines the range. I see Ouput is a float... is it -Max to +Max to include direction? Or, 0 to Max and direction comes from another var? The sample looks like --128 to 127 ... but that would ommit a bit?
I see the Axis Output modes in the Channel Config, but I do not know which one to use to have the CHAN->Output value I need. Maybe DC Servo? DAC Servo?
I do not need to be spoon fed or to have someone write code for me, but I need a few more clues.

BTW, these amps can be controlled by sourcing the signals instead of sinking by not using the +5v opto input, tying the negative (PWM-) inputs to ground and signaling the + inputs. Is there any advantage having KFlop PWM source instead of sink? For some reason, I feel better sourcing than sinking... thinking positive :-)
Either way, I bought a logic level converter to take the 3.3v to 5v in case I need to.

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> That is correct.Â You would just update the PWM duty cycle every time slice.
>
> Most all the C functions and Variables mirror what can be set and done from the Console Script commands or from the KMotion Executive Program Screens.Â So look at those places for detail descriptions.
>
> http://www.dynomotion.com/Help/CmdsCategory.htm
>
> http://www.dynomotion.com/Help/ConfigurationScreen/ConfigurationScreen.htm
>
> http://www.dynomotion.com/Help/StepScreen/StepScreen.htm
>
>
> But the available C structures, functions, variables, register definitions and such are listed in the KMotionDef.h file
>
>
> Regarding the PWM Pre-Scale: the exact PWM frequency is usually not very important.Â But you are correct that is a mistake, the 8-bit PWM counter runs at 16.6666MHz and so the frequency set by 64 would be:
>
>
> 16.66666MHz / 256 /Â (64+1) =Â 1001.6Hz
>
>
>
> Regards
> TK

Group: DynoMotion	Message: 7068	From: Tom Kerekes	Date: 3/19/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

Hi Dean,

You can set the Output Mode to No Output as you will be doing the output with your program.

Yes set max Output to 255 which will limit the Output range from -255 to +255 so you won't have to worry about clipping it in your program.

The servo Output is a signed number and will be in the variable referenced as ch->Output the same as shown in the example.

The example is for a different type of PWM than yours. It is sometimes referred to as anti-phase where the motor always sees either +Supply or -Supply. To command 0V the duty cycle is set to 50% This results in 7 bits of resolution each direction.

I think your Amp is a sign and direction - but you should research it.

Regarding sourcing or sinking: The JP4 pins that the PWM signals come out on have 150ohm pull down resistors so they sink (16ma) better than they source. If you connect the optos anode to +5V the optos may not turn off when the output is 3.3V. If 3V won't turn on the optos then you may need the level converters. To invert a PWM signal output the complement of the value (255 would be fully off).

HTH

Regards

Group: DynoMotion	Message: 7073	From: deanw1a	Date: 3/20/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

I forgot about the 150 Ohm pull downs on the first 8 JP4 I/O pins.

The -255 to +255 range and No Output mode is what I needed to know for now.

It will probably be a while before I start running wires etc.

Thank you.

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> Hi Dean,
>
> You can set the Output Mode to No Output as you will be doing the output with your program.
>
> Yes set max Output to 255 which will limit the Output range from -255 to +255 so you won't have to worry about clipping it in your program.
>
>
> The servo Output is a signed number and will be in the variable referenced as ch->Output the same as shown in the example.
>
> The example is for a different type of PWM than yours.Â It is sometimes referred to as anti-phase where the motor always sees either +Supply or -Supply.Â To command 0V the duty cycle is set to 50%Â This results in 7 bits of resolution each direction.
>
> I think your Amp is a sign and direction - but you should research it.
>
> Regarding sourcing or sinking:Â The JP4 pins that the PWM signals come out on have 150ohm pull down resistors so they sink (16ma) better than they source.Â If you connect the optos anode to +5V the optos may not turn off when the output is 3.3V.Â If 3V won't turn on the optos then you may need the level converters.Â To invert a PWM signal output the complement of the value (255 would be fully off).
>
> HTH
> Regards
> TK
>
>
>
> ________________________________
> From: deanw1a <deanwyant@...>
> To: DynoMotion@yahoogroups.com
> Sent: Tuesday, March 19, 2013 8:09 PM
> Subject: [DynoMotion] Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps
>
>
> Â
> I have looked at these references but I cannot determine how to set the KMotion settings A summary of what to do would be helpful.
> Such as: On the Config screen, set the Axis Output Modes to DAC Servo. On the Step Screen, set the Max Output to 255. This will cause the CHAN->Output values to range from ? to ?.... In code, determine the direction by checking CHAN var???
>
> I did not see how to determine direction from the CHAN struct or what determines the range of CHAN->Output. It looks like the max is set on the Step Screen and the mode set in Config determines the range. I see Ouput is a float... is it -Max to +Max to include direction? Or, 0 to Max and direction comes from another var? The sample looks like --128 to 127 ... but that would ommit a bit?
> I see the Axis Output modes in the Channel Config, but I do not know which one to use to have the CHAN->Output value I need. Maybe DC Servo? DAC Servo?
> I do not need to be spoon fed or to have someone write code for me, but I need a few more clues.
>
> BTW, these amps can be controlled by sourcing the signals instead of sinking by not using the +5v opto input, tying the negative (PWM-) inputs to ground and signaling the + inputs. Is there any advantage having KFlop PWM source instead of sink? For some reason, I feel better sourcing than sinking... thinking positive :-)
> Either way, I bought a logic level converter to take the 3.3v to 5v in case I need to.
>
> --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> >
> > That is correct.ÃÂ You would just update the PWM duty cycle every time slice.
> >
> > Most all the C functions and Variables mirror what can be set and done from the Console Script commands or from the KMotion Executive Program Screens.ÃÂ So look at those places for detail descriptions.
> >
> > http://www.dynomotion.com/Help/CmdsCategory.htm
> >
> > http://www.dynomotion.com/Help/ConfigurationScreen/ConfigurationScreen.htm
> >
> > http://www.dynomotion.com/Help/StepScreen/StepScreen.htm
> >
> >
> > But the available C structures, functions, variables, register definitions and such are listed in the KMotionDef.h file
> >
> >
> > Regarding the PWM Pre-Scale: the exact PWM frequency is usually not very important.ÃÂ But you are correct that is a mistake, the 8-bit PWM counter runs at 16.6666MHz and so the frequency set by 64 would be:
> >
> >
> > 16.66666MHz / 256 /ÃÂ (64+1) =ÃÂ 1001.6Hz
> >
> >
> >
> > Regards
> > TK
>

Group: DynoMotion	Message: 7088	From: deanw1a	Date: 3/21/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

I have been researching KFlop/KMotion and the AMC BD30A8 drives and I could use some help clarifying a few issues.

1. It seems that I will have to use 16 KHz PWM because the next highest is 21.7 KHz and the drive specification states 5-20Khz. Wouldn't 20 KHz be better? Is there a way to get nearer to 20 KHz? Is there any reason to?
2. Using thread code to change duty cycle and direction limits the changes to about 1.4 KHz to 4 KHz - varying according to the number of threads running. Can this issue cause problems? Perhaps with tuning?
3. I still cannot find information on what the CHAN->Output value is set to / how it varies according to commands. Does it simply remain 0 at rest and change magnitude for commanded moves according to tuning values?
4. Since Output is probably 0 at rest, is it possible the amps I am using do not maintain torque at stand still? I have "BD" type drives that require "Direct PWM" (See manual reference below). If Output is 0 at rest, it seems possible that amp might not provide holding torque. I cannot see why AMC would do that, but I thought I would ask just in case someone already knows the answer.

Reference:
http://www.a-m-c.com/download/manual/AMC_AnalogDrives_InstallManual.pdf
Page 10
2.4.2 PWM and Direction
PWM and Direction Input is a specialized type of command that requires a compatible controller. The controller needs two high speed TTL digital outputs to control these drives, one for PWM and the other for Direction. The PWM duty cycle corresponds to the magnitude of the output. Direct control of the PWM switching puts response times in the sub-microsecond range. Since these drives don't take analog inputs for command the need for a D/A converter for drive control is eliminated.
A PWM and Direction drive can be operated in either Direct PWM or Torque Mode PWM.

* In Direct PWM (e.g. "BD" drives) the PWM input directly controls the PWM output, giving direct control of the switching frequency and duty cycle.
* In Torque Mode PWM (e.g. "BDC" drives) the PWM input goes into a PWM-to-Analog converter. The analog signal is then used as a command into the current loop, resulting in a Current Mode drive controlled with PWM and Direction.

To recap: If these were "BDC"/Torque Mode amps I would not be worried. But since they are "BD"/Direct Mode amps I am a concerned that they may not hold torque without toggling DIR and providing some PWM duty cycle representing holding torque. That seems a little silly, but I thought I would ask. If nobody knows, that is OK. I will find out for sure when I test the amps.

Thanks.

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> Hi Dean,
>
> You can set the Output Mode to No Output as you will be doing the output with your program.
>
> Yes set max Output to 255 which will limit the Output range from -255 to +255 so you won't have to worry about clipping it in your program.
>
>
> The servo Output is a signed number and will be in the variable referenced as ch->Output the same as shown in the example.
>
> The example is for a different type of PWM than yours.Â It is sometimes referred to as anti-phase where the motor always sees either +Supply or -Supply.Â To command 0V the duty cycle is set to 50%Â This results in 7 bits of resolution each direction.
>
> I think your Amp is a sign and direction - but you should research it.
>
> Regarding sourcing or sinking:Â The JP4 pins that the PWM signals come out on have 150ohm pull down resistors so they sink (16ma) better than they source.Â If you connect the optos anode to +5V the optos may not turn off when the output is 3.3V.Â If 3V won't turn on the optos then you may need the level converters.Â To invert a PWM signal output the complement of the value (255 would be fully off).
>
> HTH
> Regards
> TK
>
>
>
>

Group: DynoMotion	Message: 7097	From: Tom Kerekes	Date: 3/22/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

16 or 20KHz probably doesn't matter unless you can hear 16KHz.

A single User Thread runs at over 5KHz. The effect is a small amount of phase lag (about 6 degrees at 100Hz). There is a way to use a Servo Call back that happens every 90us regardless of the number of Threads. But it is one more complexity that I wouldn't worry about at this stage.

The Servo Output is computed by the Servo algorithms. See:

http://dynomotion.com/Help/ServoFlowDiagram.htm

I think you have a basic misunderstanding. A Servo doesn't need any current or power to hold position. As long as there is no error in where it should be it does nothing. Outputs Zero. Only if there is an error in where it should be does it try to correct the error by commanding an Output.

HTH
Regards
TK

On Mar 21, 2013, at 9:43 PM, "deanw1a" <deanwyant@...> wrote:

I have been researching KFlop/KMotion and the AMC BD30A8 drives and I could use some help clarifying a few issues.

1. It seems that I will have to use 16 KHz PWM because the next highest is 21.7 KHz and the drive specification states 5-20Khz. Wouldn't 20 KHz be better? Is there a way to get nearer to 20 KHz? Is there any reason to?
2. Using thread code to change duty cycle and direction limits the changes to about 1.4 KHz to 4 KHz - varying according to the number of threads running. Can this issue cause problems? Perhaps with tuning?
3. I still cannot find information on what the CHAN->Output value is set to / how it varies according to commands. Does it simply remain 0 at rest and change magnitude for commanded moves according to tuning values?
4. Since Output is probably 0 at rest, is it possible the amps I am using do not maintain torque at stand still? I have "BD" type drives that require "Direct PWM" (See manual reference below). If Output is 0 at rest, it seems possible that amp might not provide holding torque. I cannot see why AMC would do that, but I thought I would ask just in case someone already knows the answer.

Reference:
http://www.a-m-c.com/download/manual/AMC_AnalogDrives_InstallManual.pdf
Page 10
2.4.2 PWM and Direction
PWM and Direction Input is a specialized type of command that requires a compatible controller. The controller needs two high speed TTL digital outputs to control these drives, one for PWM and the other for Direction. The PWM duty cycle corresponds to the magnitude of the output. Direct control of the PWM switching puts response times in the sub-microsecond range. Since these drives don't take analog inputs for command the need for a D/A converter for drive control is eliminated.
A PWM and Direction drive can be operated in either Direct PWM or Torque Mode PWM.

* In Direct PWM (e.g. "BD" drives) the PWM input directly controls the PWM output, giving direct control of the switching frequency and duty cycle.
* In Torque Mode PWM (e.g. "BDC" drives) the PWM input goes into a PWM-to-Analog converter. The analog signal is then used as a command into the current loop, resulting in a Current Mode drive controlled with PWM and Direction.

To recap: If these were "BDC"/Torque Mode amps I would not be worried. But since they are "BD"/Direct Mode amps I am a concerned that they may not hold torque without toggling DIR and providing some PWM duty cycle representing holding torque. That seems a little silly, but I thought I would ask. If nobody knows, that is OK. I will find out for sure when I test the amps.

Thanks.

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> Hi Dean,
>
> You can set the Output Mode to No Output as you will be doing the output with your program.
>
> Yes set max Output to 255 which will limit the Output range from -255 to +255 so you won't have to worry about clipping it in your program.
>
>
> The servo Output is a signed number and will be in the variable referenced as ch->Output the same as shown in the example.
>
> The example is for a different type of PWM than yours. It is sometimes referred to as anti-phase where the motor always sees either +Supply or -Supply. To command 0V the duty cycle is set to 50% This results in 7 bits of resolution each direction.
>
> I think your Amp is a sign and direction - but you should research it.
>
> Regarding sourcing or sinking: The JP4 pins that the PWM signals come out on have 150ohm pull down resistors so they sink (16ma) better than they source. If you connect the optos anode to +5V the optos may not turn off when the output is 3.3V. If 3V won't turn on the optos then you may need the level converters. To invert a PWM signal output the complement of the value (255 would be fully off).
>
> HTH
> Regards
> TK
>
>
>
>

Group: DynoMotion	Message: 7099	From: deanw1a	Date: 3/22/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

The docs for "Direct PWM" just confused me. I was thinking that they may not "Servo". Silly.. I know. :-)
I now think that the amps should work the same as +-10v amps except they use PWM / DIR and do not close the loop / provide PID/Gain. I am glad KFlop has the PWMs, can close the loop, and provide the PID to control them. IF 16Khz sings, I'll try 21.7 to see if the amp will handle it even if the spec is 5-20.

I'll quit worrying and guessing and wait until I do some actual wiring/testing.

Thanks again.

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> 16 or 20KHz probably doesn't matter unless you can hear 16KHz.
>
> A single User Thread runs at over 5KHz. The effect is a small amount of phase lag (about 6 degrees at 100Hz). There is a way to use a Servo Call back that happens every 90us regardless of the number of Threads.Â But it is one more complexity that I wouldn't worry about at this stage.
>
> The Servo Output is computed by the Servo algorithms.Â See:
>
> http://dynomotion.com/Help/ServoFlowDiagram.htm
>
> I think you have a basic misunderstanding.Â A Servo doesn't need any current or power to hold position.Â As long as there is no error in where it should be it does nothing.Â Outputs Zero.Â Only if there is an error in where it should be does it try to correct the error by commanding an Output.
>
>
> HTH
> Regards
> TK
>
> On Mar 21, 2013, at 9:43 PM, "deanw1a" <deanwyant@...> wrote:
>
>
> Â
> >I have been researching KFlop/KMotion and the AMC BD30A8 drives and I could use some help clarifying a few issues.
> >
> >1. It seems that I will have to use 16 KHz PWM because the next highest is 21.7 KHz and the drive specification states 5-20Khz. Wouldn't 20 KHz be better? Is there a way to get nearer to 20 KHz? Is there any reason to?
> >2. Using thread code to change duty cycle and direction limits the changes to about 1.4 KHz to 4 KHz - varying according to the number of threads running. Can this issue cause problems? Perhaps with tuning?
> >3. I still cannot find information on what the CHAN->Output value is set to / how it varies according to commands. Does it simply remain 0 at rest and change magnitude for commanded moves according to tuning values?
> >4. Since Output is probably 0 at rest, is it possible the amps I am using do not maintain torque at stand still? I have "BD" type drives that require "Direct PWM" (See manual reference below). If Output is 0 at rest, it seems possible that amp might not provide holding torque. I cannot see why AMC would do that, but I thought I would ask just in case someone already knows the answer.
> >
> >Reference:
> >http://www.a-m-c.com/download/manual/AMC_AnalogDrives_InstallManual.pdf
> >Page 10
> >2.4.2 PWM and Direction
> >PWM and Direction Input is a specialized type of command that requires a compatible controller. The controller needs two high speed TTL digital outputs to control these drives, one for PWM and the other for Direction. The PWM duty cycle corresponds to the magnitude of the output. Direct control of the PWM switching puts response times in the sub-microsecond range. Since these drives don't take analog inputs for command the need for a D/A converter for drive control is eliminated.
> >A PWM and Direction drive can be operated in either Direct PWM or Torque Mode PWM.
> >
> >* In Direct PWM (e.g. "BD" drives) the PWM input directly controls the PWM output, giving direct control of the switching frequency and duty cycle.
> >* In Torque Mode PWM (e.g. "BDC" drives) the PWM input goes into a PWM-to-Analog converter. The analog signal is then used as a command into the current loop, resulting in a Current Mode drive controlled with PWM and Direction.
> >
> >To recap: If these were "BDC"/Torque Mode amps I would not be worried. But since they are "BD"/Direct Mode amps I am a concerned that they may not hold torque without toggling DIR and providing some PWM duty cycle representing holding torque. That seems a little silly, but I thought I would ask. If nobody knows, that is OK. I will find out for sure when I test the amps.
> >
> >Thanks.
> >
> >--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> >>
> >> Hi Dean,
> >>
> >> You can set the Output Mode to No Output as you will be doing the output with your program.
> >>
> >> Yes set max Output to 255 which will limit the Output range from -255 to +255 so you won't have to worry about clipping it in your program.
> >>
> >>
> >> The servo Output is a signed number and will be in the variable referenced as ch->Output the same as shown in the example.
> >>
> >> The example is for a different type of PWM than yours.Â It is sometimes referred to as anti-phase where the motor always sees either +Supply or -Supply.Â To command 0V the duty cycle is set to 50%Â This results in 7 bits of resolution each direction.
> >>
> >> I think your Amp is a sign and direction - but you should research it.
> >>
> >> Regarding sourcing or sinking:Â The JP4 pins that the PWM signals come out on have 150ohm pull down resistors so they sink (16ma) better than they source.Â If you connect the optos anode to +5V the optos may not turn off when the output is 3.3V.Â If 3V won't turn on the optos then you may need the level converters.Â To invert a PWM signal output the complement of the value (255 would be fully off).
> >>
> >> HTH
> >> Regards
> >> TK
> >>
> >>
> >>
> >>
> >
> >
> >
>

Group: DynoMotion	Message: 7256	From: deanw1a	Date: 4/14/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

OK, I have a KFlop, AMC BD30A8 amp and motor setup on the bench for testing. I have found a couple of problems.

Problem #1: If I power the amp with no PWM- input (or PWM- at 5v - see problem #2), the motor does not have holding torque. So, I guess I was not off base thinking that may be the case. It seems that the PWM- input is what commands any torque/speed. I think that means I will need the PWM set to a torque value and the DIR flip-flopping at the PWM frequency / 2.
Isn't this going to be a problem?
I looks like all PWMs must be at the same frequency. I do not know how to generate a square wave at exactly half the PWM frequency.
Is there any way to do that?

Problem #2: The PWM- input to the amp expects the PWM to be inverted. IOW, it needs to be a constant high (100% PWM) for the motor to be stopped (no torque). Considering Problem #1, this does not actually seem to be a problem because I will need a PWM duty cycle at all times to maintain stand-still torque. I would not need 0% or 100%. So, I can just invert the duty cycle in code.

Thanks

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> 16 or 20KHz probably doesn't matter unless you can hear 16KHz.
>
> A single User Thread runs at over 5KHz. The effect is a small amount of phase lag (about 6 degrees at 100Hz). There is a way to use a Servo Call back that happens every 90us regardless of the number of Threads.Â But it is one more complexity that I wouldn't worry about at this stage.
>
> The Servo Output is computed by the Servo algorithms.Â See:
>
> http://dynomotion.com/Help/ServoFlowDiagram.htm
>
> I think you have a basic misunderstanding.Â A Servo doesn't need any current or power to hold position.Â As long as there is no error in where it should be it does nothing.Â Outputs Zero.Â Only if there is an error in where it should be does it try to correct the error by commanding an Output.
>
>
> HTH
> Regards
> TK
>

Group: DynoMotion	Message: 7257	From: Tom Kerekes	Date: 4/14/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

Hi Dean,

As I mentioned in the previous email a Servo motor doesn't have "holding torque". It only has torque one way or the other. It is the closed loop encoder feedback that will make it hold a particular position. So as you are testing the motor/amplifier itself you should only be concerned that you can command a torque of a controlled amount and direction. I don't think you will want to be flip-flopping the direction. The direction will be solid while the servo is accelerating toward the target. And then it will reverse at the time to decelerate.

regarding #2 yes you can just invert the duty cycle. Set the PWM duty cycle to 255 to generate on output torque, 128 for half torque, and 0 for full torque. That is the advantage of using a few lines of C to command the PWM.

HTH

Regards

Group: DynoMotion	Message: 7258	From: deanw1a	Date: 4/15/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

Thanks.

Of course... there is no holding torque when the loop is not closed. However, I thought I had the loop closed. I misunderstood the method of maintaining holding torque. I see now that the change of position would cause the controller to command torque to return to the correct position.

I have no problem commanding a torque and direction using a KFlop PWM and an output for DIR. I just have not figured out why the software is not controlling the ch->Output value. ch->Output remains 0 whether I turn the shaft by hand, use the CNC interface to jog, etc. I must be missing some steps.

Before posting another question, I will look into the docs and forum posts for info on testing encoder position tracking, etc.

Thanks.

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> Hi Dean,
>
> As I mentioned in the previous email a Servo motor doesn't have "holding torque".Â It only has torque one way or the other.Â It is the closed loop encoder feedback that will make it hold a particular position.Â So as you are testing the motor/amplifier itself you should only be concerned that you can command a torque of a controlled amount and direction. Â I don't think you will want to be flip-flopping the direction.Â The direction will be solid while the servo is accelerating toward the target.Â And then it will reverse at the time to decelerate.
>
>
> regarding #2 yes you can just invert the duty cycle.Â Set the PWM duty cycle to 255 to generate on output torque, 128 for half torque, and 0 for full torque.Â That is the advantage of using a few lines of C to command the PWM.
>
> HTH
> Regards
> TK
>
>
>
>
> ________________________________
> From: deanw1a <deanwyant@...>
> To: DynoMotion@yahoogroups.com
> Sent: Sunday, April 14, 2013 6:43 PM
> Subject: [DynoMotion] Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps
>
>
>
> Â
> OK, I have a KFlop, AMC BD30A8 amp and motor setup on the bench for testing. I have found a couple of problems.
>
> Problem #1: If I power the amp with no PWM- input (or PWM- at 5v - see problem #2), the motor does not have holding torque. So, I guess I was not off base thinking that may be the case. It seems that the PWM- input is what commands any torque/speed. I think that means I will need the PWM set to a torque value and the DIR flip-flopping at the PWM frequency / 2.
> Isn't this going to be a problem?
> I looks like all PWMs must be at the same frequency. I do not know how to generate a square wave at exactly half the PWM frequency.
> Is there any way to do that?
>
> Problem #2: The PWM- input to the amp expects the PWM to be inverted. IOW, it needs to be a constant high (100% PWM) for the motor to be stopped (no torque). Considering Problem #1, this does not actually seem to be a problem because I will need a PWM duty cycle at all times to maintain stand-still torque. I would not need 0% or 100%. So, I can just invert the duty cycle in code.
>
> Thanks
>
> --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> >
> > 16 or 20KHz probably doesn't matter unless you can hear 16KHz.
> >
> > A single User Thread runs at over 5KHz. The effect is a small amount of phase lag (about 6 degrees at 100Hz). There is a way to use a Servo Call back that happens every 90us regardless of the number of Threads.ÃÂ But it is one more complexity that I wouldn't worry about at this stage.
> >
> > The Servo Output is computed by the Servo algorithms.ÃÂ See:
> >
> > http://dynomotion.com/Help/ServoFlowDiagram.htm
> >
> > I think you have a basic misunderstanding.ÃÂ A Servo doesn't need any current or power to hold position.ÃÂ As long as there is no error in where it should be it does nothing.ÃÂ Outputs Zero.ÃÂ Only if there is an error in where it should be does it try to correct the error by commanding an Output.
> >
> >
> > HTH
> > Regards
> > TK
> >
>

Group: DynoMotion	Message: 7260	From: Tom Kerekes	Date: 4/15/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

Hi Dean,

I would start with the KanalogInitialPID.mot configuration. Run your OutputPWM.c Program. Then push "Enable" on the Step Response Screen. Your motor should try to hold position (weakly and without any damping).

Check on the Axis Screen if the Encoder counts correctly as the shaft turns.

In order to watch the servo output (which is not normally displayed anywhere) I use a trick to write it to something like an unused axis Destination by adding a line in the loop like:

ch7->Dest = ch0->Output;

If you post your C program we can check it over.

Good luck,

Group: DynoMotion	Message: 7269	From: deanw1a	Date: 4/15/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

OK, I loaded the KanalogInitialPID.mot file, changed the Output to No Output, set bit 26(PWM 0) and 30(DIR 0) to Out, powered up the amp, enabled the channel. When I switched the amp INHIBIT to enable the amp, the motor just about jumped out of its clamps and off of the work bench.
I put in a code line to limit output to 10 and that allowed me to enable the amp without the motor jumpimg off the bench.
The rotor vibrates back and forth 1/8 inch or so. I see the position in the Axis screen jumping around from -80 to 70 and in between these values. I monitored the output as you suggested and it is flipping between +-70. If I apply pressure to turn the shaft, there is a decent amount of torque (even with Output limited to 10). And, putting CW pressure on it causes the vibration to nearly stop and the commanded Output to stabilize at 7.
The encoder is set to 2048 so I get 8192 positions per rev.
I do not have Encoder 1 hooked up (it is floating) and I noticed that the Axis 1 position is incrementing every so often. I assume this is just noise because it is floating.

I do not know if any of this info is useful.
Is there anything unusual about any of this?

I guess that it is time to dig into tuning.
Any tips on things to do would be appreciated.

I am going to figure out how to attach the mot file and thread code I am using for testing.

Thanks

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> Hi Dean,
>
> I would start with the KanalogInitialPID.mot configuration.Â Run your OutputPWM.c Program.Â Then push "Enable" on the Step Response Screen.Â Your motor should try to hold position (weakly and without any damping).
>
> Check on the Axis Screen if the Encoder counts correctly as the shaft turns.
>
> In order to watch the servo output (which is not normally displayed anywhere) I use a trick to write it to something like an unused axis Destination by adding a line in the loop like:
>
> ch7->Dest = ch0->Output;
>
> If you post your C program we can check it over.
>
> Good luck,
> TK
>
>
>
>
> ________________________________
> From: deanw1a <deanwyant@...>
> To: DynoMotion@yahoogroups.com
> Sent: Monday, April 15, 2013 6:39 AM
> Subject: [DynoMotion] Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps
>
>
>
> Â
> Thanks.
>
> Of course... there is no holding torque when the loop is not closed. However, I thought I had the loop closed. I misunderstood the method of maintaining holding torque. I see now that the change of position would cause the controller to command torque to return to the correct position.
>
> I have no problem commanding a torque and direction using a KFlop PWM and an output for DIR. I just have not figured out why the software is not controlling the ch->Output value. ch->Output remains 0 whether I turn the shaft by hand, use the CNC interface to jog, etc. I must be missing some steps.
>
> Before posting another question, I will look into the docs and forum posts for info on testing encoder position tracking, etc.
>
> Thanks.
>
> --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> >
> > Hi Dean,
> >
> > As I mentioned in the previous email a Servo motor doesn't have "holding torque".ÃÂ It only has torque one way or the other.ÃÂ It is the closed loop encoder feedback that will make it hold a particular position.ÃÂ So as you are testing the motor/amplifier itself you should only be concerned that you can command a torque of a controlled amount and direction. ÃÂ I don't think you will want to be flip-flopping the direction.ÃÂ The direction will be solid while the servo is accelerating toward the target.ÃÂ And then it will reverse at the time to decelerate.
> >
> >
> > regarding #2 yes you can just invert the duty cycle.ÃÂ Set the PWM duty cycle to 255 to generate on output torque, 128 for half torque, and 0 for full torque.ÃÂ That is the advantage of using a few lines of C to command the PWM.
> >
> > HTH
> > Regards
> > TK
> >
> >
> >
> >
> > ________________________________
> > From: deanw1a <deanwyant@>
> > To: DynoMotion@yahoogroups.com
> > Sent: Sunday, April 14, 2013 6:43 PM
> > Subject: [DynoMotion] Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps
> >
> >
> >
> > ÃÂ
> > OK, I have a KFlop, AMC BD30A8 amp and motor setup on the bench for testing. I have found a couple of problems.
> >
> > Problem #1: If I power the amp with no PWM- input (or PWM- at 5v - see problem #2), the motor does not have holding torque. So, I guess I was not off base thinking that may be the case. It seems that the PWM- input is what commands any torque/speed. I think that means I will need the PWM set to a torque value and the DIR flip-flopping at the PWM frequency / 2.
> > Isn't this going to be a problem?
> > I looks like all PWMs must be at the same frequency. I do not know how to generate a square wave at exactly half the PWM frequency.
> > Is there any way to do that?
> >
> > Problem #2: The PWM- input to the amp expects the PWM to be inverted. IOW, it needs to be a constant high (100% PWM) for the motor to be stopped (no torque). Considering Problem #1, this does not actually seem to be a problem because I will need a PWM duty cycle at all times to maintain stand-still torque. I would not need 0% or 100%. So, I can just invert the duty cycle in code.
> >
> > Thanks
> >
> > --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> > >
> > > 16 or 20KHz probably doesn't matter unless you can hear 16KHz.
> > >
> > > A single User Thread runs at over 5KHz. The effect is a small amount of phase lag (about 6 degrees at 100Hz). There is a way to use a Servo Call back that happens every 90us regardless of the number of Threads.ÃâÃÂ But it is one more complexity that I wouldn't worry about at this stage.
> > >
> > > The Servo Output is computed by the Servo algorithms.ÃâÃÂ See:
> > >
> > > http://dynomotion.com/Help/ServoFlowDiagram.htm
> > >
> > > I think you have a basic misunderstanding.ÃâÃÂ A Servo doesn't need any current or power to hold position.ÃâÃÂ As long as there is no error in where it should be it does nothing.ÃâÃÂ Outputs Zero.ÃâÃÂ Only if there is an error in where it should be does it try to correct the error by commanding an Output.
> > >
> > >
> > > HTH
> > > Regards
> > > TK
> > >
> >
>

Group: DynoMotion	Message: 7271	From: Tom Kerekes	Date: 4/15/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

Hi Dean,

It is hard to say if it is just a tuning issue or a bug in your PWM control.

I think the KanalogInitPID.mot just has P Gain of 0.2. Which means if the Position is oscillating +/- 70 counts the output should only be oscillating +/-14 counts (not 70). So I don't understand that.

You can utilize the Max Output Parameter rather than adding your own code to limit the Output to 10 (I assume you mean +/- 10).

If you make a "Move" of size 0 you will see a plot of exactly what is happening (Command, Position, and Output).

With Yahoo Group's Web interface it isn't possible to attach things to posts. But you can use any email client and just send an email with attachments to the Group.

Regards
TK

Group: DynoMotion	Message: 7277	From: deanw1a	Date: 4/15/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

I replyed from gmail with attachments but I do not see it posted.
For some reason, the P was 1. I changed the P to 0.1 and set the max ouput to 10, downloaded the channel and now I get position ranging from -22 to 29 and Ouput ranging from -2 to 2.
I used the Step Response screen to Move 0 for 1 sec. The graph shows the command = 0 , and the Position and Output are oscillating 90 degress out of phase (opposite each other). I tried reversing gain and DIR, but that just makes the motor turn continuously.

The channel code is simply the KanalogInitPID.mot with P changed to 0.1 and the OutputMode set to No Output.

Here is the simple PWM DIR code:

#include "KMotionDef.h"

int out = 0; // PWM absolute Output value
main()
{
SetBitDirection(26, 1); // Set bit 26 (PWM 0 as an output)
SetBitDirection(30, 1); // Set 30 Pin as an output for DIR 0
FPGA(IO_PWMS_PRESCALE) = 2; // set pwm period to 21 KHz
FPGA(IO_PWMS + 1) = 1; // enable PWM 0

for (;;) //loop forever
{
WaitNextTimeSlice();
if (ch0->Enable)
{
out = ch0->Output; // converts to int
if (out < 0)
{
SetStateBit(30, 1);
out = -out;
}
else
SetStateBit(30, 0);

FPGA(IO_PWMS + 0) = 255 - out; // Inverted PWM
}
else
{
FPGA(IO_PWMS + 0) = 255; // Inverted - 255 = lowest torque - Note: 255 still has pulses - it is not steady state high
}
}
}

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> Hi Dean,
>
> It is hard to say if it is just a tuning issue or a bug in your PWM control.
>
> I think the KanalogInitPID.mot just has P Gain of 0.2.Â Which means if the Position is oscillating +/- 70 counts the output should only be oscillating +/-14 counts (not 70).Â So I don't understand that.Â
>
> You can utilize the Max Output Parameter rather than adding your own code to limit the Output to 10 (I assume you mean +/- 10).
>
> If you make a "Move" of size 0 you will see a plot of exactly what is happening (Command, Position, and Output).
>
> With Yahoo Group's Web interface it isn't possible to attach things to posts.Â But you can use any email client and just send an email with attachments to the Group.
>
> Regards
> TK
>
>
>
>
> ________________________________
> From: deanw1a <deanwyant@...>
> To: DynoMotion@yahoogroups.com
> Sent: Monday, April 15, 2013 4:06 PM
> Subject: [DynoMotion] Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps
>
>
>
> Â
> OK, I loaded the KanalogInitialPID.mot file, changed the Output to No Output, set bit 26(PWM 0) and 30(DIR 0) to Out, powered up the amp, enabled the channel. When I switched the amp INHIBIT to enable the amp, the motor just about jumped out of its clamps and off of the work bench.
> I put in a code line to limit output to 10 and that allowed me to enable the amp without the motor jumpimg off the bench.
> The rotor vibrates back and forth 1/8 inch or so. I see the position in the Axis screen jumping around from -80 to 70 and in between these values. I monitored the output as you suggested and it is flipping between +-70. If I apply pressure to turn the shaft, there is a decent amount of torque (even with Output limited to 10). And, putting CW pressure on it causes the vibration to nearly stop and the commanded Output to stabilize at 7.
> The encoder is set to 2048 so I get 8192 positions per rev.
> I do not have Encoder 1 hooked up (it is floating) and I noticed that the Axis 1 position is incrementing every so often. I assume this is just noise because it is floating.
>
> I do not know if any of this info is useful.
> Is there anything unusual about any of this?
>
> I guess that it is time to dig into tuning.
> Any tips on things to do would be appreciated.
>
> I am going to figure out how to attach the mot file and thread code I am using for testing.
>
> Thanks
>
> --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> >
> > Hi Dean,
> >
> > I would start with the KanalogInitialPID.mot configuration.ÃÂ Run your OutputPWM.c Program.ÃÂ Then push "Enable" on the Step Response Screen.ÃÂ Your motor should try to hold position (weakly and without any damping).
> >
> > Check on the Axis Screen if the Encoder counts correctly as the shaft turns.
> >
> > In order to watch the servo output (which is not normally displayed anywhere) I use a trick to write it to something like an unused axis Destination by adding a line in the loop like:
> >
> > ch7->Dest = ch0->Output;
> >
> > If you post your C program we can check it over.
> >
> > Good luck,
> > TK
> >
> >
> >
> >
> > ________________________________
> > From: deanw1a <deanwyant@>
> > To: DynoMotion@yahoogroups.com
> > Sent: Monday, April 15, 2013 6:39 AM
> > Subject: [DynoMotion] Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps
> >
> >
> >
> > ÃÂ
> > Thanks.
> >
> > Of course... there is no holding torque when the loop is not closed. However, I thought I had the loop closed. I misunderstood the method of maintaining holding torque. I see now that the change of position would cause the controller to command torque to return to the correct position.
> >
> > I have no problem commanding a torque and direction using a KFlop PWM and an output for DIR. I just have not figured out why the software is not controlling the ch->Output value. ch->Output remains 0 whether I turn the shaft by hand, use the CNC interface to jog, etc. I must be missing some steps.
> >
> > Before posting another question, I will look into the docs and forum posts for info on testing encoder position tracking, etc.
> >
> > Thanks.
> >
> > --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> > >
> > > Hi Dean,
> > >
> > > As I mentioned in the previous email a Servo motor doesn't have "holding torque".ÃâÃÂ It only has torque one way or the other.ÃâÃÂ It is the closed loop encoder feedback that will make it hold a particular position.ÃâÃÂ So as you are testing the motor/amplifier itself you should only be concerned that you can command a torque of a controlled amount and direction. ÃâÃÂ I don't think you will want to be flip-flopping the direction.ÃâÃÂ The direction will be solid while the servo is accelerating toward the target.ÃâÃÂ And then it will reverse at the time to decelerate.
> > >
> > >
> > > regarding #2 yes you can just invert the duty cycle.ÃâÃÂ Set the PWM duty cycle to 255 to generate on output torque, 128 for half torque, and 0 for full torque.ÃâÃÂ That is the advantage of using a few lines of C to command the PWM.
> > >
> > > HTH
> > > Regards
> > > TK
> > >
> > >
> > >
> > >
> > > ________________________________
> > > From: deanw1a <deanwyant@>
> > > To: DynoMotion@yahoogroups.com
> > > Sent: Sunday, April 14, 2013 6:43 PM
> > > Subject: [DynoMotion] Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps
> > >
> > >
> > >
> > > ÃâÃÂ
> > > OK, I have a KFlop, AMC BD30A8 amp and motor setup on the bench for testing. I have found a couple of problems.
> > >
> > > Problem #1: If I power the amp with no PWM- input (or PWM- at 5v - see problem #2), the motor does not have holding torque. So, I guess I was not off base thinking that may be the case. It seems that the PWM- input is what commands any torque/speed. I think that means I will need the PWM set to a torque value and the DIR flip-flopping at the PWM frequency / 2.
> > > Isn't this going to be a problem?
> > > I looks like all PWMs must be at the same frequency. I do not know how to generate a square wave at exactly half the PWM frequency.
> > > Is there any way to do that?
> > >
> > > Problem #2: The PWM- input to the amp expects the PWM to be inverted. IOW, it needs to be a constant high (100% PWM) for the motor to be stopped (no torque). Considering Problem #1, this does not actually seem to be a problem because I will need a PWM duty cycle at all times to maintain stand-still torque. I would not need 0% or 100%. So, I can just invert the duty cycle in code.
> > >
> > > Thanks
> > >
> > > --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> > > >
> > > > 16 or 20KHz probably doesn't matter unless you can hear 16KHz.
> > > >
> > > > A single User Thread runs at over 5KHz. The effect is a small amount of phase lag (about 6 degrees at 100Hz). There is a way to use a Servo Call back that happens every 90us regardless of the number of Threads.ÃÆ'Ã¢â¬Å¡ÃâÃÂ But it is one more complexity that I wouldn't worry about at this stage.
> > > >
> > > > The Servo Output is computed by the Servo algorithms.ÃÆ'Ã¢â¬Å¡ÃâÃÂ See:
> > > >
> > > > http://dynomotion.com/Help/ServoFlowDiagram.htm
> > > >
> > > > I think you have a basic misunderstanding.ÃÆ'Ã¢â¬Å¡ÃâÃÂ A Servo doesn't need any current or power to hold position.ÃÆ'Ã¢â¬Å¡ÃâÃÂ As long as there is no error in where it should be it does nothing.ÃÆ'Ã¢â¬Å¡ÃâÃÂ Outputs Zero.ÃÆ'Ã¢â¬Å¡ÃâÃÂ Only if there is an error in where it should be does it try to correct the error by commanding an Output.
> > > >
> > > >
> > > > HTH
> > > > Regards
> > > > TK
> > > >
> > >
> >
>

Group: DynoMotion	Message: 7278	From: deanw1a	Date: 4/15/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

I'm sorry, I should have said the position and output waveforms were 180 degrees out of phase - not 90.

--- In DynoMotion@yahoogroups.com, "deanw1a" <deanwyant@...> wrote:
>
> I replyed from gmail with attachments but I do not see it posted.
> For some reason, the P was 1. I changed the P to 0.1 and set the max ouput to 10, downloaded the channel and now I get position ranging from -22 to 29 and Ouput ranging from -2 to 2.
> I used the Step Response screen to Move 0 for 1 sec. The graph shows the command = 0 , and the Position and Output are oscillating 90 degress out of phase (opposite each other). I tried reversing gain and DIR, but that just makes the motor turn continuously.
>
> The channel code is simply the KanalogInitPID.mot with P changed to 0.1 and the OutputMode set to No Output.
>
> Here is the simple PWM DIR code:
>
> #include "KMotionDef.h"
>
> int out = 0; // PWM absolute Output value
> main()
> {
> SetBitDirection(26, 1); // Set bit 26 (PWM 0 as an output)
> SetBitDirection(30, 1); // Set 30 Pin as an output for DIR 0
> FPGA(IO_PWMS_PRESCALE) = 2; // set pwm period to 21 KHz
> FPGA(IO_PWMS + 1) = 1; // enable PWM 0
>
> for (;;) //loop forever
> {
> WaitNextTimeSlice();
> if (ch0->Enable)
> {
> out = ch0->Output; // converts to int
> if (out < 0)
> {
> SetStateBit(30, 1);
> out = -out;
> }
> else
> SetStateBit(30, 0);
>
> FPGA(IO_PWMS + 0) = 255 - out; // Inverted PWM
> }
> else
> {
> FPGA(IO_PWMS + 0) = 255; // Inverted - 255 = lowest torque - Note: 255 still has pulses - it is not steady state high
> }
> }
> }
>
> --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> >
> > Hi Dean,
> >
> > It is hard to say if it is just a tuning issue or a bug in your PWM control.
> >
> > I think the KanalogInitPID.mot just has P Gain of 0.2.Â Which means if the Position is oscillating +/- 70 counts the output should only be oscillating +/-14 counts (not 70).Â So I don't understand that.Â
> >
> > You can utilize the Max Output Parameter rather than adding your own code to limit the Output to 10 (I assume you mean +/- 10).
> >
> > If you make a "Move" of size 0 you will see a plot of exactly what is happening (Command, Position, and Output).
> >
> > With Yahoo Group's Web interface it isn't possible to attach things to posts.Â But you can use any email client and just send an email with attachments to the Group.
> >
> > Regards
> > TK
> >
> >
> >
> >
> > ________________________________
> > From: deanw1a <deanwyant@>
> > To: DynoMotion@yahoogroups.com
> > Sent: Monday, April 15, 2013 4:06 PM
> > Subject: [DynoMotion] Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps
> >
> >
> >
> > Â
> > OK, I loaded the KanalogInitialPID.mot file, changed the Output to No Output, set bit 26(PWM 0) and 30(DIR 0) to Out, powered up the amp, enabled the channel. When I switched the amp INHIBIT to enable the amp, the motor just about jumped out of its clamps and off of the work bench.
> > I put in a code line to limit output to 10 and that allowed me to enable the amp without the motor jumpimg off the bench.
> > The rotor vibrates back and forth 1/8 inch or so. I see the position in the Axis screen jumping around from -80 to 70 and in between these values. I monitored the output as you suggested and it is flipping between +-70. If I apply pressure to turn the shaft, there is a decent amount of torque (even with Output limited to 10). And, putting CW pressure on it causes the vibration to nearly stop and the commanded Output to stabilize at 7.
> > The encoder is set to 2048 so I get 8192 positions per rev.
> > I do not have Encoder 1 hooked up (it is floating) and I noticed that the Axis 1 position is incrementing every so often. I assume this is just noise because it is floating.
> >
> > I do not know if any of this info is useful.
> > Is there anything unusual about any of this?
> >
> > I guess that it is time to dig into tuning.
> > Any tips on things to do would be appreciated.
> >
> > I am going to figure out how to attach the mot file and thread code I am using for testing.
> >
> > Thanks
> >
> > --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> > >
> > > Hi Dean,
> > >
> > > I would start with the KanalogInitialPID.mot configuration.ÃÂ Run your OutputPWM.c Program.ÃÂ Then push "Enable" on the Step Response Screen.ÃÂ Your motor should try to hold position (weakly and without any damping).
> > >
> > > Check on the Axis Screen if the Encoder counts correctly as the shaft turns.
> > >
> > > In order to watch the servo output (which is not normally displayed anywhere) I use a trick to write it to something like an unused axis Destination by adding a line in the loop like:
> > >
> > > ch7->Dest = ch0->Output;
> > >
> > > If you post your C program we can check it over.
> > >
> > > Good luck,
> > > TK
> > >
> > >
> > >
> > >
> > > ________________________________
> > > From: deanw1a <deanwyant@>
> > > To: DynoMotion@yahoogroups.com
> > > Sent: Monday, April 15, 2013 6:39 AM
> > > Subject: [DynoMotion] Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps
> > >
> > >
> > >
> > > ÃÂ
> > > Thanks.
> > >
> > > Of course... there is no holding torque when the loop is not closed. However, I thought I had the loop closed. I misunderstood the method of maintaining holding torque. I see now that the change of position would cause the controller to command torque to return to the correct position.
> > >
> > > I have no problem commanding a torque and direction using a KFlop PWM and an output for DIR. I just have not figured out why the software is not controlling the ch->Output value. ch->Output remains 0 whether I turn the shaft by hand, use the CNC interface to jog, etc. I must be missing some steps.
> > >
> > > Before posting another question, I will look into the docs and forum posts for info on testing encoder position tracking, etc.
> > >
> > > Thanks.
> > >
> > > --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> > > >
> > > > Hi Dean,
> > > >
> > > > As I mentioned in the previous email a Servo motor doesn't have "holding torque".ÃâÃÂ It only has torque one way or the other.ÃâÃÂ It is the closed loop encoder feedback that will make it hold a particular position.ÃâÃÂ So as you are testing the motor/amplifier itself you should only be concerned that you can command a torque of a controlled amount and direction. ÃâÃÂ I don't think you will want to be flip-flopping the direction.ÃâÃÂ The direction will be solid while the servo is accelerating toward the target.ÃâÃÂ And then it will reverse at the time to decelerate.
> > > >
> > > >
> > > > regarding #2 yes you can just invert the duty cycle.ÃâÃÂ Set the PWM duty cycle to 255 to generate on output torque, 128 for half torque, and 0 for full torque.ÃâÃÂ That is the advantage of using a few lines of C to command the PWM.
> > > >
> > > > HTH
> > > > Regards
> > > > TK
> > > >
> > > >
> > > >
> > > >
> > > > ________________________________
> > > > From: deanw1a <deanwyant@>
> > > > To: DynoMotion@yahoogroups.com
> > > > Sent: Sunday, April 14, 2013 6:43 PM
> > > > Subject: [DynoMotion] Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps
> > > >
> > > >
> > > >
> > > > ÃâÃÂ
> > > > OK, I have a KFlop, AMC BD30A8 amp and motor setup on the bench for testing. I have found a couple of problems.
> > > >
> > > > Problem #1: If I power the amp with no PWM- input (or PWM- at 5v - see problem #2), the motor does not have holding torque. So, I guess I was not off base thinking that may be the case. It seems that the PWM- input is what commands any torque/speed. I think that means I will need the PWM set to a torque value and the DIR flip-flopping at the PWM frequency / 2.
> > > > Isn't this going to be a problem?
> > > > I looks like all PWMs must be at the same frequency. I do not know how to generate a square wave at exactly half the PWM frequency.
> > > > Is there any way to do that?
> > > >
> > > > Problem #2: The PWM- input to the amp expects the PWM to be inverted. IOW, it needs to be a constant high (100% PWM) for the motor to be stopped (no torque). Considering Problem #1, this does not actually seem to be a problem because I will need a PWM duty cycle at all times to maintain stand-still torque. I would not need 0% or 100%. So, I can just invert the duty cycle in code.
> > > >
> > > > Thanks
> > > >
> > > > --- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@> wrote:
> > > > >
> > > > > 16 or 20KHz probably doesn't matter unless you can hear 16KHz.
> > > > >
> > > > > A single User Thread runs at over 5KHz. The effect is a small amount of phase lag (about 6 degrees at 100Hz). There is a way to use a Servo Call back that happens every 90us regardless of the number of Threads.ÃÆ'Ã¢â¬Å¡ÃâÃÂ But it is one more complexity that I wouldn't worry about at this stage.
> > > > >
> > > > > The Servo Output is computed by the Servo algorithms.ÃÆ'Ã¢â¬Å¡ÃâÃÂ See:
> > > > >
> > > > > http://dynomotion.com/Help/ServoFlowDiagram.htm
> > > > >
> > > > > I think you have a basic misunderstanding.ÃÆ'Ã¢â¬Å¡ÃâÃÂ A Servo doesn't need any current or power to hold position.ÃÆ'Ã¢â¬Å¡ÃâÃÂ As long as there is no error in where it should be it does nothing.ÃÆ'Ã¢â¬Å¡ÃâÃÂ Outputs Zero.ÃÆ'Ã¢â¬Å¡ÃâÃÂ Only if there is an error in where it should be does it try to correct the error by commanding an Output.
> > > > >
> > > > >
> > > > > HTH
> > > > > Regards
> > > > > TK
> > > > >
> > > >
> > >
> >
>

Group: DynoMotion	Message: 7280	From: Tom Kerekes	Date: 4/16/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

Hi Dean,

If you aren't able to attach messages upload the files to the files section. It doesn't work well to paste C programs into emails as the tabbing gets all messed up and makes it hard to read.

But I don't see any problem with the program so it may just be a matter of tuning. It sounds like things are wired/configured properly for negative feedback because when you reversed the DIR it ran away from the target which would be what is expected.

This type of system is inherently unstable with only P gain. But usually with the gain low enough it can be made stable if there is some friction. The output is Voltage, which must be integrated to get current, to generate torque, to create acceleration, that gets integrated to velocity, which finally gets integrated into position change to reduce the error. Lots of phase lag. By the time the error reaches zero the PWM finally shuts off (zero voltage), but there is still current generating acceleration so the motor is still accelerating so overshoot, oscillation, instability is pretty much guaranteed.

Try adding D gain. Maybe a value of 50 +/- a factor of 10. D Gain introduces phase lead and has a stabilizing effect to slow things down (until too much is added and then it becomes destabilizing).

HTH
Regards
TK

Group: DynoMotion	Message: 7286	From: deanw1a	Date: 4/16/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps

I started a new thread on cnczone because it is easier for me to read and add attachments. Basically, I tried adding D and it helped but there is still oscillation at stand-still that seems excessive. Increasing the P and Max Output (with D) still causes the oscillations at stand-still to become more violent.

If it is OK to continue on the zone, there is no need to reply to this message. If you would rather discuss it here, no problem.. just reply here and I will learn how to attach.

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> Hi Dean,
>
> If you aren't able to attach messages upload the files to the files section.Â It doesn't work well to paste C programs into emails as the tabbing gets all messed up and makes it hard to read.
>
> But I don't see any problem with the program so it may just be a matter of tuning.Â It sounds like things are wired/configured properly for negative feedback because when you reversed the DIR it ran away from the target which would be what is expected.
>
> This type of system is inherently unstable with only P gain. Â But usually with the gain low enough it can be made stable if there is some friction.Â The output is Voltage, which must be integrated to get current, to generate torque, to create acceleration, that gets integrated to velocity, which finally gets integrated into position change to reduce the error.Â Lots of phase lag.Â By the time the error reaches zero the PWM finally shuts off (zero voltage), but there is still current generating acceleration so the motor is still accelerating so overshoot, oscillation, instability is pretty much guaranteed.
>
> Try adding D gain.Â Maybe a value of 50 +/- a factor of 10.Â D Gain introduces phase lead and has a stabilizing effect to slow things down (until too much is added and then it becomes destabilizing).
>
> HTH
> Regards
> TK
>
>
>
>

Group: DynoMotion	Message: 7009	From: deanw1a	Date: 3/17/2013
Subject: Using KFlop to drive AMC BD30A8 PWM Servo Amps

Group: DynoMotion	Message: 7018	From: Tom Kerekes	Date: 3/18/2013
Subject: Re: Using KFlop to drive AMC BD30A8 PWM Servo Amps