Moving from VFD to Servo Based Spindle

[mckoz]

Question - hopefully Tom is still around...

I'm finally taking the step of converting the spindle from VFD to Servo based Spindle. The primary reason is to do synchronous tapping (some folks say rigid). Here is my current system setup:

CNC Program: KmotionCNC
Kanalog
Snapamp
Kflop

The Snapamp board drives 3 Aerotech 1135LT Servos for X/Y/Z
The spindle VFD is driven off the Kanalog board (SW1)
______________

So... I have a Moog AC Servo Motor with built in resolver connected to a T200 controller. The T200 will run in Step/Direction mode or Quadrature input, and based on the max RPM and max number of steps per rev, it appears I won't exceed .8 Mhz needed from the Kflop. Here are the calcs:

Stepper Frequncy Calculations
5800 Max RPM
96.66666667 RPS
8192 Max Steps Per Rev (Range is 128-8192)
0.791893333 Mhz
_________________________

I have a number of questions as I begin this change over:

1. Will I have to use Mach3 for the tapping routines, or is one available for kmotioncnc? I noticed that Thomas Zamirski wrote one for a VFD, but I'm hoping one exists without writing it.
2. This will be an example of electronic gearing I assume, but it is going to run in effectively open loop, is that normal?
3. To expand on the open loop question, the Moog T200 has a simulated encoder output that converts the resolver info to encoder, but I'd rather use the internal resolver to T200 control instead of outsourcing it to the Kflop - thoughts?
4. Are there advantages to Quadrature vs Step/Direction or vice versa?
5. Any other thoughts?

Cheers, David

Similar Threads:

• Servo module for UHU based drives now available.
• Moving table vs. moving spindle?
• Can't start spindle (servo based) at high speeds
• Pic based servo motor drive
• I need help on PC-based for servo control

[TomKerekes]

Hi David,

Will I have to use Mach3 for the tapping routines, or is one available for kmotioncnc? I noticed that Thomas Zamirski wrote one for a VFD, but I'm hoping one exists without writing it.

You can use Mach3 or KMotionCNC. In either case the Rigid tapping is performed in a C Program in KFLOP in the same way. The only difference is how the parameters are passed and how the program is invoked. We can help you with that.

2. This will be an example of electronic gearing I assume, but it is going to run in effectively open loop, is that normal?
3. To expand on the open loop question, the Moog T200 has a simulated encoder output that converts the resolver info to encoder, but I'd rather use the internal resolver to T200 control instead of outsourcing it to the Kflop - thoughts?

Its your choice in the C program whether you will have the Z axis try to follow where the Spindle is commanded to be or where it actually is. If the Spindle controller works well with small following error then the actual position and the commanded position will match and there will be little difference. However if the Spindle response is low with large following errors then Z should be made to follow the actual measured Spindle Position. I don't fully understand your statement as I believe the T200 will use the resolver regardless of whether or not you send the exported simulated encoder position to KFLOP or not. Note that if the encoder position is connected to KFLOP it isn't necessary to use it for closed loop control.

Are there advantages to Quadrature vs Step/Direction or vice versa?

Quadrature is the preferred method. I involves only one edge per step instead of a complete pulse (two edges). It avoids timing issues such as pulse width and direction setup time. Any noise glitch or edge ringing will usually result in a step forward and backward instead of a miss-step.

HTH
Regards

[mckoz]

Awesome, thanks Tom, long time no speak. I will assume the T200 can maintain the position accurately for now. The part that's scrambling my thoughts is the timing question in an open loop, but I guess as long as teh T200 can keep up with the Quadrature signal (relatively speaking), then I should be good to go. I would like to keep using kmotionCNC, so that's the direction I will go for now.

On a tangent, do you happen to have a post processor from Fusion 360 that takes advantage of KmotionCNC? I found an old one on the web, but I'm not sure it's truly optimized.

[TomKerekes]

On a tangent, do you happen to have a post processor from Fusion 360 that takes advantage of KmotionCNC? I found an old one on the web, but I'm not sure it's truly optimized.

Sorry I don't. If you happen to find or create one please list it in the wiki here:

Known KMotionCNC CAD/CAM Post Processors - Dynomotion

Regards

[mckoz]

I will do that.

Tom, would you mind taking a look at the attached page - it's the stepper/quadrature input pinout to the DB9 on the Moog T200. I would be curious which axis and which pins you would wire from on the kflop/kanalog board if you were doing it?

I also attached the full manual, just in case a rabbit hole opened up...

Cheers, David

[TomKerekes]

Hi David,

Figure 3.47 seems to indicate the drive only has differential inputs (the + and - polarities must reverse). There is no direct way to drive these from the KFLOP single ended outputs. You might add a 3rd party line driver module such as:

https://cnc4pc.com/differential-line-driver.html

Step/Dir (or Quadrature) Outputs 0-3 can be muxed to KFLOP JP4 and JP6. You might use those through a differential driver. See the table here:

Step and Direction Setup

HTH
Regards

[mckoz]

Understood, I will get a converter. I just reread your explanation above and had a lightbulb moment:

Its your choice in the C program whether you will have the Z axis try to follow where the Spindle is commanded to be or where it actually is. If the Spindle controller works well with small following error then the actual position and the commanded position will match and there will be little difference. However if the Spindle response is low with large following errors then Z should be made to follow the actual measured Spindle Position. I don't fully understand your statement as I believe the T200 will use the resolver regardless of whether or not you send the exported simulated encoder position to KFLOP or not. Note that if the encoder position is connected to KFLOP it isn't necessary to use it for closed loop control.

So I am definitely going to connect the encoder simulation back to the flop and try and have the z axis follow it. That brings me to my next question, how would you connect the encoder output to the flop/kanalog? The diagram is shown on page 3-80 of the manual, would you mind recommending a solution?

Cheers, David

[hanermo]

Em..
I dont know for sure, but most servo drives and stepper drives can work single ended by only triggering the + or the - side of the input.

At least the chinese mige ac servo drives I use work with pokeys, and the differential 2M542 drives work with pokeys (and with CSMIO differential).
You tie either the + or the minus side of the differential signal into the drive, and only step the other side.

I have tried 3 combos and all worked.
There are connection examples online.

Just pointing this out, hth.

[TomKerekes]

Hi David,

That brings me to my next question, how would you connect the encoder output to the flop/kanalog? The diagram is shown on page 3-80 of the manual, would you mind recommending a solution?

Kanalog has Differential Encoder Inputs on JP1. So simply connect the differential Encoder outputs A+ A- B+ B- from the drive directly to Kanalog JP1.

Hi hanermo,

I think you are correct most Drives have Opto inputs like this:



can be driven by a single ended signal. They need some voltage to turn on and a lower voltage will turn them off.

But a differential input like this:



needs to see a positive voltage to switch on and a negative voltage to switch off. A lower voltage or even zero volts will will not turn the input off. It could work if the - input was connected to a voltage like 1.2V. In that case with the + input switching from 0.4V to 2.4V the differential voltage would go positive and negative. I suppose a stiff resistor voltage divider could work.

HTH
Regards