Centroid Community CNC Support Forum

Marty,

(I'll slip on my "teacher's hat", not for you, but for those who are just starting the DIY adventure.)

Yes, I've seen that shark-tooth pattern on the 'scope. At the data rates used to control a servo motor, the signal resembles an AC signal more than it resembles a DC signal. A capacitor can "pass" AC type waveforms, but it blocks DC. That's why changing DC signals have to be seen as if they were AC signals. Capacitance (and resistance) slows down a signal's transition. The more capacitance, the less "true" the waveform is to the original signal.

Let's use the example of a linear power supply that uses a large capacitor to smooth out the voltage. Without a capacitor, the voltage level coming out of a linear power supply would simply be a ratio of the voltage going into the power supply's transformer. Usually a full-wave bridge-rectifier doubles the frequency from 60hz to 120hz and cuts the voltage level in half. (The peak-to-peak voltage is actually (SQRT(2) * 120VAC * 2 or about 340V peak-to-peak) at 60hz.) With a bridge-rectifier, the peak-to-peak voltage is about 170V at 120hz. The large capacitor "stores" energy between the 120hz waves. The higher the capacitance, the less drain between cycles and the higher the average voltage. Geckodrive has an excellent white paper that gives a formula for calculating the size of capacitor needed to give a relatively smooth DC voltage to drive stepper motors.

A switching power supply often has circuitry to increase the frequency to high levels. With high frequency, the periods between charges is reduced. That type of power supply is usually much more efficient when the current required is constant. The drawback is the high-pitch squeal that poorly designed switching power supplies emit.

So, when the wire in the cabling acts as a capacitor, the wave-form is distorted. It's just a law of physics that we have to compensate for when using step and direction pulses. A pull-up resistor helps to "snap" the voltage up to the 24VDC "rail". It helps turn the "shark-tooth pattern" back into a square wave pattern. At 200,000hz or 400,000hz, there are a lot of forces pushing and pulling on the waveform.
----
I've checked my PM mailbox, but I didn't see any new messages.

All-
I’m an engineer at Teknic and came across your comment about step input timing errors and noise interference.

The ClearPath motor inputs are optically isolated and current limited. This design provides exceptional noise immunity even in the most difficult EMI environments like plasma CNC machines and applications that use a VFD. The optical isolation hardware of the ClearPath motors provide enough noise immunity that shielded and twisted pair cables between the CNC motion controller and the ClearPath motor are not required.

You might wonder, “if the noise immunity of the ClearPath motors is so good, then why is there a possibility of getting step input timing errors and why are the termination resistors required?” Great question!

The minimum length of time for a step to be “on” is 715ns. In addition, the minimum length of time for the “off” period between steps is also 715ns. The ClearPath “step input timing error” occurs when the “on” time of a step is shorter than 715ns, and/or the “off” time between steps is shorter than 715ns.

Typically this error is elicited when one or more of the following conditions occur:

- The controller has open collector outputs and the resistors specified in the ClearPath manual are not installed correctly, or not at all. Here is a link to the fractional hp ClearPath manual:(https://www.teknic.com/files/downloads/ ... manual.pdf)
An open collector output “drives” the signal to a true logic state but then the return signal “floats” back to its false logic state. This floating nature of the open collector signal means that the signal might not be “off” for at least 715ns required between steps before the next step is sent. This behavior is often exacerbated by capacitance in the control system, whether due to long cable runs or the electrical design.

The Centroid Acorn uses 24VDC open collector outputs, so installing 1k ohm resistors across both the step and direction outputs for each axis of motion is appropriate. This resistor promptly drives the signal to an “off” state between steps. In contrast, a driven TTL signal (this is a different hardware architecture than open collector signals) is less prone to this behavior because the circuitry “drives” the signal to ”on” and then “drives” the signal back to ”off”. This type of signal is referred to as a “driven signal”.

- The controller output voltage is lower than the minimum 4VDC required for the ClearPath motor and is unable to reliably turn on the optical isolated inputs. For example, some controllers only output a 3.3VDC signal.

- The controller's maximum step pulse output frequency is higher than 700kHz, or the step pulse time is less than 715 nS. These settings are typically configurable in the controller.
(https://www.teknic.com/files/downloads/ ... om=100,0,0)

- Shared return paths for outputs. The V+ and GND wires for each controller output should travel the entire distance from the controller to the motor and then back to the controller. Some systems have limited I/O connections so the output signals go individually out to the motor, but then they share a common return path back to the controller. For example, three output signal wires could go out to the motor but then the return wires are tied together and only one return wire comes all the way back to the controller. This can cause problems related to noise immunity.

As a side note, we’ve recently reviewed the Acorn schematics with Centroid and they will be updating their schematics to include this 10k termination resistor. This should address the issue that you mentioned.

I hope this helps you and other users. If you have any other questions regarding the ClearPath motors, please feel free to use Teknic’s "Contact Us' form (https://www.teknic.com/contact) or give us a call at 585-784-7454.

Best regards,
Tom T. - Teknic Servo Systems Engineer

Tom T...
What is the threshold of step timing errors that will trigger an HLFB state change?

Tom T.
Thanks for the informative explanation!

Thanks for the feedback MarkRH!

Gary,
I am not exactly sure what you mean by threshold but I can tell you that if your HLFB is set to the default "Servo On", it will go from a true logic state when the motor is enabled to a false logic state with a motor shutdown with the first step timing violation.
Tom T.

Tom, I think he was asking how much difference from 50/50 does the error kick in at? 60/40? 70/30 duty cycle?

tblough,
The actual duty cycle doesn't matter as much as meeting the 715ns minimum for "on" and "off".

MarkRH wrote: ↑Fri Feb 19, 2021 9:01 am Tom T.
Thanks for the informative explanation!

Mark,

Have you resolved the issue or are you still having problems?

You might consider looking into emi issues in the panel effecting Acorn.

Tom T., I asked this in Marty's build thread as well, but is Teknic now saying that those of us who bought SDSK motors a while ago and are using 1K resistors now should switch them to 10K for proper performance?
Tom

tkbot47 wrote: ↑Tue Feb 23, 2021 8:34 am Tom T., I asked this in Marty's build thread as well, but is Teknic now saying that those of us who bought SDSK motors a while ago and are using 1K resistors now should switch them to 10K for proper performance?
Tom

Tom answered this. It's not a matter of performance, it's more a matter of heat generated by the 1k ohm resistor. Check your resistors, change them out if heat is an issue. The performance should be the same.

Marty