One key difference between milling and plasma cutting is the feedrate (meters per minute etc). When I ran a trial toolpath in the Bridgeport, I found the current default feedrates for G0 (rapid positioning) and G1/G2/G3 (machining feed rates) were quite different, and more relevantly the G1/G2/G3 feed rate was glacial, and certainly lower than those required for typical plasma cutting.
This g-code is a simple way to move one direction in rapid G0, followed by a return at G1 feedrate.
G0 X0 Y0G1 X10 Y10G0 X0 Y0G1 X10 Y10
The default, maximum feedrates are set in the Newker Parameter > Speeds. Obvs if you ask for a higher feedrate in the g-code, it can't overwrite those hard coded parameters.
I finessed the G0 and G1 etc parameters to increase the feedrates up towards something a bit more useful. Problem was that beyond a certain combination of speed and (reasonably workable) acceleration/deceleration, one or both of the servo drives would shut down.
I'm using a Leadshine RPS4810 PSU which has fairly small output caps (being an SMPS), so it doesn't take much deceleration (aka regeneration) to cause an overvoltage of the servo drive. Like most PSUs, these have latching overvoltage protection (OVP), set between 71 and 79V:
There's no mention of an overvoltage shutdown the the DMM Tech DYN2 servo drive user manual (such a it is) but it's clearly happening, since on occasion the Y axis will stop moving while the X axis continues to respond. If the PSU was shutting down, neither axis would move.
Previously, I observed this OVP shutdown and was able to confirm via the DMM software that it was being triggered by the drives. Obvs, when an overvoltage occurs, the drive stops driving the motor, and as these motors are running above base speed ie not in the field weakened region, the overvoltage is curtailed and never rises far enough to trigger the PSU OVP or cause damage.
One solution to this issue might be to fit a larger bus cap but from previous experience I know that too big a cap can cause overcurrent protection (OIP) shutdown. On these PSUs, an overcurrent causes a latching shutdown that can only be cleared by interrupting their mains input.
Unless I'm to dial the feedrates down to a pathetically slow level, the only sensible solution is to fit a braking resistor or voltage clamp to contain the overvoltage by dissipating it in a resistor or power device.
The classical scheme for implementing a braking function is to use a comparator with hysteresis to switch a resistor on and off via an IGBT or MOSFET. It's a pretty simple circuit and is used within VFDs and also as standalone braking modules.
The cncdrive.com servo drives I used on The Shiz were ordered with one of their matching braking modules. Due to the rather sketchy construction, it managed to break down the microscopic insulation between the switching device and ground, nearly causing me to "make mud". I repaired it, reverse engineering the circuit along the way. Bizarrely, it included no hysteresis although it seems to work ok.
As for the DMM Tech servos used on the Bridgeport, I previously considered building a braking module to tackle precisely the issues I'm seeing today. Clearly lethargy, apathy or similar caused this to not happen but perhaps now's the time to get it implemented.
TBPH, I still can't be arsed to go the whole hog with that proposed design. Being a lazy fat bastard, I'm always looking for easy shortcuts. In this instance, such a shortcut would be to simply make a giant zener diode ie a linear clamp, rather than a switching circuit.
I've got some big fuck off IGBT modules (IRGTI050U06) from Internation Rectumfrier in my museum of components that will eventually turn to dust if I don't use them.In conjunction with a simple zener and some resistors, I can implement a very simple zener clamp. I can't be arsed to calculate the energy I'd need to dissipate, either from first principles or based on scope measurements. I'll simply build the fucker and connect it up. If it pops, I may think again - or replace the module and limit the peak current by fitting a source resistor.
As you can see from the schematic, the "zener" is a TL431, which allows the threshold to be adjusted, ideally with a pot. There's so little involved that I can just solder the compts on top of the module.
Let's build this circuit up and adjust the clamp voltage (using a bench PSU) somewhere between the Leadshine PSU's set voltage and the servo's OVP threshold. Then I can turn up the feedrates without the damned thing needing the maains to be recycled.
No comments:
Post a Comment