Keep watching the cncmentor.com blog for adventures with this beast!

Overview and Basic Specs

The Granite Devices VSD  comes in two flavors, the VSD-E (continuous output of 10amps, $285USD) and the VSD-XE (continuous output of 14 amps, $375USD).  They are visually distinguishable from the size and shape of the heatsink.  The VSD-XE (pictured below) has the larger heatsink and as such is able to deliver a higher current rating. For extreme duty, the drive can be easily and effectively fan cooled. Apart from this difference the drives are identical.  For that extra current handling capacity you pay about 20 percent more.  The VSD-* is well engineered to accommodates a wide range of motors from 5 to 1500W, thanks to what the manufacturer calls a high dynamic range torque control (HDRT). Two independent current sensing ranges provide ultra low noise and high torque resolution, the most critical tool in position and velocity controls.

The VSD-E features industry standard pulse/direction (ie. Mach 3), quadrature, PWM and analog control inputs in addition to a device specific SPI (Serial Peripheral Interface) protocol (impressed yet?).  Drive supports encoder & hall sensors feedback.  The VSD-E can also be easily controlled from PC or LabVIEW with SimpleMotion library and also connection to fieldbuses like EtherCAT or CANopen (more info).  All these features plus the ability for this drive to run stepper motors, brushed or brushless servo motors in AC or DC flavors truly gives it the right to be called the Swiss army knife drive!

Here is a full list of features from the product website:

• Easy setup via GDtool software
• SimpleMotion control library for PC, C/C++/C#, LabVIEW etc
• Various low level command inputs (step/dir, PWM, quadrature, ±10V)
• Easy connection to fieldbuses like EtherCAT or CANopen (more info)
• Most AC/BLDC/DC and linear servo motors supported
• Advanced sinusoidal field oriented control (FOC)
• True current control protects equipment and boosts responsiveness
• 12-160 VDC supply voltage
• 10A cont. 20A peak output for AC/BLDC motors
• 14A cont. 40A peak output for brush DC motors
• Reliability guaranteed by wide safety margins and quality components
  

• Electrical & mechanical braking of motor in faults & disable
• Fully automatic internal homing controller
• Optimized for CNC and machine automation
• Robust 200V 60A MOSFET power stage
• Panel installation, 19″ rack compatible. See example.
• Regenerative brake resistor output
• Hard to kill: short circuit, over current, over temperature, I²t, over voltage and under voltage protections
• Advanced positioning with acceleration limit and smoothing
• Wide selection of control I/O’s, see Connectivity.

Now, how does all this pan out in real life?  Do all these features warrant the cost?  $375 for a single drive is expensive by comparison to many other drives out on the market (i.e Gecko G320X at $121 per drive,  The Rutex R2030 at $198 per drive, Viper 200 at $239 per drive, or the cncdrive DG2S-16035 for $239 per drive).  Here are some photos of our unboxing experience.  Basically you get the satisfyingly hefty driver and a pile of wires, connectors, and adapters.

Additionally, the VSD-E/XE drives and VSDEPI can be easily installed to front/back panel of enclosure. Standard 19″ rack compatible stainless steel panels are available for an additional cost but may be very handy for some use cases.

There is signficant “substance” to these drives and they feel solid, massive and well made.  You don’t hold the drive and ask the question, “Where is 20amps of current going to go?”  You don’t ask the question because you know it’s going into that massive quality heatsink!

Wiring and Setup

Build quality aside, we run into our single massive complaint about the driver – WIRING AND CONNECTORS!  Pictured below is the antiquated DB25 and a DB15 ports reminiscent of the legacy printer port and gameports on computers.  Why the designers chose to these ports for their intended purposes is baffling to us?  The DB25 as the name implies is a 25 pin port but in the Granite drive is intended to be used as the power output for the drives.

Since the port was never designed to handle that kind of current capacity, the user is expected to manufacture a special DB25 cable that fuses 6 wires into one essentially turning 25 pins into 4.  Why not just put 4 terminal blocks on the board?  The amount of effort it took us to solder up a cable, the amount of effort it took to keep track of all those wires, and the inevitable troubleshooting that took place in trying to ensure we didn’t make any mistakes made us pull out our hair.  Again, why not just put four screw down terminal blocks.  It’s literally a difference of 60 seconds verses 60 minutes.  We certainly hope that future models improve upon this poor design choice.  Here are some photos of our wiring it all up.

After we got the wiring hooked up, we setup a basic test in Dual DC mode (two servo motors connected to the single drive).

Software and Configuration

The driver ships from the factory in single drive mode.  If the user wants to play around with dualDC mode (driving two motors) then they need to flash the firmware with the provided firmware tool.  It’s a relatively painless process.

To communicate with the driver the user must connect a USB cable to the board.  Like many devices, the Granite drive is not a true USB device but uses an RS232 to USB converter made by FTDI. A driver is installed in the computer that created a ‘virtual’ COM port and communicates over the USB to special circuitry in the target device. Here is an overview of some of the software configuration screens to give you an idea of the flexibility and quality of the tool.

DualDC can be configured to drive two motors from single step/dir source making it ideal for paralleled or slave axis.  DualDC features standard step/direction input for each motor. Step multiplier is user adjustable between 1 and 20×. Input smoothing filter is applied if mutiplier is 2 or more. DualDC drive can be easily connected no nearly any fieldbus including EtherCAT, CANopen and PROFIBUS.

Performance

We were very pleased with the performance of the VSD-XE.  VERY!  The Granite VSD-XE has completely smooth running and silent holding performance, even with large step multipliers. The drive features an antidither feature which aims to silence motor dithering & hunting noise when standing still.  And in our experience this is exactly what it did.  The Granite VSD has a selectable ‘input smoothing’ feature which makes the drive run with very large step multipliers as if it were running with none. This is an extremely useful feature specifically for MACH3 because it effectively eliminates the need for the smoothstepper or equivalent products in order to get the step frequency high enough to run the motors at full speed with high res. encoders. The drive is just as smooth with a 12x multiplier as with 1x even at low speeds. Additionally the Granite has an interesting configurable ‘holding zone’ feature that kicks into use whenever the motor reaches that zone.  The parameters of that zone are user configurable. The zone has its own PID settings separate from the main PID tuning. Unfortunately, like many other drivers we have tested the Granite cannot be tuned while connected to the CNC control.

 Conclusions

Overall, the Granite XE was one of the most time consuming and complicated installation of all of the drives we have tested. To some degree this is unavoidable because the Granite product can control so many different types of motors and has so many features that the available choices in connections, set ups, and tuning parameters are many times more than any of the single purpose drives.  One of the refreshing highlights of this company is the obvious attention to detail in documentation and legitimate support.  All our questions were promptly answered with good and helpful replies.  Some initial problems in our setup were quickly ironed out with help from tech support. The online manuals are detailed and helpful.  As already mentioned, the software is second to none both in terms of form and function.  This is a good “window” into the attention to detail and the overall flexibility of the product.  We highly recommend this drive and give it our highest review despite our irritation with the unnecessarily complicated wiring and a few other minor quibbles.

If you are designing a system from the ground up stop reading and buy one right now.   We get no kickback from your purchase.  We just believe in the product and the concept.   There is simply no better way to ensure reliability in your system that the SmoothStepper .  The smooth stepper will make your system ROCK solid and that is worth every penny.  In our opinion it is just silliness to invest the kind of time and money and risk loosing your work piece (at best) or ruining your tooling or machine because information is not being received properly by the motor drivers.  If you are trying to troubleshoot an intermittent problem, this may be your solution (continue reading to find out how it solved our problem).

How it works

Here is a basic overview of how the smooth stepper works and why you want one.  In a normal mach3 CNC setup your configuration looks something like this.

Note that in the diagram Mach3 is running on the PC.  Currently the SmoothStepper can only be used with Mach3 although an SDK is being developed which will enable it to be used on many other systems.  In a normal setup, Mach3 is both the controller which is responsible for taking the Gcode and converting that to step and direction pulses and then sending out a timed pulse train through the DB25 to the motor controllers.  There are severe hardware limitations to this type of configurations.

Perhaps the greatest drawback to a conventional printer port/breakout board setup is the fact that the computer functions as the pulse train generator.  Overcoming this obstacle was the main motivation behind the developing of the smooth stepper.  The reason this is such a problem is that the PC is busy doing so many things that to develop a smooth, fast, reliable pulse train is very challenging.  Code has to be written at a very low level to give it priority over other tasks and even then, mistakes happen.  In my experience, I have had the PC lock up and run away, missing steps (even in closed loop systems), and extreme slowing during heavy CPU crunching.  This again, is not the fault of the Mach3 designers.  What they have done is in fact amazing.  It is a hardware limitation.  The SmoothStepper aims to solve this problem by creating a dedicated piece of hardware that does one thing:  receive motion commands from the Mach3 controller and convert it to a smooth step and direction pulse train.  Here is a diagram of a typical SmoothStepper setup.

Notice that the SmoothStepper board in this diagram functions as a hardware datacruncher between the CPU and the motor control drivers.  In real life there is no such thing as doing two things at once.  It is simply impossible to create a 100 percent reliable pulse train AND run all the other Windows background tasks necessary to keep a computer happy.  Because of the complexity of computers and the hundreds of things that fight for the attention of the processor every single second there will eventually be a conflict of interests.  The SmoothStepper solves this problem by receiving movement instructions from the motion planner in Mach3 (illustrated as Gcode in the diagram but technically incorrect) and produces step and direction pulses for motor controllers . In addition, the SmoothStepper provides spindle control and can handle all the other Input/Output (I/O) functions that are normally handled by the two printer ports under Mach3. By taking over the task of producing step and direction signals, the SmoothStepper relieves the host computer of most of the CPU load normally involved in running Mach3. The dedicated hardware in the SmoothStepper produces pulse streams up to 80 times faster, and with more precise timing than Mach3 using the printer port.  For a detailed look into why motion control hardware is important see this article.

Amazingly the single Ethernet connection provides complete control of what would be the equivalent of 3 full parallel ports.  Practically this gives you as many inputs and outputs as you could conceivably use, saves you the money of buying additional parallel port cards and cleans up your wiring nightmare with a single cat5 cable running from the PC.  Win-win-win.  If that’s not enough the noise immunity via Ethernet is excellent and is especially helpful in noisy environments like plasma cutting.  The three parallel ports can be configured as follows:

Output/Input  Configurations:

• Port 1: 12 Outputs and 5 Inputs
• Port 2: 12 Outputs and 5 Inputs OR   4 Outputs and 13 Inputs
• Port 3: 12 Outputs and 5 Inputs OR   4 Outputs and 13 Inputs
• Full Spindle Control (PWM/Step-Dir)

 Testing and Setup

As stated earlier we tested an ethernet version of the device which differs rather significantly from its USB cousin.  Besides using a different communication port the Ethernet version adds additional input/output functionality.  Before we can really get into our setup and testing experience we need to first give a little history.  Our machine was working well “most of the time.” But for serious production work, “most of the time” may as well be totally broken.  Fixturing a part only to have the machine stall or loose steps 45 minutes into an operation is expensive, frustrating and a huge waste of time.  But what was causing the problem?  We started trying to isolate the problem

• Perhaps it was mechanical.  Perhaps the stepper motors were stalling because they were running too quickly or some part of the screw was binding.  Speeds were reduced leaving ridiculous safety margins and all mechanical components were checked for smoothness.  We still experienced intermittent failure.
• Perhaps the motors were overheating.  Current was reduced and we even talked about installing fans on motors.  Still the problem remained.  It would work flawlessly for 3-4 jobs in a row and then unexpectedly at the most inopportune time it would stutter, loose steps and the work was ruined.
• Perhaps it was the pulse train coming from the PC?  Time to try a motion control board.  We had eliminated everything else.

We tried to hook up our oscilloscope to the step and direction pins and “capture” the failure.  That would have been quite simply AWESOME.  This would have been undeniable evidence that the PC was in fact giving us a garbage signal.  As it turned out we were not able to be in the right place at the right time to do this, but we still feel very confident that had we been there, we would have been able to capture a disrupted pulse train.

So we ordered up our board and crossed our fingers.  Our initial impressions of the board were very positive.  The board had quality silkscreens, plenty of diagnostic LED’s and just had the look and feel of a well-designed, thoroughly thought out piece of hardware. Connecting the board was stunningly simple.  Simply disconnect the parallel port, use the provided parallel to 26pin header input to “splice” in the SmoothStepper  and then provide the motion control board with 5v power and that is literally it.  No more wiring.  Our complete wiring setup took 5 minutes (and we took our time).

Next comes the software setup.  First the Mach3 plugin needs to be installed.  This is simply a matter of downloading the file and running through the installer.  After that the computer’s Ethernet port needs to be configured with a static IP address.  Usually Ethernet ports are set up for DHCP, which allows a cable or DSL router to assign a unique address to a device that is requesting an address for the local network. Once we configured an ip address per the instructions here (which btw were very hard to find) we were able to communicate with the device.  Upon opening Mach 3 (with the new plugin installed) the user will notice a new dropdown menu revealing some new options.

All the default settings of the SmoothStepper worked without any modification.  That being said, is is very modifiable!  Check out the configuration screen.

Because of the nice way that Mach3 and smoothstepper play together, there is no need to change a single setting in the motor setup or the ports and pins dialog of Mach3.  Also of note, the software does have a built in device monitoring tool which is very cool.

Here’s a picture of our crude device testing setup.

Perks and Quirks

We loved almost everything about the board but we did have a few complaints.   Our biggest complaint was documentation.  Admittedly it is way more fun, designing the hardware than it is documenting it for everyone else; however, if this product is to mature and establish a solid presence in the CNC community more documentation is needed.  We will say that what documentation does exist is extremely well done and helpful.  It was just obvious that the developer kind of “ran out of steam” and needs to just sit down and pound it out.  The website was a little scant.  An overview of what the product does and how it works would be nice.  Also some more pictures of the product and usage videos would probably attract more interested customers.

Overall we were extremely pleased and would HIGHLY recommend this to anyone building a CNC machine.  We personally will never build another machine without one.

Bask in the photo awesomeness!

1. User experience.  Is it hard to get to the features?  Diving into menus for simple operations is not nice.  Is the end user’s actual usage in view?
2. Intuitiveness.  I don’t like reading manuals so please don’t make it harder than it has to be.
3. Unique features that set this software apart from other pieces of software in its class.  Why should I buy this product over the dozens of other products out there that do the same thing?

In other words, the basics of any CAM software package are expect so the majority of the content in these reviews is intended to give the end user an idea of how this software is tailored.

To understand where MeshCAM fits in the lineup of other CAM softwares on the market, I’ll refer you to a survey by Bob at cnccookbook.com.   Hundreds of visitors participated in the survey.  The survey was divided into “pro” vs “hobby” which is a distinction based solely on the pricing of the most inexpensive module in a CAM company’s lineup.  For the purpose of these reviews we will try to compare “apples to apples” and not compare hobby class software to professional class.   Here are the results of his survey:

As of December 2010 MeshCam had a 4% market share among the hobby CAM usage market.  CAM software is notoriously expensive and MeshCam is very well priced at a competitive $249.  That is very inexpensive by comparison to some other CAM packages out there.  Here are some of the entry level packages of the competition as of 2012: Sheetcam is $175, CamBam $149, MeshCAM Art $500, Vetric 2D Pro $149, MadCam $750,  VisualMill $1250, SprutCAM, $1250,  DolphinCAM $399,  BobCADCAM $2369,  MillWrite $400, TurboCADCAM $299, Heeks CAD/CAM $0

Let’s dive right in.  Upon opening the software, the user is presented with an X-Y Axis window with two large arrows (overly large in my opinion) and a user interface pane to the left.  Immediately, I felt right at home.

My first move in any CAM package is to import a file that is familiar to me.  In this case we are going to import and machine a rosette.  Pulling down the file menu, we are presented with two options:  Open and Open 2D DXF.  In my opinion this is a bit confusing and should be combined into a single menu item (i.e. Import CAD File) since there is no functional reason why they could not be meshed (pardon the pun) together  Upon opening the file the user is presented with a dialog asking for input concerning the machining  job type:  3 Axis machining, 2 Sided machining and 4 Axis.

Again, labeling here is not as clear as it could be but the pictorial aids cleared things up.  Consistent label names would have been 3 axis-single sided , 3 axis-double sided and 4 axis-indexedbut I fear my love of linguistical parallelism is coloring my perspective…;-).  One of the things that sets this software apart from others in its class is the 2-Side machining.   2 Sided machining allows users to machine both sides of a part in two separate operations giving your machine much more versatility.  This is an exciting feature with a lot of possibilities.  Perhaps the most valuable feature of this entire CAM package is the 4 Axis capabilities.  Many CAM packages have 4 Axis capability but few can rival the pricepoint of MeshCam.  It should be noted that the 4 Axis capabilities are indexed and not continuous.   Longingly absent in this list is 2 1/2 D machining.  Unfortunately, at this time, MeshCAM is not setup for pocketing and profiling, but what it is designed to do, it does very well.  For the purpose of our demonstration, I chose 3 Axis.  The file loaded beautifully.

The user can easily pan and tilt with intuitive mouse movements.  The “View->Above” and “View->Fit to Window” were essential shortcuts for reorienting me when I got lost.  Now comes the fun part – defining the toolpath.  There were several toolpath options to choose from.  The toolpath options are shown in tan.

• Define Stock:  This allows the user to input dimensions of the raw stock.  The “fit to geometry” option is helpful.  Here’s a screenshot.
• Define Supports:  This a cool feature that allows the user to add supports in strategic areas to prevent the part from tearing out once it is cut loose.  Since it is user controlled, as many or as few supports as needed can be defined as well as in specific locations.  This is unique feature and well done.
• Retract Height:  This is also known as the “Safe Z.”
• Program Zero:  Where is zero in relation to your stock?  Define that here.  Here’s a Screenshot.
• Set Machining Region:  How this differs from defining stock was not clear to me.  The only help the manual give us is the terse explanation:  ”Allows the user to define a region to machine”
• Generate Toolpath:   The defaults of every other toolpath option are typically sufficient for most use-cases; however, every machining operation requires input from the “toolpath parapeters” dialog box.  Clicking here will open a dialog that will bring you to the nuts and bolts of the program.

Here we have all the standard 3D profiling options one would expect in a CAM software package:  roughing and finishing parameters, tool selection, feed rates, machining strategy,  etc.  Everything here was very self-explanatory, the graphics were very helpful and well done.  We posted a file and ran it through our machine.   Success first time!  You’ll have to pardon the shoddy wood selection (particle board).

Here’s a video of us running our part: note the smooth Z movements.

Just for kicks we opened a file more suitable for 4 Axis machining to give users an idea of the capabilities.  In this case we chose to index the model 15 times as is indicated by the yellow splines on the faceplate.

MeshCAM Art.

The base package of MeshCAM comes with the basic 3D machining operations outlined above.  MeshCAM Art takes it to another level by allowing users to convert images to machinable art.  Everything in MeshCAM Art is image-based. You can create the base art for MeshCAM Art in any paint or vector program and import it directly into MeshCAM Art. After that, it’s easy to apply shape directly to your images and bring them to life

Since MeshCAM Art is built on the standard version you’ll have all of the machining capability you need right there- no extra programs to generate a toolpath. You provide the initial artwork and within minutes you can have a 3D object and toolpath ready to machine without leaving MeshCAM Art.

We were not given a MeshCAM art license so we could not test the full functionality but we did test the smoothing feature which is available in the standard version to get an idea of what it could do.  We started with this photo.

Once we loaded and selected a smoothness setting it magically converts our black and white image into a 3D relief.

Very cool!  Let’s run the file!  Here is a time lapse video of what the process.  A couple of things we noticed.  Unlike the 3D machining operations, the Z axis plunges were not very smooth and as such took much longer to machine even though the feedrate remained constant.  It seems that the algorithm is quite different and less efficient.  Also notice how the removal of material is much more random (i.e. it bounces around from region to region rather than predictably stepping over across the surface at a  fixed interval) and as such wastes time in rapid travel moves.

Perks:

One of my favorite parts of this program is actually a quite, low-profile corner of the software.  Under the Geometry Menu we find an option to “Slice Geometry.”   Normally this feature is used to slice model into pieces that won’t fit under the Z axis.  But, this has all sorts of cool potential to build up a 3D object in layers much like the additive process of stereolithography.  Objects like these could be made very easily with the output from this simple operation.  Kudos to the designers for adding this feature!

Quirks:

There were a few things we simply could not figure out how to do.  We tried for a long time to turn on the grid without success.  We wanted a ruler or some sort of way to get scale and calibration but could not find one.  The manual was of little help.  A search for “grid” yielded zero results.  As mentioned already the naming and organization of menus we felt could be improved.  Another thing to mention here is that the program is not very customizable which could either be by design to keep things simple or an indicator of a relatively young product.  We prefer knobs to tune things.

Conclusion:

MeshCam is a great program for anyone looking for a smooth introduction to 3D CAM software at an affordable price.  The ArtCAM features whereby a user could convert graphics into machinable products were especially cool as were the geometry slicing tools.   By far the best value comes from the 4th axis machining capabilities (indexed not continuous).  Few software packages can offer 4 axis machining for this price.  We were disappointing with the documentation of the product.  The manual was skeletal and left us wondering most of the times we consulted it.   The one video on the web was very helpful.  We would like to see more.   The location, grouping  and especially labeling of menu items could be improved but once the user gets familiar with them, this issue dissolves.    We are always glad to see a scripting language built in and MeshCAM went the extra mile to incorporate this.  It  uses the Lua language, www.lua.org, with additional commands added for MeshCAM-specific functionality.  The program was very quick and responsive and never crashed or hung.  It was not “buggy” in any way.  The recent memory management enhancements in the latest revision of the software were noticed and appreciated.  The lack of 2-1/2 D support would represent a significant handicap for some use-cases but if 3D profiling is your bread and butter, MeshCAM posts a solid performance for the price.  Meshcam offers a free demo so download a copy and give it a go yourself!  Let us know what YOU think.

Final Ratings on a scale of 1-10 (Ten being best):

• Value -9
• Feature Rich - 4
• Intuitiveness - 8
• Documentation - 3

Feel free to leave feedback on your experience with this software!  Thanks.

These plans in particular are designed for the entry level CNC enthusiast with care taken to use non-specialized parts “off-the-shelf.”  With some basic building skills, the step-by-step instructions will take you from raw materials to a working CNC machine!