I have a question about spindle speed. When turning on the spindle thru gcode is the number displayed on the VFD the actual spindle speed, or is it a number calculated based on the voltage read from my CNC controller. After I got everything set up and working I adjusted the upper (P2.0.15) & lower (P2.0.13) parameters to match the actual voltages I was getting & tweaked the potentiometer in the PWM but the speeds were still miles apart. So I kept adjusting the voltage parameters and adjusted the min speed in Gsender ($31) and I got the speeds to match very closely. But because the voltage settings in the VFD aren’t my actual voltages I don’t know if the displayed numbers are accurate. Is the displayed number the actual spindle speed?
Thanks for any info.
Hello, The VFD display, when in “run” mode, will display two things… either the calculated RPMs or the Hz. The Hz is the actual measured value and that’s sent through a calculation taking a few VFD settings into account to derive the RPMs the motor is running at.
PWM coming out of hobby CNC controllers is notoriously unreliable. This can be especially true when a digital-to-analog converter is required as in your case.
Tweaking the settings mentioned and the potentiometer adjustments are exactly the right method to get this to fix.
After doing so the vfd should be far more accurate to the desired RPMs.
Know that higher speeds will be more accurate than lower ones. Anything below 8000rpms is pretty much unusable speeds for standard woodworking spindles whose usable range is typically between 8000 and 24000 rpms.
Thanks for the info. Makes me feel better knowing that what’s displayed on the VFD is closer to what gsender says it should be.
Hi Daniel,
I would like to get clarification on your term “PWM coming out of hobby CNC controllers is notoriously unreliable.” If this is a known factor of a PWN, a spindle added to a hobby CNC I think the purchasers would need additional documentation to inform them of this incompatibility. Can you please provide a resolution to get this discrepancy to a consistent accuracy. I have had this same problem from the start. As you know inaccuracy in speeds, feeds and RPM’s will result in broken bits, projects and machines. The ramifications of this continuing problem is multiplying. I understand the benefits of running a spindle far out way a trim router. Plug and play installation has turned into tweaking and crossing fingers. If there is no resolution for the digital-to-analog converter discrepancy, please just rip off the bandaid and let us know the truth. Just looking to improve the spindle experience.
Thank you for your time and consideration on this topic.
Thank you for asking… just because a question is “hard” doesn’t mean it doesn’t deserve an answer with the best knowledge we have available at the time. I appreciate you asking for details and I’ll share what I can with the knowledge and understanding that I have. If I’m off somewhere, please clarify so I can correct my understanding and better advise customers.
As I was testing the dozen or so different hobby cnc controllers we own, the PWM signal turned out to not be as accurate as I was expecting it to be. Depending on where the voltage “comes from” or rather how it’s routed on the hobby controller circuitry, it comes out either weaker or stronger.
Most controllers I have are 0-10v PWM so I’m expecting a perfect 5.0v signal for the spindle to run at exactly 12000rpms. The VFD appears to be very accurate; if i give it a perfect 5.0v signal using a simple analog signal generator the VFD does output the expected 200Hz or 12000rpms.
The Arduino based controllers seem to be the worst offenders related to this “problem”. Ex: the arduino running GRBL on the older Mill Right Mega-V controllers. This 0-5v PWM controller will not output a perfect 2.5v signal to achieve the 12,000rpms. Instead it’ll give a 2.2v or 2.8v, thereby resulting in a a few hundred, up to a thousand, RPM difference from the actual desired speed setting.
The LongBoards (and SLB) are unique in that, they are the only controllers we own, it outputs a digital PWM signal. Unfortunately PwnCNC does not have circuit design capabilities so we must rely on off-the-shelf solutions to offer a digital to analog converter. We found the most common, tracked down the supplier, and wrapped it in an enclosure to make it easily plug-n-play.
We take each converter and hook it up to our own LongBoard and tweak it to run perfectly at 18k RPMs. This is done by adjusting the potentiometer inside the converter enclosure. Our LongBoard tells the spindle to run M3S18000, and with everything plugged up we verify the VFD is displaying the expected 18000rpms. Adjusting the potentiometer on the digital converter allows us to sync them. If we change our command to M3S12000, we’ll see it does NOT run exactly 12000rpms, but instead a hundred or two off. The further away from 18k we go the worse the variance. Note that we tweak all converters to our own LongBoard which is likely setup under ideal conditions and power. Customers LongBoards may vary and thus might require digital converter adjustments to fine-tune with their particular controllers. We are working on a video we hope to release soon related to this exact topic.
Note that there are better technologies than PWM to control spindles. Modbus is utilized by Onefinity’s blackbox and BuildBotics controllers. Unlike PWM’s “fire-n-forget” method of spindle control, Modbus is a “fire-n-verify” since that is a 2-way communication between the controller and the VFD.
PWM is a great signal if it can be fine-tuned. A few hundred rpms will not result in broken bits. In my experience it is a careful dance between the speed of the spindle/bit and the feed of the machine and even the density of the material. No one knows this better than bit manufacturers. This is why we’ve partnered with three bit companies to provide that expertise to customers. IDC Woodcraft, Cadence Manufacturing & Design, and SkyOne have the expertise to help customers perfect speeds-n-feeds using their various bits.
Alternatively we are working on a partnership with ProvenCut. I use their database of speeds-n-feeds with my HDM for carving aluminum. Although their database is heavily metal carving focused, we’re hoping to help them expand their wood carving information… and ultimately our customers experience with their machines and spindles.
We are continuing to research the PWM “variance problem” and alternatives to the converters and technologies we’ve chosen to release.
Thanks for the clarification Daniel,
I have tried to adjust the potentiometer with less than desirable results. Maybe now with a better understanding and a very welcome video on how the interactions of the components work I will get different results. I do have an SLB with 80mm air cooled PWN. My apologies for the frustrating request for more information but, my OCD kicked in. Looking forward to many more constructive conversations in this forum.
Hi Daniel. Sorry very new to this. I seem to be more like 2-3k out maybe in the United Kingdom the voltage is more extreme? Can you tell me where to go. As in what and where are the settings I should be adjusting. GSender or are there settings for the 2.2 water cooled machine you supplied? Via the Vmd. You talk about the Pwn and the potentiometer, is this what I need to adjust replace or somehow add to my machine. ?
Ps please keep it simple as I am struggling with the technical side of this. Thanks Mark
Ps I use a long mill mk2 48/30
The Longboard (and SLB) output a digital PWM signal. If you hook the VFD directly up to the controller you will not get anywhere near a usable rpm control signal.
You must use a digital-to-analog PWM signal converter. When ordering our spindle kit and you specify you have one of these two controllers, we recommend this product to include with your purchase: Control Cable – PwnCNC
One side has generic ferrules for plugging into the controller. On the other is a box with our logo on it. Within that box is a similar converter circuit to the one mentioned in the Sienci Labs article here: Adding a spindle to your LongMill? | Sienci Labs
On this circuit is the potentiometer of which i speak.
Thank you for the reply. Yes I did specify the long mill and I am using part you refer to. I am wondering if the slb will make my gsender and spindle rpms closer than the original control box. I am using your control cable with the spindle and asked for the long mill so I am sure it is correct. Thank you for your time and effort.
Daniel,
I have asked on a different thread if there is a better way to connect a SLB to your spindle. Modbus to the RS485 maybe? Is that something you could support and recommend?
I’m looking closely into this.
I’ve upgraded my Longmill with the SuperLongBoard and was immediately disappointed in the PWM performance. It seems that particular feature has taken a huge step backwards in quality.
I’m already working with Chris on figuring out how to get our VFD communicating over Modbus/RS485 in order to achieve far more accurate speed controls for all of our Spindle and ATC Systems.
That sounds great, thanks!
Thank you for the PWM discussion. The SLB added the RS485 interface which the Longboard doesn’t have. From what I understand the SLB doesn’t currently support spindle feedback via RS485 but other than that it works. I had a Haunyang vfd with a G-Penny spindle running with RS485 control before I upgraded to one of your spindle systems. I wasn’t able to configure the Chinese system to get reliable results and it dropped the VFD settings a couple of times so I upgraded to your 2.2kw spindle system. I am looking to switch to the RS485 control one my PwmCNC 2.2kw system. I have an open ticket.