Is this spindle to heavy?

Hi I am looking to buy a spindle for my stock Shapeoko pro. I was wondering if this one is to heavy for the standard z axis also will it work with carbide motion? Thanks for your time Hunter.

My understanding of the recommendation is that:

  • 800W spindles are good to go with the Z plus
  • 2.2kW are definitely HDZ territory
  • 1.5kW sits somewhere in between, so while it should work, the extra weight could induce additional wear on the leadscrew in the long run.

There have been a lot of threads around spindle choices, if you want to keep your Z axis and play it safe, you may be interested in 800W models, it is quite rare to max out spindle power anyway before some other element limits the cutting parameters anyway (deflection, chatter, chip evacuation, etc…)

Any spindle will work with Carbide Motion, as the spindle control uses a signal (“PWM”) that the Shapeoko controller generates natively, and that Carbide Motion supports (through spindle start / stop commands included in the gcode by default, even when one uses a trim router)

2 Likes

I would agree with @Julien’s view on 800w, that’s what I have (G-PENNY). It has only ever shown one weakness - peck drilling Toolpaths when running the spindle slowly for a conventional twist-drill. Available torque at much lower RPM appears to be reduced, particularly at 2-3000 rpm. I mean to experiment with the torque compensation settings in the VFD, but to date have just learned to set peck drilling Toolpaths with ‘withdraw to entry height’ between pecks as this seems to allow the spindle to cope with the high-ish feed rates the drills want with the lower torque. I don’t know what the same torque curve looks like on the 1.5kw spindles, but I can imagine it to be similar…

I think the leadscrew has a delrin block, and that is what would wear? The g-penney 1.2kw is sort of tempting except that I already own an 800w.

Maybe this is pedantry but I think that’s inaccurate: most VFDs expect a 0-10V analog control signal, like you’d get from a potentiometer hooked to the VFD’s high and low pins. Carbide Motion’s 5V PWM signal happens to be tolerated by most VFDs with the right configuration changes or some kind of converter board but I wouldn’t say that the VFD “uses PWM” so much as tolerates PWM.

5 Likes

True. But c’mon, I was just trying to manage my natural tendency to provide too many engineering details to unsuspecting passers by :slight_smile:

4 Likes

The wonderful world of electronics!! I wouldn’t mind betting that the VFD uses an A-D convertor input on its microprocessor to measure the ‘Analog In’ signal, so will use stepped dv/dt capacitive filtering techniques to calculate V down to the nearest binary division the processor uses. From digital (PWM) to approximate analogue and back again approximately to digital. I recall seeing somewhere in my manual that the Nyquist frequency of the inbound A-D conversion can be set, so clearly expecting ‘PWM-ish’ input…

3 Likes

And this, right there, is why I enjoy this forum so much. From 0 to nerdy_max in just a couple of posts :joy:

4 Likes

I was clearly failing to manage my natural tendency to provide too many engineering details…

1 Like

Hi Hunter,

I have been using a 1.5kW spindle on my Shapeoko Pro for almost a year now and it works brilliantly - go for it.

2 Likes

As a non electrically educated person, this is what I am hearing right now:

(Leslie Claret's Monologue on Donnelly Nut Spacing (Patriot) HD - YouTube)

2 Likes

And however grubby and unpleasant the ‘lowest common denominator’ analog 0 to 10V signalling might be between two digital systems, we can all thank the stars that it isn’t ModBus or the crime against all that is decent OPC-UA. You have to buy a much more expensive VFD before you go 50 years back in time with ‘professional’ industrial controls signalling :face_with_head_bandage: :pick: :cry:

2 Likes

What’s wrong with MODBUS?

2 Likes

There was nothing wrong with Modbus, in 1979 when the 8088 was bleeding edge, clock speeds were 5MHz, VMS was the latest new thing and Kernighan & Ritchie’s book on the C programming language was so new people were still wondering whether it would catch on.

In its defence, Modbus still works, Schneider did release the mouldy corpse as a proper open protocol, so better than Tridium Niagara on that front. It’s low level and fairly well suited to real time systems with time or safety critcal communication. It requires only minimal knowledge of ‘computer stuff’ from its users and it’s a few steps up from the volt free contact.

On the downside

  • Serial comms wiring with limited distance and very limited number of devices, creates an endless rats’ nest of hard to maintain and troubleshoot spaghetti, comms problems are near impossible to automatically detect and a pig to locate and fix (no dropped packet counter on the switch port for users of this museum piece)
  • Hard configured device addresses
  • Name resolution and routing came decades later (Vint Cerf was still figuring out TCP IP and token ring didn’t arrive until 5 years later)
  • No proper device level security, authentication or certificates - anything with access to the physical layer can send any instruction to anything else on the layer (ask Iran)
  • No useful encryption, no signed message bodies to validate the sender or content of a message, meaning no protection from the most basic hacking (or even comms errors)
  • Jurassic era set of data types to which you are limited
  • No metadata descriptor or even unit for values which can be read, you just have to hope you used the right version of the register spec for that firmware of that model of that device
  • Worse than SNMP for maintenance of ‘device descriptor files’ (MIBs as SNMP calls them)
  • Control outputs are called “coils” still today…

So, if you like to wear a beige tank top and green corduroy trousers and drive your VW Beetle in platform boots, I guess Modbus is great. When a few hundred devices was ‘huge’ and we had more fingers than instructions on the devices being programmed, it was a decent fit.

To channel some Irish ancestry however, if I was to start now building a securable, dependable, maintainable controls protocol, I wouldn’t start here.

3 Likes

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.