[Jesse B]

Well I found some CAD files for the pulleys, so I got those. I did a few little revisions, got something I'm fairly happy with.

I also thought that I may be able to use the pulley to retain the shank? See image below.

Would I be able to do this? I'd be able to shim it or whatever it takes to remove and linear play. Just an idea since those pulleys have clamping hubs.

Thanks,


- Jesse

[Jesse B]

Alright, found the rest of the parts I need. I figure this should be pretty much it. Can't make a final mock-up in CAD until I have all the parts I need, but this is fairly close Found some new bearings that are good up to 48k RPM, way more than I'll ever approach. Top of the spindle assembly will screw off, just to make it look a bit nicer than an open top. Using pulleys so that I can adjust the ratio for different speeds and whatnot. Anyways, here it is.



Lemme know what you think.


- Jesse

[jjobezo]

Any reason the spindle can't be a block instead of a cylinder? Just for my edification.

[Jesse B]

Any reason the spindle can't be a block instead of a cylinder? Just for my edification.

None whatsoever. I just decided to model it up using a cylinder.

[Jesse B]

Another question, more based on the control of the spindle.

The method I'm going with requires a ~5v PWM signal to determine the speed of the spindle. Can this signal be output from the software via a pin on the Parallel port, or would I have to write a small application that did it otherwise? Also, would I be able to connect an encoder and have it display the RPM in the software?

I'm also considering just making a small control box containing a microcontroller to take care of all this. I'd have some form of input method to change the RPM, and LCD to display all the necessary info, and the MCU could take care of the PWM.

Which would be preferable?

Thanks,


- Jesse

[RotarySMP]

Mach 3 can output a 5V pwm signal (assuming your parallel port is 5V, many newer ones are only 3.3V).

If you are trying to drive an RC ESC, then the PWM signal is timed into a narrow band at the start of each interval. None of the CNC programs has native support for that protocol (afaik), but there is a pretty simple circuit floating around the internet to translate Machs signal.

[Jesse B]

Mach 3 can output a 5V pwm signal (assuming your parallel port is 5V, many newer ones are only 3.3V).

If you are trying to drive an RC ESC, then the PWM signal is timed into a narrow band at the start of each interval. None of the CNC programs has native support for that protocol (afaik), but there is a pretty simple circuit floating around the internet to translate Machs signal.

Alright, thanks. Know any of those circuits off-hand?

[RotarySMP]

There was a link in the thread which Herbert posted into:

http://www.epanorama.net/documents/motor/rcservos.html

[Jesse B]

Ah, thanks. I have an idea for a circuit that I'm gonna go ahead and design, and I'll decide between the two at a later date.


- Jesse

[Jesse B]

That specific gearing would put it around 30k rpm at top speed. Should be able to get it going a fair bit slower as well just fine

[serriadh]

That specific gearing would put it around 30k rpm at top speed. Should be able to get it going a fair bit slower as well just fine

At what sort of speeds should a spindle be using direct drive rather than being belt driven? Looks like commercial spindles don't use belt drives much past 10krpm, and I'm assuming they have good reason to given the convenience of having a separate motor when it comes to manufacture and servicing the thing.

I'm assuming its a bearing life issue, primarily, but I'm only guessing here.

[Jesse B]

At what sort of speeds should a spindle be using direct drive rather than being belt driven? Looks like commercial spindles don't use belt drives much past 10krpm, and I'm assuming they have good reason to given the convenience of having a separate motor when it comes to manufacture and servicing the thing.

I'm assuming its a bearing life issue, primarily, but I'm only guessing here.

I never considered that yet; maybe somebody can chime in and let me know whether or not belts are fine at this speed. The reason I wanted to go with this setup is due to the fact that this motor will only be able to hit about 19k rpm with the power supply I was planning on using. I may be able to do a direct gear drive however. I don't know, I'll figure something out

As for the bearings, the manufacturer claims that they're good up to 48,000rpm, so I'm not too worried about that.

[serriadh]

As for the bearings, the manufacturer claims that they're good up to 48,000rpm, so I'm not too worried about that.

Bearing ratings don't tell the whole story, unfortunately. You'll have to derate that a good deal when you preload them, for example. The faster your spindle goes, the more the bearings will suffer from runout and imperfect alignment, causing them to wear out faster. Spindle engineering gets dramatically more complex when you go above speeds of 7-10krpm. Apologies if you already know this stuff!

You described Herbert Kabi spindles as overkill; I'm going to have to disagree with you there. Winding your own motors; yeah, that probably isn't for everyone ;-) but the rest of the engineering is pretty much what is necessary for the speed and power requirements he's working within. You may end up being disappointed with either the runout or lifetime (or both!) of your spindle if you don't follow broadly similar design constraints.

[Jesse B]

Bearing ratings don't tell the whole story, unfortunately. You'll have to derate that a good deal when you preload them, for example. The faster your spindle goes, the more the bearings will suffer from runout and imperfect alignment, causing them to wear out faster. Spindle engineering gets dramatically more complex when you go above speeds of 7-10krpm. Apologies if you already know this stuff!

No worries, always nice being reminded

You described Herbert Kabi spindles as overkill; I'm going to have to disagree with you there. Winding your own motors; yeah, that probably isn't for everyone ;-) but the rest of the engineering is pretty much what is necessary for the speed and power requirements he's working within. You may end up being disappointed with either the runout or lifetime (or both!) of your spindle if you don't follow broadly similar design constraints.

Alright, I'll look into making my design a bit more robust. I've got lots of time so I'm in no real rush.

Thanks for the input,


- Jesse

[serriadh]

I've just read a thread on the HSM forums covering a relatively simple high speed spindle design, which was belt driven to >30krpm. The author goes into more detail about his problems with machining his own collets (which he eventually gave up on) than into making the spindle itself, but its good to see this sort of spindle design working nicely!

See here for the whole thread. The final working design with ER11 collets is on the last page.

[Jesse B]

That's a pretty nice looking spindle. I'm fairly confident in my design. Worst case scenario, it doesn't work