So I was tired of replacing brushes and the wobble in the dewalt routers (and I’m sure the Carbide 3D router is fine, but I have nothing but issues with the collets getting stuck) so I decide to make the leap to a nice water cooled spindle.
This morning I soldered the hardest thing (VFD->Spindle power cable), connected the water cooler, the VFD and the spindle and … time for a test cut.
Unfortunately I was so used to the bitrunner by now and left the bitrunner setting enabled in carbide create. And let me tell you, the spindle is QUIET…
So it was running when it started the probing with the bitsetter and I did not realize it…
As a result I now have a nice 1/8" dimple cut in the top of my bitsetter button… but cutting metal, the spindle does.
Next step: Investigate if I can use the cable of the bitrunner as a convenient way to get PWM control into the (HY) VFD, and see if I can superglue a small metal disc on top of my bitsetter button
I forked the PWM signal between my BitRunner, VFD, and Laser unit and it has worked just fine for me so far.
The one tiny thing I have not addressed yet, is that I have a spindle warmup sequence I run at the beginning of a session in the shop, which has the spindle going through a 5-minute speed ramp, to warm-up the bearings and lubrication, starting from 1000 RPM up to 24000RPM . At about 9000 RPM, the BitRunner is triggered (as it should) which turns on my the shopvac, which is annoying as I am not cutting anything at this point. So I usually switch the BitRunner to “off” position during my spindle warmup routine. And then of course…I sometimes forget to switch it back on, and start cutting with no dust collection .
I have grown to rely on the BitRunner to start dust collection automatically so much that I don’t even think about that aspect anymore.
Maybe use adamantium, so next time this happens it won’t even scratch the button
Anyway, enjoy your new spindle !
Did you go for a closed-loop cooling circuit or an external bucket&pump / chiller ?
I went the CW3000 route (well, cw3000 labeled, mine looks slightly different and the noise level wasn’t horrible at all)
I use a festtool for dust collection, and it has an outlet built in so that once the current on that outlet exceeds some threshold, the suction goes on… and the VFD is plugged into that … so far it works great, the festtool goes on once the spindle hits like 5k rpm or so…
I need to look into warmup (so far I’ve done this manually), maybe as part of my gcode generation flow (I post process the carbide create gcode for optimization, or use my own tools to generate gcode) to keep it simple. I also need to look at which grbl parameter to change so that grbl waits longer for the RPM change to have taken effect (clearly my spindle does this slower than dewalt/makita routers).
I’m thinking of just using the bitrunner output of the control board for PWM… just need to make sure to use the right pins.
Since I only do the warmup thing once, I use a (CNCjs) macro that executes a series of G4P30.0 and M3Sxxxx, in a sequence.
Either I missed it or GRBL has not built-in param for PWM linger/dwell upon M3. I ended up adding a G4P5.0 command to my post-processor in the header part, right after the M3 commands.
It seems grbl has something called “startup blocks”… which might look perfect for the warmup sequence.
I indeed dont’ see a M3 delay option so post processor (or post-post processor I suppose) it is for that one. Adding the G4P5.0 is not complicated at least.
Actually I stumbled upon that startup blocks feature yesterday. However it seems quite dangerous to put any spindle-related (or movement for that matter) command in those blocks, I definitely don’t want my spindle to turn on automatically at the controller power-up, even at very low RPM
GRBL is designed to manage motion only. In its current incarnation, the best approach here is a separate program that does the warmup for you, and just run it at the start of your day. That is what I do also.
the “delay at M3” is more useful to do on the grbl side though
(mind you, I have the luxury to have my own gcode generators/post processors so I can and will just hack it into those… but most folks aren’t that fortunate)
Invoking chamnit? Sounds more like a spell or incantation.
The recommended way to impose a dwell is to use a post processor from your CAM after an M3. Most good ones should be able to do that. While it’s not difficult to have Grbl impose a dwell internally, I chose not to because you don’t need this for all machines (lasers); each spindle has different requirements; and Grbl has severe memory and flash restrictions on the 328p processor. Decided to omit this type of configurability for the general build.
Although it is technically allowed in Grbl, don’t put machine commands in the startup blocks. The main reason is after a crash or a bit is jammed inside the material. You don’t want the machine spindle to turn on in that scenario. Startup blocks are there as a simple way to set the initial gcode defaults like G54 or inches/mm mode.
I ended up with a 800W water cooled HY spindle with a HY VFD.
I decided I wanted to stick to 110V which limits the choices a lot…
and since I’m really only doing wood… 800W seemed plenty and my enclosure is height limited (not smart in hindsight) and a 65mm diameter limited height spindle seemed perfect
I’m also wanting to go the 110V route, just precieving doing so as easier than running a new 220V drop from the sub-panel. Can you point me toward your supplier so I might take a look?
so I would strongly suggest (since you’re in the 110V range) to get a 65mm diameter one. There are MANY MANY more dust collection options for 65mm (like the C3D option) while for 80mm they mostly seem to … be bleak.
(oh and get the C3D holder if you have a HDZ… much much better than whatever junk comes with these kits… my HDZ is one of the blue ones and the C3D 65mm blue holder is both SOLID, squared-by-machining and looks great matching)
btw I wonder if @WillAdams could suggest to the sales/product folks to start selling the cable between the board and the bitrunner, but only one side of it (so no connector on the other side) with the two wires needed for the VFD labeled and exposed (for me, brown = ground, red = PWM)… the product folks are already making these more or less and … well… convenience for us and margin for Carbide3D
.
