RapidChange ATC: ATC for ER20 (and other) spindles

I thought I’d start a thread for those of us playing with the RapidChange ATC so we can work through any issues we have.

I have no affiliation with RapidChange ATC, and I bought the magazine with my own cold, hard cash.

What is it? An ATC that works with standard ER20 (and other) collets. No need to spend a gazillion $ on a pneumatic ATC.

Website: https://rapidchangeatc.com/
Youtube Channel: https://www.youtube.com/@RapidChangeATC - you’ll want to see it to believe it!

I know @jjneeb has one and is helping with early testing - he’ll be able to provide much more intelligent input than I can.

I’ve mounted mine on the Y-axis, and have a copy of the early Mach4 script for testing.

Below is a clip of my initial positioning and movement tests on my 4848Pro The clip is without switching the spindle on, but I ran another test and confirmed spindle on/off/forward/reverse. (You’ll need to add reverse switching to your spindle - see this thread for how to do it)

I can’t check real-life tool pickup yet as I’m having an issue with my spindle speed - Mach4 reports 1000rpm, but it’s much faster. I’m not sure what the issue is - I think it must be VFD settings. I’ve changed the limits in Mach4, so it’ll let me set 1000rpm, but its clearly running quicker.

I also need to sort out the tool-setter. My spindle doesn’t conduct through the bearings, so I can’t use Avid’s touch plate for the ATC. I have RapidChange’s recommended probe, so I just have to sort out connections and positioning on the machine.

Any thoughts on troubleshooting the spindle speed would be appreciated.


Check the machine.ini file for spindle min rpm limits. It was a long time ago, but I think I had to change a variable in there when I originally lowered the speed range on mine. I know for that old spindle and VFD I didn’t have to change any VFD parameters.

1 Like

Had a look, and I must have fiddled with the machine.ini before as the lowest speed was set to 500rpm.
VFD parameters seem ok - lowest Fq is set at 33.3Hz (parameter 01.08 on the Delta), but when I set the spindle to 1000 in Mach4, the VFD is showing 139Hz.

If I try to set a lower speed than 1000rpm, it just resets it 1000rpm.

I’m not sure where to go next to try to sort it out.

Oddly, if I go to Mach4–>Configure–>AvidCNC Mach4 Configuration–>Advanced Tab, the custom spindle speed checkbox is greyed out and unselectable.

/edit - checked the avid doc and the custom spindle rpm is only for the 8Hp spindle


Had to change my Delta VFD settings (01-08 and 04-12 to 2.0) and now I can get down to 1000rpm. A bit slow, but a safe speed for testing.

Emailed Avid support and they replied very quickly with some advice :+1:

Don just shared your video with me because I was having the same problem. I had mine working at slower rpm a long time ago (at least I thought it was), but after looking closely I too am running way to fast.

Thanks for posting the VFD parameters!

I need to get a decent tachometer. When I was looking at what Mach4 said the spindle was doing, and what my cheap tacho said, they were very different. Apart from full tilt where both said it was 24000 rpm.

I’m not terribly familiar with VFD programming, but I suspect that playing around with the mid frequency settings might change the speed ramp, so maybe it would be possible to improve the accuracy of what Mach4 says is happening.

Here’s my latest test after some suggestions from Don. Spindle set at 1500rpm, and an engage feed rate of 2000mm/min. Last tool I pick up is a beast of a 1/2 inch compression bit. The tightening torque is impressive.

Not sure how much more I’ll get done before we head off to the mountains for a brief holiday. My spindle doesn’t conduct through the bearings, so I can’t use the standard Avid tool setter - I need to wire in RapidChange’s suggested 3D probe before I can go full automatic.

Got mine working correctly today. I had the same VFD issue, so after setting the min speed percentage down so I can really run down to 500rpm, its working great. I am actually running this at around 1500 rpm as well. I found that a bit of paint or white tape on the collet and then using a phone strobe light tachometer app works well to dial in your true rpm. My actual rpm runs about 12% high versus what the Mach4 setting is.

Here is a video of a 3 tool changeout with automatic tool Z measurement after each one. I’m just using my AVID touchplate in its storage location which works well. They have a nice setup file to put all the toolsetter and magazine postion settings and speeds, etc. Very flexible setup and simple to use.

You can see a double stop as it retracts out of the pocket. That is when it stops to use the IR sensor to first see if there is any tool there at all, and then on the second position checks to see if it is threaded on all the way. So far all the macros are very reliable.

The torque on the nut seems to be good. Its not as high as I get by hand, but I tend to tighten them higher than they need to be (one of these days I should get a torque wrench).

Now I think I’ll fill it up and run 100 or so tool changes and see if it ever hicups.

Can you guys share your VFD and mach4 setttings to run Spindle a 500 rpm ?
Thanks !

This is what I’m most curious about. For ER20 tightening torque specs, I see numbers like 18 to 60 ft lbs, or 25 Nm to 80 Nm (here is an example), depending on bit diameter. Avid’s 4HP spindle spec pdf shows a max torque of 1.6 to 1.9 Nm. I’ve had bits slip during a 30 minute program with medium to high cutting loads when I thought I had tightened the collet nut enough (fwiw, on a Bosch router with a 1/2" collet and a decent stick-out, so relatively demanding conditions). Because the spindle torque is so far below the tightening spec, I’m curious whether bits can slip when this system is used under moderately demanding conditions.

Ya, the torque spec is weird. Its all over the place when you start looking at different sites. I’ve been playing with this with a torque wrench, and I cannot imagine putting 60 ft lb on that nut with two 6" long wrenches (I never have used a torque wrench to install the collet in the past).

Based on what I’m testing now, I probably typically have been getting 15-25 ft lbs with my little wrenches. I’ve only ever had two bits come loose on me, one was when I totally cooked the collet because the bit shank was rubbing on the side of the workpeice (it was damn hot), and the other was with that same collet later, and I noticed it had a bunch of burrs on the inside (its in the garbage now).

The funny thing is that with the RapidChange, I seem to get higher torque values with larger bits. I don’t know why, but with 1/8" collets, I get around 12ft lb, and on the 1/2" collet I’m geting more that 16 ft lb. I also put a light amount of lube on the threads and taper with these, hopefully that gives a little more holding force at the same torque level. I didn’t realize that they recommend lube there (just not between the shank and collet), so I’ve been running them dry up to now.

Very nice, Jim - so slick with the IR tool check and the probing.

I’ll add the IR tool check once I’ve sorted out a probe.

The RapidChange youtubes has the system loading bits and cutting cast iron, so theoretically there should be enough torque for our applications.

I look forward to the 100 tool change test :smile:

That’s encouraging to hear, Jim. Looks like the spindle’s momentum adds significantly to the tightening torque. That makes sense about lubing certain parts of the collet and nut.

Could you add a stress test to your list of tests? Something like a 3/8" or 1/2" tool with a 2.5+" stickout and some moderate cutting loads for maybe 20 minutes, then remeasure the tool length. If it’s successful, it would add a lot of confidence in this system, and if not, it would be good to know its limits, maybe when to pause and add some hand tightening.

You are thinking along the same lines I am.

I have already tested a quite a few torque values after changes against a torque wrench, so I already know the range is a little lighter than what I normally do by hand tightening.

So then the question is if that is enough torque.

I was first planning on just cycling tool changes of 1/8, 1/4, 3/8, and 1/2 inch collets. Probably around 100 changes to check to see if it fails ever, and then to look for any wear on the collect/nut threads, and the collect nut outside and the gripper balls in the magazine.

After that I was going to go to some full automated Gcode files to test the whole process (I still have to figure out the tool offset stuff with the AVID probe software). The final test was going to be some big, long, dull bits mixed in with some others where I’ll mark the bit to see if its ever moved in the collet. And I’l take some deep cuts on mesquite (that is pretty tough and makes the tool work for a living) to look for slipage.

I’m open to any other good ideas for testing.

Update on my testing:


Got the new tool-setter working. I spent a few hours swearing at the thing because I couldn’t get it do the tool offsets properly. Fortunately I remembered @jjneeb muttering about the master tool in his video, and after a re-watch I had a look at my config files. Turned out I was using a very early version of RapidChange’s Mach 4 config file which didn’t have the master tool settings.
Updated everything and bob was my mother’s auntie.

Video of the initial testing once I got it sorted: Youtubes PG FUN 3D tool-setter on the Avid CNC

V1 of the instructions to add the probe attached. Uses a 12V relay module so the new probe acts just like the standard Avid CNC touch plate - ie a basic on/off switch. This has the advantage that you should be able to use any tool-setter - NO/NC/NPN/PNP etc as all it needs to do is trigger the relay module.
Avid CNC tool setter instructions V1 reduced.pdf (2.5 MB)

Adding an update to the testing. This one shows the inside of the unit and how parts can be replaced if you do something dumb like I did :slight_smile:

I did the same thing before fixing the spindle speed - actually managed to ping one of the ball bearings right out :astonished:.

Don prints them in ABS, which I don’t have (and is nasty smelly stuff to print), so I made a replacement out of PETG. Got replacement ball bearings from Amazon. (I’m in Aus, so it’s easier to make the insert myself)

We’ll see how long it lasts.

They upgraded the size recently too. The ones you can see in the picture can handle a bit about maybe 15mm in diameter. All the new ones come with cavities that will fit up to 31mm, so the 45 degree and 60 degree Amana V bits I have will now fit in this.

Just waiting now for a new software routine for the approach and double tap and I’ll start testing again. I have collected a variety of nuts and collets brands and sizes for the next round.

I thought they looked small. Mine has the bigger diameter.

Interested in seeing the new software too - I’ve had a couple of mis-unloads with larger tools and certain nuts.

I just machined mine out today, so now I have 3 of the bigger ones :-).

I think James will have the new macros tomorrow, so I’ll start cycling again after that.

I haven’t had an mis-unloads, if it was loaded by the Rapid Change, but if I loaded it manually, it often would not unload it (because I do torque them down tighter than the Rapid Change does).