Tool setter for atc

I have been manually setting my tool offsets for my atc with a gauge block but want a fixed tool setter than can take care of all that. Anyone have any experience with this one? I believe normally closed is what I need to get.

I have been trying the standard avid touch off plate for this and it works in theory its just not consistent enough. It goes through the motions correctly and adjusts the offsets, but the numbers can change .010" from one time to the next…even seen .025" of change. It may be something to do with the groundwire setup but its working so I’m not sure what else to try with it.

Yep! Sure do. They have visual defects but no functional defects. RM is typically pretty good but still get it a full test before doing long runs with it. I believe they are seconds from whatever factory that makes them and RM finds the ones that look bad but still work as intended.

I use a switch on the tool setter carriage that connects it when extended and of course disconnect when retracted. I do that so I just need a single port when using a probe. I do not want to modify the AvidCNC box which is why I did that.

Correct, that would be a mix of the accuracy and the repeatability (see the specs page) with a hair of worn pinion gear. It can also be that you are moving too fast.

Be advised that max speed is advertised as 200mm per minute, just don’t do it. 75 to 120 for first strike, 50 or less for second. Slow is accurate.

As a note, the tool setter you posted is normally closed and the AvidCNC touch-off plate is a Normally Open system. You will need to adjust that in the ESS and you can no longer use the AvidCNC touch plate.

You could get a Normally Open variant but I would strongly advise against it. NO will fail, of course open. Thus when it wears out you will crash. Normally Closed will also fail open, thus when it wears out it will fail immediately and not break anything.

Personally, I use a NC probe and tool setter. I have the RATTM Motor and its been very repeatable.

The air blower will need a solenoid. It is 5mm thread when you are looking for the fittings. So its 5mm thread and 6mm tubing. Do not use 5/32" because it really doesn’t fit. Its close but no cigar.

PS. Yes, the air blower really does make a difference.

Thanks for having this discussion, I’m working on setting up a similar tool setter and this is super helpful.

If I could ask you have expressed that the normally open variant will lead to a crash. Does that mean there is something about the probing process which disables the overtravel protection circuit? I coded mine to trigger the E-Stop, but am not sure if that code can be executed while G31 is running.


Now, if you wire that second switch (over travel) to your actual e-stop, which I did first, wow, what a mess. Once triggered you have to either unplug that sensor to restore motion or move the Z axis up by hand.

Now, if I were free to alter the AvidCNC control box, I would add another probe port to it and use that for over travel.

Right now over travel is not wired on my machine. I take this into account by not moving too fast over it.

Coded? Remember that the Lua engine is separate from the Mach4 core which is again separate from the ESS which actually executes the instructions.

Personally, I would not trust the Lua engine.

I have NOT updated to the newer install from Avid. I am testing the retracting tool setter and I was specifically trying to measure the variance for each extension and retraction action. This would be the wiggle in the bearings for the carriage as it’s pushed out to be used.

Here is a sampling, all values are machine coords set by the master tool (aka tool #1) as the value of MasterOffset. This will be the value by which all tools are measured to calculate their offsets in the tool table.

I just took the largest variance which was around 0.006mm or about 0.00023622047" of possible variance. The tool setter is the RATTMMotor NC tool setter on Amazon as stated before.

Also note that I am in fact using the 200mmpm figure for the first strike.

Sorry, taking me a while to digest everything. Programming in the mach 4 environment is very new to me.

What i have so far is connecting the overtravel switch to an input on the ESS and a simple signal script which calls mc.mcCntlEStop() when the signal goes from high to low. I was trying to figure out if these scripts work during the probe process, but haven’t actually started probing to test it. Maybe I should use a drinking straw or some other very soft item for my first few probes to make sure I don’t drive it too far.

Do you have a recommendation on how to set up the overtravel switch?

I plan on trying to document what I learned and making a post here when I am done. I can’t be the only one trying to see if a tool setter like this might make manual tool changes convenient enough that I am no longer tempted by an ATC spindle I can’t yet justify.

If I had the pin available :cry: I would set that pin as the Z axis minus limit switch. I would not use Lua scripting on something that important.

ATC is for commercial applications and hobbyist with too much money and time on their hands. Don’t get me wrong, I would love to sell you an ATC system and a tool carousel but most folks would be terribly unhappy with the added complexity.

I have that toolsetter coming(normally closed version) and would like to add an edge finder touch probe. I know people put a splitter in and use 2 normally open probes, would a splitter work to have 2 normally closed probes as well?

Nope, Sorry. :disappointed: NC precludes a splitter without some sort of selector switch.

When people ask me the pros and cons of NO vs NC that is the ONLY pro for NO.

The ATC Tool Carousel and Tool Setter system I designed can use a switch to connect a single NC probe and Tool Setter to the same wire. When the Tool Setter carriage moves forwards it actuates a switch that connects the Tool Setter to that single input. Retracted it disconnects the Tool Setter.

In that scenario you cannot use the Tool Setter and the probe at the same time, but that would not be the normal operating procedure anyways.

Are you using the standard avid control box? I have been trying to wire the setter in but couldn’t get it to work, and after some reading I found the m12 probe connection on the breakout board for the probe doesnt have an nc pin, just the power, ground, and normally open pins.

Can’t they be wired in series? Or am I missing some safety aspect?

On mine, I had to add a software selector that disabled the tool probe, because it was sensitive enough that the vibrations from using the machine would set it of…

Yes, that is true but that wouldn’t be a splitter. But you could in fact wire it in that manner. Now, my problem is I unplug the probe when its not in the spindle. That means I have to get fancy.

There is a debounce or noise suppression feature in ESS that can make some of that go away. I am not sure if that would help you though.

That little RATTMMotor sensor has a fair amount of travel before trigger and before overtravel.

What model or manufacturer are you using @djdelorie

PS. I am still jealous of your MPG. That will have to be after the flow through fixturing and “True Hole” implementation for the Hypertherm Plasma systems. My plasma friends are really crawling up my tailpipe about the True Hole. I am not sure I can pull it off with Mach4 though. Not sure if the MPG will even work for the Mach4 either and for the same dern reason.

Nope, as I’m both using linuxcnc (not mach) with a mesa card (not ESS) .

I ended up getting the DrewTronics TS1000.

Oh! That would be in the MESA board configs. I forget just how much I really do envy you.

Now I do not have any MESA boards so beyond that I might be terribly wrong.

Here’s a photo of the messy insides, to help you accept your life choices :wink:

It’s wired as such: the big rotary switch has three wires; A and B phases plus a common. The two selectors are each 2P4T and are wired as such: each pole is a signal to the computer, and there’s ONE common that’s wired to half the throws such that the two signals from each switch encode a two-bit binary number.

Eight wires total, fits in a CAT5.

The HAL code for linuxcnc has to decode the selectors, but that’s not a big deal.

The text on the front is plain paper from a laser printer, and some double sided tape. It’s more of an arts and crafts project than an engineered piece of equipment. Like the best of hacks, though, I made it with what I had, and it serves its purpose.

I can share the openscad/stl files but unless you have the same encoder (which I removed from a stereo I was throwing away), they won’t be much use without some tweaks.

Plan B was to get a bluetooth game controller and use that instead.

1 Like


Day one, AvidCNC, got the Xbox 360 wireless controller. Hit the Z axis button, started moving it down, let go and the machine kept moving and did not stop until I hit the e-stop.

Never used the Xbox thing again.

I have the media jog wheel thingy. It is USB and works great but I would love to switch over to steps and thus why your project is of such interest. I was hoping you had a product coming :rofl:

The mesa 7i76e is pretty cool, but in the end, it’s mostly an industrial-grade parallel port with a bit of timing-critical offloading. It’s the HAL stuff in LinuxCNC that makes it useful, and that would work with a vanilla parallel port (which is what my old cnc did) too.

You know what is truly fascinating to me? ChatGPT has a huge amount of detail on LinuxCNC. You are a Tech guy like me so you know its not going to do everything for you but is a crazy good search engine for the topic.

Funny of funnies, it knows Mach3 and Mach4 exist but has zero knowledge beyond that. If it were a human it would snort before answering on the topic :stuck_out_tongue: