Carbide Motion - Endmill load position over wasteboard

I’d appreciate having the endmill load position moved back slightly so that it’s over the wasteboard. This would reduce the risk that it falls three feet to the floor in the event that it drops out of the collet while installing or removing the endmill.

If it’s relevant to this discussion: SO3 XXL with bitsetter

Thank you!

1 Like

Hi there. No idea which spindle you use but this should work for most designs that use two spanners (wrenches). If you place them close together (position the handles 10 minutes apart on a clock face) with the undoing handle on the right, then you can use one hand to loosen the collet nut while holding the endmill.

Tightening the endmill in the collet requires reversing the process. Doing up handle on the left while you hold the endmill in place and tighten the collet nut with one hand. This is useful if you insert the endmill as far as the index mark, you can hold it while tightening.

In both use cases, you are just squeezing the handles together.


I put a plastic tray under the load/unload tool position. If my bit falls it hits the plastic tray and keeps from damaging the bit. However usually I reach up and loosen the collet with one hand, hold the lock button with my second hand and with my third hand I hold pressure on t he side of t he bit to keep it from falling. Really I use my second hand little finger to try and put pressure on the bit but sometimes it falls anyway so the plastic tray is in place.


Thank you both for the tips. My usual approach is to have the 13mm wrench in my left hand, get that into position, then use another finger on that hand to hold the bit. My right hand has the 22mm wrench that engages the collet nut. Similar to guy, I have a piece of foam that I always lay under the bit when changing tools. It doesn’t have sides like a tray does (that’s a good idea!), but at least it’s soft if/when the bit lands on it. I’ve never actually had one fall to the garage floor, I’m just paranoid it’ll happen one of these days. Moving the load position back 4-6 inches (10-15 cm) would mitigate that risk without making it any harder to load.

1 Like

Having the wrenches in either hand is where the issue begins… I happen to think it may apply unequal forces laterally. I think the forces exerted by one hand rotating both wrenches around the centre of moment are less likely to pull something laterally via unequal forces. I will grab an illustrative image for you later today.

1 Like

HiTravis: As promised, the images of the wrench positions to help illustrate what I have said.

The first image shows how far apart you would usually place the ends of the wrenches so that you could hold the handles easily without your hands touching each other. This may be slightly too much. I placed the wrenches in this manner to illustrate that tightening or loosening would be quite difficult from this position. The red circles denote the collet and spindle so that the torque moment required is obvious. With the wrenches far apart there is some possibility to introduce a bending moment at the spindle centre and pulling or pushing the router spindle off axis.

The second image shows how I place the wrenches so that I can tighten or loosen the collet nut with one hand. The torque moment is probably more efficient because it feels much easier to apply to the collet nut. I have not felt the same instability and tendency to apply a bending moment as with the handles widely separated.

Large wrench on the left equates to tightening the collet nut. Large wrench on the right equates to loosening the collet nut. One advantage when loosening the bit is that you are left with a free hand to hold the tool bit you are releasing for a change. When tightening the collet nut, you can support the tool bit at the right length. I usually can place my thumbnail at the level of the index mark to ensure that the tool is not inserted too far into the collet.

I clean the collet, collet nut and collet taper at every tool change, using a soft brush which is inserted to the whole length of the collet taper. The tool, the collet and the collet nut are cleaned via a blast of air. Only the users can judge whether the technique is useful or better for them. I quickly found that I preferred this method and it felt on all fours with my own mechanical sympathy.

A very short video (please excuse the low resolution and the amateur production) illustrating the technique for a tool change.


I think this is a good technique.

However, it’s still possible to accidentally drop the bit with the other hand (which is the issue in question).

I think having the “change tool” location a user-specified option is still a reasonable request to make accidental drops less likely to be drops to the floor.


I think we may part company on this point, Gerry. :smile: I am not sure that it is a fault within the tool change routine, or the design of the SO3, if the operator is likely to drop the bit. The possibility of an accident (probability theory is not my strong suit) is ever present.

It would be arguable as to whether the provided software along with the current mechanical solution to changing a tool, leaves the user open to accidents that would not have been occurred if things were different. While I see a pathway of continuous improvement from SO1 up to SO4 and Pro, it is hard to predict how one may misuse a piece of kit until we see the situation.

There is the old saw that states that just when you make something absolutely idiot proof… along comes a better class of idiot. :rofl:

I guess I am trying to say that although the current routines (for tool changes in this case) could be examined for inefficiencies during a production run, I don’t believe it is possible to prevent handling issues which emanate in the first instance from the operator mishandling the tool.

I think that there could be benefits to a change tool location that was user specified. I have not dropped one tool to date and I like the fact that the X axis moves to the front of my SO3 to facilitate the tool changes. Maybe a small accessory such as a swivelling tray which is foam lined, could be swung out from under the framework at tool change time and sit under the tool change location.

During my video demonstration, I found it awkward to reach to the back right hand home position for the tool change routine. My reason for leaving the spindle there was so that I could find enough space in the diagonal, that would let me park my phone to take the camera footage. I could not have rested the phone easily, if I had brought the spindle to the normal tool change location.

Another point to consider is the folk who have larger machines than the standard SO3, would possibly not be able to reach to the inbuilt home position. I wonder if it would be reasonable to change the home position to front left. :thinking:

Sure. The option would not be for you. That’s fine and good to know. I would use it.

1 Like

I see both sides and freely admit that my technique can and should improve. Thanks, Jeff, for the clear explanation and visuals! I agree that squeezing two wrenches/spanners is less likely to apply a radial force to the router compared to using two hands. In the former case, I’m using only my grip; in the latter, my arms enter the equation and it’s harder to be certain that the force I’m applying is purely tangential to the router.

That said, having the endmill over the wasteboard is inherently safer than extending beyond the front edge, and 10 cm won’t affect its accessibility to users. In safety-speak, teaching someone to use proper technique is an “administrative control.” It’s good, and I should follow it as standard procedure. Moving the load position back 10cm would be an “engineering control,” taking the human’s ability to follow procedure out of the loop. Administrative controls should be enough, but engineering controls are there for when people screw up. As a brief aside, Guy’s plastic tray suggestion is another engineering control – there to literally catch my mistakes. If controls don’t add cost or hurt productivity, the more the better!

Photo 1: Current tool change position, y=0
Photo 2: Preprogrammed south jog position, y = +43.75 mm (better, but still close to front edge)
Photo 3: Suggested tool change position, y = +101.6 mm (4 inches)


Hello Travis; thanks for the images. They were helpful for understanding the precise conditions under which you work. I think the main component parts to this issue can be usefully broken down into current practice, desirable practice and enforcing standards. I would like to speak to them all so please excuse the prolix post.

Current Practice:
This concerns the methods employed by the provided software for negating the limitations of the supplied hardware. The endpoint is to produce a working solution that suits most of the user base.

Desirable Practice:
This concerns the methods of work we would all choose were we to pool the collected hive knowledge. We would choose to apply it to the hardware development cycles as well as the software used to drive the machinery. It would be efficient and configurable for as many use cases as we could imagine.

Enforcing Standards:
This concerns producing a machine that will not kill the customers if they obey the rules for usage. It will also prevent the foolish from coming to harm. The standards would ideally be consonant with whatever standards are applied in the industry for employing and supervising machinists who work with CNC machines.

Current Practice:
Often people will decide to purchase a CNC machine without prompting (other than a little advertising) and they do their own research. After this examination of existing choices and capabilities, they will settle on some CNC machine which they believe will suit their needs.

The manufacturer cannot control many of the variables that attend this process. Who buys a CNC machine? What do they intend to accomplish with it? What level of knowledge, skill or experience does the purchaser have? Will they be sensible when using the machine? Will the machine be neglected and routine maintenance forgotten so that it becomes a dangerous artefact when it is first used after a long period of shutdown?

Desirable Standards:
The idea of including software and hardware features with which to create and enforce standards is an attractive one. A router bit spinning at 30,000 rpm is always going to be a very dangerous tool. It can cause a lot of damage to human soft tissues in the unwary. The danger could only be removed by preventing the spindle from turning on. Or we could educate users in basic safety when using a router.

Basic safety includes ear protection, eye protection, knowing the sounds which indicate something is wrong, never leaving a working machine unattended and prime rule number one… NEVER place any part of yourself or any tool inside the cutting area while the router is working. We could arrange guard fences and electronic interlocked doors for large machines. It is probably not practical or very easy to implement such systems on a hobby sized machine.

I support the notion that we sometimes may prevent people from learning important lessons if we always do their thinking for them. A simple example is my once two year old child was for always walking up the staircase and jumping down the stairs. The game was great fun and after about 6 months of this behaviour, my wife asked me to fit stair gates to the top and bottom of the staircase. We were responsible parents and could be seen to actively try to prevent accidents to our young son when he played on the stairs.

The first day that I had fitted the stair gates, my son fell down the stairs. He had not liked the restriction to his freedom to climb and jump on the stairs as he had always done. Anyhow, he climbed over the stair gate and had his very first accident on the stairs because of my removing his ability to choose to play on them. Immediately after that accident, I removed the stair gates. My son never had another accident on the stairs.

I hope this illustrates what I understand can occur when we take away the ability of the user to learn from their own mistakes. In the instant case under discussion, a simple swivelling tray that occupied the region immediately surrounding the tool change location would be a good cheap solution. Tool bits can be expensive and I am sure that users would appreciate the ability to prevent accidental damage.

Having a tool change location that could be programmed by the user would be a good solution. That way you could choose to set your X axis back from the front of the machine by 4 inches. I am sure it would be possible to apply the 4 inches at the factory and have the issue solved without the necessity for the users to consider it. Not dropping tool bits would also be a good rule to follow.

What the future presages in the case of domestic CNC machines, we do not know. I don’t know how many manufacturers of technical machinery can afford to listen to a large and very disparate user base with widely differing requirements. The notion that users may suggest software and hardware changes to the people who produce the machinery is attention grabbing.

What is nearer to reality is that if the suggestions fall easily within the company’s intended development roadmap, some lucky users may have their ideas implemented. Only the manufacturer knows what is planned and only they are privy to all of the use cases and they can either accept them or reject them as unimportant whenever they feel the need. This kind of committee approach to how the products are used is quite common and it is not usually very responsive to ideas and demands from the user base.

The knowledge required to lay down what would be a desirable practice exists within the user base. I would not wish to see any sort of proxy nanny state but it is clear that professional machinists could have a lot of very useful input into the design and use cases of the CNC machines manufactured by Carbide 3D. I don’t know if it is ever likely to happen (a users expert committee feeding into Carbide 3D) but it is likely to draw up plans which are going to be very slow to implement.

The legal position is also interesting. What responsibility would a manufacturer have if some users hurt themselves, where the manufacturer had followed the suggestions of an external expert committee? That sort of nightmare does not bear thinking about and the vicarious liability for a harmful happenstance must always rest with the company. Primarily because it is unlikely that any civilised legal system would permit a manufacturer to divest their company of its legal responsibility for harm caused to a user, howsoever it was initiated.

Enforcing Standards: The enforcement of standards can produce a situation where it is exceptionally difficult to rigorously apply set standards. Carbide 3D would need to positively vet all of their potential customers and pass them as fit or unfit to purchase, own and run Carbide’s CNC machines. The way forward is training rather than regulation. If you ever get around to thinking that training is very expensive… you should try ignorance.

It behoves the companies who make and sell dangerous machinery, to ensure that the user receives adequate instruction in the use of the products. With webinars, printed materials, internet lessons, adequate support mechanisms et al., it should be possible to train each purchaser to a known standard. It would also underline to the prospective purchaser that you should not buy a CNC machine (for example) if you are not serious about learning how to use one safely.

Hi Jeff, thanks for the thoughtful discussion.

I by no means intend to suggest what Carbide3D should do. They can take my thoughts, your thoughts, Guy’s thoughts, Gerry’s thoughts, and anyone else who wishes to add thoughts to discussion, then make the right decision for their company/product. They’ve heard a few different perspectives here, and no doubt have heard from others in the past and have informed viewpoints of their own. For all I know, they prioritized people’s ability to load extra-long endmills, and so decided that loading over the front edge was better.

I also agree that controls/restrictions are not a replacement for training and common sense. Regarding use of the word “safety,” though, I probably wasn’t clear. I meant safety for the endmill, not safety for the user. I wholeheartedly agree with you on everything with regards to operator safety. I was tempted to share an anecdote of my own, but I don’t want this thread to drift too far off-topic.

1 Like

Travis, I’m personally in agreement but since I’m not the programmer I can’t say if/when that will change. In the mean time though, one hack I use is to tell Carbide Motion my Y travel is shorter than it actually is. The tool loading position is relative to what CM thinks the machine’s max Y travel is. If you don’t ever use the overhang area in front of the machine, then you can trick CM into stopping short of the edge of the bed. I do this a lot because I designed my enclosure too short, and the Shapeoko + dust boot will punch the doors open when the machine comes forward, which can be mildly startling if you don’t expect it…


Thanks Winston, that’s a clever workaround. On the rare occasion that I need to cut something overhanging the front of the machine, I can easily change the soft limit back.

If I recall correctly, the bitsetter coordinates are specified independently, so this shouldn’t interfere with its function. I’ll play around with it tomorrow and see how it goes.

Good day, Travis. You are welcome. I don’t see much point to discussions which are not considered.

I did not get the impression that you were attempting to force a change in Carbide’s policies.

It is entirely possible that I was not clear. I had understood that you meant endmill safety. I thought about the issue and concluded that the rationale behind an effective solution was, of necessity, going to be multi-factorial. I thought that I would try to broaden the discussion a little to see what fell out of it.

At base, I find myself treating my SO3 as I would treat all of my other tools. I pay for all of my tools with my personal money. This means that I do not lend them to anyone. I clean and maintain them to prolong their accurate, and satisfactory service life. I buy the best tools I can afford (and understand how to use) and I do my very best to care for them. The idea of dropping a tool bit is anathema to me.

I use some outstanding Pfeil chisels and I would no more drop one by accident than fly to the moon. It does not take place because I do not let it. I understood very quickly that what I had in the SO3 was not a replacement for me thinking or taking care to treat the tool and the bits with respect and care.

This is the point at which I differed with @Gerry. People should accept what they are working with and not expect the solutions to rest with other parties.

My internal response to that position is hold onto the tool bit tighter and then you will not (can not?) accidentally drop it. I don’t mean that to come across as me being rude or uncaring. I try to accept the responsibility for my own actions.

I am not trying to say that I am perfect and so never would have an accident with my tools. I am just trying to show that I take as much care of my endmills as I do with any of my other tools. I see that it is my responsibility to assess each situation I am faced with then act accordingly. This makes the act of me dropping a tool bit quite unlikely. In a commercial production environment, or where employees were trusted to use the machinery, I think it likely that my priorities would change.

I didn’t make (the opposite of) that point.

People are accepting and working with what they have. Some are finding that on occasion there’s a mishap, and can see an simple solution, and are asking the supplier if that solution is reasonable to implement. You’ve done this same thing yourself on this very forum.

Thanks @wmoy, it works! No adverse impact on bitsetter. And it’s trivial to change the y soft limit if I need the extra 100 mm (4") at the front of the wasteboard. @gdon_2003’s plastic tray with foam solution (held with magnets to my supplementary steel supports) is working well, too. I may end up making the foam tray my go-to solution for now, but it’s good to know that I have the ability to move the load position back if/when the situation calls for it by sacrificing the front 4" of y travel. I’ve also started incorporating @jepho’s tool change hand positioning/technique, reducing the odds that I ever need to depend on the foam or the tool position as fallbacks. It’s wonderful that there’s so much collective knowledge/experience to draw from on this forum!

Winston Moy’s suggestion is a good one for your positioning over the wasteboard. The only draw back is you cannot jog past the limits you set in the configuration with CM. You could always change the configuration if you needed to jog out over the front of the bed. When you set the limits in the configuration CM is controlling your jogging limits. Soft limits are not used with CM. If you use a different gcode sender then the GBRL settings are adhered to but CM overrides the settings and sets its own will on the Shapeoko. So you can increase travel and/or decrease travel with the CM configuration and you can change the configuration anytime you want but if using CM it is in control of your Shapeoko.

Earlier I mentioned that you could not jog past the limits set in the configuration with CM. However if you send gcode that tries to exceed the physical limitations of your Shapeoko the gcode will try to execute resulting in the physical limit being hit and your stepper and belt will grind away. Do not confuse what CM does and what gcode does.

1 Like

You’re right! I’ve used the soft limits before in another GRBL system I built/programmed, so I was editing the travel dimensions in settings and clicking “Send Configuration Data” each time, assuming that they were soft limits. After reading your message, I ran a test where I set it y_max = 850, sent config data, initialized, jogged to center, then set y_max=750 and clicked “OK” without sending config data again. This ensured that GRBL was unaware of the change. And sure enough… when I jogged to the front of the machine, it stopped at y=750 mm.

Thanks for pointing that out! You saved me having to reinitialize every time, and I learned something new about the software. Now it’s trivially easy to change limits, even when loading a tool for a job overhanging the front.

1 Like