Boxford Improvement Thread

[Toby Williams]

Lot's of threads about the Boxford about, however, none really seemed the right place. 

Last night, Richard and I spent a couple of hours untangling (literally and metaphorically) the wiring in the boxford. This included finding a DFA relay. Overall, got a very good understanding of the wiring and voltages involved. We then returned it to full working state and closed it up.

Next Tuesday we are going to try and drive one of the axis around, using Richard's spare smoothie clone and some electronics wizardry (the logic level of the stepper drivers is 24V, incompatible with the 3.3 of the smoothie).

If that goes well, we'll look closer at things like e-stop and spindle control next.

All that are interested are welcome to help out in our adventure.

Toby

[Richard Ibbotson]

I added information in an electronics section to the Boxford VMC Wiki Page on what we found:

http://rlab.org.uk/wiki/Boxford_260_VMC

 

No big surprises, but good to get it down and to share. Maybe Toby can add some of his photos too.

 

Richard

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.
To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Richard Ibbotson]

Separate post to avoid mixing this opinion with the fact work done by Toby and me.

 

Why should we make changes to the Boxford?

 

The first question I guess is whether it is useful as it is? The current setup is fine for learning, but of much less value to produce any finished article. Those using the machine have all learnt a massive amount, but still struggle to follow a basic design process from concept to finished product. Limitations in the software nearly always result in compromises of the design to meet the limitations of the machine, OK for learning, but not for getting stuff made. So as is, it might remain a project or novelty, but not a tool for the space. In time, I think eventually it would be hard to justify the space taken compared to the value as a tool.

 

I think it important to understand, that CNC Milling is a complex process, a CNC Mill will never be an easy tool to use. Like other 3D CNC tools, we know that competency of 3D design technique and software is a prerequisite. Also, knowledge of milling in general and appreciation of manual milling is required. Many hackspaces do not induct on CNC mills until users are trained on manual mills, this is usually true in industry too. What I am saying is that if we make the Boxford easier to use, there is still a significant learning curve for new users. If we were to improve the VMC260, there will still be an entry learning level to usage (maybe 10 to 50 hrs), and so limited capability to make parts for casual users.

 

Based on my evaluation of the VMC260, which has a solid mechanical structure with CNC control, my opinion is that improvements can be made to turn the VMC260 into a productive tool, all be it that the entry level for users may be high.

 

Costs of the improvements may be possible at around £100, but I think a figure of £500 is more realistic.

 

Open to discussion here on group, or at rLab.

 

Richard

[Laurence Rochfort]

I completely agree with all of that, Richard.

In particular, I don't think that designing for or running a CNC mill will ever be a process that requires no experience, intuition or experimentation.

It will never be the laser cutter, which for all intents and purposes is a printer/plotter.

People should adjust their expectations and take that into account when considering the cost of options re the mill.



Thus far on the Boxford, I've made some small brackets and adapters, nothing complicated, just some facing, pockets, channels and holes; maybe 4 or 5 tool changes. However, every time the Boxford software hung and didn't prompt for later tool changes.

It's also not repeatable setting up datums, which would make swapping the vice a pain.


Whilst I think swapping the innards of the Boxford is doable I question the return on investment.

The hardware is much more capable than the software to be sure, but how capable is it compared to £500 replacement hardware and professional control software?

The software is critical, as demonstrated by the Boxford software.

Alex showed me the de facto open source mill control software and frankly it looked like a teenager had thrown a UI toolkit at the screen. Complex processes don't require horrendous UX, in fact the inverse is true.

We don't want to be in the same situation as with the Piranha where the software options are limited, so for me it's a priority that the replacememt is well known and can just plug in to Fusion or at least understand new/standard dxf or gcode.


As much as this is a hack space, it is also a maker space and I don't think the Boxford in any configuration will ever be a maker's machine. I don't see how I'd make more complicated projects on it than I already have. I also don't see how projects wouldn't require a great deal of time.

Ian has done outstanding work downstairs where every tool just works if you have the skill, with no hackery or maze of accomodations to navigate.

As you say, consisldering the floor space of the Boxford I don't think it meets the standard of the other tools downstairs.

Laurence

[Malcolm Napier]

> However, every time the Boxford software hung and didn't prompt for later tool changes.
> It's also not repeatable setting up datums, which would make swapping the vice a pain.

For me these are showstoppers to me using the Boxford.

> The hardware is much more capable than the software to be sure, but how capable is it compared to £500 replacement hardware and professional control software?

Is that £500 plus the cost of professional control software?

> We don't want to be in the same situation as with the Piranha where the software options are limited, so for me it's a priority that the replacememt is well known and can just plug in to Fusion or at least understand new/standard dxf or gcode.

I am not clear what you are saying here. I have had issues with the Pirhana (that are documented elsewhere in this Google Group) but they are with the hardware needing maintenance/like for like replacement of components. I have never found the software toolchain to be a limiting factor and am not aware of major issues other than it does not seem possible to move forward from RDWorks 7 to RDWorks 8.

However, I do think that the continued presence of the Boxford in the space needs review. The considerable amount of work that people have done with this machine means that any decision is likely to be an informed one. However, has anyone contacted James to establish his views? It is his machine and, as far as I am aware, is on loan. So we need to make arrangements with him to return it.

I have found contacting him to be a hit and (mostly) miss proposition over the last 12 months or so, but as far as I know he is still around - albeit very busy. Is he still a card holding member?  

[Laurence Rochfort]

I am not clear what you are saying here. I have had issues with the Pirhana (that are documented elsewhere in this Google Group) but they are with the hardware needing maintenance/like for like replacement of components. I have never found the software toolchain to be a limiting factor and am not aware of major issues other than it does not seem possible to move forward from RDWorks 7 to RDWorks 8. 

Rdworks 7 doesn't support dxf newer than r15 and there are issues with that.

There's nothing wrong with rdworks, it's a workable toolchain dxf issues asside, but being locked to a single version of a single piece of software isn't a desirable situation for a new machine.

--
You received this message because you are subscribed to a topic in the Google Groups "Reading Hackspace" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/reading-hackspace/u0ICTTql_a4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to reading-hacksp...@googlegroups.com.

[Malcolm Napier]

OK. I hadn't appreciated that. Thanks.

What does r15 get me that r14 doesn't?

I agree with the principle that being locked into a single piece of software isn't desirable. However, the Pirhana illustrates that such a situation can be acceptable (or good enough).

Which brings us back to my question - Is that £500 plus the cost of professional control software?

And a supplementary:- If we need professional control software (which will presumably cost £ or even £££) won't the developers be trying to lock us in?

[Laurence Rochfort]

Newer dxf has better expressions for curves and metadata. However, it's more that we're now on release 2007 of dxf and CAD software is increasingly not producing r15.

All good questions, and I don't honestly know.

If we bought a used "pro" machine it could come with pro software, which would be the same situation as the Piranha.

I think it's worth evaluating other open source software because there could be some that are user friendly.

[Mr E]

Hi,

Back to the Boxford.  The way I see it is, it's physically a more then good enough machine.  The question of if it is worth spending £500 on converting it to a more modern software tool chain - Well assuming James doesn't object and I very much doubt he would.  There is nothing you can buy for £500 that comes close in terms of being able to mill metal in a CNC fashion.  With regard to costs - last time I heard anything, rLab's finances are less then poor - I think it would be a worthy upgrade, to as you say - get to the point, you can drive it from Fusion 360.

I don't see why we need to get rid of a usable machine, which people are engaged with.  Me, Gavin, Toby have all succeeded in making something functional with it.  In terms of being a "entry level" machine - no I don't think it is, but I don't think we need to shy away from that either, I think machines like this, help pull rLab'ers together in more team style to get things done, when you don't fully understand things.

That said I'm from the 3D printing world, and have managed with help from people here, and reading and sheer experimentation to "get things done", so I don't think it's beyond the whit of other members too.

All the best.

Rupert

To unsubscribe from this group and all its topics, send an email to reading-hackspace+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hackspace+unsubscribe@googlegroups.com.

[Alex Gibson]

Please read the below as a positive attempt to help get a great outcome for rLab, using the enthusiasm that clearly exists to improve our cnc milling capability:

 

I am very concerned that before any serious game plan is developed and agreed with rLab members, and the owner of the tool, risks are already being taken by opening it up and making changes, however competent and well intended. 

 

People may be unimpressed by the VMC260 as it currently is, but

a)      It works

b)      It is being used by several members, on what looks like an increasing curve of productivity.  

c)       It is in a fungible state – it can be sold as a Boxford VMC260 in working order, and was bought in that state by James.

 

Meanwhile, the Shapeoko is sitting there, still an incomplete project but very nearly back operational, much better than it was before, and could very shortly be complete and useful in its own right, but also the perfect staging environment to test any new toolchain being developed for the Boxford. 

 

I suggest those who are interested in improving the Boxford join in on the Shapeoko completion first, and use it to get solid, practical knowledge, especially of the END-TO-END practicalities of the toolchain.  This could then be transferred to a solid, planned overhaul of the Boxford, which could be reasonably expected to ‘just work’ upon completion.

 

I want the same great outcome others do – a working, reliable, easy(ish)-to-use milling machine, that we can treat as a dependable tool.  I am not heavily invested in the Boxford, I’m not interested in slowing anyone down here.  But let’s work in a positive, risk-aware way.  We have to learn from past mistakes, recognise that unless planned and to some extent practised in advance, what seems like a simple change in theory will often throw up its own issues, turning a tool into a project, no longer standard, no longer usable even to learn on, and at constant risk of parts being lost, people wandering off because life outside rLab happens, and then the inevitable calls to throw it away and put something newer and shinier in its place.

 

At minimum I would ask that James is aware, and that every single change we make is documented, leaves the machine in a usable working state (as much as it is right now), and can be immediately rolled back if that’s not possible.

 

Anyone working in IT will recognise this as good change/release management practice.

 

On the plus side, I can see the real potential for this to be a machine we can drive and maintain ourselves with open source hardware, and a powerful, straightforward integration with tools I can use at home like Fusion360.

 

I remember the Denford micro-mill, my first attempts to make things on it, and think of all the things I could have made on it if we had left it as it was.  Let’s ensure we don’t repeat this, and remember the C17 revision of the Hippocratic oath:

“First, do no harm”

--

You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

[Malcolm Napier]

I have a couple of other reasons to try to get in touch with James. I have tried to do so this morning and will do so again over the next few days. However, I have no problem with others trying to reach him as well. I know that my email addresses have fallen foul of his spam filtering on more than one occasion. If anyone does succedd can they tell him that I am trying to get in touch and would appreciate a call.

I am not qualified enough to opine on the quality of the Boxford hardware but I suspect it is physically a more than good enough machine - as long as it is regarded as "large enough" by the community that wish to use it . (My issue with the Shapeoko is that is is too small for my requirements (annoyingly by 20-50mm) but no-one sought my opinion before it was bought.)

However, given other (including recent) discussions over the history of the Shapeoko and the vending machine, I think that we will need a clear plan of what the modifications will involve and how long they will take. We will also need James agreement to those modifications or they will need to be made in a reversible manner and reversed should James request his machine back. 

--

You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

[Richard Ibbotson]

Alex, that is a load of misinformation and pure rubbish.

 

The VMC 260 does not work to a level to be an effective Hackspace tool which can be integrated to Fusion 360. It is being used for experimentation and learning, when that is over we should reconsider the value if we do not upgrade. I did recommend this was done before it was brought into the hackspace, but was overruled at the time.

 

The Shapeoko has been a project management and documentation tragedy from the beginning, and for all your cries below, you are still considered significantly to blame for that. It certainly does not look, “very near back operational”, has it ever been operational? Why was good practice not followed on this?

 

There is already more documentation on the wiki about the Boxford than the Shapeoko. It is up to date including what we found out last Tuesday. Our work on Tuesday was to find out as much as we could about the state, make it safer, document ready for further experiment, and to create a recovery state.

 

We are good with the mechanical part of the machine, but will next determine if the existing stepper drivers and the motor controller will work with alternate controllers like Smoothie or Mach3. Then we can make a sensible upgrade proposal to members and to James.

 

We have tried and continue to try to contact James. His own entry on the Wiki already suggested an upgrade should be considered.

 

Not sure what you are talking about on the Denford. It was even more restricted than the Boxford. It was never modified, just sold to buy the Shapeoko. I continued to use it with the old software till it disappeared.

 

From: reading-...@googlegroups.com [mailto:reading-...@googlegroups.com] On Behalf Of Alex Gibson
Sent: Saturday, June 24, 2017 2:32 PM
To: reading-...@googlegroups.com
Subject: RE: [RDG-Hack] Boxford Improvement Thread

 

Please read the below as a positive attempt to help get a great outcome for rLab, using the enthusiasm that clearly exists to improve our cnc milling capability:

 

I am very concerned that before any serious game plan is developed and agreed with rLab members, and the owner of the tool, risks are already being taken by opening it up and making changes, however competent and well intended. 

 

People may be unimpressed by the VMC260 as it currently is, but

1. It works

1. It is being used by several members, on what looks like an increasing curve of productivity.  

1. It is in a fungible state – it can be sold as a Boxford VMC260 in working order, and was bought in that state by James.

 

Meanwhile, the Shapeoko is sitting there, still an incomplete project but very nearly back operational, much better than it was before, and could very shortly be complete and useful in its own right, but also the perfect staging environment to test any new toolchain being developed for the Boxford. 

 

I suggest those who are interested in improving the Boxford join in on the Shapeoko completion first, and use it to get solid, practical knowledge, especially of the END-TO-END practicalities of the toolchain.  This could then be transferred to a solid, planned overhaul of the Boxford, which could be reasonably expected to ‘just work’ upon completion.

 

I want the same great outcome others do – a working, reliable, easy(ish)-to-use milling machine, that we can treat as a dependable tool.  I am not heavily invested in the Boxford, I’m not interested in slowing anyone down here.  But let’s work in a positive, risk-aware way.  We have to learn from past mistakes, recognise that unless planned and to some extent practised in advance, what seems like a simple change in theory will often throw up its own issues, turning a tool into a project, no longer standard, no longer usable even to learn on, and at constant risk of parts being lost, people wandering off because life outside rLab happens, and then the inevitable calls to throw it away and put something newer and shinier in its place.

 

At minimum I would ask that James is aware, and that every single change we make is documented, leaves the machine in a usable working state (as much as it is right now), and can be immediately rolled back if that’s not possible.

 

Anyone working in IT will recognise this as good change/release management practice.

 

On the plus side, I can see the real potential for this to be a machine we can drive and maintain ourselves with open source hardware, and a powerful, straightforward integration with tools I can use at home like Fusion360.

 

I remember the Denford micro-mill, my first attempts to make things on it, and think of all the things I could have made on it if we had left it as it was  Let’s ensure we don’t repeat this, and remember the C17 revision of the Hippocratic oath:

“First, do no harm”

 

 

From: reading-...@googlegroups.com [mailto:reading-...@googlegroups.com] On Behalf Of Laurence Rochfort
Sent: 24 June 2017 13:03
To: Reading Hackspace
Subject: Re: [RDG-Hack] Boxford Improvement Thread

 

Newer dxf has better expressions for curves and metadata However, it's more that we're now on release 2007 of dxf and CAD software is increasingly not producing r15.

[Alex Gibson]

What a thoroughly unpleasant ad hominem response.

[Mr E]

Gentlemen - This is the Boxford thread, not the general bitching at each other thread, or who's project's gone less well.  We might have to hand out some handbags if this keeps up.

On 24 June 2017 at 16:04, Alex Gibson <al...@alexgibson.net> wrote:

What a thoroughly unpleasant ad hominem response.

 

 

From: reading-hackspace@googlegroups.com [mailto:reading-hackspace@googlegroups.com] On Behalf Of Richard Ibbotson
Sent: 24 June 2017 15:27

To: reading-hackspace@googlegroups.com

To unsubscribe from this group and all its topics, send an email to reading-hackspace+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hackspace+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hackspace+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hackspace+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hackspace+unsubscribe@googlegroups.com.

[Gavin]

My views on the Boxford:

I think that we are still in a position where we are trying to determine the causes of some of its issues, and assess its ultimate value to the hackspace. We are making rapid progress and the next logical step is to try patching in some different driver electronics and try and isolate out the Boxford software which is not great and may be the cause of some of the issues.

Whether it currently is / is capable of being an effective hackspace tool depends on your expectations - it would be useful to know what people’s expectations are.

- it’s not (and no milling machine will ever be) a walk-up, turn-on and make-thing tool - it requires training, it requires some reading on milling, and it requires some practice. Using it in manual mode first to understand the basics is a good idea;
- the spindle speed is currently limited to 2000rpm, though is notionally capable of 3500rpm. We know if you put some thought into feed-rates it produces pretty good results with ali, will cut wood and foam. But we are limited the size of tools we can use, we have not tried mild steel and do not have a proper coolant system set up. A faster spindle would make producing things faster and allow us to use a wider range of materials;
- we think that what it produces is dimensionally good, but as already mentioned, there is something wrong with the datum setting process. We don’t yet know what is causing this - perhaps the aged electronics / wiring, perhaps the serial connection, perhaps the software, perhaps something else. Investigations with an alternative controller will likely shine light on the issue. As it stands it is not a showstopper in terms of making something, but is a bit irritating and requires work to set up if the vice is moved.
- the machine is mechanically sound and could likely be used as the basis for a more capable machine, but we obviously need to assess the cost and benefits of doing so;
- we have a working Fusion 360 post-processor based on the generic Boxford one, but modified to trap some G-code commands that are incompatible with the VMC260, and deal with some issues with rapid movements into the workspace.

The control, rigour and collaborative nature of the work done to commission and test the machine has been the best that I have seen at rLab. There are five people actively engaged in work on it, and others have contributed advice / parts. Progress is being reported to the list, software on GitHub, wiki kept up to date (though I need to update the section on the Post Processor).

Depending on response from James, I think we need to assess the value of the machine as it stands and/or the value of an upgraded machine. Personally I think it adds value and given (i) the work that is begin done on it (there’s no sign of appetite to work on it waning), and (ii) its potential, I think it is earning the floorspace that it is sitting on at the moment. I accept that needs to be kept under review though.

[Richard Ibbotson]

I do judge the Boxford to be a physical good machine too. Probably comparable with something like this:

http://www.arceurotrade.co.uk/Catalogue/Machines-Accessories/CNC-Milling-Machines/Sieg-KX3-CNC-Hobby-Mill/Sieg-KX3-Hobby-CNC-Mill

 

Large second-hand mills come available, but it can be a minefield for support, spares and compatibility with modern software packages

 

Gavin and Toby have worked on fully understanding the Fusion 360 Post Processor. This is a powerful Javascript module which converts the Fusion 360 paths to g-code. Very nice and open. Most of work to date has been to make g-code tame enough for the Boxford, but co-processors are available for Smoothie and Mach3 too.

 

Toby and I are investigating if the existing stepper drivers and DC motor controller are suitable for use with other logic controllers. It is likely they are, but if not, we would have a cost of around £250 to upgrade. The existing stepper drivers only offer full or half step, but that does not appear to be a limitation. Existing power supply likely adequate

 

We understand there is a later version of the Boxford Software, but we consider it unlikely to offer much improvement. Also support from Boxford and the community is poor.

 

I guess you already understand that the tool chain is 3d Software(Fusion 360), G-Code converter( Fusion 360 Post Processor), gcode sender and control panel, and finally a firmware control board in the mill. Much the same as in 3D printing, though some components might be shared in the same application.

 

So far, we have only considered Fusion 360 as the 3D CAD Package.

 

For the other parts, we have looked at MACH3 + USB/Ethernet control board, and bCNC plus Smoothie. Both appear to have the full functions we are looking for including tool change, multiple workpieces, and probing. They both appear to support the spindle and the steppers we have and possible suitable alternatives

 

Mach3 has the advantage of lots of things having been made using it, great support community, responsive to improvements. Disadvantage is that it is not open source, though scripting gets around quite a lot of that practically. We don’t have experience of Mach3 ourselves and are looking for someone who can advise. The Mach3 software is around £150, and a compatible interface and control board (USB or Ethernet) around the same. We would not use parallel port control!

 

bCNC has about the best reputation right now as an open source gcode sender and control panel for CNC. It is open source and written in python. We checked it out with the latest CNC Smoothieware release on an AZSMZ board and display. The two appear well integrated, previously I had only used Smoothieware with LaserWEB. We will use this configuration to test the stepper and spindle drive next Tuesday. The AZSMZ costs around £60 including display and is also open source and highly configurable. Grbl is not really an option due the spindle drive and feedback requirement.

 

We are considering a mechanical upgrade to add probing for the tool length.

 

All are welcome to join discussion or activity.

Richard

 

 

From: reading-...@googlegroups.com [mailto:reading-...@googlegroups.com] On Behalf Of Mr E
Sent: Saturday, June 24, 2017 2:04 PM
To: Reading Hackspace <reading-...@googlegroups.com>

Subject: Re: [RDG-Hack] Boxford Improvement Thread

 

Hi,

--

You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

 

--

You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

[Toby Williams]

The ER25 collets and associated tool holders have been brought to the space, although the company sent one ER32 and ER25. I will get the ER32 sorted out.

I have also purchased a 16mm end mill to make better use of the 16mm End Mill holder now that we don't need to use the adapters for smaller bits.

The next session to look at the Boxford is Tuesday night. Everybody is welcome to join in and contribute. The Boxford will be returned to a working state before the end of the evening.

[Gavin]

Wow. £5k!

[Ben Scammell]

It is nice though!

Sent from my iPhone

[Richard Ibbotson]

Just an additional point on the Mach3 and bCNC software options.

 

Mach3 runs under Microsoft windows only. As we would use the USB or the Ethernet interface, then the latest versions of Windows 32 and 64bit appear to be OK.

 

bCNC runs on Microsoft Windows and linux providing Python 2.7 and PySerial are installed. It is also said to run well on a Raspberry Pi, though we will have to try that.

 

I will put this into the wiki with some links.

 

From: reading-...@googlegroups.com [mailto:reading-...@googlegroups.com] On Behalf Of Richard Ibbotson
Sent: Saturday, June 24, 2017 4:38 PM
To: reading-...@googlegroups.com
Subject: RE: [RDG-Hack] Boxford Improvement Thread

 

I do judge the Boxford to be a physical good machine too. Probably comparable with something like this:

http://www.arceurotrade.co.uk/Catalogue/Machines-Accessories/CNC-Milling-Machines/Sieg-KX3-CNC-Hobby-Mill/Sieg-KX3-Hobby-CNC-Mill

 

Large second-hand mills come available, but it can be a minefield for support, spares and compatibility with modern software packages

 

Gavin and Toby have worked on fully understanding the Fusion 360 Post Processor. This is a powerful Javascript module which converts the Fusion 360 paths to g-code. Very nice and open. Most of work to date has been to make g-code tame enough for the Boxford, but co-processors are available for Smoothie and Mach3 too.

 

Toby and I are investigating if the existing stepper drivers and DC motor controller are suitable for use with other logic controllers. It is likely they are, but if not, we would have a cost of around £250 to upgrade. The existing stepper drivers only offer full or half step, but that does not appear to be a limitation. Existing power supply likely adequate

 

We understand there is a later version of the Boxford Software, but we consider it unlikely to offer much improvement. Also support from Boxford and the community is poor.

 

I guess you already understand that the tool chain is 3d Software(Fusion 360), G-Code converter( Fusion 360 Post Processor), gcode sender and control panel, and finally a firmware control board in the mill. Much the same as in 3D printing, though some components might be shared in the same application.

 

So far, we have only considered Fusion 360 as the 3D CAD Package.

 

For the other parts, we have looked at MACH3 + USB/Ethernet control board, and bCNC plus Smoothie. Both appear to have the full functions we are looking for including tool change, multiple workpieces, and probing. They both appear to support the spindle and the steppers we have and possible suitable alternatives

 

Mach3 has the advantage of lots of things having been made using it, great support community, responsive to improvements. Disadvantage is that it is not open source, though scripting gets around quite a lot of that practically. We don’t have experience of Mach3 ourselves and are looking for someone who can advise. The Mach3 software is around £150, and a compatible interface and control board (USB or Ethernet) around the same. We would not use parallel port control!

 

bCNC has about the best reputation right now as an open source gcode sender and control panel for CNC. It is open source and written in python. We checked it out with the latest CNC Smoothieware release on an AZSMZ board and display. The two appear well integrated, previously I had only used Smoothieware with LaserWEB. We will use this configuration to test the stepper and spindle drive next Tuesday. The AZSMZ costs around £60 including display and is also open source and highly configurable. Grbl is not really an option due the spindle drive and feedback requirement.

 

We are considering a mechanical upgrade to add probing for the tool length.

 

All are welcome to join discussion or activity

Richard

It's also not repeatable setting up datums, which would make swapping the vice a pain.


Whilst I think swapping the innards of the Boxford is doable I question the return on investment.

The hardware is much more capable than the software to be sure, but how capable is it compared to £500 replacement hardware and professional control software?

The software is critical, as demonstrated by the Boxford software.

Alex showed me the de facto open source mill control software and frankly it looked like a teenager had thrown a UI toolkit at the screen. Complex processes don't require horrendous UX, in fact the inverse is true

[Toby Williams]

I'm putting bCNC on a raspberry pi as we speak, will report on how I get on.

Toby

[Toby Williams]

It went well

[Toby Williams]

I'm heading down to the space now, everybody is welcome to come join in.

[gavi...@gmail.com]

Be down later

 

From: 'Toby Williams' via Reading Hackspace

Sent: 27 June 2017 18:37
To: Reading Hackspace
Subject: Re: [RDG-Hack] Boxford Improvement Thread

 

I'm heading down to the space now, everybody is welcome to come join in.

 

--

[Toby Williams]

Boxford Update 27/6

A good evening of hacking took place. The Boxford was restored to original working state at the end of the night.

We achieved:

• Ran the spindle using a potentiometer, Richard measured the speeds and frequency of the pulses. 50 teeth on the encoder and we can get the spindle up to ~4000 rpm (although it makes hell of racket).
• Mostly successfully drove around the x-axis using smoothie. There where definite problems with accuracy and consistency, we believe this will be fixed by fine tuning the pulse width and other such parameters. 
• More documentation of wiring (including the mystery pin outs on the Gemini)

Next steps

• Sort out the configuration of the smoothie board to control the steppers correctly. May involve testing with more modern drivers. Until this is properly resolved, no permanent changes to the Boxford will be considered. 
• Eyeball the belt drives, at least on the X. 
• Swap over/replace the x-axis limit, as its the only one that appears to be failing to be accurate.
• More bCNC/Smoothie tool change research - z-height probing

Thank you to all that joined in, was fun!

 

[Alex Gibson]

Cool!  Thanks for the update Toby.  The upfront comment about restoring original state was smart ;) Great to hear about what sounds like a fun and productive night! 

What sort of limit sensor is currently on the x – and is it the same as Y?

 

From: 'Toby Williams' via Reading Hackspace [mailto:reading-...@googlegroups.com]
Sent: 28 June 2017 12:23
To: Reading Hackspace
Subject: Re: [RDG-Hack] Boxford Improvement Thread

 

Boxford Update 27/6

--

[Gavin]

This is great progress - thanks all! 

[Gavin]

Same inductive sensors on x and y, On x looking for notch in the underside of the ways, on y looking for the end.

http://www.seltec.co.uk/products/details/10617.html

[Alex Gibson]

I’ve got a couple of those in stock – will check the exact specs of mine, is that the exact brand/model on the Boxford, or just one of similar spec Gavin? 

 

If we’re confident the issue is only/predominantly on X, which matches what Laurence & I saw, then physically swapping over X and Y sensors and re-testing would be good – if we can make X good and Y become less reliable, it positively confirms the sensor is wonky. Other possible outcomes are then also useful diagnostic indicators.

[Gavin]

It’s that particular one fitted to the machine

[Richard Ibbotson]

I am seeing that sensor as having PNP output. Is that correct? Had thought it NPN. Can support either, just need to confirm for interface.

Sent via Richard's Phone

--

[Richard Ibbotson]

I looked up some stuff on the steps per mm, you may want to do some more experiment or investigation.

 

The data on the wiki says a resolution of .005mm, which would be 200 steps per mm.

 

A thread on modifying a VMC190: http://www.mycncuk.com/threads/10394-Boxford-VMC-190

Says they found 2.5mm pitch and possibly 20:16 or 20:12 gearing on belt from stepper to ball screw. This would be 200 steps per mm or 266 steps per mm.

 

We were getting inconsistent results and used 212 steps per mm.

 

The spec also says max 2000mm per min. Which should be in the capability of the stepper control.

 

Not really seeing a mechanical issue, though I think we might improve measuring setup. Maybe we should look further at the interface signal for integrity and noise. Maybe make a gcode file to move back and forth, to see if we are missing or adding steps.

 

Sorry to all those who not have all background to this. Toby or others will fill you in, or you may want to join more tests (Toby has Smoothie(AZSMZ)) Just wanted to share this before my vacation.

Richard

 

From: reading-...@googlegroups.com [mailto:reading-...@googlegroups.com] On Behalf Of Richard Ibbotson
Sent: Wednesday, June 28, 2017 2:31 PM
To: reading-...@googlegroups.com
Subject: Re: [RDG-Hack] Boxford Improvement Thread

 

I am seeing that sensor as having PNP output. Is that correct? Had thought it NPN. Can support either, just need to confirm for interface.

[Toby Williams]

Next week I hope to try hooking up a different driver and/or decoupling the motor from the pulley. It would be nice to isolate the source the horrible grinding noise (it's not that bad) when the axis are moving slowly (happens when the boxford is controlling the motors as well). It might be that, that is what those steppers sound like when driven slowly at half step.

Given the ease of access to the x-stepper, it would be easy to do this experiment and reassemble the axis in one evening.

Toby 

[Gavin]

Can’t be certain it is exactly the same part. Certainly the right manufacturer and part number as Laurence and I noted when we took it off the machine to inspect, but there maybe variants.

[Laurence Rochfort]

The code on the part is IS-3002-ANOG. As Gavin says, not sure about variants.

http://www.ifm.com/products/us/ds/IS5003.htm

http://www.seltec.co.uk/products/details/10617.html

Can you test whether its syncing or sourcing with a multimeter? If it registers a voltage change would that make it PNP/sourcing?

[Laurence Rochfort]

Hi all,

I came in to do some milling today and the Boxford is having some movement issues.

The only direction I can get any of the Axis to move is right or X-, but very slowly with no response to the feed rate knob.

I don't have time to look at it now because I've got to run to the office, but thought I should mention it.

Cheers,

Laurence.

[Toby Williams]

Thank you for bringing up, this is weird as we were driving the axis round just fine on Tuesday night. Rupert did have some concerns over the integrity of the x-belts, however that wouldn't affect the other axis. 

Given we only touched one connector on Tuesday night, it shouldn't be hard to diagnose. I probably won't be in until Monday night to look at it though. 

Another option could be that the x-limit switch has died and the boxford will only let the x-axis move whilst the limit is active. 

If anybody gets a chance to look at it before me, please report back.

T

[Laurence Rochfort]

The Z axis isn't responding either, so I'm inclined to think it's electrical and just a glitch from poking at it.

I'll see how things pan out at work. If I have free time this afternoon I'll have a look.

--
You received this message because you are subscribed to a topic in the Google Groups "Reading Hackspace" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/reading-hackspace/u0ICTTql_a4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to reading-hacksp...@googlegroups.com.

[Toby Williams]

I suspect you know, but when the limit switches are triggered, the software stops all the axis form moving apart from the axis that has been triggered.

T

[Mr E]

I know when at the end of the night, after we'd disconnected everything, I did move all the X,Y and Z around using the buttons on the front, at varying feed rates successfully.  So I don't think we broke anything.  This was because we seemed to be stalling the X motor, and I wasn't sure if it was stalled or spinning because we'd buggered up the drive belt.  As it worked ok - I think now we probably were driving it too fast/wrong, rather than a mechanical problem.

I have seen before this kind of thing - as Toby mentioned - resetting/power cycling, and see if you can get it to move (I usually do the Z first), then it all comes alive.  Odd little control firmware that it is!  Roll on the new one, which hopefully will be a bit more sane.

Good luck!

Rupert

On 30 June 2017 at 10:40, 'Toby Williams' via Reading Hackspace <reading-...@googlegroups.com> wrote:

I suspect you know, but when the limit switches are triggered, the software stops all the axis form moving apart from the axis that has been triggered.

T

--

You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hackspace+unsubscribe@googlegroups.com.

[Laurence Rochfort]

I tried the usual power cycle and tapping all the direction buttons in case it was at overtravel like usual, but no joy.

The x axis moved right, but not left. Nothing at all on Y or Z.

The spindle responded fine.

I don't think anything's broken, but I've made 10 parts on it now and it's not behaving like it normally does, quirks and all :-)

As an aside, are you saying there's a drive belt between the X-axis motor and the drive screw?

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to a topic in the Google Groups "Reading Hackspace" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/reading-hackspace/u0ICTTql_a4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to reading-hacksp...@googlegroups.com.

[Mr E]

> As an aside, are you saying there's a drive belt between the X-axis motor and the drive screw?

Toby apparently has had it apart, and yes - much to my surprise - you'd really have thought that would be geared.

On 30 June 2017 at 12:35, Laurence Rochfort <laurence...@gmail.com> wrote:

I tried the usual power cycle and tapping all the direction buttons in case it was at overtravel like usual, but no joy.

The x axis moved right, but not left. Nothing at all on Y or Z.

The spindle responded fine.

I don't think anything's broken, but I've made 10 parts on it now and it's not behaving like it normally does, quirks and all :-)

As an aside, are you saying there's a drive belt between the X-axis motor and the drive screw?

On Fri, 30 Jun 2017, 11:57 Mr E, <excha...@gmail.com> wrote:

I know when at the end of the night, after we'd disconnected everything, I did move all the X,Y and Z around using the buttons on the front, at varying feed rates successfully.  So I don't think we broke anything.  This was because we seemed to be stalling the X motor, and I wasn't sure if it was stalled or spinning because we'd buggered up the drive belt.  As it worked ok - I think now we probably were driving it too fast/wrong, rather than a mechanical problem.

I have seen before this kind of thing - as Toby mentioned - resetting/power cycling, and see if you can get it to move (I usually do the Z first), then it all comes alive.  Odd little control firmware that it is!  Roll on the new one, which hopefully will be a bit more sane.

Good luck!

Rupert

On 30 June 2017 at 10:40, 'Toby Williams' via Reading Hackspace <reading-hackspace@googlegroups.com> wrote:

I suspect you know, but when the limit switches are triggered, the software stops all the axis form moving apart from the axis that has been triggered.

T

--

You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hackspace+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to a topic in the Google Groups "Reading Hackspace" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/reading-hackspace/u0ICTTql_a4/unsubscribe.

To unsubscribe from this group and all its topics, send an email to reading-hackspace+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hackspace+unsubscribe@googlegroups.com.

[Mr E]

Next thought - we were using it with the back open - I wonder if one of the connectors popped off when we closed it up?

[Toby Williams]

Yes, there's a belt.

Toby

[Andy Noyes]

Gears = backlash, which would negate having ballscrews.

Accurate machines always have toothed belts, sometimes even dual belt reduction.

On Friday, 30 June 2017 12:47:48 UTC+1, Mr E wrote:

> As an aside, are you saying there's a drive belt between the X-axis motor and the drive screw?

Toby apparently has had it apart, and yes - much to my surprise - you'd really have thought that would be geared.

On 30 June 2017 at 12:35, Laurence Rochfort <laurence...@gmail.com> wrote:

I tried the usual power cycle and tapping all the direction buttons in case it was at overtravel like usual, but no joy.

The x axis moved right, but not left. Nothing at all on Y or Z.

The spindle responded fine.

I don't think anything's broken, but I've made 10 parts on it now and it's not behaving like it normally does, quirks and all :-)

As an aside, are you saying there's a drive belt between the X-axis motor and the drive screw?

On Fri, 30 Jun 2017, 11:57 Mr E, <excha...@gmail.com> wrote:

I know when at the end of the night, after we'd disconnected everything, I did move all the X,Y and Z around using the buttons on the front, at varying feed rates successfully.  So I don't think we broke anything.  This was because we seemed to be stalling the X motor, and I wasn't sure if it was stalled or spinning because we'd buggered up the drive belt.  As it worked ok - I think now we probably were driving it too fast/wrong, rather than a mechanical problem.

I have seen before this kind of thing - as Toby mentioned - resetting/power cycling, and see if you can get it to move (I usually do the Z first), then it all comes alive.  Odd little control firmware that it is!  Roll on the new one, which hopefully will be a bit more sane.

Good luck!

Rupert

On 30 June 2017 at 10:40, 'Toby Williams' via Reading Hackspace <reading-...@googlegroups.com> wrote:

I suspect you know, but when the limit switches are triggered, the software stops all the axis form moving apart from the axis that has been triggered.

T

--

You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to a topic in the Google Groups "Reading Hackspace" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/reading-hackspace/u0ICTTql_a4/unsubscribe.

To unsubscribe from this group and all its topics, send an email to reading-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

[gavi...@gmail.com]

I might be able to peer inside later – what should I be looking for?

 

From: Mr E
Sent: 30 June 2017 12:48
To: Reading Hackspace

Subject: Re: [RDG-Hack] Re: Boxford Improvement Thread

 

Next thought - we were using it with the back open - I wonder if one of the connectors popped off when we closed it up?

On 30 June 2017 at 12:47, Mr E <excha...@gmail.com> wrote:

> As an aside, are you saying there's a drive belt between the X-axis motor and the drive screw?

 

Toby apparently has had it apart, and yes - much to my surprise - you'd really have thought that would be geared.

On 30 June 2017 at 12:35, Laurence Rochfort <laurence...@gmail.com> wrote:

I tried the usual power cycle and tapping all the direction buttons in case it was at overtravel like usual, but no joy.

The x axis moved right, but not left. Nothing at all on Y or Z.

The spindle responded fine.

I don't think anything's broken, but I've made 10 parts on it now and it's not behaving like it normally does, quirks and all :-)

As an aside, are you saying there's a drive belt between the X-axis motor and the drive screw?

 

On Fri, 30 Jun 2017, 11:57 Mr E, <excha...@gmail.com> wrote:

I know when at the end of the night, after we'd disconnected everything, I did move all the X,Y and Z around using the buttons on the front, at varying feed rates successfully.  So I don't think we broke anything.  This was because we seemed to be stalling the X motor, and I wasn't sure if it was stalled or spinning because we'd buggered up the drive belt.  As it worked ok - I think now we probably were driving it too fast/wrong, rather than a mechanical problem.

 

I have seen before this kind of thing - as Toby mentioned - resetting/power cycling, and see if you can get it to move (I usually do the Z first), then it all comes alive.  Odd little control firmware that it is!  Roll on the new one, which hopefully will be a bit more sane.

 

Good luck!

 

Rupert

 

On 30 June 2017 at 10:40, 'Toby Williams' via Reading Hackspace <reading-...@googlegroups.com> wrote:

I suspect you know, but when the limit switches are triggered, the software stops all the axis form moving apart from the axis that has been triggered.

 

T

--

You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to a topic in the Google Groups "Reading Hackspace" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/reading-hackspace/u0ICTTql_a4/unsubscribe.

To unsubscribe from this group and all its topics, send an email to reading-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

[Laurence Rochfort]

The limit switches come in through grommets in the dividing wall and onto the board on the left as you look from the rear.

The stepper drivers are in a row on the left with large heatsinks. The cables for the steppers are also here.

[Mr E]

---------- Forwarded message ----------
From: "Mr E" <excha...@gmail.com>
Date: 30 Jun 2017 20:28
Subject: Re: [RDG-Hack] Re: Boxford Improvement Thread

To: "Laurence Rochfort" <laurence...@gmail.com>
Cc:

Fixed! 

One of the orange connectors on the large board on the left-hand side, had come off. 

Reattached and its back to itself :) 

Rupert 

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hackspace+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to a topic in the Google Groups "Reading Hackspace" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/reading-hackspace/u0ICTTql_a4/unsubscribe.

To unsubscribe from this group and all its topics, send an email to reading-hackspace+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hackspace+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hackspace+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to a topic in the Google Groups "Reading Hackspace" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/reading-hackspace/u0ICTTql_a4/unsubscribe.

To unsubscribe from this group and all its topics, send an email to reading-hackspace+unsubscribe@googlegroups.com.

[gavi...@gmail.com]

Great – cheers!

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to a topic in the Google Groups "Reading Hackspace" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/reading-hackspace/u0ICTTql_a4/unsubscribe.

To unsubscribe from this group and all its topics, send an email to reading-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to a topic in the Google Groups "Reading Hackspace" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/reading-hackspace/u0ICTTql_a4/unsubscribe.

To unsubscribe from this group and all its topics, send an email to reading-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

[Laurence Rochfort]

And they said your electronics PhD wouldn't come in handy!

Thanks for looking at it. Teach me to try and sneak in before work :)

[Toby Williams]

Good news!

[Toby Williams]

Boxford wiki updated with more findings.

[Toby Williams]

More Boxford work tonight, all hands welcome.

Key things to be worked on:

• Driving an axis with smoothie
• including potentially using a more modern driver to do some A-B comparisons with the existing drivers
• removing the stepper motor to check on the belt and try driving the step unloaded.
• Investigating the limit switches
• is the x-axis limited actually fcuked?
• Home for the boxford
• In the middle of the floor, completely in the way is unacceptable. Where can it be given more of a home in the short/medium term.
• Small air compressor/lubrication system upgrade/purchase
• The oil can needs to be returned to it's home.

[Laurence Rochfort]

I started milling a piece today for Jono, and after finishing the facing operation the Boxford just kept saying it was at axis overtravel, even though it really wasn't.

I tried several jobs that have worked find previously and it always just said axis overtravel.

Probably worth checking if it's there's something awry going on before you start with your investigations.

--

[Toby Williams]

Ok, first point of investigation then.

[Toby Williams]

Boxford has been given a home.

No x-Axis investigation I'm afraid, definitely an area I could use help with.

Some extra collets dropped off, spare 6mm and 4mm

[Toby Williams]

I'll be heading down to the space tonight for Boxford things.

Things to be worked on:

• x-Axis limit switch
  
• new controller with Richard
• giving the Boxford a home, v2.0 - i.e. getting through the door into the back room.
• messing around with material sizes/zero points.
• test cuts/ShapeOko z-axis milling.

[Richard Ibbotson]

I will be down.

 

Have replacement limit switch for X if needed.

 

Added the other 3 buffers drivers for Y and Z axis for smoothie to drive existing stepper drivers.

 

Found problem with smoothie if you set “ microseconds_per_step_pulse“ greater than 8 ( we had 10).  8 actually gives 10us, so expect we will now be OK.

 

Did not do the spindle speed yet, but now have schematic for the complete interface.

 

From: 'Toby Williams' via Reading Hackspace [mailto:reading-...@googlegroups.com]
Sent: Tuesday, July 11, 2017 1:14 PM
To: Reading Hackspace <reading-...@googlegroups.com>
Subject: Re: [RDG-Hack] Re: Boxford Improvement Thread

 

I'll be heading down to the space tonight for Boxford things.

--

You received this message because you are subscribed to the Google Groups "Reading Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

[Toby Williams]

Awesome. 

[Ian Petrie]

I think that if you

1. Take the door off the small room
2. Take the door stops off both sides
3. Cut through the door casing above and below the height of the widest point on the Boxford and remove the section

Then the Boxford will just go through the door and the door can be easily reinstated afterwards.

Just watch for the power cable on the back of the Boxford. That might need a grrove cut into the door casing on the other side.

[Toby Williams]

Boxford Update - Awesomeness

The Boxford is all closed up and tested software/driving around the head and the limit switches.

The smoothie ran the Boxford great. Here is some video that I took of a test. There is 'some' tool deflection. This means we can square up the remaining proof of concept to do list:

• Integrate the limit switches with the Smoothie.
• Persuade the Smoothie to produce 0-10v for the speed controller
• Take the pulses back in from the encoder

Once these things are done, we'll have a fully functioning cnc mill, operating off the Smoothie. We can then (not before!) begin to think about how we make the call, to swap the Boxford controller for Smoothie.

Richard is blasting through the electronics so far. However, if we make the change, then we will need to do things like make a new front panel (allowing us to keep the existing front panel intact) and coolant solenoids for which more help will be greatly appreciated.

T

[Laurence Rochfort]

This sounds great, I'd love to help with the front panel, but unfortunately it looks like I'll be in California until the end of August.

Look forward to seeing the reincarnated Boxford without the Boxford software! :)

[Daniel hansten]

Hi All

Sorry forbarging in on your group, just want to share my 2 pennys.
I am a member of a Makerspace in Norway who owns a Boxford 260 VMC witch we plan to upgrade to modern electronics.
We just finished converting our Boxford 125 TCL lathe to LinuxCNC with Pathpilot by emulating the electronics in Tormachs Slant Bed 15L Lathe.
The Plan is to emulate the Tormach PCNC 1100 machine for the Boxford VMC.
Our current BOM is as following:

Mini PC with PCI-E slot - Donated/Free
Mesa 6i25 FPGA Controller card - 1500kr
3x Wantai DQ860MA Stepper drivers - 1600kr
Honeywell spindle encoder - 120kr
DB-25 Breakout board (output - Mach3 type) - 40kr
DB-25 Breakout board (inputs) - 25kr
Mills & holders - 4000kr

-Daniel

On Wednesday, June 21, 2017 at 3:42:45 PM UTC+2, Toby Williams wrote:

Lot's of threads about the Boxford about, however, none really seemed the right place. 

Last night, Richard and I spent a couple of hours untangling (literally and metaphorically) the wiring in the boxford. This included finding a DFA relay. Overall, got a very good understanding of the wiring and voltages involved. We then returned it to full working state and closed it up.

Next Tuesday we are going to try and drive one of the axis around, using Richard's spare smoothie clone and some electronics wizardry (the logic level of the stepper drivers is 24V, incompatible with the 3.3 of the smoothie).

If that goes well, we'll look closer at things like e-stop and spindle control next.

All that are interested are welcome to help out in our adventure.

Toby

[Alex Gibson]

Hei Daniel

 

Takk for at du kom i kontakt, og velkommen til rLab - jeg håper vi kan hjelpe hverandre!

 

Alex

--

[Toby Williams]

Daniel, 

Thank you for your thoughts. If you are ever in the UK and interested, please come down to rLab.

We'll probably stick with our current path, it looks like it's going to work out really well. Although I am not 100% sold on bCNC as the controller software, I will need to spend some more time getting familiar with bCNC and looking at other options.

Definitely jealous of PathPilot!

Toby

[Daniel Hansten]

Thanks for the norwegian welcoming Alex. And i will deffently be visiting rLabs the next time i am in the UK. The same goes to you guys, please come visit Creator Makerspace the next time you are in Stavanger.
I understand that you are sticking to your plan. We work with the tools we have and know, LinuxCNC is a bit hairy to get running.

I am hoping we can help each other out. Our machine is not operational because we are missing the control computer, so we are "a little in the blind" about how the machine operates. So a couple of questions:
What current and voltage are you seeing on the steppers under load on your machine? We are considering using 7A 36V stepper drivers.
How does the toolchange/carusel work?
In the manual they mention the tool offset table, is this automatic or do you manualy measure the lenghts of each tool?
How does the cooling work? We are considering upgrading to flood coolant.
Is the spindle motor and the VFD powerful enough? Apperantly the VFD is a weak spot on this machine until Boxford changed to Mitsubishi VFDs.
How big mills can you use? Do you use the EasyChange toolholders?

[Toby Williams]

Daniel, 

I have spent a lot of time in Sweden but never Stanvanger, it is definitely high on my list.

I love the double approach, hopefully Smoothie will provide a solid basis however LinuxCNC will provide a more expansive solution. We have written up our findings on the Boxford here.

The steppers are being driven at 30V our drivers are 2A each. Richard is our lead guy on this, understands far more than I do.

We don't have a tool carousel, but appear to have a driver for it.

We haven't had much success with the ;Boxford' tool offset setting process. I have developed a custom process that hopefully will quickly become redundant. This involves manual g-code to show where the machine thinks the tools is relative to the work piece. (Before committing carbide to Ali)

Compressed air (with an oil mist) is doing great for coolant, we definitely aren't constrained by this yet.

We have a 400 watt motor and (appears) to be appropriately sized speed controller. At the moment, it doesn't feel like we are asking too much of the motor/vfd. My hope is that we try and take a "high speed" machining approach and lean our hackers towards Fusion's adaptive tool paths. 

We have a single 16mm end mill that I brought to play with, we do use EasyChange tool holders. We have just brought 2x ER25 collect holders to help out with spotting, drilling, etc. I've also got a 8mm roughing bit to play with. 

[Toby Williams]

Hopefully, tonight, Richard will be demonstrating the spindle control circuit on the Boxford. This would get us pretty close to a fully integrated demonstration.

We might even get as far as having smoothie command the Boxford to cut something.

[Toby Williams]

Tuesday night we got Smoothie controlling the spindle, this involve uncovering the mystery of pin 3 and 9 on the controller. We also found that the spindle controller uses some kind of feedback loop to do speed control even with out the encoder.

This means that we are ready to do an 'integrated test' with the Smoothie controlling all elements of the Boxford. My current thinking is to do this on the 13th August (Sunday) from 2pm (subject to Richard confirming availability).

Anybody interested in joining in our Boxford party, please shout or turn up on the day. All welcome.

Toby

[Alex Gibson]

I’d love to but will be getting back from France at that time L

 

From: 'Toby Williams' via Reading Hackspace [mailto:reading-...@googlegroups.com]
Sent: 04 August 2017 10:27
To: Reading Hackspace
Subject: [RDG-Hack] Re: Boxford Improvement Thread

 

Tuesday night we got Smoothie controlling the spindle, this involve uncovering the mystery of pin 3 and 9 on the controller. We also found that the spindle controller uses some kind of feedback loop to do speed control even with out the encoder.

--

[Richard Ibbotson]

Yes, I will be down for this
Richard

[Toby Williams]

Boxford Integrated Test

For the main part, the integrated test went really well. The Smoothie controlled the Boxford into making some cuts and functioned exactly as normal except with a much better controller. 

We didn't test the limit switches, however, they will be easy. 

The major sticking point is how the tool changes are implemented, bCNC relies on sending an M0 command to feedhold the Smoothie. The Smoothie ignores this command and carries on anyway. Once we have the tool change sorted, we will be 100% confident that we could make the change, if we wanted to.

T

[gavi...@gmail.com]

Great news. Just shout if you need a F360 post modified

 

Gavin

 

From: 'Toby Williams' via Reading Hackspace
Sent: 15 August 2017 09:42
To: Reading Hackspace
Subject: [RDG-Hack] Re: Boxford Improvement Thread

 

Boxford Integrated Test

--

[Richard Ibbotson]

As Toby said, we are now confident we know enough about the Boxford to change the controller. We can keep the existing stepper drivers and the spindle driver, saving cost and making the conversion simpler. They could be replaced with modern parts if they fail.

 

I do like the Smoothieware software with bCNC, but it has the feedhold issue as Toby said.

 

It was one of our early considerations in the software/firmware to be able to do feed hold, especially when doing a tool change. bCNC had the best support for tool change, however it now appears this uses some grbl features not present in Smoothie. We are looking at options of other software and grbl-lpc. We might also look to make changes to the software or firmware as they are both open source.

 

The core of the problem is that bCNC sees itself as a gcode sender, and so not responsible for a pause state, while Smoothieware also does not want to be responsible for a paused state (want me to stop, cease telling me to do stuff!).

 

If anyone want to get involved please join in, especially Python people.

 

From: 'Toby Williams' via Reading Hackspace [mailto:reading-...@googlegroups.com]
Sent: Tuesday, August 15, 2017 9:42 AM
To: Reading Hackspace <reading-...@googlegroups.com>
Subject: [RDG-Hack] Re: Boxford Improvement Thread

 

Boxford Integrated Test

--

[Richard Ibbotson]

I decided that rather than do a plug-in shield for the AZS MZ to contain the interfaces for the Boxford, I would make a build a more generic CNC version of Smoothie.
Based on LPC1769 Expresso, and will be open source hardware.
It can run the standard CNC  build of Smoothieware or grbl-lpc. It supports the AZS MZ graphic display, and has Ethernet interface.
Designed for a variety of external stepper drivers with buffered outputs.
Opto coupled inputs for weird sensors and analog inputs too. Spindle control by voltage output, or RS485.
I didn't put the relays on the board to avoid 240V AC on the board.

3D design of a lid is clearly going to be more of a challenge than the base!

To unsubscribe from this group and stop receiving emails from it, send an email to reading-hackspace+unsubscribe@googlegroups.com.

[Toby Williams]

You made that? Holy ****, that's some next level stuff.

[mikethebee]

This looks like a really great controller for many projects. What is the cost of the BOM? -Mike

[Richard Ibbotson]

For small quantity, the full BOM cost is around £55

This includes processor board, Ethernet, LCD graphic control panel, Spindel(RS485 and 0-10V), and SD card.
External stepper drivers are needed of course.

[Toby Williams]

Long time, no post. 

Boxford Update.

Last time, we successfully tested the Smoothie running the Boxford in every respect (apart from spindle encoding). However, due to a discrepancy in how pauses are implemented, we couldn't do any tool changes.

Since then, we have success in persuading the CNCjs developer to include some functionality to do tool changes, and this week we did a test using a preview version which worked great. Once CNCjs 1.9.11 is launched, we will be in a good place to do another integrated test with hopefully, tool changes supported.

Toby