Contemplating some upgrades to my CNC Router

[Tool-n-Around]

Just changing to a Acme screw and anti-backlash nut will be a huge improvement.

The inexpensive 16x5 and 16x10 are plentiful but knowing what standard you are buying is a bit nebulous. I can probably guess they're lower end rolled screws, just would like identify the ball nut specs preload/clearance.

I’d install a ballscrew, but before you order one make sure that there is enough room for the bearings as this may be the reason that these were selected in the first place.

Design layout work almost complete for Z. I've used a 16mm ball screw which is quite a bit larger to package than the existing screw but not much so compared to a 12mm BS. 

It would be good if you can supply a pic of the drives as some of the cheaper ones have less micro-stepping choices, before splashing the cash make sure that you can get access to the GRBL setup parameters for the millright software, if not then download and install the free LaserGRBL software (not for the laser part) but the GRBL parameter gui is the best and easiest one to use. Once you have changed the parameters to suit the new leadscrew close down the software and use the millright one as usual. If you have access to the command line in the millright  software then you can just type in the command in the second pic. If your leadscrew is 5 mm pitch and you have set the drive to X 10 microsteps  so 2000 steps per revolution, then in parameter $102 set it to 400. = 2000\5 = 400 steps per MM. About 4.5 m|min feedrate about 177 inches\min

This is what I need to understand. Also forgot to check if existing screw is LH or RH. Got some work to do on this and will do some more investigation before I buy the screw.

Many thanks.

Best,
Kelly

[Dragonfly]

LH or RH do not matter as you can always get the right direction through GRBL setup. Or directly reverse current motor direction by simply swapping two wires of one of its windings.

FWIW I am using a 16x5 single start ballscrew with a Nema 23 type stepper (4.2A max current), a driver Leadshine like on Dave's picture and 1/8 microstepping. This with a heavy 2.2 kW water cooled spindle. Max speed (G0 speed) is set to 2500 mm/min though it can do a bit more. Never happened in my practice to use plunge and retract speed higher than that. 10 mm pitch screws are actually double start with 5 mm pitch between the channels.

[Tool-n-Around]

....FWIW I am using a 16x5 single start ballscrew with a Nema 23 type stepper (4.2A max current), a driver Leadshine like on Dave's picture and 1/8 microstepping. This with a heavy 2.2 kW water cooled spindle. Max speed (G0 speed) is set to 2500 mm/min though it can do a bit more. Never happened in my practice to use plunge and retract speed higher than that. 10 mm pitch screws are actually double start with 5 mm pitch between the channels.

Thanks Dragonly, good to know. Think I've got the mechanical hardware sorted out. Will be a few days before I can dig into the machine control. I've got a couple project to complete first.

Best,
Kelly

[dave benson]

Just to be clear, the quoted feedrates (4.5 m/min) was not about what you should set the
the rapids to, it was in answer to this question asked earlier in the thread  post 11.

For example, if I selcted the 10mm lead, would I just expect to make the stepper calibration adjustment and expect it could keep up with my max X&Y rapids?

If your router + backing plate + (carriages or rail) depending on where you have mounted
them is less than 8 Kg then a Nema 23  3 N/m will be fine, if it’s 10Kg then a 4 N/m motor
would suffice. If the weight is greater than this then you would have to drop the feedrates
and accelation.

Something to note is that for a Z axis where the travel stroke is short, setting the velocity
is not he secret to getting a good performance, it’s the acceleration.

Dave

[Tool-n-Around]

Just to be clear, the quoted feedrates (4.5 m/min) was not about what you should set the the rapids to, it was in answer to this question asked earlier in the thread  post 11.

Understood. Also, the available ball screw leads were 4 or 5 (not 4.5) or 10. 5 & 10 have interchangeable interface dimensions.

If your router + backing plate + (carriages or rail) depending on where you have mountedthem is less than 8 Kg then a Nema 23  3 N/m will be fine, if it%u2019s 10Kg then a 4 N/m motor would suffice. If the weight is greater than this then you would have to drop the feedrates and acceleration.

The moving mass on the Z will be <<10kg, probably <10lbs and comparable to that of the existing machine hardware. Compared to the existing machine, the stroke increases from 3.75" to 8"+, the router power doubles, the centerline distance from spindle to beam centerline is slightly less, and the rigidity for the first 4" of stroke should be greatly improved. The only time the machine would ever see the upper regions of the stroke where rigidity could be more of an issue is for cutting foam, so it shouldn't be challenged. I was happy to be able to achieve all that. The height of the entire cutting surface can be repositioned or use of a platform will be used to control the height under gantry and accommodate different active travel cutting rates, but I don't expect to need to change this often as 4" under gantry with 2" cut height covers the vast majority of how the machine is used.

The X&Y axis will remain R&P drive for now. The moving mass on the Y axis will increase more significantly but it has two N23 steppers driving it.

Something to note is that for a Z axis where the travel stroke is short, setting the velocity is not the secret to getting a good performance, it's the acceleration.

I guess in my mind, I was thinking for a given motor torque a 5mm/rev ball screw would have at least 5x the mechanical advantage in linear force as an acme screw that had 5x that lead/rev. I just didn't know how much of the Z-accel was dictated by this and the traveling Z mass versus the stepper motor's internal inertial mass, stepper driver, and controller. Sounds like we're going to find out. I think you'll like it.

I'm not optimistic about being able to access the command lines on the proprietary MillRight control board but that is yet to be confirmed. That could expand this project a bit. Hoping to build hardware over the holidays and have plan of action on this before I pull the existing hardware apart.

Just for my learning, when feed rate is set in a CB MOP, that is just for X&Y axis, and the Z is a calculated/along-for-the-ride as determined by the tool path, correct? Hard to say what Z accelerations will be reuired since it's so dependent on feedrate and profile shape, but if Z cant keep up, it'll just require backing off feed rate, but it's hasnt been a problem to date on my high(er) feedrate foam pattern cutting.

Best,
Kelly

[Dragonfly]

Z feed rate (plunge speed) is also set in the G-code. Retract is usually done at maximum (set by the machine config) speed. 'Plunge height' is also an important parameter as it sets the distance above the stock to which Z will move down at full speed and then switch to the set slower plunge feed rate to bite into the material.
Every program which represents a synchronous motion controller, GRBL included, has the task to move the three axes synchronously so that they arrive at the commanded X,Y, Z coordinates at exactly the same time. And it does this taking into account the slowest of the three. So if the Z feed rate is not enough to comply with the faster X/Y drive the controller actually slows down their speed accordingly.  This applies mostly to real 3D engraving of 'organic' shapes when one needs maximum available speed for a finishing pass with a small cutter which could take several to tens of hours.

[Tool-n-Around]

Z feed rate (plunge speed) is also set in the G-code. Retract is usually done at maximum (set by the machine config) speed. 'Plunge height' is also an important parameter as it sets the distance above the stock to which Z will move down at full speed and then switch to the set slower plunge feed rate to bite into the material.
Every program which represents a synchronous motion controller, GRBL included, has the task to move the three axes synchronously so that they arrive at the commanded X,Y, Z coordinates at exactly the same time. And it does this taking into account the slowest of the three. So if the Z feed rate is not enough to comply with the faster X/Y drive the controller actually slows down their speed accordingly.  This applies mostly to real 3D engraving of 'organic' shapes when one needs maximum available speed for a finishing pass with a small cutter which could take several to tens of hours.

If I'm interpretting that correctly, that's a pretty fundamental thing I didnt understand, and I can see how that would probably be inconsequential for most of the 2D & 2.5D machining I do, but for 3D MOPs of, as you say of organic shapes, it may mean I was only achieving modest gains (or no gains) in run time reduction by increasing (XY) feedrates without making corresponding increases in plunge rate......depending upon shape. Thank you for that explanation.

Best,
Kelly

[dave benson]

I want to preface what I say next with this, I’m not throwing shade on GRBL or the
ATmega2560, I’ve used both in many projects and hacked on the code.
It’s amazing what functions the GRBL team have managed to shoehorn into an
8 bit controller running at 16 mhz.

It’s the entry level controller of choice, and for my Laser, 5 Axis grinding machine,
the AI tool Changer for the lathe, and for countless other projects it’s great.

However there are many features that could not be incorporated into such a modest
micro-controller.

For classes of controllers like Mach3\4 , Linux and  the newer USB and Ethernet off board controllers, the
motion planners are very sophisticated in comparison.

Just one example of the many differences.

Mach3 reads ahead and plans its moves to avoid the toolpath having to slow down more than in necessary.
Lookahead determines the number of lines of GCode that the interpreter can buffer for execution.
It does not normally require tuning. 

For GRBL it’s 10 Mach3 default 200 can be more.
There’s also things like when the machine comes to a fast stop it doesn’t jerk or
lurch about.

I’m not suggesting that you upgrade to one of those controllers, as for what you are doing
mainly (foam patterns) light fast cuts it’ll be ok.

If your finding the 3D models are taking too long to cut, and if they are simple shapes
then doing them in part or whole as a 2.5 D job offers the greatest time savings.
This can save literally hours on jobs.
The next best option is to judiciously select the machining strategy, Horizontal, vertical
or waterline.
Use the biggest tool diameters for roughing as practicable.

Sometimes even when you have done some or most of these things, it still takes too long
then you can use the feed rate adjusters which work by achieving the highest ‘average’ speed
around the tool path.

I was going to mention that moving the spindle further away from the gantry beam would
cause a bending moment around the Y axis but you were on to it.

Dave

[Tool-n-Around]

I do use most of those strategies when it's worth it to do so. I also previously was only conventionally cutting but now when roughing I rip off big DOCs in both directions. For 3D foam cutting, I was cutting at 200in/min, but I wasn't paying any attention to the plunge setting (typically was <60in/min) so dont know how that may have been affecting some of the larger 3D projects, but I have some big(for me) surface area projects on the horizon. When you're making one-offs and have no tool changer, some times it just makes more sense to keep the CAM simple, design for a single tool, and let the machine do it's thing.

I have no delusions about my machine. It is and will always be a hobby machine, but compared to what most call hobby machines, especially for a 35x35x8+ cutting envelop, it's probably better than most and with PiD speed control and the mechanical upgrades, I expect it will be better and expand it's capabilities within the hobby realm. Good enough for who its for....

I think I've bitten off enough for now, unless I must make further controlller changes. A little over a year ago, I'd never heard of a post processor, barely knew what GCode was, and had never written or executed a CAM program of any kind. Now I know enough to be dangerous! LoL

Always appreciate your replies....

Best,
Kelly

[dave benson]

You have come far in a short time, I used you as an example, saying (to my son) that if
you have a passion (the casting thing) driving your need to lean cnc’ing then half the
battle is over, I’ve been trying to get him interested in cnc’ing for years and so far he
has resisted steadfastly, preferring gaming on the pc or the phone using unsocial media.

Dave

[Tool-n-Around]

.....I’ve been trying to get him interested in cnc’ing for years......

Seems like the hook for the young'uns is 3D printing.

I like 3DP. It's very useful for pattern work and smaller investment/shell casting projects, but there's no printable evaporative casting filament that works for, or has any where near the economics, of (machined) lost foam, for the size and nature of parts I make......so cnc it is for me. If I also get some other modest capability out of the cnc router, that's just icing on the cake.

On another front, was advised there would be a 2-6 week delay on shipment of Super PiD due to parts shoratges ....Sign of the times I guess. On the bright side, I should have the motion control hardware for the machine upgrades in hand yet this week. I have an intake manifold project to finish up first but looks like the hardware upgrade may preceed the S-PiD.

Best,
Kelly

[dave benson]

Seems like the hook for the young'uns is 3D printing.

Yes, I had hatched a cunning plan maybe a year or 18 months ago and bought
an Ender3 Just an entry level one, with the idea that, because 3D printing is an order of
magnitude easier to set up and use, that this would have some appeal to him.
No interest shown yet, although the V8 inlet manifolds and a trebuchet 
I printed a while ago have made an appearance on his computer desk.

I on the other hand love it, and have made parts that are unpractical to do with subtractive
machining.
I’ve resisted the urge to upgrade it except for adding another Z Axis lead screw and have
the hotend suitable for carbon fiber.
Something new I saw the other day was water soluble support material.

I use Freecad for the cad, and one of the benefits of using a parametric modeler is that
I have one drawing of a ‘Jiffy box’ for electronic projects where the width, length and
height is in a companion drawing spreadsheet, the fantastic thing is that if I want a
box with varying dimensions I just change those parameters in the spreadsheet and
the box is re-drawn for me without ever having to touch the original drawing.
This is a massive time saver.

There’s a very intriguing story behind the reasons for the chip shortage, this young
man nailed it, Why Buffett Bought TSMC - YouTube
The likes of TSMC are opening up new fabs in Arizona and Intel\IBM
are also opening up fabs in the US somewhere, my opinion is that offshoring these critical
industries were a very bad idea in the first place.

If you have a dial indicator and magnetic stand, installing the new running gear
should be pretty straight forward.

Dave

[Tool-n-Around]

I've started on my Machine upgrades. Increasing the Z-Axis travel from 4" to 12". Of course, I had to incorporate some lost foam castings....because that's what I do  . 20mm Linear Rails, 16mm Ball Screw. From the attached pictures, all three axis will be converted to 20mm linear rail. If you look back through the thread, it's quite an improvement to my existing machine hardware.

In the pictures I show retaining the rack and pinion drive on X & Y but for +$150 I can convert the X & Y axis to 20mm ball screws. If so, at this point, the only thing left from the original machine will be the steppers, controller and extruded aluminum beams. I'm having evil thoughts of just selling my machine and upgrading those too. I'd have a a much better more rigid machine and probably money in my pocket from the sale of old machine.

Glad the site is back in good health.

Best,
Kelly

[Tool-n-Around]

More Pictures

Best,
K

[Bob La Londe]

PC690?  I did a lot of work with those before switching finally to the 3 phase liquid cooled spindles.  I burned up more than a couple of those PC690s running in aluminum.