Author Topic: Contemplating some upgrades to my CNC Router  (Read 4701 times)

Offline Tool-n-Around

  • Storm Trooper
  • ***
  • Posts: 243
    • View Profile
Contemplating some upgrades to my CNC Router
« on: November 28, 2022, 17:06:08 pm »
I’ve been kicking around some machine upgrades. All comments and input are welcome.

90% of what I do with it is machine polystyrene foam patterns for lost foam casting, so the machining loads are light and desired feed rates high. The machine is fairly lightweight so the smaller steppers do ok. It’s a relatively large work envelop but I need more Z travel (more on the z-design later) for pattern work and will be modifying that  to 8”+ under spindle. To get the additional height I plan to just raise the entire gantry system and for work where I need more rigidity I’ll build a platform that minimizes Z extension and may build wet cutting fluid retention/return features into the platform for harder materials.

Here’s the machine as it stands. There are of course likes and dislikes about the machine but all things at a price point. IMO, It’s a decent hobby grade machine and at the time I bought I have about $2000 into it. It’s the larger XL model. There’s pictures at the link and you can take in whatever you have the appetite for.

https://millrightcnc.com/product/millright-cnc-mega-v-router-bundle/

Some of my future pattern work may have very long run times. I need to get the machine to the point where it can reliably and at relatively low risk run unattended. I’m looking at replacing the POM rolling wheels with HGH linear track and SFU ball screws on all axis. Although I’ll likely pick up some accuracy, another major reason is debris tolerance. The v-grooves gather statically charged foam which can accumulate to the point to cause loss of steps. The rack and pinion has a similar issue to a lesser extent with foam but worse when collecting hard materials. The wipers on the linear rails and screws seem to do a good job, and as they collect a little foam dust even better. It may also reduce rolling and actuation friction a bit which can’t hurt with my smallish NEMA 23 steppers but may actually be needed with the increase Z-axis travel and weight due to not only the larger travel but also the possible addition of a spindle instead of the router.

The manufacturer of my machine more recently has offered what they call a “pro-version” that has similar linear rail approach but retains the rack and pinion drive. I’d replace the R&P with ball screws. Here is a link to that machine.

https://millrightcnc.com/product/mega-v-pro/

I can buy import versions linear rail and ball screw the such to do the entire machine for about $350.
I’m also looking at upgrading from the smallish DeWalt 611 VS Router to a spindle. I don’t think I need a lot of power, but I do want the ability to handle ½” shank bits and ER20 collet spindles seem to be minimum 2.2kw or better and quieter running would certainly be welcome. There’s also the matter of air vs water cooled. Of course, I’m looking at the 4-bearing Chinese imports which can be had with VFD for under $300. I’m very leery of purchasing the Chinese sourced VFD without first seeing a manual first to make sure I can understand it. I seem to recall a thread somewhere (cnczone maybe?) that covers this very well. Any reference sources would be appreciated. Any opinion on water versus air cooled?  Water is just more overhead but not that big of a deal to engineer.

While I had everything apart, for added rigidity I’d fill the gantry aluminum extrusions, at least the stationary two, and maybe the traveling one, I’d use the existing extrusions because they are very easy to retrofit. I could replace them with steel but at that point about the only thing I’m using from the original machine are the steppers and controller, which also could be better. I know the extrusions are not as straight/true as you’d want for linear rail but I see it done all over the place with apparent success.
On the z-axis, it seems the practice is to mount the rails on the stationary Z-plate and bearings on the travel portion. To get the under gantry clearance the router would have to be mounted significantly below the lower linear bearing packs. I was going to mount the rails on the traveling plate and make the traveling plate a stiff as possible, that way it would remain quite stiffer in the retracted position where heavier cutting could be done on platform. When extended, I’d probably only be cutting foam so the loads would be very low and inconsequential. Any other pros/cons to the two approaches?

In addition to my original purchase price, +$650 for ball screw and linear rail along with a spindle seems like good value for money, compared to $7.5-$15k to buy another more rigid machine with similar features……and yah I suppose the upper end of that could potentially buy a used (up?) real cnc machine, but I don’t have the space or appetite for that.

Best,
Kelly

Offline lloydsp

  • CNC Jedi
  • *****
  • Posts: 8875
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #1 on: November 28, 2022, 17:46:05 pm »
The newer of my two 4' x 8' bed routers has an HSD brand 3-phase spindle, model AT/MT1073-140.  It's air-cooled, whisper-quiet, and has been perfectly reliable for over five years.  I periodically check its runout, and it's still within factory tolerance.

I cut mostly PVC rigid foam sheets for fireworks tube base plates.

Lloyd
"Pyro for Fun and Profit for More Than Fifty Years"

Offline dave benson

  • CNC Jedi
  • *****
  • Posts: 1773
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #2 on: November 29, 2022, 02:53:32 am »
Kelly I don’t have the link at hand, But here is a snap of the front page, I was
made aware of this when doing my studies years ago, as a mature age student,
I had to do a primer year to get up to speed with the new math.

I’ve just tried to load the pdf into open office and export it from there
(from 10 mb to 8.5mb)  so no joy there either.
If you struggle to find it in the MIT library, shoot me a PM with a link to sent it to.
There are some good resources on the web including the Mycncuk website
Where it’s for the most part dedicated to routers and they have some resource
libraries including frame stiffness calculators, some interesting build blogs including
one from our very own Eddycurrent who built a large format router.
The web link in my profile is dead and I have not used it for years, so don’t
send anything there as I’ll never get it.

I did go and have a look at the Millright machine and the linear bearings on the
Y Axis are the type I would have chosen, and will give a trouble free service life
and will probably outlast you if you can manage to cover them with a bellows
like cover, and lubricate them on a regular schedule with a light lithium based
grease, avoid using a graphite fortified grease.
The Masso controller is good, a little spendy though.
The 4th Axis is a toy, as it’s using a planetary gearbox, so it must have some lash
to work, I made a rigid tapping head for my mill using a planetary gearbox too,
and spent a little bit more and got a low lash one, but for a 4th Axis, where I would
be making components out of steel it would be unacceptable as it would have lots
of chatter and poor positional accuracy. A better choice would be a Harmonic Drive
and you can get them on Ebay second hand (out of south Korea) at a reasonable price.

With the spindle bigger is not always better, as if it’s a lot heavier than the old one
it will affect the machine dynamics; The sweet spot is to get one that has enough power
to cut the materials you plan on cutting, but is not so heavy as to destroy your acceleration
or rapids or over stress the frame.
If you plan to let the jobs run for many hours and live in either a very hot (cooling) or cold place
(heaters in the tank) then water cooled spindle bearings last longer, If you live in a temperate
climate (and all else is equal) it doesn’t matter that much.
Make sure that the VFD has inputs compatible with the output from your controller.


Dave




Offline Tool-n-Around

  • Storm Trooper
  • ***
  • Posts: 243
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #3 on: November 29, 2022, 16:21:56 pm »
The newer of my two 4' x 8' bed routers has an HSD brand 3-phase spindle, model AT/MT1073-140.  It's air-cooled, whisper-quiet, and has been perfectly reliable for over five years.  I periodically check its runout, and it's still within factory tolerance. Lloyd

Thanks Lloyd. Looks like a really nice spindle but at $1750, a bit out of my league, especially for a guy that mostly machines Styrafoam……LoL!

Kelly I don’t have the link at hand, But here is a snap of the front page, ……..If you struggle to find it in the MIT library, shoot me a PM with a link to sent it to.

Will do.

The Masso controller is good, a little spendy though.

Definitely a big improvement over my machine, but like all things, performance at price point.

If you plan to let the jobs run for many hours and live in either a very hot (cooling) or cold place (heaters in the tank) then water cooled spindle bearings last longer, If you live in a temperate climate (and all else is equal) it doesn’t matter that much.

My machine lives in climate controlled shop…..no problem there.

Best,
Kelly

Offline Tool-n-Around

  • Storm Trooper
  • ***
  • Posts: 243
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #4 on: November 29, 2022, 16:23:31 pm »
With the spindle bigger is not always better, as if it’s a lot heavier than the old one it will affect the machine dynamics;

This is potentially a deal breaker for a spindle upgrade. I noticed immediately that they are listed as 10kg. I’ll weigh my little router and compare, and I could even weight my Z to observe the performance impact, but with NEMA 23s and weight added for additional travel, I doubt I’ll get that far. I’m waning a bit on enthusiasm for the spindle upgrade and may just use a slightly larger router I have with ½” collet and make/buy a controller for it, preferably one I could drive from CamBam…..seems easy enough to drive a Universal motor.

Make sure that the VFD has inputs compatible with the output from your controller.

$300 for a spindle and VFD sounds too good to be true and may very well be. The reason I added the VFD/Spindle upgrade to this thread was hope that someone that had done it and uses an import could point me to a reliable source and comment on reliability. Huang Yuang sells on eBay but there are many knock-off of the knock offs, and though they are inexpensive HY has been a round for a long time.

There’s 41 page thread over on cnczone about the HY VFD and spindles. Some complain of failures in a few weeks, others years of reliable service. The most common issue is poor documentation which leads to difficulties in properly configuring the unit…..like Chipowner’s thread below in this sub forum. The cnczone thread has a collection of English manuals and 41 pages of examples of issues and fixes. Though fighting through that isn’t particularly attractive, one thing that is a big concern to me is EMI from a cheap VFD.

I have a couple (KB Brand) VFDs in my shop, and even though they are filtered and isolated, they generate large amounts of EMI, both conducted and radiated. I solved most with additional filters for everything except the DRO scales on my lathe which were iGaging brand magnetic scales. The problem was lack of sufficient ground plane in the pcb on the controller and it just wasn’t fixable (without major pcb surgery) if it was plugged into any source with the VFD running (which ran the lathe!). So I run the DRO displays on batteries and they last a good long time.

Introducing a VFD to my CNC scares me to deat, let alone running power cables anywhere near my control cables or controller. I had horrible unexplained crashes and errors when I first started with my machine and it was all caused by static build/discharge from dust collection. Ground the piss out of everything solved that problem. Between that and the lathe experience I lost a large portion of my life I’ll never get back and don’t wish a recurrence! Oddly, I didn’t see one single complaint of such in the 41 page anaconda thread at cnczone on the HY VFDs.

I did go and have a look at the Millright machine and the linear bearings on the Y Axis are the type I would have chosen, and will give a trouble free service life and will probably outlast you if you can manage to cover them with a bellows like cover, and lubricate them on a regular schedule with a light lithium based grease, avoid using a graphite fortified grease.

I had to search the naming convention for the import stuff and found they're based upon the Taiwanese manufacturer standards.

https://circuitist.com/hiwin-linear-rails-guide/

Ya-know, set me straight if you think otherwise, but I was thinking HGH series linear rails are a better choice than the ones MR uses because they look identical to what they use on the Z-axis on my machine and those are or the equivalent of MGN series linear bearings. The latter are two-point contact and former four-point, which supposedly have both higher load and precision. They are only slightly larger for a given rail width and the cost difference on the imports is negligible for a comparable size. I’d bet the MillRight rails are 15mm. I could get HGH in either 15mm or 20mm for a slight price difference. With 8 linear bearings guiding each axis, I suspect load capability would be fine with smaller.

Thanks for all input and replies.

Best,
Kelly
« Last Edit: November 29, 2022, 16:28:22 pm by Tool-n-Around »

Offline Tool-n-Around

  • Storm Trooper
  • ***
  • Posts: 243
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #5 on: November 30, 2022, 15:05:12 pm »
To simplify this (Maybe :))

I scrapped the idea of a spindle in lieu of using Super PiD and Router to get more speed range and CAM control of speed. I’ll start a thread on that when I get to installing the hardware.

I still like the idea of linear rail and ball screw conversion of my machine, but what I really need is a better Z-axis with more useable z envelop and travel.

The advertised travel for my machine was 3.7” but there was no way to achieve anywhere near that as received because of the height of the MDF bed and the router ran into the Z-stepper motor when retracted. After I dropped the MDF bed and router in the mounting clamp and can realize about 3.5”. I need to be able to machine stock at least 4” thick and preferably 6”. Right now there is 6.75” under the gantry rail, 5.75” under the stationary Z-Plate and wheels. I think with a properly designed Z-axis, I could get nearly all that 5.75” as available travel and Z work envelop.

It seems that most if not all cnc router Z-axis with linear rails are designed such that bearing blocks are on the traveling Z-plate and the linear rails are mounted to the stationary Z-plate. I was thinking of doing it the other way around with the rails mounted to the traveling plate so the traveling plate could extend below the fixed Z-plate and gantry to cut thinner stock with short cutters over the entire work envelop without inserting a platform, re-tramming that surface, etc. Obviously, the linear rails get stiffened by being attached to the stationary plate, but the same could be done by attaching them to a (stiff) traveling plate. It does potentially increase the weight of the traveling plate, which is of course undesirable but I think this could be minimized. Then it’s just a matter of where you attach the ball screw bearing and length of the screw to make it work at desired stroke range.

In theory, it could be designed such that you could extend the traveling Z plate or at least the collet all the way to bed, but in practice would only need to approach the height needed for the shortest cutter ever used.

The reason for the additional extension is to be able to machine very soft materials (polystyrene foam) across the entire work envelop without (a large) platform to the elevate the work, then I could use a smaller (than 35x35 which actually would need to be more like 44”x44” to use the work envelop) platform for machining hard materials with the machine in its most rigid position.
 
Do you follow? Anyone see anything fundamental wrong with this train of thought and approach?

Best,
Kelly

Offline Dragonfly

  • CNC Jedi
  • *****
  • Posts: 2610
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #6 on: November 30, 2022, 20:01:48 pm »
This is a very funny design - rack and pinion drive which is very rigid an durable (unless made from plastic, of course) combined with those funky rollers on the gantry. While the Z slide uses true prismatic rail guides ...
Even supported round steel guides with slotted sliding bearings will perform better.

Offline dave benson

  • CNC Jedi
  • *****
  • Posts: 1773
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #7 on: November 30, 2022, 23:36:30 pm »
Kelly
If it’s not that much extra to get the 20 mm then do so.
The four-point contact bearings (Gothic Arch) have even less compliance than
the two-point contact type, and the cylindrical type, even has one
more degree of freedom altogether. This can be an advantage though.
Each one has a different type of minimum surface prep to operate
happily. And each type has a superpower.
I’m not suggesting you do this, it’s just an example of “Horses for courses”.

For example, in the photo, you can see a cylindrical type linear bearing
operating on a Beam Stand, the beam stand is made of square hollow section
seam welded mild steel 40 mm x 40mm by 2 meters long.
If you pick up the two-meter section in the middle and shake it vigorously, you can
displace the ends Maybe 50 mm or more. A wet noodle.

The steel tubing's dimensions and surface finish will be far worse than an Al extrusion.
There’s two of these sections that run that parallel to each other 400 mm apart
supporting the linear rails over 2 meters that a robotic carriage runs on.
There are two reasons that the cylindrical bearings were chosen.
Price played no factor in the selection criteria.

1…. Using the cylindrical shaft and foot added more than double to the stiffness
of the assembly because it increased the beam depth by a factor of 2.

2…. The tolerances in the bearings plus the extra degree of freedom had enough
compliance for the mechanism to operate over the whole length off the beam
without any binding.
The ones I used here have adjustable pre-load.

The mechanism is broadly analogous to a pick and place machine where instead off
placing a resistor or a capacitor it inserts a new colour in the yarn.

It uses a module one rack and pinion, the teeth spacing are PI apart.
As delivered and stacked end to end 1.8 m they ran out 0.4 mm
happily, they were 0.4 mm too long and after dressing the butt together ends
got it down to about half that.
For a standard rack properly installed, there will be a bit of clearance
required. In a perfect world you want your teeth to roll over each other rather than slide.

There are a few ways to reduce the lash in the system using a rack and pinion.
One is a spring-loaded split pinion, another which consists of two helical gears
bolted back-to-back forming a herringbone gear where you can adjust the phasing
between the teeth.
My mill has NSK High Preload Linears and carriages Tsubaki ground ballscrews
and the Z axis column was sent away and ground.
However, it will remain a fool's errand until I replace the dovetails that the milling table
rides on with linears, as I can claim no better positioning accuracy (X\Y) than I had before.
Ballscrews are good, when driven very fast they can whip, on some long bed routers the
ballnut is rotated instead of the screw.
As best you can, make sure that the machine frame is level and square before you install the
Linears. Some simple tools can help, a string line and three identical gauge blocks, a machinists
square and a machinist's level (when needed I can borrow one that is graduated in tenths).
Dave

Offline Tool-n-Around

  • Storm Trooper
  • ***
  • Posts: 243
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #8 on: December 01, 2022, 21:22:33 pm »
Some great points Dave. Thank you very much for that. Last night I took some measurements to see the art of the possible with my existing hardware.

The sketches below use 20mm HGR/HGH style linear rail and bearings. 2 bearings per rail. One rail for each Y-Beam, and two for the X. I call the traveling beam the X. It uses my existing machine's beams, steppers, and rack & pinion drive. I'd need to make the mounting hardware, but it would be very inexpensive easy fab for me with materials I have on hand. This could all be reversed but can't see why I'd ever revert.

I could also very easily convert the X & Y rack and pinion to ball screw actuation too.

For the Z, I'd use the existing stepper but I'd build the rest from scratch using the same 20mm HGR/HGH linear rail with a pair of bearings on each and a ball screw actuator.

But, the weight increase for the traveling X-axis will be about 18lbs, 11lb of it due to the linear rail and bearings. The extra Z axis travel desire and larger router will also add some weight, but I'm guessing <5lb. The Y is powered by two steppers (one on each beam) but the X only one. I dont know the total weight of the existing traveling beam assembly or what percent invrease that is but the majority of the weight increase would at least be borne by the two-stepper driven Y-axis.

Looks like ~$250 for as shown which seems like good value for money to me and another $150 to add 16mm ball screws on X&Y. Not sure about that. What do you think, BallScrews or R&P? I see a rack beside your rail.

Best,
Kelly
« Last Edit: December 02, 2022, 01:29:17 am by Tool-n-Around »

Offline dave benson

  • CNC Jedi
  • *****
  • Posts: 1773
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #9 on: December 02, 2022, 03:36:13 am »
Yes, the 16 mm ball screws would be good, not any bigger though.
With the ballscrew vrs rack thing.
To reap the full benefits of installing a ballscrew on the Y axis for the benefits of better positioning
accuracy and no backlash then you have to do the X.

The good thing about racks is that they are light and robust and can be fast.
For example, say you have a 5 mm pitch leadscrew, so for one revolution of the motor
you get 5 mm distance travelled, for that module one rack with it’s 20 tooth pinion one
revolution of the motor equates to 62.83 mm.

The torque curve for a stepper, say a standard Nema23 motor falls off a cliff after 300 rpm
so 5 mm x 300 rpm = 1500 mm a minute G1 cutting speed at full torque from the motor
You be able to rapid G0 around faster than that though.
 A module one rack at 300 rpm with a 20-tooth pinion = 18849 mm a minute.

If most of your work is foam patterns, where the material is very easy to cut and your
happy with the surface finish, then the racks are practical, robust, tolerant of little
or no maintenance and can stand a little foreign debris and they can be very fast.
If you have a hankering to do some detail work using small tools (1/16)
and smaller or want to cut metals\hard plastics then, the ballscrew\rectangular section linear
shafts and carriages are the go.

I used a rack where the backlash would have been a problem if the machine was a cnc
machine tool (you can use backlash compensation in Mach3) but at best it is a hack.
So the trick was to make the mechanism that selected (grabbed and or pinned) the yarn
sweep in from both sides in a pincer movement so the yarn could be only roughly aligned
+- 2mm where upon it was pushed into a guide slot, the backlash in the rack was something
like 0.05 mm, so I able to install the rack with the proper centre distance allowing very high
feed rates all while being able to completely ignore the backlash and not having to do
any screw (rack) mapping.

In the pic there is a type of coupling to use, it provides shaft alignment with no backlash.
the spider type with a flexible member (common on ebay) will compress under cutting
load leading to some backlash.
In the link a guy is making a desktop milling machine where he is using a machining levelling
compound (epoxy), this provides a good level surface for the rails, is easy to do and costs
about 120-150 bucks here and makes the alignment process easier and much faster.

I’ve installed machines on the factory floor where a 5 meter by 3 meter by 2 meter deep
section was cut out of the concrete floor and an isolated slab poured 40 mpa, when the slab had cured usually a month, the machine frame was maneuvered onto the slab and set level with
the jacking screws, a moat was created around each foot of the machine and the epoxy
self levelling compound was poured, and after maybe a week we would undo the jacking
screws and bolt the machine to the slab.
This was the recommended installation procedure, when I was an apprentice the
the first job I ever participated in was done with molten lead.

BUILD LOG: Desktop CNC 'Thing' (mycncuk.com)

Dave

Offline Tool-n-Around

  • Storm Trooper
  • ***
  • Posts: 243
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #10 on: December 02, 2022, 13:07:15 pm »
I hadn't thought about backlash from the racks as a possible source of the poor finish on the heavier cuts/harder materials but that makes sense. I was pretty focused on the wheels as the source, but any displacement there could also affect rack clearance. For now, I think I will stay with the R&P drive on X&Y. For the approach I'm taking to incorporate the linear rails, the X&Y could fairly easily be retrofitted to ball screws in the future. That might help isolate whether I'm getting flex/displacment from the wheels under heavier loads.

As far as cutting foam patterns and foundry tolerances, the accuracy is already good enough. It's really all about speed.  The R&P seems quite tolerant of what little foam fluff it sees, but the rolling wheels are not and cumulatively gather more than the R&P. The new positions of the R&P should also better shield them from debris. Since the Z axis already has lubricated linear bearings, I'm quite confident the additional linear rails and bearings will be tolerant to the foam dust. In use, the existing ones pick up a little foam dust on the wipers which makes them even more effective.

I will definitely use a ball screw for the Z. It is currently has a fairly small acme screw.

Best,
Kelly


Offline Tool-n-Around

  • Storm Trooper
  • ***
  • Posts: 243
    • View Profile
Nema 23 Stepper and Ball Screw Lead
« Reply #11 on: December 02, 2022, 16:52:16 pm »
Was looking at ball screw upgrades and concluded my steppers would need to step much faster for the same feed rates on either the rack and pinion X&Y or Z.

The acme screw/nut on my Z has a lead of 30mm. Most of the economical ball screws are only available in 4,5, or 10mm lead. High lead ball screws seem to be much more precious. The lower lead seems like it would be helpful to actuation force and positional accuracy but what determines how fast a NEMA 23 stepper can step when operating within rated torque/load?

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?

Best,
Kelly

Offline dave benson

  • CNC Jedi
  • *****
  • Posts: 1773
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #12 on: December 03, 2022, 02:11:43 am »
Kelly
Where circling the precipice of the rabbit hole.
This was my area (motion control) and also my hobby, so I’ll try not to ramble on.

I don’t think it’ll be 30 (the lead) in the pic you can see an Acme screw that’s for sale
from another manufacturer. I went to the millright site and couldn’t find the
exact specifications.
It’s a ½ inch diameter with a pitch of 0.2 and 5 starts.
Nomenclature:
    • Pitch is the distance between the adjacent threads.
    • Lead is the distance that a nut will turn with one rotation, and;
    • Start is how many starting (or thread entry) points at either end.
So, for a leadscrew, with a pitch of 2 mm, if there is only one start to the screw then the lead is the same as the pitch. However, if there are four starts to the screw, then the lead will be 8 mm. If there are two starts to the screw, then the lead will be 4 mm. And so on.

Just to put things into context and to give a big picture overview.
You machine is a type called moving gantry and is made from Al extrusions.
It’s powered by a Atmega2560 using the millright fork of GRBL.
I didn’t see what type of Stepper driver they are using, but it’ll be the
minimum entry level.  Maybe something like the TB series which use a chip
that was originally used to drive dot matrix printers.

It’s the minimum entry level machine, and for milling foam patterns it’s quite acceptable.
The minute you want to machine metal or use very small\Large diameter tools you’ll run
up against machine design limitations.
moving gantry aluminum construction, GRBL, Drives, motion system ect.
It’s very easy to over-capitalize a machine like this, by buying all the cnc bling
and still have mediocre results.

I’m not saying the you should not do any modifications, just some where they
will provide good value for the money spent.
For example replacing the drives (if they are the tb series) will offer a performance
upgrade, in the pic you can see a Leadshine DMM 522 drive, This drive is the ‘bees knees’
for your steppers they are excellent drives and they are not very expensive offering
good value for money. Don’t by the knock offs. There are plenty out there.

In the larger pic you can set the side of the driver where there is a table
to set the micro stepping value.
A standard 1.8 degree stepper motor takes 200 pulses to step the motor 1 REV.
however you can set the drive to 400,800 ect.
For your machine you would set 2000 pulses per rev.

This would equate to (with a 10 mm single pitch lead screw) 0.005 mm resolution.
However, we live in the real physical universe with real drives and motion components
so this is a theoretical number, it’ll be more like in the order off 0.025 for your motion system.
If I remember correctly GRBL can produce 30khz steps
How fast can it go? (grbl.org)
At 30khz per channel this equates to roughly 4.5 M/min. Running on the 8 bit controller
GRBL can run on a ESP32 where I’ve seen figures quoted as high as 100K this is overoptimistic though and it will be more like 70khz.

It’s already a long thread, so I’ll try to address at least one issue.
By looking at the cut on the periphery of the air cleaner cover, it’s obvious that the tram
of the tool is changing (this is causing your tool to dig in), this can be to some degree a
little slop in the Z Axis (backlash), but in this case I believe it’s the gantry flexing in Z and
rotating in Y caused by the cutting forces needed to cut the polycarbonate.

You could beef up the gantry by making it stiffer and replace the wheels with linears
or if it’s a life and death situation and you have to make the part right now with the machine
you have right now, then the option would be to use lube (reduces the cutting forces a lot)
or run the machine with much reduced feedrates.

Something to note is, that the coefficient of linear expansion for Aluminium to twice
that of steel.
The manufacture of bi-metalic strips as used in thermostats operate on this principle.
As the two dissimilar metals heat up a bend is induced in the strip closing or opening
a contact.

Dave
« Last Edit: December 03, 2022, 02:15:25 am by dave benson »

Offline Tool-n-Around

  • Storm Trooper
  • ***
  • Posts: 243
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #13 on: December 03, 2022, 02:52:44 am »
Thanks Dave, helpful as always but as usual a good part of it is over my head. I'd probably distill it down to I'd probbaly be no worse off than I am now if I changed to a 10mm lead ball screw than the one I have.

Aside from pitch I simply rotated my Z lead screw one revolution and measured the displacment on a scale, not a dial. I just wanted a general idea, but it was very close to 30mm for one revolution, of that I'm certain.

The MillRight literature might say the Z is .5" diameter acme screw buit I can assure you it most certainly it is not. It's very close to 8mm and the anti-balash mechanism is simply a split in the polymer nut preloaded with a cantilever spring. As far as the stepper drives, I dont know. No documentation from the manufacturer is available and no repsonse on their user forum so I suppose as usual I'll just have to open the controller, remove one, and identify it.

Admittedly, I'm out of my element here but I can economically buy a ball screw with a lead of 10mm/rev and was just trying to assess if the stepper could likely keep up at 3x rotational speed to maintain similar linear Z speed and if I could expect to change the Z-motion calibration that much in the machine set up/calibration. It's been a 18 months since I did the calibration and my memory isn't what it once was. I suppose I can just take what I get as far as Z.

If the machine never did anything more than machine foam patterns for me I'd be OK with that and take what I can get for the rest. As far as I'm concerned the upgrade to linear rails is to avoid missing steps due to accumulation of foam debris on the wheels on long unattended run times and any other benefit is just icing on the cake.

I need more Z-axis travel for my pattern work, and frankly wouldn't care if I used as similrar acme drive for my pattern work, but the cost difference compared to effort expended is deminimus, so I figure why not use a ball screw? Sometime it sucks not knowing what you're doing but I've always tried to not let that stop me.

Best,
Kelly



« Last Edit: December 03, 2022, 03:14:08 am by Tool-n-Around »

Offline dave benson

  • CNC Jedi
  • *****
  • Posts: 1773
    • View Profile
Re: Contemplating some upgrades to my CNC Router
« Reply #14 on: December 03, 2022, 04:18:55 am »
Kelly

That type of lead screw in commonly shipped 3D printers.
It’s very likely to be contributing to your  Z axis woe’s.
Because of it’s basic design and high helix angle it will be
easy to backdrive under load compared to an Acme or 5 mm pitch
ballscrew.
GRBL will be more than able to supply the pulses.

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

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.

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


Dave