CNC advice please

[user 43593]

This is my main point take a look at the 1010 build there might be on the best side of £200 worth of ally extrusions, £2 worth of chipboard in the bed, £120 router, it all runs on little wheels on top of the extrusions, maybe £500 would be a realistic price,
This is what I mean by giving the Wrong advise out as to what a cnc is, now if you wanted to give your little child something for Christmas to keep him from under your feet then this would fill the space for two days.

When starting out many folk don't know what components are required. In my opinion, that site provides its own insight into components to consider.

I have used expensive tooling from UK, from USA and have settled on using Carbide End Mills Associated Production Tools yes china made but you try these cutters and tell me they are rubbish, they certainly work well for me in Wood, Metal, and stone and are a good price (disclaimer: I am only a customer of this company)

Looking at the specs and the prices (broadly similar for some cutter configurations) I will stick with tools made by a long standing British manufacturer (since 1929) so that I can support a British company. It is partisan I know, but I see no compelling reason to abandon a British manufacturer who makes great tools and sells them at reasonable cost. I have dealt with John Rennie on several occasions and to say that he is helpful is to understate his willingness to go the extra mile.

Shopping list to build a cnc: (disclaimer: I am only a customer of this company)
www.wmh-trans.co.uk

Good link. Thank you.

 

[worn thumbs]

For my hobby machine I buy the cheapest carbide cutters I find on ebay and they work,its no disaster if they break either.For building a machine,there is no way I would consider small section extrusions,I'd rather buy steel box sections and bolt them together.As it was,I used MDF box sections for this one as I had doubts about my ability to create a functioning machine.Having used commercial machines since the late 1990's,I had every reason to believe I could create the code and I knew very well how rigid a machine needs to be.A 600W router can't develop the sort of loads that a 7.5KW machine deals with but could certainly make a piece of extrusion deflect.I continue to be bemused by the number of home designed machines where the proposer adds a rigid web along the machine to brace the gantry and does nothing to brace the uprights against the gantry beam.Why do they imagine there is no force to resist?There is a useful repository of knowledge at www.cnczone.com but the level of activity has dropped off a lot in the last year or so.I think there is much to be said for building the machine to fit the space and the user's requirements and gaining the depth of understanding to be independent of a supplier or service engineer.

 

[Phill05]

I have come across some guys not too far away from me that have spent big money on toys like this only to find out they wasted their hard earned wages, bit of dust builds up under the little wheels and bang jumps off extrusions router still turning takes a chunk out of hold downs frightens the living daylights out of him then they have ear ache from family when it has to go in the bin, or gets e-bayed to some other unsuspecting person.
You can see why I don't like this cr-p being sold to unsuspecting honest people that all they want to do is make something for them and their family to be proud of.

I have no problem how or where you chose to buy your cutters the point I am making is there are some very good cutters coming out of china at a price that a new to cnc person does not have to worry about braking a few.

To me they are a joke end of rant:

 

[user 43593]

For my hobby machine I buy the cheapest carbide cutters I find on ebay and they work,its no disaster if they break either.

Probably not a great problem, realistically speaking, when working in wood because tight tolerances will be difficult to maintain in a workpiece that may twist, expand and contract. Machining other materials to a specific tolerance is where the easily broken bits can hurt the workpiece... and a worker where sharp pieces of metal whirling around at 20,000 rpm may surprise the unprepared. I have on occasion milled 20mm thick 6061 aluminium and my last tight tolerance job was milling 10mm thick acrylic, where the job called for uniformity in design and tolerances in multiple sizes. (see pix)

There is a useful repository of knowledge at www.cnczone.com

Thank you for the informative link

pic 1 - I was asked if I could replicate these wooden figures in black and red acrylic in 4 different sizes
pic 2 - Here are 3 different sizes of finished items. There were 130 in each of the four sizes and two colours totalling 1040 individual pieces.
pic 3 - Large sized shadeteken cut with tabs in 10mm acrylic that was secured in between two modular vices
pic 4 - Consistency of smallest pieces

 

[worn thumbs]

I may be a bit cynical about the origin of most router cutters because I don't believe most of the people that sell them have the equipment on their premises to grind shaped cutters or to attach a tungsten carbide tip.I do suspect that they are quite capable of opening a pack of cutters that have arrived from China/Taiwan and attaching a label to the package.My approach cuts out the middle man.As far as breakage is concerned,I have had a breakage,but not during cutting.I was establishing a part datum (G54 for those who might be interested) by touching off on the side of the blank and I had gone the wrong way from the previous incremental setting i.e. from 0.5mm to 1mm instead of to 0.1mm when I touched the mouse key the 0.75mm cutter snapped.I wasn't too happy.I have yet to break a 1/4" shank cutter of any size on my hobby machine.

I don't believe the tiny machines,built from extrusions,or those with less than 300 watts turning the cutter are useful for much more than learning the principles,while machining foam or similar.Once we move on to cutting wood,I don't have much faith in toothed belts for moving the axes or unsupported round rails.Something like an SBR12 rail and ballscrews is nearer the mark.I don't doubt that many users of lesser hardware are happy enough with what they have got and what they produce.

 

[user 43593]

I may be a bit cynical about the origin of most router cutters because I don't believe most of the people that sell them have the equipment on their premises to grind shaped cutters or to attach a tungsten carbide tip.I do suspect that they are quite capable of opening a pack of cutters that have arrived from China/Taiwan and attaching a label to the package.

You may be right. When cutting Aluminium bar (20mm thick) sheet (6mm and 2 mm thick) I found cheaper cutters reveal rough edges from obvious chatter. Rennie's own carbide single flute bits left smooth cut lines. I purchased a jobber's set of Rennie's cobalt drill bits (1/8") of which there were ten bits. They looked for all the world like a typical Chinese package as found on Amazon.They cost me £7 for 10 drill bits plus a small amount for postage. I had to make 1/8" slots in hardened steel for a particular project to modify my belt tensioners. I was surprised at how effective these drill bits were. In the event I used just one bit for 30 holes and the drill bit felt like it would be capable of drilling another 30 holes at the end of the job. You may find the history of Rennie interesting. I had read that many of their endmill bits were made in Germany. I cannot find that article at the moment.

https://www.rennietool.co.uk/blogs/news/the-history-of-rennie

My approach cuts out the middle man.As far as breakage is concerned,I have had a breakage,but not during cutting.I was establishing a part datum (G54 for those who might be interested) by touching off on the side of the blank and I had gone the wrong way from the previous incremental setting i.e. from 0.5mm to 1mm instead of to 0.1mm when I touched the mouse key the 0.75mm cutter snapped.I wasn't too happy.I have yet to break a 1/4" shank cutter of any size on my hobby machine.

Right, I accept that your removal of the middle man works for you. It may also be a policy worth adopting if I run into supplier problems.

I don't believe the tiny machines,built from extrusions,or those with less than 300 watts turning the cutter are useful for much more than learning the principles,while machining foam or similar.Once we move on to cutting wood,I don't have much faith in toothed belts for moving the axes or unsupported round rails.Something like an SBR12 rail and ballscrews is nearer the mark.I don't doubt that many users of lesser hardware are happy enough with what they have got and what they produce.

I agree with your point about small extrusions and low wattage spindles. The initial belt set up on my SO3 was a 9mm wide Gates GT2 belt that I had trouble keeping adjusted. Carbide had opted for a right angled bracket to hold the doubled over belt at the end. It was the devil's own job to keep the belt under tension when changing it and the adjustment was iffy. Infact, I could not see the mechanical advantage of trying to tension the belt by pulling both up (initial) and then down (to hold fast) to get the belt through 90 degrees.

I spent a lot of time researching belts versus lead screws and was surprised to learn that belts have less backlash and respond quicker to stepper motor input. I changed the tensioning mechanism for one which kept the belt straight at all times. I learned to tune it by using a vibration based tuner for guitar. It turns out that 135Hz is the ideal frequency. After tuning each belt which takes me around 20 seconds rather than the previously possible 20 minutes for each belt. The belts stay stays tensioned accurately for around 9 months of regular use two or three times per week. I have managed on one occasion with new belts to machine 6mm aluminium 6061 sheet to a tolerance of ±0.0004". I can usually rely on a ±0.001" tolerance when the belts are freshly adjusted. The frequency for the tension was established as a part of a very good article and discussion by Liam Newcombe on the Shapeoko forum.

https://community.carbide3d.com/t/measuring-belt-tension-squaring-and-calibration/24712
EDIT: Spelling and grammar/context corrections

 

[Ollie78]

I have not used CMT but they do look to be comprehensive... AFAICT from their web pages

Look at scott and sargeant, they are often on a good price deal, I have several single flute compression bits, they do a really nice 3 flute engraving bit that I like and I also have a great big surfacing bit with replaceable inserts.

Ollie

 

[user 43593]

Look at scott and sargeant, they are often on a good price deal, I have several single flute compression bits, they do a really nice 3 flute engraving bit that I like and I also have a great big surfacing bit with replaceable inserts.

Ollie

Thanks for the recommendation. I will see what they have in the new year. I have a 1 and 1/8" Amana surfacing bit that I like a lot but replaceable inserts would be an additional blessing.

 

[BradyS]

I only managed to read some of the posts, not the enire thread and, being in a similar situation with 'everything needs to come through the front door', I will share my journey. The CNC machine was on my bucket list and, after a lot of research, I chose the 1010 Workbee to be my machine of choice. At that time there were 3 popular versions: Ooznest for the UK market, Openbuilds for the American market and Bulkman 3D for the Asian market but all 3 selling everywhere. It was a hard choice. I mainly chose Openbuilds because of the proved and tested American quality. The UK was great, even improved, but the price difference from USD to GBP was pretty prohibitive for me. Bulkman's was cheaper but I was hesitant to make such a purchase from them. This was also 2019 and since then Bulkman built a name for themselves.

Out of nowwhere, in 2020 a Workbee dealer appeared in the Romanian market. I contacted them, stated my interest in buying the CNC and got no reply for a long time. Eventually I managed to find out that they didn't have the actual machines, they were distributing the CNC routers sold by a company from Portugal, by the name of RatRig. I also found out that RatRig were innovators experienced in 3D printing technology, a group of passionate people who added CNC machines to their portfolio. Not only added it but, after consulting with the community, added their own improvements.

After researching into it, I ordered their version of the 1010 WorkBee CNC. Soon after my purcahse, they launched their own design of a much improved version of the WorkBee, this time with linear raild and steel plates, called the KillerBee. I was annoyed for not knowing about it but it was my fault since I wasn't following their community at that time. After much consideration and saving the money required, I purchased the upgrade kit for my machine. Soon after my purchase and now owning a superb machine, they launched their Stronghold Pro CNC which is 100% their design, concept, integrating feedback from the community as well. It was again my fault because now, even if I was following their community, I had not paid attention to their updates...

This Stronghold Pro machine might be a good candidate for your needs in relation to the stated conditions. Might be worth checking it out. This is the link to their machines Routers - CNC Machines and here is the version I was referring to Rat Rig StrongHold PRO CNC - Standard Kit

They have a great customer support. And they also offer very solid tables and other accessories for the machines. Further, they have a helpful community where you can find all info you need before making a purchase decision.

Merry Christmas and happy holidays!


Brad

 

[BradyS]

One more thing that I forgot. Rennie end mills are on my bucket list. They seem exceptional (to me). In the meantime I'll leave a link to what I currently bought but didn't have the time to test them yet. They are Chinese. The build quality left me speechless, remains to be seen if the durability will have the same effect.

Official website: Milling Cutter, Boring Cutter, Threading Cutter Suppliers - Fengchi
AliExpress store: https://dreanique.aliexpress.com/store/2166082

 

[sploo]

I spent a lot time researching belts versus lead screws and was surprised to learn that belts have less backlash and respond quicker to stepper motor input. I changed the tensioning mechanism for one which kept the belt straight at all times. I learned to tune it by using a vibration based tuner for guitar. It turns out that 135Hz is the ideal frequency. After tuning each belt which takes me around 20 seconds rather than the previously possible 20 minutes for each belt. The belts stay stays tensioned accurately for around 9 months of regular use two or three times per week. I have managed on one occasion with new belts to machine 6mm aluminium 6061 sheet to a tolerance of ±0.0004". I can usually rely on a ±0.001" tolerance when the belts are freshly adjusted. The frequency for the tension was established as a part of a very good article and discussion by Liam Newcombe on the Shapeoko forum.

(Almost) any meshing of gears/teeth etc needs to have a little bit of gap in order to not bind, but I assume the rubber teeth on belts will have a very close fit with the teeth on pulleys, so I can see how that would result in very little backlash.

My concern would be stretching in the belts over longer distances (especially under load); though I guess a lot of belts contain metal wire that reduce this.

I am impressed with those numbers though (9 months tensioned, and around a thou of accuracy).

 

[sploo]

This Stronghold Pro machine might be a good candidate for your needs in relation to the stated conditions. Might be worth checking it out. This is the link to their machines Routers - CNC Machines and here is the version I was referring to Rat Rig StrongHold PRO CNC - Standard Kit

Got to admit that Stronghold Pro looks pretty good for the money; rails, ball screws, 3hp spindle with VFD. It doesn't look to be the most rigid machine, but that's not really a disadvantage vs to the competition at this price range.

 

[user 43593]

Got to admit that Stronghold Pro looks pretty good for the money; rails, ball screws, 3hp spindle with VFD. It doesn't look to be the most rigid machine, but that's not really a disadvantage vs to the competition at this price range.

My major misgiving would be that there is no space under the Y rails. This may limit the length of workpiece. In my experience, not many people have room to spare in the Y direction. This limits the length of the workpiece whereas my machine will accept any length (under 40mm in thickness) of workpiece under the Y rails, as can be seen in pic 4 above.

 

[user 43593]

(Almost) any meshing of gears/teeth etc needs to have a little bit of gap in order to not bind, but I assume the rubber teeth on belts will have a very close fit with the teeth on pulleys, so I can see how that would result in very little backlash.

It results in a significantly quicker initial movement presumably bacuse there is less inertia pulling on a belt, rather than turning a normal leadscrew with a square cut acme thread. I have not seen a ballscrew set up in action so cannot speak to that. I am assuming that the mass/inertia of any steel threaded rod will be greater than any rubber belt.

My concern would be stretching in the belts over longer distances (especially under load); though I guess a lot of belts contain metal wire that reduce this.

I thought I saw a lot of stretching initially but I had no idea how to set up the belts and this may just have been my own ignorance in that I did not tension the belts sufficiently. Once I started using a reproducible method of tuning the belts, that issue does not seem to bother the belts. Interestingly, Gates provide a tuning app, although not for CNC machine belts and it provided the seeds of tuning by notes. The Gates app relied on sound and was difficult to hear or get consistent results in any environment that was not completely silent. My use of a vibration based guitar tuner solved the problem easily.

I place wooden blocks under the belt at points I have marked so I am measuring a set distance of belt, which is suspended a set distance above the rails. I attach the guitar tuner to the end stops of the X and Y rails depending on the belt tension being measured. I pluck the belt using a wooden plectrum so that additional resonances are not created. I tighten a single screw and read the tuner until I get exactly a middle C (135Hz) reading on the tuner. According to Liam Newcombe's tests, I can tension the belts anywhere from 100 ~ 150Hz and the 135Hz point appears to keep the belts from stretching, almost like tuning a stringed instrumment to concert pitch.

I am impressed with those numbers though (9 months tensioned, and around a thou of accuracy).

The upshot of learning to tension the belts properly is that I am very much happier with the SO3 machine. The original Shapeoko SO3 was not an easy beast to tame. After junking the MDF baseboard, along with various MDF workholding solutions, the vastly improved rigidity, along with modular vices, has permitted me to access many more of the abilities of this hobby level machine. For me the key points were accurate belt tension, baseboard and machine rigidity, modular vices and reproduceability of workpiece holding.

EDIT: formatting correction and multiple spelling errors

 

[sploo]

My major misgiving would be that there is no space under the Y rails. This may limit the length of workpiece. In my experience, not many people have room to spare in the Y direction. This limits the length of the workpiece whereas my machine will accept any length (under 40mm in thickness) of workpiece under the Y rails, as can be seen in pic 4 above.

TBH I'd think that's quite a rare thing to have (space under the rails along which the gantry runs). My K2 has the work area "free" but if you put really wide stock on it then it would foul the uprights of the gantry as it passes. Long stock can be fed through under the gantry though; like the RatRig.

 

[sploo]

It results in a significantly quicker initial movement presumably bacuse there is less inertia pulling on a belt, rather than turning a normal leadscrew with a square cut acme thread. I have not seen a ballscrew set up in action so cannot speak to that. I am assuming that the mass/inertia of any steel threaded rod will be greater than any rubber belt.

I'd assume that the mass of the gantry, spindle, and any loads due to cutting forces would be much greater than any inertia issues with a ball screw vs a belt. Recirculating ball screw nuts on a decent ball screw are incredibly smooth. Much, much more expensive than a belt, granted.

 

[Ollie78]

Got to admit that Stronghold Pro looks pretty good for the money; rails, ball screws, 3hp spindle with VFD. It doesn't look to be the most rigid machine, but that's not really a disadvantage vs to the competition at this price range.

The trouble with the machines in this style is going to be similar.
If you just take a look at the machine construction you can see them already fighting thier own flaws. the way the ratrig has tackled the problems of underbuilt gantry sideplates is to make them very short, this will likely work to a degree to reduce possible flexing but at the cost of clearance under the bit. Oddly they do not supply wires just the end connections .
I think it is a good step above some in the similar range and has many little improvements. I do believe it could be beaten by a self build machine within the same budget.

Ollie

 

[user 43593]

TBH I'd think that's quite a rare thing to have (space under the rails along which the gantry runs). My K2 has the work area "free" but if you put really wide stock on it then it would foul the uprights of the gantry as it passes. Long stock can be fed through under the gantry though; like the RatRig.

Yes, the rarity value is possibly because a pair of suspended Y rails may introduce some resonant frequency vibrations if not held adequately. The X rail is identical to the Y rails. The hardened steel endplates of each rail read 9 on the standard wire guage (3.658mm). A sense of the hardness is that a 1/8" cobalt drill bit required about 3 minutes to cut through the metal.

The space under the Y rails is 45mm and the width of all rails is 55mm with a height of 80mm. which implies a height to width ratio of 16:11. The aluminium box section rails are made from an extrusion that is a constant 5mm thick. I have no idea how hard the rail is (possibly 5083 or 6061) but it seems to resist flexing well. I deliberately chose a standard machine so that the 1 metre long rails would help to curb the rail flexibility, when compared with the larger rail lengths.

 

[user 43593]

I'd assume that the mass of the gantry, spindle, and any loads due to cutting forces would be much greater than any inertia issues with a ball screw vs a belt. Recirculating ball screw nuts on a decent ball screw are incredibly smooth. Much, much more expensive than a belt, granted.

Somewhat controversially, I think I tend to disagree. Not with the overall premise that the combined mass of moving metal parts, loads and cutting forces will make a contibution to the inertia which must be overcome before the spindle starts to move. My discomfiture is with the notion that the mass and inertia of the belt and lead/ballscrew component parts would be discounted.

Assuming everything was equal, only the difference in mass between a belt and the lead/ballscrew would make a difference to the inertia in that part of the system... I think. The mass of the components doesn't cease to be positive just because the whole system has greater mass than its component parts. It may not amount to very much but is should be possible to isolate the mass for each component and work out the inertia for each.

My schoolboy arithmetic is letting me down at this point. Any mechanical engineer should know how best to express the problem, and the solution, of the question of whether a belt develops less inertia than a lead/ballscrew, when dragging the same weight of machinery at the same speed.

 

[sploo]

Somewhat controversially, I think I tend to disagree. Not with the overall premise that the combined mass of moving metal parts, loads and cutting forces will make a contibution to the inertia which must be overcome before the spindle starts to move. My discomfiture is with the notion that the mass and inertia of the belt and lead/ballscrew component parts would be discounted.

Assuming everything was equal, only the difference in mass between a belt and the lead/ballscrew would make a difference to the inertia in that part of the system... I think. The mass of the components doesn't cease to be positive just because the whole system has greater mass than its component parts. It may not amount to very much but is should be possible to isolate the mass for each component and work out the inertia for each.

My schoolboy arithmetic is letting me down at this point. Any mechanical engineer should know how best to express the problem, and the solution, of the question of whether a belt develops less inertia than a lead/ballscrew, when dragging the same weight of machinery at the same speed.

I can't argue because I don't have any data either

I'm just assuming that ball screw inertia will be a fairly small component of the total load on the motor. For a large machine, I expect the ball screw inertia would be significant, but then the moving masses of the rest of the system would also be large.