559,954 active members*
2,502 visitors online*
Register for free
Login
Page 1 of 2 12
Results 1 to 20 of 22
  1. #1

    Join Date
    Sep 2022
    Posts
    12

    DIY UHPC 7 Tonne CNC Bridge Mill

    Hello all,

    As the title suggests, I am building a large bridge type CNC mill from DIY UHPC.

    I am currently part way through the build, so there is too much to cover in one post. I have already built the UHPC frame components, and intend to work on leveling and assembly of these parts while awaiting delivery of the needed components. I am now at the stage of looking to select and order components, such as linear guideways, ball screw systems and servos etc. I am hoping to get some advice regarding this aspect of the build.

    The work envelope of the machine will be appoximately 1250mm x 1050mm x 600mm, and the design mindful of the ability to upgrade to 5 axis capability in the future. It is intended as an all round prototyping machine, with the possibility of doing light production work to get going. This means cutting everything from foam to steel, though I do not yet know what I am likely to do most. I am considering the possibility of using interchangeable heads, though I intend to begin with a cheap spindle, as my first project won't require any hard material and this will allow me to learn CNC machining without risking a valuable spindle.

    I have encountered various difficulties with developing a DIY UHPC, but have read well over 100 academic papers on the subject and have used software to help me design the mix, so may be able to offer some valuable information on this subject. The result of my mix design is a coarse UHPC, which can be mixed effectively in a drum mixer (though a pan mixer is preferable) and is self compacting - can give an excellent, almost glossy finish. I have also utilized a thermal curing regime based on employing the concrete's own heat of hydration, the principles of which were tested and proven in an academic study. This serves to bring an end to long term shrinkage (confirmed by numerous papers). Many people have used proprietary UHPC mixtures for DIY CNC builds, often using nanodur compound 5941, however this is based in Germany and not readily available elsewhere. there is nothing particularly special about nanodur compound however, it is still a portland cement based binder (albeit with careful particle grading and a pozzolan).

    I imagine more UHPC info will naturally arise in the thread, so will save that for later.

    Here is a basic CAD visualisation of the design. I will show some other progress photos in coming posts.


  2. #2

    Join Date
    Sep 2022
    Posts
    12

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    The first thing I needed to sort out for this build was a means of handling the heavy parts. The bridge weighs approx 3000kgs and the base approx 2500kgs, and needed to be cast face down. This meant flipping the parts 90 and 180 degrees respectively. I settled on a gantry crane for this job. I therefore designed a gantry crane of appropriate height and ran simulations in Fusion360 (this was before the simulation capabilities were removed from the free licenses). I based the simulations on a load of 5 tonnes in the center of the beam and worked to a safety factor of 3. Once the design was complete, I purchased the steel box sections and cut the mitres. I hadn't attempted any welding before, so had a friend teach me how to do it - he is experienced with TIG welding, so this is what I used - I bought a stahlwerk DC TIG welder with a 7 year warranty and began welding the parts.

    Here is the design of crane pictured with the CNC design:



    Here are some pictures from the welding stage:















    Also tried my hand at stick welding the 10mm cross beam gussets to the girder (TIG welded the roots).







    More to come.

  3. #3

    Join Date
    Sep 2022
    Posts
    12

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    The next stage of the build was to design and produce the moulds. I haven't necessarily modeled all features, but modeled enough to create the needed cutting lists for the melamine faced mdf. Here are some cad pictures of the bridge and base moulds respectively:





    The profile for the bridge was created by stacking mdf boards beneath the melamine faced boards like so:



    From these designs I created cutting lists and ordered melamine faced MDF boards, and began to build the moulds:









    Here is the bridge mould complete and insulated (more on this later) - there are numerous large anchors welded to the steel plates, which are not visible in these pictures:



    The next stage of the process was the pour of the UHPC. I did not take pictures or video footage during this process as it was laborious, however I did take pictures immediately after, so will show these in the next post.

  4. #4
    Member
    Join Date
    Sep 2005
    Posts
    1185

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    big machine ... :cheers:

  5. #5

    Join Date
    Sep 2022
    Posts
    12

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Initial testing of my UHPC recipe was not successful. I had procured various samples of the constituents, but attempts to mix these in an ordinary drum mixer simply never properly mixed (this was in part due to using too little in the mixer due to the limited samples). There was also a significant difference between the 3 PCE based superplasticizers; though initial trials were unsuccessful with each of them - UHPC mixes tend to be extremely sensitive to changes in the mixture or mix procedure.

    I ultimately purchased a small planetary kitchen mixer (kenwood K-mix) as the papers I had read indicated that a lack of sheer force when mixing in a drum mixer was a likely culprit (most UHPC recipes aren't able to be successfully mixed in drum mixers). I attempted the same mix again in the planetary mixer, and it failed once again. I therefore decided to carry out 'stepwise' testing. I began with just the cement and a saturation dose (determined experimentally) of the 3 superplasticizers I had - one came out the clear winner with my chosen cement (52.5r grade). I then added the various other constituents and ended up with a self compacting mix (mini slump flow tests performed to confirm). The mixture looks very similar in consistency to Nanodur coarse UHPC, but is slightly more flowable.

    I found that I was also able to adopt a mixing strategy to successfully batch the mixture in a drum mixer, which proved to be invaluable, as hiring a large pan mixer near to me proved near impossible (I was even considering that I might have to build one; thankfully not).

    I also had great difficulty in purchasing the larger quantity of ingredients for the full scale pour. Elkem, who had provided a sample of silica fume (microsilica 940u) free of charge, initially offered to supply a bulk bag as a sample, if I paid the shipping charge (380 euros). I agreed to this but the representative simply kept promising to send an invoice and never did it. There was almost no-one else to purchase silica fume from in the UK and a new variant would additionally require a repeat of all the testing, so this caused me to waste months! Eventually I was told that without a company registration number, I could not get it. I eventually found an alternative product and supplier and purchased the needed product, though I had to purchase a lot more than I needed, so will probably sell some of this on.

    I also couldn't just purchase the superplasticizer from the manufacturer, who sent the sample. I ended up suggesting going through a retailer and managed to purchase it this way, although it cost more.

    With all the ingredients procured, I hired 3 diesel site mixers (again easier said than done) and batched the UHPC in batches of 83 litres per mixer. This meant doing 12 batches for the base mould (4 full batches of each mixer) and 14 batches for the bridge mould. these were completed in 5-6 hours each including cleaning out the mixers at the end (not between batches) and were done a few days apart.

    After completing the pours, I placed 6" thick insulation boards over the mould and placed a oven thermometer on-top of the UHPC to record the air temperature in the moulds. The Base reached approximately 80 degress Celsius and remained above 70 degrees C for around 4 days. The bridge was recorded at 97 degrees C but may have reached even higher; This remained above 90 for about 50 hours and above 70 for over 4 days. Excellent temperature-time histories for thermal curing of UHPC. This method gave slightly better results than laboratory steam thermal curing in the following study, on which the method was based:

    https://www.researchgate.net/publica...t_of_Hydration

    The benefit of thermal curing from a machine frame perspective is the retardation of long-term shrinkage. Numerous studies indicate that thermal curing accelerates initial shrinkage to occur entirely during the thermal cure (2 days). It also significantly increases compressive strength, durability and other properties (elastic modulus, tensile strength etc.). Some studies indicate complete end to shrinkage, others indicate that long term shrinkage is virtually eliminated or negligible. See for example: https://www.fhwa.dot.gov/publication...103/chapt5.cfm

    Some might be concerned about delayed ettringite formation, which can be a problem and crack risk in concrete when temperatures above about 70 degrees C are experienced. This is because those temperatures prevent ettringite formation or cause its breakdown. In normal concrete ettringite then forms later, (up to 2 years) hence Delayed Ettringite Formation. Various papers report that Delayed Ettringite Formation (DEF) isn't a concern in UHPC because it requires water to form and UHPC uses almost all of its mix water almost immediately, and is too dense to allow exterior water to penetrate.

    Anyway, here are some pictures from this part of the process:

    There are 3 main support feet (arranged as a triangle for levelling) and 7 support feet.







  6. #6
    Member
    Join Date
    Jul 2018
    Posts
    4351

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Hello Mike - What tensile and compressive modulus are you expecting to achieve ? Peter

  7. #7

    Join Date
    Sep 2022
    Posts
    12

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Hi Peter,

    In terms of tensile/compressive modulus, all of the papers I have encountered simply refer to elastic or Young's modulus and haven't distinguished between whether the modulus is determined in tension or compression; I note that they would theoretically be the same, though in practise, for concrete can differ (tensile modulus seems usually to be greater see for eg. https://www.hindawi.com/journals/ace/2012/391214/).

    As far as the elastic modulus goes, almost all of the UHPC designs which report the elastic modulus place the figure between 40 and 60 Gpa. There are certain trends which can help to indicate where between these figures mine is likely to lie. For instance, generally, higher compressive strength coincides with a higher elastic modulus; thermal curing increases it; higher paste volume decreases it; coarse aggregate increases it and stiffer aggregates have a particularly strong effect - refractory aggregates such as calcined bauxite can give an elastic modulus similar to solid aluminium - see Durigid by Durcrete.

    If you look at the following papers:

    https://www.eirichusa.com/images/dow...Aggregates.pdf

    https://www.extension.iastate.edu/re.../UHPC_ID66.pdf

    you can see that the elastic modulus reported ranges from approx 48-58 Gpa. My mix design is actually quite similar to the designs in the second paper, which give E module in the 53-57 Gpa range. My design employs certain materials which are less ideal for maximizing elastic modulus, i.e. I employed granite which is less stiff than the basalt aggregate employed in this paper; I also used ultra-fine limestone powder rather than quartz powder, which would have been stiffer, though has other advantages. Some of this reduction in stiffness will have been offset by my use of thermal curing however.

    I would therefore expect my E module to sit somewhere near 50 Gpa, however I would probably employ a design factor lower than this, maybe 40-45 Gpa. Given the large sections, the exact modulus isn't particularly important to me for this application.

    Something interesting (to me at least) is noticed when cracking open samples of my mix. I broke open several samples of my UHPC and found that the split/crack had propagated in every case, through the granite aggregates rather than around them in the inter-facial transition zone. This was even true of a sample I split just 18 hours after casting, indicating very good early strength, and utilization of the granite aggregate. This phenomenon has also been reported in other UHPC papers.

    Mike

  8. #8

    Join Date
    Sep 2022
    Posts
    12

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    After about a week in the insulated moulds, I was able to demould the parts. The melamine separated fairly easily ( I did not use a mould release agent). Unfortunately, I had also used some standard MDF without melamine for certain parts where I didn't have appropriate melamine faced boards to hand. I painted these in an attempt to avoid separation problems, however this was ineffective. As the pour was done in batches, I used a rake to amalgamate the each new batch with the previous one, however it was difficult to get right up to the faces of the mould, so there are some signs of the different layers and imperfect surface finishes at points, however these are just aesthetic and will not affect the function of the parts. They will be painted in any event. The surfaces cast face down are generally excellent. Here are some pictures from this part of the process.
















  9. #9

    Join Date
    Sep 2022
    Posts
    12

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    After de-moulding I placed the bridge on the base for a test fit. The bridge will have to come back down for flattening work on the faces and to prepare the mating surfaces which is unfortunate because it was more difficult than anticipated to maneuver the gantry with the bridge hanging from it. Here are some pictures of the frame components in position:







    The van in the last picture is my father's Vito to visualize the scale.

  10. #10
    Member
    Join Date
    Jul 2018
    Posts
    4351

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Hi Mike - Thanks for the detailed response. I spent considerable time like you on researching UHPC and similar materials. The concrete industry traditionally has not placed much emphasis on modulus for design. Recently with high rise and very high rise buildings the compressive modulus has become important for buckling of the lower stories members so they are starting to look at this more closely (unless you are a pyramid builder). I have not yet looked at the references you quote, but will. The compressive modulus as you say is highly dependent on the aggregate used and its volume ratio. Its straightforward to get E=30Gpa out of concrete, then using basalt get 40GPa then using alumina 50GPa then specialty fillers to get to 60Gpa. I think the 60-70GPa is in compression only where the aggregate is "touching" and therefore the "concrete" is not really in the load path. I think the tensile modulus would be poor. Being a machine designer I'm very keen to use this sort of material but have been disappointed with my own tests and with info supplied by local concrete engineers and chemists. I was aiming at 60GPa plus tried for 70GPa I've found a local source of CSA engineering grout which is very stable and available out of a bag as a premix (E-35GPa). Adding basalt will get it to 40GPa. I spoke to Durigid but they would not supply to Australia so via the SDS's I tracked the material suppliers to China then back to Germany. The company chemists where helpful and since alumina is made in Oz sourced suitable aggregate here. Made a few samples but have not had them tested. I have many years experience in the design of carbon fibre and composite structures with associated testing. With std CF its easy to get to 70GPa, 80 even 90GPa. With glass fibre we regular test at 30GPa tension and compression and flexure so I've sort of stopped going down the concrete path as I've concluded CF/GF is a better path than concretes. Also we test the strength of glass and carbon at 600MPa plus (tension compression and flexure) so strength is far above concrete, although that is not a major factor in machine design. One thought is to use CF or CF for a thick skin then backfill with CSA. Will be following your build with great interest. Very nice work.... Peter

  11. #11

    Join Date
    Sep 2022
    Posts
    12

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Hi Peter,

    Thank you for your interest. You might be interested in the following page by Durcrete (https://durcrete.de/wp-content/uploa...ine_E_0615.pdf), which shows various data about their E45 and E80 (84.5 Gpa static modulus) concretes. It is worth mentioning that the E80 variant appears to be employed more for specialist applications where the high modulus is particularly needed, such as when extremely high natural frequencies are sought. For machine frames it is my understanding that the E45 with natural aggregates is typically preferred despite the lower elastic modulus - this may be due to the significantly higher damping ratio as well as thermal properties which are more similar to steel.

    Carbon fiber clearly has excellent properties, particularly for the axes where weight reduction is sought; I understand that some commercial high speed CNC manufacturers are experimenting with it. I have no experience with it but would expect it to be prohibitively expensive for my application and may pose difficulties with attaching things to it. I would ultimately like to make use of carbon fiber for future designs, and use my machine to make the plugs/moulds, so i'm sure it will appear in my future projects.

    Kind regards,

    Mike

  12. #12
    Member
    Join Date
    Jul 2018
    Posts
    4351

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Hello Mike - The E80 UHPC material is very attractive but they could not tell me the tensile modulus. The E80 is the compressive modulus. The E80 they are reluctant to sell for machine use and they do not sell in small quantities and do not sell any product outside of the EU. E45 is relatively easy to achieve and my local concrete chemists can help with that. My other aim with developing a cold casting material was that it be machinable so surfaces can be finished accurately. Concrete can be machined with hard tooling and a wet machine. There is a local stonemason here with such a machine that expressed interest in machining my stuff but when it came to actually doing it he baulked due to the accuracy required. Headstones don't have to be very accurate.... Importing granite from china and having them machine it or locals machine it was an attractive route. But I decided against that method. Prefer to make my own moulds and use GF or CF to optimise shapes.

    Carbon fibre ticks all the boxes and I use diamond tooling on my router for machining composites dry or wet. I don't think its expensive relative to the project. Make the parts with thick skins 10-15mm thick then back fill with CSA grout (or UHPC) I think is a cost effective approach. The CF I regularly make and test at E80 for yacht masts and industrial springs. So confident in that number (tension comp and flexure). Would make it at E70 to improve its shear stiffness and it would be a good all round material. You will need to get familiar with infusion to go down the CF path. Wet lay up and hand layup will not achieve the consolidation required to get to E50 or better. Peter

  13. #13

    Join Date
    Sep 2022
    Posts
    12

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Thanks Peter, I will give carbon fiber some thought (may be a future upgrade).

    This is as far as I have got with the build thus far. It is worth mentioning that I have dedicated a disproportionate amount of my budget to the frame, as this is integral to achieving the rigidity of the machine and will not be upgrade-able later. I am now looking to purchase the linear motion components, and will need to economize here. I am anticipating that I may look to upgrade various components in the future, once the machine can start paying for itself, but I don't want to buy junk either. To this end, I am hopeful to solicit the opinions of members here on the selection of these parts. In an ideal world, I would have 45mm roller rails, but I'm leaning towards 35mm rails. I priced up some Hiwin HGR35 rails and carriages in the UK at around £2050 (45's were £1000 dearer). I would certainly be interested in buying from China if the savings were substantial, though I do have concerns about the quality - I would welcome any comments on this. It is very difficult to find much information on sizing LM guides, the best I have come across is to select a guide similar in size to the ballscrew (though there is little available on sizing these too). Does anyone have any such information - has anyone come across any worked through examples?

    In terms of motors, I did use an online calculator by oriental motor which has been shared here before to get an idea of sizing. I was initially considering Clearpaths, but for the size, I cannot afford these. I also saw some suggesting DMM were a better choice (I think this may have been claimed because of encoder resolution, though I cannot recall with any confidence). I was considering the 1.8kw servo and associated driver for axis control (the spindle will be dealt with later), though I think the price has increased since I originally looked at them - I've also read some comments of people regretting their purchase due to poor/slow . I have seen some here use Servo's from China, which appear to offer substantial savings - can anyone offer any guidance on the quality of these, are they worthwhile?

    Any suggestions/advice would be greatly appreciated.

    Mike

  14. #14
    Member
    Join Date
    Jul 2018
    Posts
    4351

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Hi Mike - Hiwin and other manufacturers publish design manuals on their products. They include worked examples. It's also the case you will need to determine the precision grade and the preload spec of the rail and car to meet your precision requirements. This will impact cost. Hiwin and others also have design manuals for ballscrews. The workflow is like this. 1) design the structure so it exceeds target rigidity. Then you know the mass of everything including your max part mass 2) Now you can pick the initial track and car size based on static conditions 3) Pick a design acceleration as this will determine motor sizes and inertial loads 4) pick max rapid speeds and cutting speeds. This will determine motor sizes and if the system can be steppers or has to be servos to achieve the speeds 5) Now you are getting into inertial conditions you can pick the ballscrew pitches to match the speed requirements 6) Now you can review all prior decisions based on inertial loads. If OK ie meets speeds and accels required then you can use these loads to size the ballscrew diameter. This is driven by the max forces trying to buckle it vs its length. NOW you review the entire puzzle and cost it then step back a little and make cost/performance decisions. Since you want to make $$$ on this project speed is a big consideration. I doubt you will up grade the machine in future. There are too many things connected. If you change something like rails their stack heights are different and then all sorts of things have to be changed. "Upgrade" usually means build another machine using the original machine and on sell or keep using the old machine... as the biz has grown and needs capacity.

    If you are designing this machine as a general-purpose machine this is difficult. Find a commercial machine that does the job and use this as a benchmark/reality check. If it is a special purpose machine optimise it entirely on that purpose to maximise its payback potential. Many make the mistake of making it bigger so they can do 1 job a year on the machine. This wastes floor space, makes things like ballscrews more costly as they are longer so have to be bigger, and bigger means slower. The ballscrew diameter is the biggest inertia in the system so minimizing it makes $$$ sense in terms of capital cost and performance. High performance means you get your money back earlier and you attract the right job for the biz due to your efficiency and good value.

    Sometimes cutting a corner on cost initially is a poor decision. In 2-3 years the extra 1000$$$ spent on servos for instance will pail away in terms of being able to do a job in half the time so your earning capacity is double. Do not buy anything until you have worked through the entire puzzle and covered all bases, sourced all stuff and figured the logistics. Better to be a failed paper exercise then have extra parts you can't use or find you can't get something machined because its 60mm longer than the local machinist's machine... Hobby machines are full of compromises, a business machine is a several levels above hobby if you want to earn $$$ and generally you only have one good shot at it. Peter

    HG Series - Block On Hiwin Corporation

  15. #15
    Member
    Join Date
    Jul 2018
    Posts
    4351

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Hi Mike- Here's is my motor sizing spreadsheet - Peter
    Attached Files Attached Files

  16. #16
    Member
    Join Date
    Nov 2013
    Posts
    2095

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Hi,
    Clearpath servos are over priced and under perform. Their standard encoder is only 800cppr, and their high-res encoder is still only 6000cppr, whereas my B2 Delta servos are 160,000cppr
    and even that is considered entry level with 10,000,000cppr common amongst the top brands like Yaskawa. They have only one digital output, no auxillary encoder output, but a few digital inputs,
    no analog velocity or torque modes. A 400W Clearpath is a little over $500 whereas my 750W B2 Delta is $438 (servo, drive and cables), so nearly twice the power for less.
    No.....Clearpath are not an economic solution.

    DMM (Canadian brand made in China) and Delta (Taiwanese made in China) both perform very well, good support and documentation, and most importantly free set-up and tuning software
    at reasonable prices. You are correct there are cheaper Chinese brands, and by and large they perform well, but are let down by poor or questionable support, and crap documentation
    and in most cases NO set-up and tuning software.

    Setting up and tuning servos is quite a learning curve, I would not make that more difficult than need be by buying el-cheapo Chinese. An extra $100 per servo will get you either Delta (my preference)
    or DMM and you'll be delighted you did. If you are a past master of setting up servos then the cheaper Chinese stuff is worthwhile, as all the techniques you've already mastered will overcome the poor
    documentation/ lack of tuning software but if your new to servos get ones with setup and tuning software.

    When it comes to servo sizing you might be surprised just how fast a 750W servo can swing a heavy load. My new build mill has 5mm pitch ballscrews and the axis weight represents 7.5% of
    the inertia of the axis. Once you get some numbers, like ballscrew pitch, ballscrew diameter (mucho important) and axis weight then we can do a 'broad sketch' type calculation. You may well
    find that you can achieve your desired accelerations and cutting thrust with rather smaller servos than you believe.

    When I built my mill I was lucky enough to find some top quality near-new doublenut C5 ground ballscrews by THK. I paid $1800NZD including shipping for three sets of ballscrews and support bearings
    second hand, that would have cost $3000NZD EACH new. I can certainly sympathise with trying to economise on linear motion parts but they are the heart of your system. Try second hand, you'll get C5's or even better
    C3's for less than you'll buy crappy Chinese C7's. Start here:

    https://www.ebay.com/str/industrialp....m47492.l74602

    Craig

  17. #17
    Registered
    Join Date
    Jan 2008
    Posts
    1353

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Look at Hiwin rails from BST Automation on AliExpress. BST have an excellent reputation and sell genuine Hiwin products (as well as their own cheaper brand).

    I don't think you will want to be upgrading rails later as hole spacing etc may be different. I.e. cry once.

    Have a look at Ebay etc for new old stock. I was lucky to pick up some new old stock Bosch Rexroth 45mm Roller rails for very cheap (approx $200 USD each set of a metre of rail and two cars). They are sitting patiently waiting for me to get the time to build a machine somewhat like yours, though substantially smaller.
    Often listed under brands and odd names without things like "Linear rail". Search for THK, Bosch Rexroth, Star Rexroth, etc

    Your build is looking great!

    I am very interested in UHPC and CSA based concrete. However, I have been wary as there is a lot less academic literature about machine tools built with UHPC compared to epoxy granite. I suspect DIY concrete may be more reliable than DIY epoxy granite.
    7xCNC.com - CNC info for the minilathe (7x10, 7x12, 7x14, 7x16)

  18. #18

    Join Date
    Sep 2022
    Posts
    12

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Hi all,

    Thank you for your comments, these are very helpful (and for the spreadsheet, thank you Peter).

    It is probably worthwhile elaborating somewhat on what I want/expect from this machine. The purpose of the build is really to allow me to prototype my own varied product ideas. I have no intention of becoming a professional machinist or building a job shop, but want a fairly capable prototyping machine for my personal use. I may use the machine to do small series runs but I do not expect the machine to be running all-day every day. A good analogy might be using it the way that something like a Tormach is suggested, as for hobbyists, start-ups and light industrial use. If I find that I am doing very significantly more of one type of work, I could rebuild the machine using the same frame.

    In terms of upgrades, I anticipated the rail size choice to be largely locked in i.e. an upgrade might be going from hg35 to rgw35 of a higher precision but with bolt spacing and sizing being equal. I considered that I might replace a rolled ballscrew with a ground one of equal dimensions for example. I will certainly be on the look-out for higher quality bargains; if I can find bargains like some of yours first time round I will be thrilled. (Nice tip on the search terms, thanks pippin.)

    I will take a look at the delta Servo's Craig, thanks for the tip. When I originally did very rough estimating of the motor requirements, I targeted very ambitious performance. for example, for the table axis, approx 1050mm travel, I targeted 25-30 M/min (1000-1181 inches per min) max travel speed and around 1g acceleration. I used the Oriental motor guide to get some idea of motor sizing, and assumed a 32mm diameter ballscrew for this purpose. Most of the motors I had viewed had max speeds in the range of 2500-3000 rpm, so I assumed a direct coupling and a 10mm lead. Cutting force was estimated at 1000N based on a paper which designed a small gantry mill and employed a 6kw (3.3nm and 12000rpm) spindle - this is a similar power rating to what I am considering ultimately (possibly up to 7.5kw). I cannot find the original paper I read where the author actually built the machine, but an abridged version is available at: "https://www.semanticscholar.org/paper/Improvement-on-a-CNC-Gantry-Machine-Structure-for-Sarhan-Besharaty/03d6d99c9194e1242d38956ba2ab370484f22f78"

    The weight of my table is likely to be around 400kgs and I assumed a max load of 1000kgs. The bridge crossbeam axis is likey to weigh approx 350kgs (approx 1500mm travel) and the z axis potentially over 200kgs (700mm travel) (might it be a good idea to reduce this ballscrew pitch to 5mm?) The 700mm travel on z was to allow a tilting axis upgrade later, otherwise it would have been 600mm. I have also considered that it might be preferable to favour a larger motor on the z axis rather than a counterbalance system - any thoughts on this? I may also employ a belt due to height constraints.

    I have since seen that some high end machines use far less acceleration (Hermle c400 has 0.6g axis acceleration and DMG Moris DMU 200 Gantry has 0.5). I can therefore reduce this target - 0.5-0.6g would clearly be excellent, and downsizing the motors is attractive. I'll see what I can do about calculations and finding appropriate parts.

    With regard to using UHPC in machine tools, I have noticed that epoxy granite seems to be far more widely represented, however there are some examples. Kern has moved to UHPC for the frames of their new MicroHD series for example. I also came across a paper (though I do not have it to hand) which compared a sample of UHPC to 2 samples of different epoxy granites. It found that the epoxy granite samples underwent about twice the degree of creep deformation when compared to the UHPC, which clearly has advantages in machine elements. I would be happy to offer any guidance I can on DIY UHPC if you want to do something similar.

    Thanks for your help and I would welcome any thoughts.

    Mike

  19. #19
    Member
    Join Date
    Jul 2018
    Posts
    4351

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Hi Mike - Setting these very high acceleration targets for a prototyping machine maybe taking you to a costly place. 0.3 to 0.4 would still be very very scary fast and it will reduce motor size and component size to an economical level. You have picked very expensive and very precise machines as your examples. Even if you design on 0.5g you still have to realise it with precision in every mechanical part. Plus your electronics/software will need to be the expensive type to cope with the jerk at those figures. It's not simple to start and stop large masses accurately at high g's

    Ideally your making swarf not cutting air so the machines absolute accel capability is not important. It is important when you are doing 100's of parts and the air time becomes dominant...

    Those machines and machine parts are made on toolmaker grade machines which are a cut above those machines. Not in speed but in precision. But keep at it you have to aim high.

    In regard to epoxy and creep. I have been involved with this sort of stuff for over 35 years, I'm creeping myself. Epoxy granite is a great product if you use the correct materials and the correct processes. Same as UHPC and CSA. You have post cured your UHPC for good reason. Most hobbyists (and many research stuff I have read) don't post cure their epoxy, they use the wrong type of epoxy and they certainly don't dry their aggregate and I don't expect the paper on creep did either. Epoxy and water do not mix. Water is a plasticiser in epoxy and the aggregate will have a huge amount of water in it. Depending on the environment in the next year or so after casting will determine what the part does. Also many people get the creep and relaxation mechanics mixed up with plastics, thats another topic. UHPC, CSA grout and mineral cast done right are just as stable as metals, probably more so, just more stuff to learn and get confidence with. Peter

  20. #20
    Member
    Join Date
    Nov 2013
    Posts
    2095

    Re: DIY UHPC 7 Tonne CNC Bridge Mill

    Hi,
    my new mill has an acceleration using rated torque of the servos of 0.27g, which I find more that adequate. If you genuinely want 1g accelerations then you are looking at very serious servos.

    If you use 32mm screws of 5mm pitch, then direct coupled at 3000rpm works out to 15m/min. My Deltas are rated to 3000rpm, but can (with field weakening) go to 5000rpm or 25m/min.
    I detune my machine as I find these speed just too scary fast and the machine (800kg) starts lurching around the workshop....abut 12.5m/min and 0.15g are my normal settings. If I have a longer
    job, or a repeat job with a well proven toolpath I'l take it up to 0.25g and 25m/min.

    Let make some ballpark type calculations;

    Assume 32mm diameter screws of 5mm pitch, 1m long.
    Assume a servo inertia of 2 x 10-4 .kgm2 (My 750W Delta is 1.16 x 10-4.kgm2, so this assumption will be fair in the range of 750W to 1.8kW)
    Axis weight of 1000kg

    Jscrew=(0.032/2)2.PI. 1m. 8000 x 1/2 x(0.032/2)2 (8000kg/m3=density of steel)
    =6.43 (kg) x 0.000128
    =8.23 x 10-4.kg.m2

    Jlinear= 1000 x 0.0052 / (2.PI)2
    =6.33 x 10-4 kgm2

    Thus the total inertia is:
    Jtotal= (8.32 + 2 + 6.33) x 10-4 kgm2
    =16.65 x 10-4 .kgm2

    The inertia ratio (of inertial load vs servo inertia) is (8.32 + 6.33)/2= 7.3:1, a very achievable ratio.

    Note that the momentum of the linear axis is less than the rotational momentum of the ballscrew alone, and as a consequence the rotating components, namely the ballscrew and the armature
    are going to dominate (61%) the acceleration equation. This may be hard to fathom, but physics does not lie. In truth the linear mass, the axis, moves fairly slowly while the ballscrew and armature spin
    really fast, and so dominate the momentum.

    Just as a comparison my 750W servos have a rated torque of 2.4Nm and so with this combination:

    d2w/dt2= 2.4 / 16.65 x 10-4
    =1441 rad/s2

    Which is equivalent to a linear acceleration of 1.147m/s2 or 0.11g.

    I would suggest to you that 0.1g is a little low for your design, but a 1.5kw servo would result in a 0.2g acceleration which is very adequate for toolpath following in my experience.
    If you want 1g accelerations then you'll need something like a 20Nm servo.....maybe 5kW........serious and eyewatering money.

    Craig

Page 1 of 2 12

Similar Threads

  1. How to build a UHPC 4 Ft x 7 Ft CNC Router?
    By rajhlinux in forum CNC "do-it-yourself"
    Replies: 13
    Last Post: 12-19-2021, 03:24 PM
  2. My DIY Stone Bridge Saw
    By DennisCNC in forum Glass, Plastic and Stone
    Replies: 46
    Last Post: 07-27-2020, 03:29 PM
  3. DIY H-bridge
    By medved in forum General CNC Machine Related Electronics
    Replies: 14
    Last Post: 06-26-2020, 10:05 AM
  4. Replies: 2
    Last Post: 10-21-2019, 11:21 AM
  5. CNC bridge mill for cutting aluminum.
    By DICKEYBIRD in forum Vertical Mill, Lathe Project Log
    Replies: 27
    Last Post: 03-05-2017, 03:32 PM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •