DIY UHPC 7 Tonne CNC Bridge Mill

[mactec54]

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

For a Machine build like you are doing you would be fine with DMM AC Servos or Delta the next step up would be Yaskawa, DMM will do everything you want to do and more at a very good price, would not consider any of the Chinese servos you are looking at, they have zero support and can't be considered for a build like you are doing

For Linear Rail Sizing and Ballscrew there is plenty of information that you can use from any of the Top manufacturers THK have lots if information

35mm Linear Rails would be a good starting point to check out the load rating that you will be expecting them to see and carry, there are only a few Chinese manufacture which have a brand name for Linear Rails and Ballscrews, TBI MOTION is in Taiwan and most likely the only one I would look at for a serious build like you are doing

[MikeGM]

Hi Mactec,

Thank you for your input.

I am currently trying to source the various components for this build; so far I have managed to pick up some brand new THK SHS45's with extra long carriages and the highest pre-load available that were originally destined for a DMG Mori mold making machine - surplus stock found on ebay. These will be used for my Z axis, I have also just bought a ballscrew from ebay that will be used on the Z axis; this is a 38mm diameter with 10mm lead. I don't know the accuracy grade, but the threads are ground and I am informed that it has come from a grinding machine, so am hopeful it will be reasonable. It also comes with bearings for end support. I am awaiting delivery of this. I was originally looking at 32mm diameter 10mm pitch ballscrews for the other axes but critical speed calculations appear to greatly limit the potential travel speed for such screws, so I am looking at larger diameter and pitch. I missed out on a 3820 screw for the bridge axis recently which is unfortunate, but I did notice that a picture of the nut appeared to show an oil line crimped off, so it may have been starved of oil. Unfortunately, I am not having much luck finding the longer linear rails that I need for the other two axes, and I am now leaning towards 45's for these as well. I did make an offer on some from Korea, but the shipping is prohibitively expensive it seems. I will keep looking for the time being, as I have other things to arrange, but may have to consider companies in the UK or cheaper rails from china until I can justify the expense of upgrading them. I'll take a look at TBI - do they sell directly?

In terms of the motors, I am considering the CTB servos used by badhabit in his excellent epoxy granite mill build here on the forum. He posted a manual of these servos in his thread which appears to be written in good quality English, and they are significantly cheaper than DMM and offer more power options - I was considering the 2.2kw motors (which happen to be the same power as those used by badhabit). I note that the delta servos only go up to 1.5kw in single phase (they do offer larger but I understand these are three phase). I intend to ask badhabit about his experience with these servos and whether they have any auto-tuning software but just haven't got around to it yet.

I hope to post an update soon regarding the next stage of this build concerning my approach to flattening the rail mounting surfaces, as I have some ideas for a slightly unusual approach.

Mike

[joeavaerage]

Hi,
finding linear motion components in the sizes and lengths you require is going to be extremely hard UNLESS you buy new. THK, NSK and others will very happily make
whatever size and length you require....at a price. As you have discovered grade C5 and C3 ballscrews are eye wateringly expensive, linear rail and cars only modestly less so.

When I was planning my new build mill I secured the ballscrews and linear rails/cars before I even started on the frame. In particular I was able to get three 32mm diameter,
750mm overall THK BNFN double nut C5 ballscrews including FK25 support bearings for $1000USD including FedEx shipping to New Zealand from Korea. The ballscrews are
nominally second hand but seem to have done no work. The linear rails and cars are TKH HSR's new old stock at a fraction of new price. The linear components cost a little under
$1500USD landed in New Zealand, whereas if I had bought new these components would have cost nearly $10,00USD.

Once I had those components then I designed and built the machine to make the most of them. You on the other hand have made the frame and now have to economize on linear
motion components to meet the dimensions imposed by the frame, or alternately pay the big dollars and order the right components new. Did you ever consider pricing components
before you started building? If you had, you would soon have found that getting good linear motion components at a reasonable price is actually the bottle neck.

I have also just bought a ballscrew from ebay that will be used on the Z axis; this is a 38mm diameter

38mm is a big screw. Remember the calculation I posted earlier regarding the rotational inertia of a ballscrew. This is the first moment of a 32mm screw 1m long:

J_screw=8.23 x 10^-4kgm2

For a 38mm screw, the same 1m long:

J_screw=(38/32)⁴ * 8.23 x 10_-4kgm2
= 16.36 x 10^-4kgm2


So by increasing the diameter by only 4mm you have doubled the moment of inertia, or in effect halved the acceleration of the axis with the same servo.

This leads to the next problem:

I note that the delta servos only go up to 1.5kw in single phase (they do offer larger but I understand these are three phase)

You can in fact run 2kW and 2.2kW Delta servos from single phase....but you are on the limit. Around 2kW is the practical maximum for a single-phase servo from any manufacturer.
Single-phase servos draw very high currents from the supply, around 20A peak. You're going to need a huge single-phase supply if you expect to be able to run three
2kW plus servos AND a spindle.

For the sum you have already invested in the frame and the further thousands you will invest in linear motion components, servos and a spindle would not having three-phase installed
to your workshop make sense? Then you could choose both servos and spindle to suit your requirements WITHOUT restriction.

Craig

[mactec54]

Hi Mactec,

Thank you for your input.

I am currently trying to source the various components for this build; so far I have managed to pick up some brand new THK SHS45's with extra long carriages and the highest pre-load available that were originally destined for a DMG Mori mold making machine - surplus stock found on ebay. These will be used for my Z axis, I have also just bought a ballscrew from ebay that will be used on the Z axis; this is a 38mm diameter with 10mm lead. I don't know the accuracy grade, but the threads are ground and I am informed that it has come from a grinding machine, so am hopeful it will be reasonable. It also comes with bearings for end support. I am awaiting delivery of this. I was originally looking at 32mm diameter 10mm pitch ballscrews for the other axes but critical speed calculations appear to greatly limit the potential travel speed for such screws, so I am looking at larger diameter and pitch. I missed out on a 3820 screw for the bridge axis recently which is unfortunate, but I did notice that a picture of the nut appeared to show an oil line crimped off, so it may have been starved of oil. Unfortunately, I am not having much luck finding the longer linear rails that I need for the other two axes, and I am now leaning towards 45's for these as well. I did make an offer on some from Korea, but the shipping is prohibitively expensive it seems. I will keep looking for the time being, as I have other things to arrange, but may have to consider companies in the UK or cheaper rails from china until I can justify the expense of upgrading them. I'll take a look at TBI - do they sell directly?

In terms of the motors, I am considering the CTB servos used by badhabit in his excellent epoxy granite mill build here on the forum. He posted a manual of these servos in his thread which appears to be written in good quality English, and they are significantly cheaper than DMM and offer more power options - I was considering the 2.2kw motors (which happen to be the same power as those used by badhabit). I note that the delta servos only go up to 1.5kw in single phase (they do offer larger but I understand these are three phase). I intend to ask badhabit about his experience with these servos and whether they have any auto-tuning software but just haven't got around to it yet.

I hope to post an update soon regarding the next stage of this build concerning my approach to flattening the rail mounting surfaces, as I have some ideas for a slightly unusual approach.

Mike

The most important part of any Servo is the Encoder anything below 17Bit and you are wasting your time 20Bit is the norm for any new build like what you are doing, don't cheap out or you will have an unhappy experience

[MikeGM]

Hi Craig,

Yes, I did consider pricing components before starting the build, however I did intend to begin with relatively cheap components, i.e. rolled ball screws etc from china, with a plan to upgrade later as and when I needed to. I note that I don't yet have a project in mind that will require very high tolerances. In fact I originally planned to build a router for making plugs for molds but have various ideas for projects that will require metals and don't want to have to build a new machine almost straight away; I also only have the space for one machine, so versatility is significant for me. I have noticed that rigidity is overwhelmingly the most common shortcoming for many DIY builds, so have designed my frame components such that I will have a large work envelope and they will also be very rigid. I can then upgrade components without needing to rebuild the frame. I also plan to start with a very cheap spindle, as it will satisfy my immediate requirements. I can see the value in your approach of purchasing the ball screws and sizing the work envelope accordingly, however for my application, the work envelope is actually more valuable to my immediate needs which will require a large envelope straight away.

With regard to the ball screw, I actually used the calculator available at oriental motor to gauge the motor demands for this screw, compared to others, and it indicates that for a load of 250kgs (oriented vertically), the total torque requirement to achieve a given linear acceleration is lower for this screw with 10mm pitch than for a 25mm diameter screw with a 5mm pitch. I understand the 10mm pitch makes the screw component of torque-demand far lower than for a 5mm pitch (while increasing the load component), as it's rotational acceleration only needs to be half that of the 5mm pitch screw to achieve the same linear acceleration. My understanding is that, as the lead of the screw increases, changes to the diameter of the ball screw will have less of an effect on the overall acceleration capability of the axis. Whilst going from say 32 to 38mm might double the first moment of inertia of the screw; the inertia of the screw is only a relatively small portion of the inertial components of the axis in this case (e.g. the motor's inertia is circa 30 x 10-4 kgm2) then there is the load component, so whilst it certainly makes a dent in the the acceleration of the axis, it is far from halved with the same size servo even with the same pitch screw. Just as an example, I input an axis speed of 30m/min and an acceleration time of 0.2 seconds; a 250kg load in vertical orientation and a 32mm diameter, 10mm pitch screw of 1000mm total length into the oriental motor calculator and it indicates an acceleration torque-demand of 6.984nm (RMS 7.795nm). If I change to a 38mm diameter, 10mm pitch, the acceleration torque-demand rises to 8.247nm (RMS 8.359nm). This amounts to an 18% increase in torque-demand for overall axis acceleration; certainly notable, but far less than the 100% indicated in your last post. If I decrease the pitch to 5mm, this demand rises dramatically (12.48nm for the 32mm screw and 15nm for the 38mm screw).
I have to consider, in this case, competing interests of maximizing acceleration and acquiring decent (hopefully) quality components at reasonable prices when they become available.
Incidentally, I have changed the design of my z axis slightly to reduce the travel from 700 to 600mm, and increased the spacing of the carriages, so I may be able to reduce the weight of this axis somewhat to achieve the desired rigidity.

In terms of opting for a three phase power supply, whilst this would certainly be attractive, I believe this would be prohibitively expensive where I am. I have heard of people receiving very high quotes to supply 3 phase power, so I put this out of my mind.

Hopefully, this elucidates my thinking for my design and component selection choices somewhat.

[MikeGM]

The most important part of any Servo is the Encoder anything below 17Bit and you are wasting your time 20Bit is the norm for any new build like what you are doing, don't cheap out or you will have an unhappy experience

Hi Mactec,

Thank you. I would definitely need to check, but I believe the CTB servos have 17 bit absolute encoders. I will certainly bare that in mind.

Mike

[joeavaerage]

Hi,

With regard to the ball screw, I actually used the calculator available at oriental motor to gauge the motor demands for this screw, compared to others, and it indicates that for a load of 250kgs (oriented vertically), the total torque requirement to achieve a given linear acceleration is lower for this screw with 10mm pitch than for a 25mm diameter screw with a 5mm pitch.

You think?

The mechanical advantage of a screw is the ratio of circumference to pitch:

For a 38mm daimeter screw of 10mm pitch
Mechanical Advantage= PI x 38 /10
=11.93
For a 25mm diameter screw of 5mm pitch:
Mechanical Advantage= PI x 25 / 5
=15.71

To achieve a given thrust the fine pitch screw does so with less torque. Thrust is a force and F=ma, ie the higher the thrust the higher the acceleration.

A 38mm screw 1 m long has a first moment of inertia of 16 x10^-4 kg.m²
A 25mm screw 1m long has a first moment of inertia of 2 x 10^-4 kg.m²

The angular acceleration for a given torque is 8 times better with the smaller screw.

You are correct if you choose a servo with an armature of 30 x 10^-4 kg then the screw inertia is a modest, even small proportion of the total momentum, which
begs the question why you'd choose such a servo?

Just as an example, I input an axis speed of 30m/min and an acceleration time of 0.2 seconds; a 250kg load in vertical orientation and a 32mm diameter, 10mm pitch screw of 1000mm total length into the oriental motor calculator and it indicates an acceleration

That is your mistake, you have assumed a high speed, namely 30m/min. A 10mm pitch screw at 3000 rpm is 30m/min. A 5mm pitch screw would require a 6000 rpm servo, and to accelerate to 6000rpm
would require more torque. Do the same calculation but with the 5mm pitch screw choose 15m/min, ie the same 3000rpm at the servo. You sacrifice speed for acceleration.

If I change to a 38mm diameter, 10mm pitch, the acceleration torque-demand rises to 8.247nm (RMS 8.359nm). This amounts to an 18% increase in torque-demand for overall axis acceleration; certainly notable, but far less than the 100% indicated in your last post.

That is because you have chosen a servo with such a large armature that doubling the ballscrew inertia increase the sum by only 18%, again that points out you are choosing the wrong servo.

Isn't Newtonian physics a b*****h..

Craig

[mactec54]

Hi,



You think?

The mechanical advantage of a screw is the ratio of circumference to pitch:

For a 38mm daimeter screw of 10mm pitch
Mechanical Advantage= PI x 38 /10
=11.93
For a 25mm diameter screw of 5mm pitch:
Mechanical Advantage= PI x 25 / 5
=15.71

To achieve a given thrust the fine pitch screw does so with less torque. Thrust is a force and F=ma, ie the higher the thrust the higher the acceleration.

A 38mm screw 1 m long has a first moment of inertia of 16 x10^-4 kg.m²
A 25mm screw 1m long has a first moment of inertia of 2 x 10^-4 kg.m²

The angular acceleration for a given torque is 8 times better with the smaller screw.

You are correct if you choose a servo with an armature of 30 x 10^-4 kg then the screw inertia is a modest, even small proportion of the total momentum, which
begs the question why you'd choose such a servo?



That is your mistake, you have assumed a high speed, namely 30m/min. A 10mm pitch screw at 3000 rpm is 30m/min. A 5mm pitch screw would require a 6000 rpm servo, and to accelerate to 6000rpm
would require more torque. Do the same calculation but with the 5mm pitch screw choose 15m/min, ie the same 3000rpm at the servo. You sacrifice speed for acceleration.



That is because you have chosen a servo with such a large armature that doubling the ballscrew inertia increase the sum by only 18%, again that points out you are choosing the wrong servo.

Isn't Newtonian physics a b*****h..

Craig

His machine build would require a minimum of 32mm Ballscrew, 25mm is not going to cut it no matter how you calculate it, machining centers use / start at 32mm Ballscrews and larger, a regular Bridgeport cnc Knee Mill uses 32mm x 5mm pitch Ballscrew a 750w AC Servo motor can run it at 450IPM @ 1:1 Bridgeport machining centers use 32mm x 12mm pitch Ballscrews and a 1200w Yaskawa AC Servo motor @ 1:1

Haas Mini Mill machines use 32mm Ballscrews run from 6mm Pitch to 10mm Pitch, depending on what machine model, IPM range from 600IPM for the 6mm Pitch and 1200IPM for the 10mm Pitch Ballscrews all using the Same 850w Yaskawa AC Servo motor with this AC Servo motor they have 2000ft lbs. of torque on X Y Z axis they are all 1:1

[mactec54]

His machine build would require a minimum of 32mm Ballscrew, 25mm is not going to cut it no matter how you calculate it, machining centers use / start at 32mm Ballscrews and larger, a regular Bridgeport cnc Knee Mill uses 32mm x 5mm pitch Ballscrew a 750w AC Servo motor can run it at 450IPM @ 1:1 Bridgeport machining centers use 32mm x 12mm pitch Ballscrews and a 1200w Yaskawa AC Servo motor @ 1:1

Haas Mini Mill machines use 32mm Ballscrews run from 6mm Pitch to 10mm Pitch, depending on what machine model, IPM range from 600IPM for the 6mm Pitch and 1200IPM for the 10mm Pitch Ballscrews all using the Same 850w Yaskawa AC Servo motor with this AC Servo motor they have 2000ft lbs. of torque on X Y Z axis they are all 1:1

I need to correct this Post to 2000ft lbs. of thrust not torque per X Y Z axis with the 850w Yaskawa AC Servo motor

[joeavaerage]

Hi,
OP has said he wants or possibly needs to economize on linear motion components yet he is using outsize ballscrews (38mm) which require outsize servos. For a screw 1m
in length 25mm x 5mm would be adequate or 32mm x5mm perhaps better.

In the examples you have mentioned they have all used servos between 750W and 1200W, all of which will have way less inertia than the 30 x 10^-4 kg.m²
quoted by OP.

Craig

[peteeng]

Hi Mactec - You mean 2000 Lb force (907kgf)not 2000 ft lbs (torque value)? Peter

[mactec54]

Hi,
OP has said he wants or possibly needs to economize on linear motion components yet he is using outsize ballscrews (38mm) which require outsize servos. For a screw 1m
in length 25mm x 5mm would be adequate or 32mm x5mm perhaps better.

In the examples you have mentioned they have all used servos between 750W and 1200W, all of which will have way less inertia than the 30 x 10^-4 kg.m²
quoted by OP.

Craig

No, the 25mm would not be adequate for his machine, and screw length, the 38mm he said was for the Z axis which would be fine and will not be an issue, your lack of experience is showing again

A 1Kw servo with a 2:1 ratio will drive that 38mm Ballscrew and have plenty to spare, if he uses a counterbalance for the Z axis, he won't have any problems

[mactec54]

Hi Mactec - You mean 2000 Lb force (907kgf)not 2000 ft lbs (torque value)? Peter

Yes, correct 2000 lbf. or 8896 N thrust

[MikeGM]

Hi,

You think?

The mechanical advantage of a screw is the ratio of circumference to pitch:

For a 38mm daimeter screw of 10mm pitch
Mechanical Advantage= PI x 38 /10
=11.93
For a 25mm diameter screw of 5mm pitch:
Mechanical Advantage= PI x 25 / 5
=15.71

To achieve a given thrust the fine pitch screw does so with less torque. Thrust is a force and F=ma, ie the higher the thrust the higher the acceleration.

Using, in isolation, the equation you have provided here, you could be encouraged to increase the diameter of the screws to say 50mm and achieve, in the case of the 10mm pitch screw, exactly the same mechanical advantage as the 25mm diameter 5mm pitch screw. In the case of the 5mm pitch screw, if you double the diameter to 50mm, the formula you have provided indicates that you would double the mechanical advantage to 31.42. Your statement would then lead the reader to conclude that this would result in more thrust from a given torque and therefore a higher acceleration of the axis; however, if you were to input this example into the calculator for the conditions described earlier, despite the massively increased mechanical advantage, you would roughly triple the acceleration torque-demand of the overall system.
It therefore appears that looking at mechanical advantage in isolation would be a poor means of estimating the overall system demand. The calculator evidently takes into consideration other variables.

You are correct if you choose a servo with an armature of 30 x 10^-4 kg then the screw inertia is a modest, even small proportion of the total momentum, which
begs the question why you'd choose such a servo?

In terms of the motor selection, I am certainly not fixed on a motor, but this rotor inertia seems similar to other servos I have checked in the roughly 1.8 -2.2kw class with rated speed of 1500rpm. You may recall from earlier in this thread the 1.8kw dmm servo I am considering had rotor inertia of 23.8 x 10-4kgm2 and the 1.8kw leadshine ELM servo cited had 30.15 x 10-4kgm2, so for 2.2kw and 14nm rated torque, 30 x 10-4kgm2 rotor inertia doesn’t seem abnormal. I note that, in the calculations you performed earlier in the thread, they showed the aforementioned Leadshine servo with 30 x 10-4kgm2 rotor inertia, when coupled to a 10mm pitch screw, achieved slightly better linear acceleration and equal travel speed when compared to the more powerful 2kw delta with only 4.45 x 10-4kgm2 coupled to a 5mm pitch screw of equal diameter.

I am also under the impression that servos are often able to operate with inertia ratios of say 10:1, but that it is preferable to be closer to 1:1 for the servo to be best suited to control the loads most effectively (I have been considering this on the X and Y axes, where that ratio would still be well above 1:1). As I say, I am not settled on a particular servo, but for the purposes of inputting details into the aforementioned calculator, I have started with a motor that I have already looked up (originally for the table axis) and kept this the same for the variable conditions i.e. screws, load etc. I understand this is typical. I have had to look at larger screws than 32mm diameter, and with a larger pitch than 10mm for the X and Y axes, due to the limits imposed by critical speed calculations. (For a 1700mm unsupported length 3210 ball screw, with a root diameter of 27mm and fixed-supported configuration, I would be limited to around 1100rpm, assuming an 80% safety factor of critical speed – this seemed like it would be too slow).

That is your mistake, you have assumed a high speed, namely 30m/min. A 10mm pitch screw at 3000 rpm is 30m/min. A 5mm pitch screw would require a 6000 rpm servo, and to accelerate to 6000rpm
would require more torque. Do the same calculation but with the 5mm pitch screw choose 15m/min, ie the same 3000rpm at the servo. You sacrifice speed for acceleration.

Your proposal to equalise for ball screw rpm and angular acceleration rather than the result I’m ultimately looking for i.e. linear speed and acceleration seems odd to me. Clearly, if you accelerate and spin the two screws at the same rate, the system employing the larger screw would demand more torque but I don’t see why you would do that. If you equalise for linear speed and acceleration instead, the aforementioned calculator still shows the system employing the 38-10 screw demanding less torque at 15m/min max speed for the given conditions. If you equalise for rpm and angular acceleration then you would have one screw achieving 15m/min max speed and the other achieving 30m/min, and accelerating to this higher linear speed in the same time. If you’re satisfied with 15m/min and want to increase linear acceleration, you would still only need to rotationally accelerate the larger screw at half the rate of the lower pitch screw and could employ a belt gear reduction to double the torque delivered to the screw. (I will be using a belt on this axis in any event due to height constraints.)

I took a look at the formulas provided for calculating torque requirements for ball screw driven systems on linearmotiontips.com found here: https://www.linearmotiontips.com/cal...e-ball-screws/. It shows the following:



If I input the figures for the system while employing the 3810 ball screw and the aforementioned 1.8kw dmm motor, I get the following:

Jm = 23.8 x 10-4kgm2
Js = 16.36 x 10-4kgm2 (your calculation)
Jl = 6.34 x 10-4kgm2 - 250kg x (10/6.28)2 x 10-6
Therefore J = 23.8 + 16.36 + 6.34 = 46.5

?’ = 2pi x 1500 / 60 x 0.1 = 1570.8 rad/s2

Tacc = 46.5 x 10-4kgm2 x 1570.8 = 73,042.2 / 10000 = 7.3nm

If I do the same but change the screw to the 3205 screw you recommend, and equalise for linear performance, I get:

Jm = 23.8 x 10-4kgm2
Js = 8.23 x 10-4kgm2 (your calculation)
Jl = 1.58 x 10-4kgm2 - 250kg x (5/6.28)2 x 10-6

Therefore J = 23.8 + 8.23 + 1.58 = 33.61
?’ = 2pi x 3000 / 60 x 0.1 = 3141.6 rad/s2
Tacc = 33.61 x 10-4kgm2 x 3141.6 = 105,589.2/ 10000 = 10.56nm

You could certainly make the case for the rotor inertia being disproportionately high, particularly in the example with the 5mm lead screw, but as I said, the high rotor inertia motor was a starting point due to previous consideration of the other axes and the calculations you provided earlier. Looking at the 2kw delta you referenced earlier with only 4.45 x10-4kgm2, and 6.37nm of torque, I get the following for the system with the 3205 ball screw:

Jm = 4.45 x 10-4kgm2
Js = 8.23 x 10-4kgm2 (your calculation)
Jl = 1.58 x 10-4kgm2 - 250kg x (5/6.28)2 x 10-6
J = 4.45 + 8.23 + 1.58 = 14.26
?’ = 2pi x 3000 / 60 x 0.1 = 3141.6 rad/s2
Tacc = 14.26 x 10-4kgm2 x 3141.6 = 44,799.2/ 10000 = 4.48nm

Employing this same motor with the 3810 ball screw I have purchased, I get the following:

J = 4.45 + 16.36 + 6.34 = 27.15
?’ = 2pi x 1500 / 60 x 0.1 = 1570.8 rad/s2
Tacc = 27.15 x 10-4kgm2 x 1570.8 = 42,647.2/10000 = 4.26nm – much closer now with the lower rotor inertia motor, but still lower than the smaller 32mm diameter screw with 5mm lead/pitch. In this configuration, it would also be possible to employ a 2:1 reduction to effectively double the torque at the screw from the same motor, and still achieve the same travel speed as the 5mm pitch screw.

These figures are all the same as the results I get from the aforementioned calculator, subject to minor variance due to rounding differences. The calculator is clearly much more convenient.

Mike

[joeavaerage]

Hi,

It therefore appears that looking at mechanical advantage in isolation would be a poor means of estimating the overall system demand.

You are missing part of the argument, mechanical advantage is only a part of the thrust equation. If you apply 1Nm of torque to a screw, that is by definition 1N at 1000mm
then at the periphery of the screw the force is 1000/radius. So the total equation for the thrust is:

Thrust(per Nm applied) = 1000/r x 2 pi x r/p where r is radius in mm and p is pitch in mm
= 1000 x 2 pi/p

Note that radius drops out of the equation, the sole remaining variable is pitch.

For instance you say your Z axis weighs 250kg or 2500N. The torque required to hold that 250kg stationary without a counterbalance is:
1) 38mm x 10mm screw Torque=3.9Nm
2) 32mm x 10mm screw Torque=3.9Nm
3) 25mm x 10mm screw Torque=3.9Nm
4) 25mm x 5mm screw Torque = 1.98Nm
5) 32mm x 5mm screw Torque =1.98Nm
6) 38mm x 5mm screw Torque=1.98Nm

The fine pitch screws have advantage irrespective of diameter.

I have checked in the roughly 1.8 -2.2kw class with rated speed of 1500rpm. You may recall from earlier in this thread the 1.8kw dmm servo I am considering had rotor inertia of 23.8 x 10-4kgm2 and the 1.8kw leadshine ELM servo cited had 30.15 x 10-4kgm2, so for 2.2kw and 14nm rated torque, 30 x 10-4kgm2 rotor inertia doesn’t seem abnormal.

The inertia of the servo is determined by the diameter and mass of the armature. The servos that you are looking at are mostly 130mm or 150mm frame size, whereas you can get the same power output from a smaller servo.

For example:

https://www.fasttobuy.com/delta-2kw-...nc_p33936.html
This is a 2kW servo 100mm frame size 6.35NM at 3000rpm and an inertia of 4.45 x10^-4 kg.m²

https://www.fasttobuy.com/delta-2kw-...ng_p33939.html

This is a 2kW servo 130mm frame size 9.55Nm at 2000rpm and an inertia 14.59 x 10^-4 kg.m²

Thus you have a choice, you could have a low inertia servo OR a medium inertia servo but with the same power output.

I am also under the impression that servos are often able to operate with inertia ratios of say 10:1, but that it is preferable to be closer to 1:1 for the servo to be best suited to control the loads most effectively

In fact modern servos can with things like notch filters and other tuning aids can have an inertia ratio of 20:1 or even more, but by and large you are correct, the norm is between 5:1 and 10:1.
If you went down to 1:1 the 50% of the energy you put into the servo is actually the kinetic energy of the servo itself? Why would you consider that benefical?. Stated another way...the maximum
acceleration is determined as much by the inertia of the servo rather than the load.

At 7.5:1 then 86% of the energy you put into the servo actually gets to the load with the servo kinetic energy being only 13%. Stated the other way the maximum acceleration is determined
in large (86%) manner by the load NOT the servo.

You can certainly use an outsize servos......but is that the most effective use of your money?

In the examples that Matec has provided the servos range from 750W to 1200W, and yet they are machines as big as yours. Do you really need 1.8kW- 2.2kW per axis? The calculator you are using
is leading you that way because you are using such large ballscrews, which are hugely expensive in themselves.

As an example my new mill has 32mm x 5mm screws 750mm long. With a 150kg axis mass and a 750W Delta servo it accelerates at 0.27g or 2.7m/s². Note this is just rated torque,
not overload torque or anything like that. At rated speed (3000rpm) its axis speed is 15m/min, but I run them to 5000 rpm for 25m/min. The axis can accelerate to max axis speed in 0.15sec.
Remember this is only a 750W servo, $438USD......so pretty good value. So I could have double the acceleration by using a 1.5kW servo at twice the price......but why? In practice I use
only about 25%-50% of the max axis speed, and about 60% of the max acceleration. Only when I have a perfectly repeatable toolpath and multiple parts to make that I ever push it out
to the max. Even if I had more axis power I doubt I would use it all. My budget is better spent on a better spindle or better coolant pumps and filtration.

You have posted that you want or need to economize on linear motion components....so do so. Using smaller diameter and fine pitch ballscrews may well save you dollars right there but will
also require servos in the 750W-1kW class and that will save you dollars also. Machine acceleration of 0.25g is very adequate as are machine speeds of 15m/min for a machine in the home
workshop.

Craig

[Momentz]

All you pictures are broken, nothing shows up.

Sent from my SM-G781B using Tapatalk

[vnuEndru]

Hi,
I'm making small benchtop cnc mill and also decided to go with uhpc. This is first time I see someone making his own uhpc mix like me. I thought I'm the only one xD. I'm super interested in detail of your uhpc mix - proportions, what exact components did you use (especially silica fume and PCE), what your w/c ratio? Also as I can understand you didn't test your mix? I tested mine for compressive strength and it resulted in 110Mpa which is low, but its with high w/c. I succeeded in lowering w/c, samples are curing rn.

[peteeng]

Hi VNU - Do not put much interest in compressive strength, but you need to test for its young's modulus. There are many commercial engineering grouts available and they are economical. If you only get 30GPa modulus you may as well use commercial mixes that get 40-50GPa modulus. I can buy non shrink grout with over 120MPa comp strength and 56GPa modulus at the local hardware in 25kg bags so why invent it yourself? Its water ratio by weight is 6-7%. Why are you using fumed silica in concrete? Hope your wearing a serious mask, silicosis is bad news.... Peter

[vnuEndru]

Hi VNU ...

Hi,
Yes, you right, as far as I know young modulus correlates with compressive strength in uhpc, so I use compressive strength as indirect indicator for young modulus for now. I will incorporate some alumina aggregate and test for young modulus next ofc. Unfortunately I don't have any mixtures commercially available in my country. The only options is import (which is very expensive) or diy.
Silica fume is not the same as fumed silica. Silica fume is pozzolanic supplementary cementitious material which is usually used in all this commercial mixtures such as ducrete to achieve theirs superb characteristics. There are others pozzolan like metakaolin, but as far as I know SF is the most efficient. I know about dangers of working with materials like SF. I do my best in safety regards.

[joeavaerage]

Hi,

Yes, you right, as far as I know young modulus correlates with compressive strength in uhpc,

What sort of evidence do you have for that belief?

The reason I ask is that I've often wondered why for instance mild steel which has a Youngs Modulus of 205GPa while hardened and tempered 4340 has about the same modulus but
probably three to four times the strength. It would appear that the tensile strength does not correlate to Youngs Modulus. Even maraging steel up to 2650MPa tensile is still about 200GPa Youngs Mod????

Craig