Initial design advise wanted

[driftspin]

Hi,

I'am new to cnc and looking to build a steel gantry type machine, capable of aluminium, and wood.

Hobby use, although you never know.
3d work is a wish.


After reading build logs for months now i would like to bypass build mk 1 an mk 2 [emoji1] [emoji1] and go for a reasonably good machine.

Work area +/-1250mm x +/- 700 mm x +/- 150 mm (bed adjustable in height)

Constructed mainly out of steel box section (120x80x4 and 80x80x4)

Zaxis aluminium

Designed around 20 square rail , 1605 / 1610 ballscrews, 3nm nema 23 steppers. Watercooled 1.5 / 2.2 kw spindle.

I am reading up on electronics... i am not up to speed yet.
Digital drives separate stepper processor pc / software

I am also not sure if i will need flood, mist, spray or just airblast cooling.
Or how to put this in my design at this point.

I have not yet desided if the machine will be table top or on legs of its own.


One sides of the frame support will be removable to accept sheet material 4x8 sideways.

Not sure about belt or direct drive for y and x or slave motor for twin ballscrews.

I think y and x will see some epoxy before mouting the rails.
Although for the gantry (X) i am still considering having the rail and block (Y) mounting surfaces machined parallel and reasonably flat.

I have access to solidworks (had no training) and basic tooling.

My mig co2 welder is a 155 amp cebora 4mm is about the max it will do.

I will have to buy some measuring equipments like

a basic straight edge 1500mm,
90deg square angle and
0.01/0.001 mm indicator.



This is my first real post and or design please be gentle ☺.



http://uploads.tapatalk-cdn.com/20160819/52555aff9e1fae75d51412c5c5ff2927.jpg


http://cloud.tapatalk.com/s/57b7638dc5df3/total.JPG


http://cloud.tapatalk.com/s/57b76398b6b67/Gantry%20Z.JPG



Verstuurd vanaf mijn SM-G7102 met Tapatalk

[driftspin]

http://cloud.tapatalk.com/s/57b76ac7a1c6c/Gantry%20Z.JPG

http://cloud.tapatalk.com/s/57b76adb090cc/total.JPG


Verstuurd vanaf mijn SM-G7102 met Tapatalk

[Clive S]

Hi driftspin and welcome to the forum. I can tell that you have been doing your homework and have got off to a good start.
Ask as many questions as you think fit. As usual the general advise is don't buy any electronics etc until you have a final plan sorted and keep away from kits. Good luck with the build.

[driftspin]

Thanx for having me 😅

For now i guess the ballscrews are going to be typical C7 rolled Chinese.

The 1st question i have at this point.
what would give best results.

A) Slave the 1600mm balscrew steppers or;
(i am not up to speed on how this actually works but i do get the general idea)

B) Connect the ballscrews through a belt and a single stepper.

And is then still a nema 23 3nm ?

Total length of this belt would be +/- 2300mm.




Verstuurd vanaf mijn SM-G7102 met Tapatalk

[Clive S]

what would give best results.
A) Slave the 1600mm balscrew steppers or;
(i am not up to speed on how this actually works but i do get the general idea)
B) Connect the ballscrews through a belt and a single stepper.
And is then still a nema 23 3nm ?

Well that question has been debated a lot on the forum and my answer would be non generally is better than the other.

Option B You would possibly need a nema 34 with a high voltage drive like 230V and also you are guaranteed never to rack the gantry. You would also need idler pulleys on the belt.

Option A. you use two drives similar to AM882 (they have stall detection built in) you also need two homing switches so that the controller can make sure the gantry is squared with the homing routine.

The screws in my opinion are best driven by short belts.

Both systems are used widely on the forum.

[driftspin]

Ok thanx for this awnser.

It looks like i can still place the screws so that both options stay possible.



Verstuurd vanaf mijn SM-G7102 met Tapatalk

[driftspin]

I have a sound proofing question.

So i have this typical backyard shed.
Its roughly 4 by 6 meters.

Constructed with concrete floor.
Walls are made by stacking 44mm wooden interlocking planks.

I live in an urban area.. so lots of neighbours nearby. Row housing... yup its not ideal for a cnc hobby.

So what works best on insulating cnc cutting noise... maybe maybe rockwool and plywood /dry wall on the inside?

Anyboy got better solution / idea? Instead of moving of course :-)




Verstuurd vanaf mijn SM-G7102 met Tapatalk

[Nr1madman]

Very good question! I have just built a shed for my future router. I went with 120mm insulation in the walls , then windprotection and 22mm woodpanels. Have not closed the walls on the inside, should I add something more before osb boards and drywall?

[driftspin]

Haha it will never be cold in there 😊😊😊

I would definitly go for a layer of wood on the inside of the wall... maybe drywall over it.

The wood makes mounting stuff easy...



Verstuurd vanaf mijn SM-G7102 met Tapatalk

[routercnc]

Hi driftspin

Noise reduction is related to how heavy the walls are. Look up 'mass law sound reduction' for more info.

There are some basic calculations on this site:
http://www.engineeringtoolbox.com/sound-transmission-massive-walls-d_1409.html

Let's say with the cnc machine running noise levels inside the workshop are 90dB. If the surrounding urban noise levels near the workshop are 50dB (with the machine off) then you need to achieve 40dB reduction.

With concrete at 2300kg/m3 and 0.2m thick this gives surface mass of 460 kg/m2. This gives an average noise reduction of 52dB, which would be enough.

With wood at 400kg/m2 and 0.044m thick this gives surface mass of 17.6 kg/m2. This gives and average noise reduction 28dB, which is not enough on its own.

Assuming you are stuck with wood, this can be improved by adding another wall with an air gap in between. Using the air gap you get more noise reduction than if you just stuck the walls back-to-back with each other - but they must not touch each otherwise the vibration will pass easily from one to the other and the benefit is lost. So mount them on separate vertical studs. The larger the air gap the better the performance, but workshop space is lost.

Finally you can add absorption material in-between the walls which takes out the high frequencies (it does nothing for mid to low frequencies which is why you need the double wall first). This can be anything fluffy, so rock wool is fine.

Make sure you seal any gaps otherwise lots of the noise benefit is lost. Make sure the roof, door and window specs are considered as well (plus floor if it is just suspended wood).

[JAZZCNC]

The 1st question i have at this point.
what would give best results.

A) Slave the 1600mm balscrew steppers or;
(i am not up to speed on how this actually works but i do get the general idea)

B) Connect the ballscrews through a belt and a single stepper.

And is then still a nema 23 3nm ?

I can tell you thru much experience that Twin screws are much more reliable if connected with timing belt. They remove any chance of screws becoming out of sync.
The only down side is they require larger motor and drive. If the machine is wide then they can also get a little unweildy. In this case Twin motors can work best. At the size you plan they will work great.

Using screws joined with belts my prefered setup over slaved motors.

[driftspin]

Thank you routercnc.

This is what makes this site so great for me.

You guys give excellent advice.
Backed with theory explained in a simple way.

So is the number 90db a real world guess? I have no clue on this...

It might be a deal breaker btw....


I guess high freuquent sounds s the ones that are most annoying.

Your example says 20cm concrete as the ideal solution.. i think i agree, but can't do that though...



so rockwool air gap sounds like the way to go... i will look in to this.






Verstuurd vanaf mijn SM-G7102 met Tapatalk

[mekanik]

Have a look @ this
http://www.fermacell.co.uk/en/docs/fermacell_Orange_Book_2014_082014__s.pdf
It gives construction methods for various grades of soundproof stud partition wall. nice stuff to work with(nothing like standard plasterboard)
and you can get it in one man boards(1200 x 1200)
Regards
Mike

[Boyan Silyavski]

I have tried to soundproof my garage. Almost wasted effort. I used 30db or sth sound isolation material on garage door inside. Nahhh.

The best soundproofing material is Asphalt sheet heat glued to whatever door you have . Noise will not come from walls. It will come through windows or through doors. Basically you need all outside possible vibrating surfaces to be hard dampened, not just covered with sth. Of course depends where you live though. I live in extremely quiet neighbourhood where people pay big $$ to come for 10 day at sea and quiet, so my challenges are bigger than yours. I simply can not work at night, but daytime you can not hear it from neighbours.

[driftspin]

Mike, thanks for the farmacell link.
I can get their product here locally too.

Boyan sorry to hear you feel like you failed to soundproof your shop.

Sounds like soundproofing aint that easy.

What i read about the farmacell panels sounds promissing though.


Maybe build a free standing 2.5 x2.5 x 4 meter box with a solid wooden door ( it will be the weakest point) and make a the pc station outside.
Or make 2 compartiments having 2 doors between machine and inside of the shed.



I feel like when the machine is working in a (smaller and insulated) closed box, temperatures might rise quickly since most electrical energy used by the machine will be turned efficiently [emoji1] [emoji1] [emoji1] in to heat.

So soundproofed ventilation is the next requirement.











Verstuurd vanaf mijn SM-G7102 met Tapatalk

[driftspin]

Ok... so i am in a bit of a struggle... i am not sure about my gantry beam.. i have not found a way to calculate deflection for the beam.. Iam not a strucural engineer ☺

So my beam is 940mm wide between the rail blocks. It is a 120h x80w x4t mm profile.

The maximum z extension is about 25cm on the tip of the tool.
Planned working extension maybe 15cm


The tool is extended about 15 cm.

Can sb point me to a tool to guestimate deflection based on real world cutting forces?

Is read on this forum cutting forces in any given direction do not exceed 400Newton? In soft metals.


I need this info to guestimate the total weight of the gantry and continue on to electronics selection...


Any help is welcome here.

Bert.

Verstuurd vanaf mijn SM-G7102 met Tapatalk

[Boyan Silyavski]

cnc machine stiffness calculator (http://www.mycncuk.com/threads/2214-cnc-machine-stiffness-calculator)

for soft metals 120x80 x4 is not rigid enough . Weld 2x / 1 thin side and one wide side 10mm steel plate/ and you will have reaper

[driftspin]

Ok sb is good with excel... wow.
This is exactly what i needed


Boyan.. thanks for the link.

Hmm deflextion is within 0.01 mm not sure cutting what material.

Will have a better look tomorrow.

I think i better step up to a bigger size gantry beam maybe 120x120x10, there will be more meat to mill flat also.

BTW i cant weld more than 4mm.
Oh... endcaps ... hmmm well...
Need to figure out that one

I will bolt the gantry beam to the bearingblock mounts i guess then.



Verstuurd vanaf mijn SM-G7102 met Tapatalk

[Boyan Silyavski]

If its to cut wood and occasionally aluminum its perfectly fine.

But apart from deflection, there is ringing, vibration, when cutting metal. Ideally when an engineer is designing a machine he will know frequency of the tool - material and space bearings of Z and all other parts so it does not synchronize so all starts ringing like a bell. And i suspect use much more weight than needed just to be sure .


I remember what Dean said about one commercial machine , an expensive one, that if you hit it with a hummer it would "thump" not ring. So thats also important especially with metal machining.


In a DIY situation after designing a couple of machines as light as possible and as rigid as possible / blamed to be OTT even / i believe using 10mm plate as an additional steel bracing here and there we could imitate at least 5-8 times heavier machine properties.


For example my machine /1300x2600x200/ weighs 700kg all included and is at least as strong as 4 ton commercial one, if not more. You could lift it and hold it from one side, other side in air and deflection will be less than 0.05mm if not less than 0.01/


So remember the magic word- " bracing" with thick 10mm steel in all possible directions to imitate solid shape properties and at the same time keep things light. If you are making a machine to work aluminum. For wood your problem always will be not this but the spindle KW and how fast the machine could move.

[driftspin]

Ok.. so i did some calculations.

Based on my original design the total moving gantry weight is +/- 65kg

~32kg z
~65kg z+y


Aditional bracing would add some 35kg.

Can this still be run on twin nema 23 / 3.1 nm gearing 1:2/ 1610 screws (y)

These numbers are a bit overwhelming to be honest. ..




Verstuurd vanaf mijn SM-G7102 met Tapatalk

[Clive S]

Ok.. so i did some calculations.
Based on my original design the total moving gantry weight is +/- 65kg
~32kg z
~65kg z+y
Aditional bracing would add some 35kg.
Can this still be run on twin nema 23 / 3.1 nm gearing 1:2/ 1610 screws (y)

These numbers are a bit overwhelming to be honest. ..

Verstuurd vanaf mijn SM-G7102 met TapatalkI am pretty sure they can do you have a drawing you can put up.

[driftspin]

Hmmm initial design is in post 1

Verstuurd vanaf mijn SM-G7102 met Tapatalk

[JAZZCNC]

Can this still be run on twin nema 23 / 3.1 nm gearing 1:2/ 1610 screws (y)

Yes it will be fine. Your good to around 90Kg

[Boyan Silyavski]

You mean additional bracing will add 35kg to the 65kg? It will become a bit heavy for steppers then. Problem starts that it's ok for steppers at 1:2, but steppers are not like servos to spin reliably at 3000rpm or even 1000. So i think then it could be better to use bigger motors/ that will need bigger and much higher voltage drives /


Then just brace with 1 plate on the 80mm side from bottom of the gantry beam. That should be enough. You keep the weight right, also the rails themselves when bolted on are very strong. As bend from forward backward movement could be the biggest eventual problem and that will prevent it . You could always bolt sth at the back later if needed.

Or just don't brace at all, i like a bit OTT so you may not listen to me. When build my first machine and my current one, at the end i decided to make everything strong as i think it must be and then think how i will move it.

Have you thought for servos? I assure you i will never go back to steppers again. I could point you to servos that will not break the bank, still untested though by me. Direct from manufacturer. I think they must work well. Around 600$ 4 x AC servos with drives, you have to add transformer and make PSU for 100$ more.

Steppers will be half the price though and work fine more or less up to normal 5-8k mm/min feeds depending on screws and other details. Servos 1:1 could move up to 30k mm/min, which is ultra nice. BUt at the end i moved my machine today for first time with 16k mm/min, so all depends on what your final purpose is.

So to simplify, or stick with your initial plan, or build as you like and go with servos.

[Clive S]

Hmmm initial design is in post 1

Verstuurd vanaf mijn SM-G7102 met TapatalkYes but is that the design with the extra bracing you want to put on?

[driftspin]

Well the extra bracing will add 35 kg to 65 kg...

Boyan suggest this for the purpose of anti ringing while cutting metal

At first i thought a heavier profile would do the same but not sure now...

Stuf to think about...

Not sure if i am ready for the extra cost involved for servos.. which i am sure are nice for high acceleration and heavy gantry




Verstuurd vanaf mijn SM-G7102 met Tapatalk

[driftspin]

Yes but is that the design with the extra bracing you want to put on?

Clive eh no, the initial gantry design is 65 KG with no extra bracing (80x120x4). The profile used is 940mm between the linear blocks.


193311933219333

The profiles on the gantrysides will have inserts so the m5 bolts can still be reached but sides ways flex is stopped.

Maybe i need to brace the Z slide on the sides ... aluminum can be flexi i heard.


What i am thinking about is maybe step up the gantrybeam to 120x100(or120x120)x5mm .... the linear rail wil be mounted close to the edge of the profile, for maximum stifness.
It will also generate some extra space for cable management, i have to look in to that too for the long axis also.








Some diagonals in the legs. maybe panel icw kabinets.. to be determined.
Filling the Highlighted parts with sand may do some good anti vibration and ringing?

[driftspin]

Electronics ... headache, questions...

Steering in the direction of 65volt dc PSU, 80volt drives 500 Watts or so will do....
And will keep me on track building a reliable machine...

But....
I dont want to go OTT, but dont want to under engineer either.

So after reading a lot of stuff i still can`t decide what voltage PSU/Drivers i would need to run my machine well.
Online-electronics-kits all list 36volt-PSU and 50volt dirers, that is confusing.

So here are a few questions.

The best guestimate i have for the total moving weight of the gantry is about 65-70KG
(including steppers, rail+blocks z axis 2.2kw spindle, steel, alu plate etc)

The best guestimate i have for the total moving weight Y+Z = +/-35 Kg.

The best guestimate i have for the total moving weight Z = +/-20 Kg.

X/Y 2:1 stepper : 1610 Ballscrew
Z 1:1 stepper 1605 Ballscrew

Main questions:
Q Y-axis: 2X slaved Nema 23 3nm 4Amp should be good, or OTT for my guestimated weight?!? 65-70KG
Q X-axis 1X Nema 23 3nm 4Amp should be good, or OTT for my guestimated weight?!? 35 Kg.
Q Z-axis 1X Nema 23 3nm 4Amp should be good, or OTT for my guestimated weight?!? 20 Kg.

Q What is the real advantage of a DIY Toroidal PSU, just noise?
Q What DC PSU voltage should i be running?
Q Stepperdrivers, Leadshine? 50-80Volts?

Q 1 or more PSU`s?
If 2 PSU`s. what setup, like X and Z on 1 PSU,
and the slaved-X set on 1 PSU? Or mix it up?

I think for now my initial PC setup would look like this for now.

- Dedicated networked windows 7 PC.
- Mach 3/4 through PP.
Q 64Bit or 32Bit?
Q PCI or mainboard PP`s?
Q PCI PP, what to look for when buying one.
Q Any Minimum video card requirements?
Q Maximum length PP cable to BOB.


Thanks in advance, Bert

[njhussey]

What works well, and is proven, for this size setup is 80V drivers, 3Nm steppers coupled with a diy torroidal psu that gives 60V - 72V after rectification. Advantage of the diy transformer is you can make it for the exact voltage you want, it gives smooth power and the capacitors absorb and back emf from the steppers on deceleration.

You will need 2 psu's one for the stepper drivers and one for your bob, home & limit switches, relays etc.

These days I'd be more inclined to ditch the pp and go USB or ethernet, more reliable.

Sent from my HUAWEI VNS-L31 using Tapatalk

[Clive S]

I think for now my initial PC setup would look like this for now.

- Dedicated networked windows 7 PC.
- Mach 3/4 through PP.
Q 64Bit or 32Bit?
Q PCI or mainboard PP`s?
Q PCI PP, what to look for when buying one.
Q Any Minimum video card requirements?
Q Maximum length PP cable to BOB.Bert From what I understand is that the pp does not work with win7 64bit.
For the system you are building you need a motion controller not via the pp. there are various ways of doing this.

Plus one with Neil on the power supplies. use 24V for the estop, limits etc .

Have you decided on whether to run one nema 34 with a belt between the two motors or 2 nema 23 and slaving them there are cons and pro between the two systems I use slaved I think Neil uses one motor.

[driftspin]

So... choosing between slaving 2 nema23 and a single nema34 ...
Or maybe a servo like Boyan said.

As far as i understand is a debated item...

Lookslike it comes down to speedy acceleration but complicated setup vs more robust setup but more expensive parts... and maybe inertial acceleration is slower due to heavier motor (bigger diameter) and 2 balscrews instead of 1 to drive..

Am i right about the above?

Verstuurd vanaf mijn SM-G7102 met Tapatalk

[driftspin]

BTW...

I am adding up prices based on 4 nema23 and linearparts from chai.
Steel frame and all shielded cable, magnetic limit switches.
4 psu's
2x65v dc diy toroidal
1x24v dc din rail
1x5 v dc ?!? Samsung smartphone USB charger hahaha

Waterpump for spindle.

Paint, 105 209.

Aluminum for Z
Fastners
Smal tools like spiral taps and stuf

Chinese 4 bearing vfd/2.2kw spindle.

But no motion controller ...yet...
Figured... mach3/4 is needed anyways... and a BOB too...
So a motion controller can be added later when room is spared in the controls cabinet.

+/- 3500 euro... does this sound like the right ball park figure?



Verstuurd vanaf mijn SM-G7102 met Tapatalk

[Clive S]

So... choosing between slaving 2 nema23 and a single nema34 ...
Or maybe a servo like Boyan said.

As far as i understand is a debated item...

Lookslike it comes down to speedy acceleration but complicated setup vs more robust setup but more expensive parts... and maybe inertial acceleration is slower due to heavier motor (bigger diameter) and 2 balscrews instead of 1 to drive..

Am i right about the above?

Verstuurd vanaf mijn SM-G7102 met TapatalkServos need a lot of setting up I believe and I don't think there are many on here with them.

The price is about the same for two motors or one as you only need one drive if using one motor.

Not sure what you mean. You will need two screws either way

[driftspin]

Servos need a lot of setting up I believe and I don't think there are many on here with them.

The price is about the same for two motors or one as you only need one drive if using one motor.

Not sure what you mean. You will need two screws either way
Clive.. i understood the single nema34 setup has a disadvantage ... it is bigger in diameter and has to run about 2.4 meter of belt between the screws.

The inertia of a nema 34 is higher than a slaved nema 23 setup, right?


A single long belt nema34 on 2 ballscrews vs a slaved dual nema23 software controlled solution.

Hmm sounds like i just made my choice [emoji1] [emoji1] [emoji1] aaaargghhh... no.

Is a stepper +/- 6nm nema34 still an option or is the twin ballscrew 70+ kg gantry asking for a servo solution.

BTW i dont want to start a discussion here on single 34 vs slaved 23.

Grtz Bert

Verstuurd vanaf mijn SM-G7102 met Tapatalk

[Clive S]

2x65v dc diy toroidal
1x24v dc din rail
1x5 v dc ?!? Samsung smartphone USB charger hahaha

Waterpump for spindle.
A 50v 0 50v AC toroidal when rectified will give you about 72Vdc if you are using AM822 drives that would be good.

A 24Vdc din rail PS something like this;- https://www.aliexpress.com/item/DR-30-24-Din-Rail-Switching-power-supply-30W-24VDC-1-5A-Output/32367113537.html?spm=2114.40010308.4.22.RcR90i

water pump https://www.aliexpress.com/item/Practical-Adjustable-Submersible-Micro-Brushless-Water-Pump-For-Aquarium-Fountain-Fish-Tank-Pump-DC-12V-3W/32632842083.html?spm=2114.01010208.3.161.uqHUUk&ws_ab_test=searchweb0_0,searchweb201602_2,searchwe b201603_1&btsid=1a27f1c9-6a16-4d0a-bd12-37daba8373ea

motion controller needs to be at the top of your list some have the bob built in.

Don't buy any powers supplies until you know what controller you are getting as some run on 5v others need 12v or 24 v.

In fact leave all the electronics until the frame is built.

[JAZZCNC]

With 70Kg + Gantry then I'd advise you go with servos if you want very best performance. Setup isn't so much an issue these days modern servo drives are much more forgiving. The Main downside is cost.?
With servos it's not just the Motor/Drive cost you have to factor in the that to get the very best out of them you'll need Motion controller that can provide the required Pulse frequency. Most of the cheaper USB Controllers like UCCNC etc cannot provide the required pulse rate to get the full speed and resolution out servos. Electronic gearing is required and this comes at the cost of resolution. However for most routers this is sacrifice that can be made without affecting machine performance. It just defeats the point of one main advantage of servos.

If your going with belts then 8Nm Stepper motor will be minimum and best run on Mains voltage drive to get best from it and give decent speed.
I've just built two Routers with 1250 length cut that runs Belt drive twin 2010 screws with 4.5Nm 34 motors running 72Vdc AM882 drives with Gantry aprox 45Kg giving 700mm width cut and they will happily run at 10,000 mm/min with acceleration 1200s/s.

Now with bigger 8nm motor you won't get those speeds with only 72Vdc but with Mains voltage drive then you will and probably more. Obviously the higher weight and inertia of the Gantry will play bigger part in lowering performance but still you'll get at least 8000mm/min with 10mm pitch screw which is more than enough for most routers at this level/size.

Like Clive mentions Don't cut corners on the Motion controller. Avoid USB it's too twitchy. Ethernet is by far more stable.
There are several suppliers of Ethernet Motion controllers depending on what control software your thinking to use. My prefered choice and only controller I'll use with Mach3 is from Cslabs. However the new Ethernet card and UCCNC Software from CncDrive are very serious challenge to Mach3 so worth close look. Not used this Combo yet so can't comment properly but I'm hearing good things from People who know there stuff and who's opinion I trust. I Will be testing shortly along with the Cheaper Stand alone controller.

Then you have the stand alone controllers like may have seen mentioned lately on Forum (but I don't advise you take Boyans lead and connect two drives to one Signal Output.) Not sure if the 4 axis version will allow Slaving but if your using belts this won't matter and could use 3 axis version.
Again Can't comment or give an informed opinion but I will shortly have one to torture so will post something at some point.

hope this helps.

[driftspin]

So i have been reading a lot in the meanwhile.

I found this calculator for critical speed.

https://www.zappautomation.co.uk/index.php/ecalculators.html#
And for comparison
http://www.thomsonbsa.com/pdf/bsa_engineering_catalog_section.pdf


With about 1500 mm between the bearings, (long axis),
And about 900 mm between the bearings, (short axis),

SFU 1610 @ +/- 1500 mm (+/- 10mm core dia* = +/- 500RPM max * 10mm pitch) = 5 mtr/min
SFU 2010 @ +/- 1500 mm (+/- 15mm core dia* = +/- 800RPM max * 10mm pitch) = 8 mtr/min

SFU 1610 @ +/- 900 mm (+/- 10mm core dia* = +/- 1400RPM * 10mm pitch) = 14 mtr/min



It looks like 2010 is a good investment for the Long axis

I made a redisign of the frame to trade off some stiffness and complexity of the gantry vs frame.

Ill post soon since steel is on the way :-)

[driftspin]

ok so does the ballnut on the gantry count as a bearing for length calculation?

Verstuurd vanaf mijn SM-A320FL met Tapatalk

[driftspin]

Epoxy 105 209 questions



What would be a sufficient epoxy thickness and width mounting surface for hiwin type 20mm rail?

My box section is 120x80mm so the flat surface width of the 80mm side
is about 60 mm wide... would 50mm wide and 5 mm thick be enough?

After welding the overall flatness of the frame will be in the +/- 1.5 mm range... so 3.5 to 6.5 mm epoxy thickness after optimum leveling by spirit level..

I am planning to do a single epoxy pouring of the x rail mounting surface and bed supports.. this will give me a good start for a level bed vs x rail

Total needed epoxy amount will be in the 3.0 dm3 range at 5 mm thickness

Verstuurd vanaf mijn SM-A320FL met Tapatalk

[Boyan Silyavski]

Epoxy 105 209 questions



What would be a sufficient epoxy thickness and width mounting surface for hiwin type 20mm rail?

My box section is 120x80mm so the flat surface width of the 80mm side
is about 60 mm wide... would 50mm wide and 5 mm thick be enough?

After welding the overall flatness of the frame will be in the +/- 1.5 mm range... so 3.5 to 6.5 mm epoxy thickness after optimum leveling by spirit level..

I am planning to do a single epoxy pouring of the x rail mounting surface and bed supports.. this will give me a good start for a level bed vs x rail

Total needed epoxy amount will be in the 3.0 dm3 range at 5 mm thickness

Verstuurd vanaf mijn SM-A320FL met Tapatalk


40mm width and 3mm thickness of the epoxy are the minimum for 20 size Hiwin rails. 5mm thickness as all around is ok. 6x6 or 8x8 steel bar welded to frame will provide nice strong cheap channel for pouring in the epoxy. If not a continuous weld, then prime all well to check for leaks. Then pour on second step.

[fifa]

Drive should be defined based of the performance.

Gantry mass play role, however if acceleration is 1200steps/second assuming 200 steps/rev this means 6 rev/s, acceleration of the gantry will be 60mm/s2, assuming gantry's mass is 100 kg - results 6N inertia force ... Correct designed drive can run additional load of human body (100kg) without problem - I am testing drive in this way.

For better understanding: side force of 6mm mill can not exceed 40N, typically it is bellow 25N.


For hobby machines main inertia forces (torques) are defined with spindle diameter. Relation between Spindle Inertia and its diameter is power of 4 (double diameter requires 16 time higher torque), same rule is applying for pulleys and gears. All others relations are linear.

Do not expect stepper motor to run more than 10m/min. It is hobby machine. 10m/min means 1000rpm (assuming 10mm pitch), which is ca 7kHz with half step mode. Not an issue for cheap drivers. For 2 motors you need ca 40kHz cheap Arduino type breakboards can do the job (96kHz).
However they are not good enough if controller need to do also micro-stepping.

Motor performance depends on motor inertia. Therefore motor rotor inertia shall be as big as possible (rotor accumulates an energy). With servo you can afford ratio up to 4 with stepper it shall be always less than one (to avoid risk of loosing steps)


Start designing from "top" , do not go too much in details.

2kW main spindle is quiet (no cooling fan), it is strong enough to continuously cut oak with 10 mm dia 8 mm deep speed aprox 4m/min.

I am assuming you have in mind cheap Chinese spindles. Front bearings are small (15mm dia), consequently bearings are loaded "eccentrically" - not all the balls take a load, some of them are unloaded. If ball is unloaded it start skitting (due to friction between ball and lubricant), consequently bearings worn out relatively quick.

Regarding noise: Whatever you will do, most likely workshop has a window, which will work like horn. The best way is to "absorb" as much noise as possible (for example carpets on the wall - but then walls becomes also dust collectors...). Machine housing (envelope) will do the job.


regards

[driftspin]

Drive should be defined based of the performance.

Gantry mass play role, however if acceleration is 1200steps/second assuming 200 steps/rev this means 6 rev/s, acceleration of the gantry will be 60mm/s2, assuming gantry's mass is 100 kg - results 6N inertia force ... Correct designed drive can run additional load of human body (100kg) without problem - I am testing drive in this way.

For better understanding: side force of 6mm mill can not exceed 40N, typically it is bellow 25N.


For hobby machines main inertia forces (torques) are defined with spindle diameter. Relation between Spindle Inertia and its diameter is power of 4 (double diameter requires 16 time higher torque), same rule is applying for pulleys and gears. All others relations are linear.

Do not expect stepper motor to run more than 10m/min. It is hobby machine. 10m/min means 1000rpm (assuming 10mm pitch), which is ca 7kHz with half step mode. Not an issue for cheap drivers. For 2 motors you need ca 40kHz cheap Arduino type breakboards can do the job (96kHz).
However they are not good enough if controller need to do also micro-stepping.

Motor performance depends on motor inertia. Therefore motor rotor inertia shall be as big as possible (rotor accumulates an energy). With servo you can afford ratio up to 4 with stepper it shall be always less than one (to avoid risk of loosing steps)


Start designing from "top" , do not go too much in details.

2kW main spindle is quiet (no cooling fan), it is strong enough to continuously cut oak with 10 mm dia 8 mm deep speed aprox 4m/min.

I am assuming you have in mind cheap Chinese spindles. Front bearings are small (15mm dia), consequently bearings are loaded "eccentrically" - not all the balls take a load, some of them are unloaded. If ball is unloaded it start skitting (due to friction between ball and lubricant), consequently bearings worn out relatively quick.

Regarding noise: Whatever you will do, most likely workshop has a window, which will work like horn. The best way is to "absorb" as much noise as possible (for example carpets on the wall - but then walls becomes also dust collectors...). Machine housing (envelope) will do the job.


regardsDear fifa,

I have kinda refined the ballscrew selection.

Next station would be steppers and so on.

looks like its gonna be steppers for economical reasons, or servos if i decide to go on a spending spree [emoji16]

I am eyeballing the uccnc controller.


I hope mechanical stuff is done by end of this year.

ESo electronics in 2018 !

Grtz. Bert.

Verstuurd vanaf mijn SM-A320FL met Tapatalk

[fifa]

No, every ball-nut allow
angular misalignment, and ballnut is not designed to support radial loads.

You shall not drive screw up to critical speed, 70% limit is recommended, this is due straightness tolerances (screw is not ideally straight)...

2010 robust, however, I would say for such type of machine too big. You will not be able to work with high speeds. Frame is too flexible.

I am assuming you are from Netherlands, threrefore German language shall not be an issue:

Check this site:


Search for Minerlaguss...

kr

[driftspin]

No, every ball-nut allow
angular misalignment, and ballnut is not designed to support radial loads.

You shall not drive screw up to critical speed, 70% limit is recommended, this is due straightness tolerances (screw is not ideally straight)...

2010 robust, however, I would say for such type of machine too big. You will not be able to work with high speeds. Frame is too flexible.

I am assuming you are from Netherlands, threrefore German language shall not be an issue:

Check this site:


Search for Minerlaguss...

krDear fifa,

Thanx for taking interest in my project.


"2010 robust, however, I would say for such type of machine too big. You will not be able to work with high speeds. Frame is too flexible."

I have actually selected the 20mm over 1610 ballscrew based on lenght and critical rpm vs speed, based on the zapp automation calculator.
And some other docs i found to verify outcome.

1610 screw looked a bit thin on a 1500mm + span.

Thanx for clearing things up on using the balnutt for a virtual span reduction.


Since i am unsure about max speeds i need... i designed towards >5m/min speeds. Did i mis something here?

Although i might only need that speed for rapids?

2010 will be harder to get moving though..
but is more rigid.
trade offs..

You might be right on maybe not being able to take advantage of 2010 over 1610 on my machine.
Do you feel it is over the top?

Will i lose much speed on small movements?

1610 screw looked a bit thin on a 1500mm + span

Or will i need more powerful motors: step up to servo's.

Now working towards diy 68v psu /80volts digital drivers using 4 x nema 23's 2:1 belt reduction 3.1 / 4nm... probably zapps. not final on this.

Or will better motors be ott compared to "flexible" frame design?
Is 120/80 x80x4mm box section considered flexible? 🤤

About my design.. there will be additional diagonal bracing towards the legs.

I am starting to think i need a 50ton battle tank chassis to get a " rigid enough " base. :-)



Please specify.
Don't be gentle.


The whole idea was to skip a typical mk1 and mk2 build and go straight for a mk3 quality and size machine.



Grtz Bert.








Verstuurd vanaf mijn SM-A320FL met Tapatalk

[fifa]

It is your design, your money.

5m/min with 10mm pitch is 500rpm.

It is better to choose 2020 screw...

I do not understand why you need "rapid" speed. It is hobby machine, and it is CNC. You do not need to be with machine to do the work...

It is relative very easy to "limit" the spindle "wobbling" - 2 nylon bushes do the job (bushes must be splitted axially and loaded with spring to omit radial clearance...

Check the motor torque curve - steppers are not good at high speeds, therefore belt pulley reduction does not work....

For whatever reason link do not works. I am suggesting to do the "bed" form cheap "cast" (epoxy concrete - cheaper than welded structure), gantry must be designed to be torsional stiff, however it is better that laterally it is flexible - to reduce loads caused by rail misalignment...

regards

martin

[driftspin]

It is your design, your money.

5m/min with 10mm pitch is 500rpm.

It is better to choose 2020 screw...

I do not understand why you need "rapid" speed. It is hobby machine, and it is CNC. You do not need to be with machine to do the work...

It is relative very easy to "limit" the spindle "wobbling" - 2 nylon bushes do the job (bushes must be splitted axially and loaded with spring to omit radial clearance...

Check the motor torque curve - steppers are not good at high speeds, therefore belt pulley reduction does not work....

For whatever reason link do not works. I am suggesting to do the "bed" form cheap "cast" (epoxy concrete - cheaper than welded structure), gantry must be designed to be torsional stiff, however it is better that laterally it is flexible - to reduce loads caused by rail misalignment...

regards

martin

Dear martin.

It is as always a trade of isnt it?

stepper : ballscrew pully's 1:2 results in 400 steps per 10 mm right .. so... resolution would be 10/400=0.025 mm per full step right...

Microstepping does not count for repeatability or precision right?

I did actually think 1:2 reduced 1610 and 2010 was generally accepted as the better solution here for alu and hard wood machines diy builds on this forum.
Or am i wrong here?

When one would use 20mm pitch ballscrews... what would you suggest the reduction should be like and keep some resolution?

If i would not need speed i better go for 1605? low rpm and low speed high resolution.


Using 2010 ballscrew on a 1500mm span looked like the best trade of to me. speed vs critical speed vs inertia price.

And yes... it is gonna be expensive [emoji52]

For controls it looks like uccnc 300eth has the affordable price quality i like... not decided on this though...

The most amount of money is not in C7 ballscrews anyway.. it is in linearguides and and bearings ... shipping and tax...


Do you feel i should make a radical move / design change?


Scrap the frame buy an used tank chassis? [emoji16][emoji16][emoji16]


And if yes... can you suggest what you would change at this point in the build.



I am open to any suggestions...
The build of the machine is part of the hobby right :-)



Grtz Bert.





Verstuurd vanaf mijn SM-A320FL met Tapatalk