you'd want at least 20bits but they got 23bits so that's even better, price difference between 17 and 23 is negligible and those ballscrews look really good by the way. would go well with 23bits servos, their pitch is 10 so with 23bits it would be a very well balanced combo of speed and precision. Just wake sure the ballscrews are long enough for what you want to do.
Hi,
No I did not check, I mean how would you do so? I don't have the measuring equipment to come anywhere close to measuring this stuff accurately enough to be considered a check.He said it difficult to give such a guarantee. (understandable). Did you check the accuracy of the parts you ordered from him?
The 32mm THK BNFN C5 screws that I use for my current machine had the original hand signed test certs from THK when they were manufactured.
The bottom line here is that I trust dyGlobal to represent their sale items correctly and the manufacturer of those items to be honest about their products. With all the major and leading players in
the market like THK, NSK, Bosch Rexroth, Koruda etc I'm happy that is the case. I have made over a dozen purchases from dyGlobal and all have been exactly as advertised.
Naturally new old stock items are the best. There is no wear and the prices are commonly half or less than direct from the manufacturer. I would say they are ideal, both quality and price. When I bought my screws IWhat are your thoughts on the price? I can use this same one for all three axis.
found three identical units, and so I bought them. That is why my machine (current machine, not my first mini-mill) has 350mm x 350mm x 350mm travels, there are three identical ballscrews.
I see the screw length is 333mm and the nut length is 47mm. That will mean that the absolute max travel of those screws is 286mm, and probably 270mm is a more practical maximum. Is that enough
for the sort of machine you want?. Lets assume that you are happy with 270mm, then the rails will need to be 270mm plus the distance (between centres) of the cars, say 150mm and the length of the car,
46mm for a 15mm car if I recall correctly, for a total of 466mm, and you should add another 20mm or so for the sake of practicality so minimum 486mm.
My first mini-mill had 410mm long 15mm rails, and had 150mm between cars (centres) and 46mm cars for an absolute max travel of 214mm. In the event I settled for 180mm x 180mm x 180mm travels
as practical.The 15mm wide rails were adequate, the limitation on rigidity was the C-frame, despite being all cast iron and steel. I can assure you that building a frame rigid enough to match even 15mm
rails is a challenge. Of course 20mm rails would be better and 25mm better again....but where do you stop? Especially given that its likely the lack of rigidity of the frame that is the limiting factor....not
the rails at all.
One, I doubt you'll ever wear out preloaded screws and two re-balling screws is a major. Don't bother, buy light, med or heavy preloaded screws at your wish and they'll last under hobby conditions forever.Does this mean I can still use the Z0, let the balls wear out and then fill with new larger balls for medium pre-load? I could not see and instructions on how to do that in the manual however.
If you are cutting steel then yes, you will require slow speed but high torque. That is why its important to decide early if cutting steel is what you want your machine to do.......it makes a HUGEI guess I really do not need to go that slow.
difference in the rigidity required of the machine, and the spindle seem suddenly to have cost four times the amount! If a machine has rigidity R to do wood, plastics etc, then it'll need 5R
to do aluminum, and 20R or more to do steel. Cutting steel separates the men from the boys BIGTIME.
The surface speed for carbide tools in steel ranges from about 50m/min for stainless up to about 150m/min for coated carbide in mild steel.
A 12mm tool:
S=50 m/min RPM= 1326
S=100 m/min RPM=2653
S=150 m/min RPM=3979
High speed (12000 to 24000) asynchronous spindles and VFDs used by hobbyists are NOT well suited to low speed operation. For that purpose a servo driven spindle is probably a better bet.
This guy is not the only one to report that Lichuan work well, but they have Chinglish manuals and no set-up software. I repeat; I do not have any particular complaint about ANY of the cheap Chinese made servos in termsCheck lichuan servomotors, there's a guy on mycncuk that goes by the name jazzcnc, he makes a living building routers using lichuan servos and steppers and says he never had any issues with them.
of quality or performance, but they lack good documentation and set up software. My Delta servo have over 500 hundred parameters, programming the drive by pushing buttons is a joke, and a bad joke at that.
If you have had no experience with setting up servos then I'd recommend that you get servos that DO have set up software. There are plenty of posts on this forum alone about people whom have struggled to set up
the cheap servos. It seems they get there in the end but hugely frustrated by the process. It should not be like that.
Craig
Don't know about manuals or software which is why it's best to ask him directly since he was able to build a business around them, I doubt he'd do that if the setup was a nightmare, I also know he's a helpful fellow that goes way beyond normal courtesy to answer peoples questions over at mycncuk.
Hi,
for many years an encoder with 10,000 counts per rev was considered state of the art. For even more years before that, before encoders when servos were equipped with
resolvers with a typical accuracy of 10 arc min and yet CNC machines of those days still produced fantastic work.
In the current day 10,000 count encoders are common in cheap Chinese servos, but 16 bit, DMM for example, 17 bit, Delta B2 for example, 20 bit Delta A2 for example and 23/24 absolute and 24
bit multiturn, Delta B3 for example, are now common and cheap.
I use Delta B2's which have an encoder of 160,000 count per rev.......and I can't use more than a small fraction of that resolution. I support the notion that smoother movement can be had with high
resolution encoders but for us hobbyists even 10,000 count per rev is still world class. My 1.8kW Allen Bradley servo, circa early 2000's has an 8000 count per rev encoder and is still mega accurate.
I think the argument between 20 bit and 23 bit is pointless, either is truly superb and world class....we'll never come close to taxing that!
Craig
Hi Sus - Yes use the Z0 until worn. Then decide on new balls or entire new cars. Changing balls is not in the design manual but you may experience this in the build. Those balls can get out and they do give you a merry chase around the workshop table/floor/kitchen. So when handling have them over a tray or have the slides in place... Peter
Hi,
do yourself a favour and 'NEVER UNWIND THE NUT OFF THE SCREW'. The balls are slightly oversize and very hard to put back in. Many of the double nut types the balls are not oversize
but one nut is wound tight against the other for preload, and they are bloody hard to put back together also. The bottom line is NEVER REMOVE THE NUT. Some of the two circuit types
have two ball circuits one with a 4.999mm pitch and the other circuit with a 5.001mm pitch....and guess what they are bloody hard to put back together, the bottom line is NEVER REMOVE THE NUT.
You did not pay big dollars for the best of screws only to try to wreck them by disassembling them. If that means you have to devise a special means of fitting them so you don't have to remove the
nut then do it.
The good news is that in all my designs there is only ever one loose nut....and that's me....it's never, absolutely never, one of my precious ballnuts.
Craig
Fully agree on that. I did this for a Chinese ball screw mistakenly and was only able to put back part of the balls back - But it still works but surely not accurate at all. Good enough for my existing 6040 router though.
However what I meant was to re-ball the HIWIN linear rails as my current ones are Z0 - which means they are lightly pre-loaded. Even HIWIN recommends mid or high pre-loaded ones for CNC mills. My question was whether it is feasible to just use the Z0 ones until they become too loose to be usable and then maybe re-ball them manually with slightly larger balls to get higher preload?
I only realized this now. This may be ok for Y, not for X and Z. I asked the seller if he has longer ones.
The maximum I could do would be the size of Stef's machine. Do you think that is enough for a 550 X 350 X 350 machine?
I thought with high speed milling, we are cutting small amounts at a time, but faster achieving similar or better material removal rates. For example the machines by DATRON High-Speed CNC Machines - DATRON Dynamics. Is it not the case?
Certainly, Will verify that whatever the servos I buy have the software at-least and possible a good enough manual.
Hi Sus - Use the Z0's anywhere but not the Z axis. The z axis is the highest loaded area and closest to the impact loads so they wear first. Any bearing outlet will have suitable balls. Getting the right size will need a part number or size from your car maker. HSM techniques are moving the spindle speed upward, the issue is steel cutting makes the tool load greater than the cheaper spindle makers expect. Gpenny have models for metal work but its description is engraving and general work. Some texts say milling.
Water Cooled Spindle Motor Vfd | Metal Work Spindle Motor | Metal Working Spindle - Machine Tool Spindle - Aliexpress
Datron rarely show steel being milled. They have copious info on aluminum, they have 60,000rpm spindles for that. Many people are using low cost spindles to cut steel - light and fast... but do your research based on the things you want to make.... gpenny also have a BT30 spindle.... Peter
Re: HSM and high speed spindles. In theory if you know the correct chip thickness you can scale up the feed rate and rpm linearly to whatever your machine can do. However Power=force x velocity so if you double your feed then you need twice the power to achieve this. Peter
addition- here's a snapshot using wswizard for HSM mild steel. generally your up against the recommended surface speed of the tool so need to use smaller tool to keep it ion the range. But then you can feed faster to make up for it. In the attached the 6mm tool has a bugger MRR then the 10mm tool. I'd have to check if 452m/min is OK for a coated tool surface speed for instance...
this is how datron does with steel
https://www.youtube.com/watch?v=BV6KJJ-cjo0
this is how cutting steel might work for you, dental mills do it like this even with titanium.
https://www.youtube.com/watch?v=7zYLuMbCEA8
I'm aiming to do stainless at 100mm/s with a DOC of 1mm for my machine.
https://www.youtube.com/watch?v=gBfkNZZSqDE
Hi,
Yes, that works, but a machine capable of such high speeds and particularly high acceleration is very VERY VERY hard to achieve on a hobby budget.I thought with high speed milling, we are cutting small amounts at a time, but faster achieving similar or better material removal rates
I, like you thought that if I had a high speed spindle and took small enough cuts it would work......FAIL. There is a critical surface speed at which a tool will cut properly.
It depends on both the material to be cut and the the tool. For uncoated carbide cutting mild steel the sweet spot is 100m/min. If you try to use a significantly faster speed,
say with a high speed spindle you'll overheat the tool within seconds. Once carbide turns red hot its stuffed!
You might get away with surface speeds of 200m/min or even more with CBN tools....but can you afford CBN? I can't.
Craig
Hi - With some coated tools 300m/min is recommended. I've been trying to find a chart that just handles coatings, tools and surface speed with no luck. The info is buried with other data. It will be a juggle. HSM uses smaller tools for the lower surface speed and faster feeds. And then there's lubrication. Flood or mist? Datron uses pure ethanol as it evaporates completely. I have used metho which has 5% water in it so that gets left behind. No big deal. Trying to organise a mister system at the moment... Then if you use a high speed spindle and ask the vendor is it suitable for milling steel they will say NO. So pick your poison. Peter
Hi,
I've never had any success with carbide, coated or uncoated beyond about 150m/min, the tool turns cherry red and that's the end of it. I run my four flute 1/8th inch Raptor Destiny Tools endmills under coolant
at 15000rpm in mild steel, and that equates to a surface speed of 148m/min. Even then I'm pushing the envelope and they don't last as long as I would like...and they're not bloody cheap!!! Uncoated tools
like 1/8 th Kyocera Tycom tools fry up after a few minutes, coolant or not, no trouble in aluminum or brass but fry-up in mild steel.
In 4140 I slow down to 80-90m/min with coated tools, and with 4340 I slow down to 75m/min even with coated tools.
You might try running them faster....but I can't afford to fry them up all the time.
300m/min is in CBN territory.
The bottom line is that OP needs to be aware that high speed spindles are not recommended for ferrous metals. No matter how many feeds and speeds data sheets I look at they all say
100-300 ft/min (32-100m/min) for uncoated carbide in mild steel. Free cutting steel, and you can up that by 50%, for tougher alloy steels you need to reduce by as much as 50%.
Craig
Hi All - In my coatings vs speed search I came across this chart from Suttons. Its the friction co-efficient of the coating against steel. I'm surprised they are so high. That explains the heat generation issue. Peter
A quick question:
For a machine that has approximately 550mm X 300mm X and Y travel, what is the recommended Z travel? Would 300mm be enough?
If this is the case I can buy two the ball screws from e-bay seller and find something else for X.
Hi,
my machine has travels of 350mm x 350mm x 350mm and Its adequate. I have fitted a fourth axis, and in a month or twos time a fifth axis, I'll unbolt the Z axis and shift it up 200mm or so
to recoup the Z axis travel lost. Ardenum does not like C frame mills, seems to think they lack imagination, but what they do allow is to shift the Z axis upwards at a later date.
In the attached pic you can see that the frame extends some 300mm above the Z axis bed, thus I can remove the Z axis and shift it upwards. This was always my intention and part of my design
from the start.
Its always nice to have more travel....but.....and its a big but......the cost goes up and up and up. Not only does trying to find good ballscrews and linear rails get harder, they NEED to be bigger
and more rigid and thus so must the frame be bigger and more rigid. So the price is exponential. Double the size of the machine but the cost goes up ten-fold.
To be honest I think those ballscrews you linked to look good, they are new, from a reputable manufacturer, and you can get a complete set for a small fraction of the cost of buying new.
You might consider the travels somewhat limited....but its what you can get and can afford, and that often trumps 'what you'd like'.
I've had this battle/decision to make over and over and over again......I can often not afford what I want, so I have to find and make do with what I can get and afford.....OR.....go without.
Craig
That's right, I think they've overstayed their welcome by a decade or more. If you offset the saddle onto a wider column you get a much stiffer machine, you can load heavier blocks onto the table, your frame is bulkier, your spindle can move faster than a saddle+table and your z travel is not compromised. All in all better stiffness better dynamic movement and the design doesn't change that much from a c frame. C frame is great for belt driven spindles I give it that.
For these types of spindles a lightweight fixed double column with a moving table would probably be best.
Hi,
BS.All in all better stiffness better dynamic movement and the design doesn't change that much from a c frame.
Craig