Ready Steady Eddy

Jonathan said he won't use BK12 bearings and I think he's half right.
The housings seem to be okay provided they are dimensioned correctly for the bearings used, but it's the supplied bearings themselves that are extremely suspect.
I just stripped a couple out, one had Japan bearings the other China bearings, one had the bearings facing opposite directions, one had both bearings facing the same direction (keep in mind these are angular contact bearings), the China bearings had no grease whatsover in them.
So it's a good job I stripped them down but now they have to be sorted out.
First I had to understand about AC bearings so found this;

Bearings for universal matching
Internal clearance and preload
(click on the 'table' links near the bottom)

Now I understand I need to reassemble according to those diagrams but can't decide if Fig2 or Fig3 arrangement (see first link above) is best, opinions please.

Edit: After reading this
http://www.mycncuk.com/forums/linear...her-apart.html I'm thinking Fig2.

and this
http://www.mycncuk.com/forums/linear...end-float.html

Back to Back Eddy.

.Me

That's Fig2 then, thanks.
I also decide to get some new bearings for the blocks because the backlash on the China bearings was serious. I fitted a shim between the outer races but with the locknut fitted there was still slight play, the shim was the thickness of a Guinness can which in this application is thick. I also took a skim off the block to make sure the outer races were being clamped.

Also I'm sure I have a bent ball screw on the Y axis, it's not something I've done myself, on the last 1/3rd of it as you rotate it by hand it goes, slack,tight,slack,tight,etc. and using inside calipers between the screw and the gantry the gap goes big,small,big,small,etc. I'm not 100% yet but it might be the turned down section on the end is off centre.

I may have to get into this

http://www.roton.com/screw_shaft.aspx

As requested layout info for panel, please refer to schematics posted previously.

Starting top left.

Panel isolator, AM882's, ABB Inverter
PSU for stepper drivers, FL2, FL1, CB2, 5V PSU, 24V PSU
CB1, K1, K2, K5, Breakout Board with spindle board and K6, Smooth Stepper (mounted on panel side (aargh I hate that))
Terminal rail, K3, K4, PILZ Safety Relay

Attachment 11816 Attachment 11817

On the close up photo you can see 3 wires (77,75,24N) going to a piece of vero board located on a stand-up on the bob, that's K6 the transistor relay for flashing the lamp (see schematic)
It got a bit crowded and I've not got the fans in yet ! the grey wire hanging down is just the spindle I have lashed in temporarily.

Added a limit override switch (it's a spring return type so you can't accidentally leave it on), still need to label up everything on front.

Attachment 11818

Quote:

---

I also decide to get some new bearings for the blocks because the backlash on the China bearings was serious. I fitted a shim between the outer races but with the locknut fitted there was still slight play, the shim was the thickness of a Guinness can which in this application is thick. I also took a skim off the block to make sure the outer races were being clamped.

---

I think I remember Dean saying some time ago that the tubular spacers on the BK bearings are not always long enough and that he had made fresh ones. It is there to enable the lock nut to push the inner races together.
I might be talking crap though :very_drunk: ..Clive

Clive, you're not talking crap but in this case the spacers are okay, it's the bearings themselves, the inner and outer races are dimensioned wrong so there is no gap for preload. I've been reading that the bearings should be obtained as a pair.

After reading this I've just been looking closer at one of my fixed end bearing blocks and it has play in the bearings. The floating end has a NSK Japan bearing in, not sure about the fixed as I've not taken it apart yet but I can bet my bottom dollar that they'll be cheap Chinese bearings...

Yep....chinese bearings (ungreased) and face to face so the inner races were not being tensioned...need a shim and some grease and see how they fare...

Attachment 11820

Looks like Z axis ball screw has issues too. Noticed while driving it up and down, with the front plate off, the ball nut and bracket were fine at the bottom but the more they reached the top the more they oscillated side to side. I took it all to bits and found two things, the turned down section at the driving end was slightly off centre and the the whole turned down section was bend to one side. With the ball screw resting in two V blocks it was rotated and a DTI measured a deviation of 0.8 mm at the end where the pulley sits, this reduced to 0.35 where the BK12 bearing sits. So I did what any self respecting person would do, put it in the vice and knocked seven bells of shite out of it ! to say I was upset would be an understatement. Actually I put it in my mini lathe chuck with a DTL fitted and 'coaxed' it back straight. It's a lot better but not perfect due to the off centre turned down section. Next I'll have to see what can be done with the Y ball screw. Who's idea was it to get stuff from China ?

Who did you buy from ?

Quote:

---

Originally Posted by Lee Roberts

Who did you buy from ?

---

I bought them from Chai

Whilst we're close to the topic...for a while now I've been dismantling new ballnuts I've obtained from Chai, as it's not uncommon to find debris in them - be that just general dirt or bits of metal.

On the machines I've made to get zero backlash (lathe, mill & router), I preloaded the angular contact bearings using disc springs and shims between the bearings. That way there's less risk of damaging a bearing by over-tightening the nut as you know how much force you're applying from the spring deflection and it avoids buying matched bearing pairs. I also use bearings with a bigger outer diameter than the ones in the BK blocks - i.e. use 7201. They're less delicate...

If you want an easy option for a standard CNC router, then I don't see much wrong with just using double row angular contact bearings (52xx series), as although they have a small amount of axial play, the amount is much smaller than the single ballnut backlash so overall the effect isn't very significant. If you start pre-loading the nuts, then naturally the bearings need pre-loading too.

Also, be very careful with opening out the bearing blocks to better fit the bearings, as going too far may introduce backlash (depending on quite a few factors).

Yes I think you are correct about the double row bearings, I was looking at those yesterday. Getting the nut set right on the BK type is a bit of a fiddle to say the least.

You make me laugh you lot.!! . . . . Your quite happy to pay peanuts for the parts but then whinge and whine when they aren't top quality.!
WTF do you expect when You couldn't even buy one decent Angular contact bearing in this country for the price payed for both Blocks with bearings.

I strip, grease and shim every BK/BF bearing. It Takes minutes to do and other than that they are mostly fine considering the price payed.

Yes I could make my own better but unless chasing true Zero Backlash like Jonathan mentions then I'd just be wasting time and money because the BK/BF blocks when fettled work perfectly fine for any normal router machine.

Quote:

---

Originally Posted by JAZZCNC

You make me laugh you lot.!! . . . . Your quite happy to pay peanuts for the parts but then whinge and whine when they aren't top quality.!

---

Totaly wrong, I don't expect top quality but I do expect 'fit for purpose'

Quote:

---

Originally Posted by EddyCurrent

Totaly wrong, I don't expect top quality but I do expect 'fit for purpose'

---

I know what you mean but come on it's 10min job at tops and it's not difficult or hard. There has to be some affect of cheap prices and If folks have the abilty to build the machine in first place then in the grand scheme it's a drop in the ocean to fettle these bearings.

Getting all anal about bearing's and whether one came from China and other Japan is mostly pointless unless your chasing very tight tolorences because your own workman ship will introduce a lot more error than a couple of cheap or mismatched bearings ever will.!! . . . . . End of the day it's a router for cutting Wood and provided the bearings are greased and shimmed to remove lateral movement then they last for years.! . . . I've just after 5yrs changed my bearings, yes I've adjusted and greased them couple of times in that period but that's just normal service like any machine needs.

Well, not even that, I just want to buy them and fit them knowing they will work as expected, is that asking too much ? I don't think so.

Anyway I stripped the Y axis ball screw out and laid it on a flat surface, sure enough it was bent about half way. Using a feeler gauge I slid it along underneath to find the bend position, then I put it in a vice with wooden jaws, put a pipe over the other bit and 'leaned' against it until it was straight. Now it's about right however the ball nut is a bit crunchy, I pumped some grease in and the metal particles that came out was alarming, it's still a bit rough and sticks now again so I see no alternative than to strip it out and clean, that's assuming it will go back together.

Quote:

---

Originally Posted by EddyCurrent

Well, not even that, I just want to buy them and fit them knowing they will work as expected, is that asking too much ? I don't think so.

---

Well you need to dig deep then and pay the price.! . . . Can't expect to pay £15-20 for something that cost's £80-100 over here and not have something to complain about.!! . . . BUT I suspect we'll have to agree to disagree. :stupid:

Been away for a few days, got back and now the PC graphics card has gone ! more expense.

Took the ball nut off like I mentioned earlier, emptied the contents into an old but clean washing up bowl I use for occasions like this.
The ball nut has 3 ball circuits and each ball measured about 3.15mm in diameter, maybe they are 1/8" inch ? anyway the number of balls that came out was 53 so right away you know there is something amiss. Before removing the ball nut the problem was that it felt like it was dragging and looking at grease patterns on the ball screw it looked like the centre ball circuit was stuck because the grease was being totally removed from the ball screw in those threads.

I watched this video on how to repack the ball nut and it was no problem putting it back together.
Linearmotionbearings C7 Ballnut Flipping and Repacking - YouTube

He explains in the video that 16 balls should be in each circuit which leaves a gap, I found that 17 balls left a similar gap so that's what I put back in and it's now nice and smooth running.

Edit: 10/12/2014 It was later confirmed that the correct number of balls for a 1605 or 1610 ball screw is 17 balls per circuit 51 total. (so that's good news for me !)

So the problem was too many balls in the ball nut so that they were jamming instead of rotating.

No one wants jammy balls!

Interesting video, thanks for the link.

Eddy - I'm looking to build a steel-framed machine and yours looks like a great candidate for stealing the design! I am looking for a slightly larger cutting area although I don't think enough to make any significant difference. Before I start, do you have thoughts on what you might do differently if you did it again?

The only thing I would do different would be to make sure the ground was level that the welding was going to take place and I would not weld the 10mm plates to the bottom of the top beams because it warped them like a banana, I would bolt them to the beams instead with a thicker plate inside to beef up the 3mm wall thickness of the box section.

I had to strip it down a bit today because there was still a BK and FK bearing to sort out, both had bearings in facing the same way and no grease, made shims, greased, reassembled.
More progress, ready to order some energy chain and start wiring and I'm very pleased the panel is already built.

Attachment 11883

Limit and home switch for Z axis, there is one switch and two strikers. For the home switches I mounted a piece of 6mm thick mild steel to aluminium angle to act as the target.

Attachment 11884

Motor mount for Y axis

Attachment 11885

limit switch and adjustable striker for Y axis, there's another striker at the other end of the gantry, so one switch with two strikers.

Attachment 11886

Eddy I would move that bottom switch as the swarf and resin builds up very quick around the bottom of Z axis and you'll be cleaning it constantly.

Quote:

---

Originally Posted by JAZZCNC

Eddy I would move that bottom switch as the swarf and resin builds up very quick around the bottom of Z axis and you'll be cleaning it constantly.

---

Noted, however I did have a struggle to find a better position for the switch so that it could be operated at both ends of the Z travel. If it becomes a problem in use I'll move it and remember I'm cutting wood.

Quote:

---

Originally Posted by EddyCurrent

Noted, however I did have a struggle to find a better position for the switch so that it could be operated at both ends of the Z travel. If it becomes a problem in use I'll move it and remember I'm cutting wood.

---

Wood is even worse than Aluminium for sap and resin etc, honest you'll be surprised just how much it collects at the bottom of Z axis.
Also with Z axis why would you need a Limit on the down because often the bed is the limit and it just stops. Most Z axis only have limit switch at the top so put it at top.!

Quote:

---

Originally Posted by JAZZCNC

Wood is even worse than Aluminium for sap and resin etc, honest you'll be surprised just how much it collects at the bottom of Z axis.
Also with Z axis why would you need a Limit on the down because often the bed is the limit and it just stops. Most Z axis only have limit switch at the top so put it at top.!

---

There's no sap or resin in any wood I cut but I'll see how it goes once cutting starts.
If there was no down limit the front plate would 1/2 leave one of the hiwin blocks before the ball nut hit the end, also I like the idea of a limit at each end.

Thought I'd have a play with other stuff while waiting for the energy chain to arrive.
As I'm a hoarder I've had an old TV wall bracket in the attic for years, it can be extended out quite a long way and swivels side to side. As it is, the keyboard can be accessed easily from each side of the machine and from the front.
The wireless MPG has magnets built into the back allowing it to 'stick' to the panel.

Attachment 11904 Attachment 11905 Attachment 11906

Very nice set up Eddy. I assume the keyboard and monitor swing around to the side so you don't have to lean over the bed to get at them?

Looks like you are into woodwork as well, if I'm not mistaken that's a Jet 16" drum sander in the background. I'm planning on making a drum sander once the CNC (and workshop) is built. It'll probably turn out to be cheaper to buy one but where's the fun in that.

Looking very smart Eddy !

More playing about today, lashed some wires in for the two X axis stepper motors and got it moving, A is a slave off X
Settings as follows;

AM882 - microstep 8 giving 1600 steps/rev
Mach3 - steps per = 160 (using 10mm pitch ball screws), Velocity 10000 mm/min, Acceleration 3000 mm/sec.

Tried velocity up to 15000 but it stalled quite often, so the question is, how do these numbers sound ?

It's not clear how to set things like 'Kernel Speed' when using the Smooth Stepper, I set it to 35000 Hz anyway.

Quote:

---

Originally Posted by EddyCurrent

It's not clear how to set things like 'Kernel Speed' when using the Smooth Stepper, I set it to 35000 Hz anyway.

---

Doesn't use Kernal Speed has that's for Parallel port but it's recommended to set it to default 25000Hz. What is the Freq in the SS plug-in.? This is the equivalent of the Kernal speed but you can go much higher. That said don't go too high and you won't need to go much higher than 128Khz.

Your accel is too high really at 3000 try 1500-2000, having Accel high will make for jurky movements and potential stalls or lost steps. You'll probably find you can go higher on Velocity without changing any thing other than lowering Accel.

To be honest thou 10mtr/min is fine for rapid speeds, 10mtr/min is 1000Rpm on 10mm pitch so going higher risks chance of missed steps from lower torque as motors reach the corner speed and gantry weight and inertia push the motors.!

In practice you'll rarely cut at those speeds so I wouldn't go chasing any more speed really.!! . . . .Your after smooth stable running.

Smooth Stepper is set to 1MHz, the range of settings is 32kHz to 4MHz so I set it to upper mid range.
Acceleration is high only because I read somewhere to keep it in line with Velocity, I did have it at 1500 and noticed no difference so I've just set it back to that now.

Quote:

---

Originally Posted by EddyCurrent

Smooth Stepper is set to 1MHz, the range of settings is 32kHz to 4MHz so I set it to upper mid range.
Acceleration is high only because I read somewhere to keep it in line with Velocity, I did have it at 1500 and noticed no difference so I've just set it back to that now.

---

No point having SS set that high because both the drives and the BOB can't work at that pulse rate so going higher than 256Khz is a waste and can potentialy cause problems or make system unstable. Personaly I would drop it back to 128Khz has you'll never need higher than that anyway.

Quote:

---

Originally Posted by Wobblycogs

Very nice set up Eddy. I assume the keyboard and monitor swing around to the side so you don't have to lean over the bed to get at them?

---

Yes it does, it can swing out to either side

Quote:

---

Originally Posted by Wobblycogs

I'm planning on making a drum sander once the CNC (and workshop) is built. It'll probably turn out to be cheaper to buy one but where's the fun in that.

---

I keep an eye out here for customer returns Engineering Chuck Centre Chisel, Woodwork Woodturning Cutting Craft items in Axminster Tool Centre Devon Tools Power Provider store on eBay!

Yes, I've been known to have a look in that store from time to time. Got a rather nice P/T from it a few years back. I little bit of cosmetic damage saved me several hundred.

I can't believe this Igus direct ordering, I order some energy chain (less than 3 metres) on Thursday 20th March and the estimate for delivery is Monday 31st March, so how does this video make sense ? I must have ordered part number 80,001

http://www.youtube.com/user/igusUK

It happens.
They can't keep every possible item in stock, so sometimes you have to wait for it to be made and/or shipped across from Germany. If you're desperate for something, it's better to give them a call, as they can tell you exactly what they have in stock, and can suggest alternatives.

I know what the problem is.?. . . . . .The stripes on there ties go in wrong direction. . Lol

On 21st I ordered some parts from China and the tracking number has them in Uk now and should be with me Thursday/fri. This order includes Energy chain Which IMO is every bit as good as Igus stuff but only costs me £18.50 for 2mtr.!!

I'd not by items from Igus just because of those ties! I can't say fashion is high up on my list of "things wot I know something about" but even I could do better than that.

I completely agree though Jazz, as long as you're prepared to do your homework the stuff coming out of China is just as good as the locally sourced stuff at a fraction of the price and often with similar delivery times. I'm sure Igus are making plenty of money at the moment but I can't help feeling they (and a lot of European companies) are leaving the door open for far eastern companies to come over here and eat their lunch.