Most of the flex comes from the attachment of the column to the base - in particular that big springy washer under the big nut, along with compression of the column. Put a solid support tube inside the column, around the big bolt, and toss the big washer, replacing it with a solid plate that contacts the column only at the corners. Epoxy-filling the column for the bottom few inches will also help some, and will hold the tube in place. Do that, and no external supports will be necessary, or even particularly helpful.Originally Posted by renatopub
Regards,
Ray L.
Yes, technically "backlash" refers to the amount of motion lost during screw rotation reversal due to clearance between a nut and its mating screw. The movement measured by my method will be, for the assemply in question, the actual reversing error. To record the actual backlash, the screw would have to be removed from the machine. Assuming the builder had used dual opposed angular conact bearings was a bold decision on my part
Joe
SPECS!!!!
I e-mailed linearmotionbearings2008 for the specification and to post the specs in his ebay store for these ballscrews, here is the responce:
"Dear Robert,
Thanks a lot!
The ballnut of 1605:
The ballscrew assembly (Ballnut and Spindle assembled) Precision rolled, class 7 accuracy; JIS (0.050/300mm) (0.002"/12")
Ballnut: Single ballnut with zero backlash (<0.015mm) (0.0006").
You can see the specs online in my store.
Best Regards to you and all your family!
Chai"
Robert
Is it really necessary to remove the screw? What I did was to measure the difference in displacement of the table when I come from CW screw rotation against CCW rotation (that showed to be 50um).
After that, I repeated the process of CW / CCW difference but what I measured was the position of the screw in respect to the bearing beads (that is, trust bearing backlash). For my configuration (trust bearing + self-aligning with preload) I could not measure anything, meaning it was smaller than 5um.
Now, if we can measure separately the table total backlash and the trust bearing backlash, would there still be any advantage in dismounting the screw?
Well, I measured two of the ball screw sets I bought and got around 50um (0.050mm) backlash consistently on both. Now, there are might be 3 explanations:
1- I am a very unlucky and got the only 2 bal screws that were out-of-spec
2- My measuring technique is wrong (it was posted in this thread and I would appreciate comments on that).
3- This claim of < 15um backlash in unrealistic.
Did anybody else tried to measure the backlash of these 1605 Chinese ball screws besides me and rwestbr???
The way you measured backlash is fine; however, i think you got the accumulation of variations because the way the dial was mounted.
If I remembered correctly, X2 has the Xaxis screw attached to the table and the nut bolted to the Ybase. So to measure Xaxis nut's backlash, the indicator should be mounted to measure axial variation of Ybase to the screw.
I would mount the dial to the table and rest the tip inline to the end of the screw, the measurement should be the thrush bearing backlash, and I think you said that you did it and saw result < 5um.
However, I would also mount the dial to the Ybase and also rest the tip inline to the end of the screw (Since Ybase should not move while measure X movement, I would just mount the dial to the mill base). This measurement should give a fairly accurate X's nut backlash.
Yaxis has the nut bolted to the Ybase and the screw is attached to the mill base, so the dial also has to be mounted accordingly for measurement.
ht
Last edited by htrantx; 05-29-2009 at 11:36 AM.
Guys:
very interesting thread.
Could anyone enlighten me on how to measure backslash?
thanks a LOT!!!
Backlash with Chinese linear motion BS
Hi folks
I have just purchased two ball screws from LinearMotionBearings in China via ebay. Thanks to previous contributors to this thread and other on these BS. I have measured the bacllash as follows.
1. Clamp the ball screw in a ER collet in the lathe head stock.
2. Pre load the projecting screw using the tail stock to remove all the movement due to bearings in the lathe.
3. Mount the ball nut which is mounted on the screw halfway between the tail stock and the collet holding the screw using a tool makers clamp held in the tool post.
4. Attach a sensitive DTI attached to the lathe bed using an adjustable magnetic stand to measure movement of the ball nut when pressure is applied.
Results are as follows:- 12 mm ball screw gives 0.006mm total movement excersion under what I judge to be 25Kg force. The pressure being applied to the ball nut firstly in the direction of the head stock and then in the direction of the tail stock to give the total excersion on ther DTI. (0.006 mm = 0.0002 inch for the non metric fraternity) This screw runs very easily and appears to have PTFE wipers and is ideal for the cross slide on my 7x12 lathe.
Using the same measuring method on the 16 mm ball screw I cannot generate enough push to detect any movement. I had to re adjust the tail stock locking so I guess the applied force was approaching 100 Kg with my feet agains the bench legs for additional purchase. This screw runs nicely but has more friction than the 12 mm screw. There is no sign of any increased resistance on reversing rotation which I think means the balls are OK in the tubes as I have not added any lubricant.
The support from China has been excellent in getting the ends machined to my requirements.
I will take up the tip of replacing the washer on the mast and adding a tube to stop the casting flexing. I can see movement between the mast and the base casting when applying a good hefty push to the top of the mast. I am using an electronic DTI from Amerdale in the UK which is very sensitive - checked the repeatability by lifting the punger and letting it return - gently to see if the reading shifts which it does not. Very good for a reasonably priced DTI - thanks Hugh for good value.
I need information on the backlash in toothed belt drives as I would like to use a rotating ball nut in the X axis of the lathe. Has any one any experience of the lost motion in such drives. Can two belts be used with wide pulleys with an adjuster on each belt. The idea being to take up the slack between the belt and the pulley tooth by pulling the belts in opposite directions i.e. separate tension adjusters either side of the belt path. I am thinking this will give a similar action to the split gears used in anti backlash servos where two gears mesh with a common pinion. One gear being fixed permanently to the drive shaft the other with limited freedome to rotate but loaded with springs parallel to the circumference. The action is that each tooth in mesh is loaded on both surfaces. This works well for light drives at modest speeds for precision servos. I am guessing that belt wear would be an issue at high speeds but the cross slide is not a high speed mechanism so this should not be too important. Any practical experience or comments or am I barking up the wrong tree????
Thanks to all the previous posters for encouraging me to take the plunge as I was going to stick with the old lead screws.
Kind regards - stay safe and have fun with CNC.
(Quote)
"Using the same measuring method on the 16 mm ball screw I cannot generate enough push to detect any movement. I had to re adjust the tail stock locking so I guess the applied force was approaching 100 Kg with my feet agains the bench legs for additional purchase. This screw runs nicely but has more friction than the 12 mm screw. There is no sign of any increased resistance on reversing rotation which I think means the balls are OK in the tubes as I have not added any lubricant."
These Ball Screws are 1st Rate and a real bargain!!! I am totally satisfied with them....
Thank You,
Robert
Chinese ball nuts & screw on Ebay
Hi
I am another happy customer from http://stores.ebay.com/linearmotionbearings ,Chai was great help.Air freight only took 6 days from paying to delivery.He charges $48 mtr (about $15/ft or $1.25/inch US) for the 2505 screw and $28 each standard nut.I ordered 2x 1350 and 1 x600 screw and 8 nuts for $630 Australian.The price from Australia for the same amount of Thompsom screws and nuts was around $1800.
Davo
Pat,
That sounds like a reasonable way to measure backlash.
Isolating the screw eliminates the effect of lost motion. It also eliminates any spurious measurements caused by racking.
It surprises me that so many people are concerned with backlash, which is relatively easy to fix or at least, compensate for. And .0002 backlash sounds great to me.
I also ordered some supported linear shafts and bearings from this seller. I asked if he could substitute 8 short blocks for the stated 4 long blocks, and he was happy to do it. At no additional cost.
Some time ago, I tried to buy some shaft support rails from another online merchant. I received a curt reply and a phone number when I emailed a request for info. Well, I called the phone number, and the conversation went something like this:
Me: Hello, I'd like to order 4 shaft support rails, the SR-16 's.
Phone Lady: Uhmmm. What company are you with?
Me: Just me. No company.
Phone Lady: ( sounding agitated ) I'll have to find the catalog numbers...
I can't find them right now. If you give me your phone #, I'll get back to you.
Me: That'll be fine. My number is 555-5555. Thanks.
Well, I waited for her call. It never came.
So, I placed my order with our friendly Chinese merchant. Not only did I get my supports, but I got the shafts, too. Drilled and tapped. And the bearings. For roughly what I would have paid for just the shaft supports. If the witch had taken my order.
Diplomacy is the art of saying "Nice doggie" until you can find a rock. - Will Rogers
I ordered their ball screws and I ordered them with 2 ball nuts each and end supports. I was pleased with the screws and nuts, the end supports required shimming to get rid of excessive backlash. with the bearing retainer tightened all the way down there was still about .010" movement in the bearings and the bearings had a additional .005 play I ended up putting a .010 shim between the bearings and that eliminated all the movement in the bearings and allows me to adjust the slop out of the bearings.
also with the pair of ball nut and some die springs i am able to adjust out any backlash in the nuts.
Seems like Chai is a great fellow.
So what is the deal with the preloaded nuts, I don't see him offering any.
I mean the auction says antibacklash nuts but I am not understanding this.
Is he stating that the single nut is zero backlash via over sized balls or what is meant by that?
I always though 2 nuts with preload between them was needed for removing excess backlash.
Any one know the cost for end machining?
I think if he can end machine and supply the support blocks for a decent cost, its far better then doing it your self A - save time B less chance of foul up!
THANKS!
i also reccomend that you make sure that the bearings are in the right orentation, the thrust surfaces (LARGER side of the race) should be facing the larger side of the second bearing. one of mine had both bearings in the same way and they dont last long with pressure on the wrong end of the bearing..
thanks for the reply diycnc.
Can you go into more depth on how you measured the backlash in the bearing supports?
So from what I gather the single ball nut is not zero backlash and a pair is needed.
Much like what I had planned with a floating center piece and bellville washers between the two.
Any pictures of your preloaded setup?
thanks
I used a dial indicator on the end of the shaft... the shaft moves in and out .015" on the block I have... when I took the block apart I noticed that the bearings were installed the both in the same direction, and basically shot from being loaded in the wrong direction, I also noticed that the total depth of the bearing bore plus the protrusion on the bearing clamp down plate was more than the width of both bearings, thus leaving the bearings floating in the holder by .010"
Basically if you can’t slide the thinnest feeler gauge you have between the retainer plate and the main body of the support you know the retainer plate isn’t clamping on the bearing.
Now what I did was remove both bearings and placed a .015 shim between both bearings also make sure that the bearings are installed in this fashion
// shim \\
\\ shim //
And now you can also adjust the preload on the cheapo china angular bearings.
I have got my end support to have zero backlash but am still experiencing .005 backlash in my table. I don’t understand where this is coming from as I have a double nut and the second nut is preloaded with 150lbs..
i have one nut mounted rigid then the other one is mounted with 6 die springs holding it down.
i will try to post pics tomorow but it may not help as i still have backlash..
Hey diycnc,
Thanks for the reply, Rather delightful!
Ok I understand how you are getting the axial backlash.
How do you calculate or how do you apply X lbs of f to the AC bearings?
If you get a chance to take pictures of the double nut and end support that would ROCK!
In your opinion is worth buying these China end supports?
It sounds like it might be better to buy my own brand name AC bearings and go from there?
OR but the end supports and just ask him to deduct the cost of the bearings, and Apply your own quality bearings?!
Thanks !!
I am cheap so I would buy them and make them work. Looking at Canadian vendor for end supports I see that they want over 170$ shipped per set of supports. I paid far less than that for the screw, 2 nuts, custom machining per my drawing, and both end supports. They even offered to through in couplers.
Also looking at some skf price lists for the ac bearings they are as much money as the Canadian vendors end supports..
Shim stock is cheap and I used my die grinder to grind the hole in the shim then cut them round with my tin snips then smoothed the OD with my industrial belt sander, then cleaned the burrs with my wire wheel...
I figured the preload of the springs using this chart
http://www.flat-stock.com/catalogs/SDS_DIE_SPG_RED.htm
My spring is 1" long and 1/2" dia part no 342-0804 I bought them for Fastenal. There are 6 per loaded ball nut and they are preloaded .15" which equates to about 25 lbs of pressure each x 6 = 150lbs
For the ac bearings I just used my dial indicator and slowly tightened the nut until there was less than .0005 play in the bearing. Basically I tightend the nut up then backed it off a bit. Too tight the bearing would be short lived.
Hi diycnc
I have just recieved three 5201 2RS (double row angular contact bearings). These are 12 mm bore x 30 (correction 32) mm OD x 15.9 wide full information can be found on the SKF site and they appear to be suitable for mounting C7 type ballscrews. As the information on bearings is not easy to find here is the link to the SKF page:-
http://www.skf.com/skf/productcatalo...odid=124003401
Clicking on the further information will get you the manufacturing tolerances for the normal internal clearances and all the data needed to check out the sums for load capability.
Since I also want maximum value I decided to take a punt on the 5201's which are readily available and probably differ in quality and bought the cheapest I could find! (A tad over £8 each in the UK.) I have just measured the radial play and find I have one that has 0.002 mm and the others 0.001 or less end float. I estimate I have applied a 5 to 10 Kg push / pull. The bearing was clamped to a face plate mounted on my lathe with the DTI also mounted on the face plate via a magnetic block to ensure there are no errors in the measurement. The outer race was held free of the face plate using a washer with a hole big enough so as not to restrict the inner race from moving.
Since mounting the bearings will further reduce the internal clearances I am well pleased as the installed end float should be less than the unmounted 0.002 mm total push to pull radial float I have just measured. My fall back strategy was to use ball thrust races with a single radial ball race - one thrust race with two deep groves and the other of the floating type to avoid and miss tramming effects of having two trust races less than perfectly mounted.
I now intend to use the 5201 bearings as I can easily meet the minimum load conditions. This is important to keep the balls firmly tracking in their groves. With a matched pair pre-loaded the pre-load minimum achieves the effect without the external load.
In another thread there is concern expressed about the use of double row AC bearings but the links point to a bearing size that does not match with the bearings I have nor with the SKF catalogue. I guess it is buyer beware.
Good luck and have fun
Pat
Last edited by wildwestpat; 07-27-2009 at 03:58 AM.
i have the -15 end support with 20mm screw the bearing is 15 x 32 x 9mm, 2 of them together would be 15 id 32 od 18mm wide
Posting Permissions
Reply with Quote
