PDA

View Full Version : Backlash in the gearhead of the Oriental Motor PK296A2A-SG7.2


Gerald D
Sun 13 July 2008, 04:19
Backlash in the gearhead of a new Oriental Motor PK296A2A-SG7.2 was measured today at 0.43 degrees. On a machine with 30 tooth DP20 pinions, that will be about 0.005" (For the metric people with 36 tooth module 1 pinions, that will be 0.13mm).

Note that the gearbox is SG7.2.......that is Standard Gearbox (with SH gears) 7.2 ratio. They do also have gearboxes with less backlash.

http://catalog.orientalmotor.com/attributeimage?&plpver=11&assetid=26062&attname=Gear%2f+Shaft+Type

Gerald D
Sun 13 July 2008, 10:41
To clarify; OM have gearboxes with a spring-loaded taper gear to reduce backlash. For standard production, these gearboxes are supplied with their Alpha-Step series motors, and not with the plain stepper motors in the PK series.

For those mailing me asking what to do about eliminating the backlash, the stock answer is belt-drives. However, for a big machine slicing up boards, .13mm [0.005"] is negligible.

J.R. Hatcher
Sun 13 July 2008, 21:10
:):):)

liaoh75
Thu 17 July 2008, 02:06
Man, I wished I had this information two weeks ago. What is the most powerful recommended direct drive motor from Oriental Motor that will work with the Gecko G203V drive? Is the concensus now that the very expensive SG7.2 geared drive motors are inferior to non-geared steppers with belt drives??? I think I'm opening a can of worms here but, J.R., would you be so kind as to do some backlash measurements for comparison?

J.R. Hatcher
Thu 17 July 2008, 04:06
David if my mechmate has any backlash at all it's too small to measure :). My main concern when I was considering belt drives was belt failure. It just looked like the weak link in the chain. Well I'm here to tell, those little belts are as tough as nails. I may have a belt failure today but it will be caused by something extremely stupid on my part. The reason I say extremely is because I've already done some of the stupid stuff, like running the X axis into the physical stops 3 times in less than 5 min @ 1375 ipm :o (what a sound that makes :eek::eek:). If you have the resources I suggest installing belt drives. My opinion may differ a year from now although I don't feel like it will.
When the motors with gear boxes start showing wear and need to be replaced I think it's well over $1000 for 4, compared to around $140 for 8 pulleys and 4 belts.
I'll participate in an experiment if several others will. We would select a small file decide the parameters (so everyone is cutting exactly the same) cut it and mail it to one person and let them compare all and report back and include pictures. This will help each person determine the best setup (7.2:1, 3.6:1, 3:1, 1:1 etc) for their needs, in other words I may only need 1:1 for what I cut. Does this sound like something y'all might like to try, or to quote my mom "Is this too much candy for a nickel"

Gerald D
Thu 17 July 2008, 04:17
Is the concensus now that the very expensive SG7.2 geared drive motors are inferior to non-geared steppers with belt drives???

Not at all.

No belt drive will give you 7.2 to 1 resolution. No belt belt drive will cost you under the maybe $100 premium you spend at OM to get a motor plus gearbox.

What the OM gearbox does is to give you a fine finish on your board edges - that is very important for big boards on a big machine. More important than the little bit of backlash....for the people cutting boards.

A cheap motor with a single belt drive will not produce as smooth a finish as the cheap 7.2 to 1 gearbox, because the ratio can't really get better than 4 to 1.

liaoh75
Thu 17 July 2008, 09:28
Thank you Gerald and J.R. for the response. My friend just got back from the U.S. with my PK296A2A-SG7.2 motors. and other goodies two days ago. I was so excited I drove to the airport to pick him up. We sat down to have dinner on the way home and I'm embarrassed to say that I was like a 2nd grade school boy on Christmas morning opening my presents. After spending all that money, I just didn't want to make any mistakes that could have been avoided. I now see where the benefits of the gear boxs are. However, J.R., I think your propsal if played out, would be very interesting to see the results.

Gerald, after reading this thread and my initial post, I did a little research and discovered that you can actually compensate for backlash through a Mach3 setting. Can anyone confim this?

Gerald D
Thu 17 July 2008, 09:58
Yes, Mach3 can compensate for backlash to some extent, but I don't know of anyone using this feature for router work.

Just realise that 0.13mm [.005"] is a tiny amount in the bigger scale of things. A salesman would say that is only +/- 0.065mm [+/- 0.0025"]. A slender router cutter bends by this amount.

Gerald D
Thu 17 July 2008, 10:04
I'll participate in an experiment if several others will. We would select a small file decide the parameters (so everyone is cutting exactly the same) cut it and mail it to one person and let them compare all and report back and include pictures. This will help each person determine the best setup (7.2:1, 3.6:1, 3:1, 1:1 etc) for their needs, in other words I may only need 1:1 for what I cut. Does this sound like something y'all might like to try, or to quote my mom "Is this too much candy for a nickel"


JR, your experiment idea definitely has merit. We need to define a standard test piece and supply a standard cut file (inch & millimeter) for folk simply to test their machines in general, not only for comparing belts/gears.

Richards
Thu 17 July 2008, 18:04
I have used 1:1 (no gear box), 3:1 (belt drive with original OM Alpha motors) and 7.2:1 (gearbox on replacement OM Alpha motors).

The 1:1 on my machine was not acceptable for the kinds of work that I do.

The 3:1 was a vast improvement, but some sanding was still required. Most of the "chatter" that was still visible after installing the belt-drive seemed to mirror the irregularities in the V-rails.

The 7.2:1 gave a very slight improvement over the 3:1 belt-drive, but only sometimes. That is probably due to the fact that the 3:1 belt-drive gave me about 0.001 inch resolution (1" gear diameter, 1000 steps per shaft revolution) and the 7.2:1 gave me about 0.0006 inch resolution (1.5" gear diameter, 1000 steps per shaft revolution).

With the same gear diameter, a Gecko driven motor would give 2X better resolution than I get with the OM Alpha steppers - but there are things that nullify the resolution.

As Gerald pointed out, some cutters flex much more than the basic resolution of the gear train. Minor imperfections in the V-rails will "bounce" the cutter. The grain in the wood will push or pull the cutter off its programmed path.

smreish
Thu 17 July 2008, 19:57
Much agree with Mike. I get far more cutter deflection than would ever be seen in the gearbox. Recent QA checks on my machine yielded the following: I cut 8 identical circles - 8" DIA, @ 190 IPM 3 passes, then the same again at 90ipm 2 passes. All 1/4" cutter. Material 3/4" birch 7 core ply.

The average diameter cut was 8.005. The mean low was 7.9975, mean high was 8.015. Best resolution @ speed and passes was 8.0015, in 1/2" acrylic, 3 passes, 60ipm. Sometimes slow and steady wins the race when you need perfection. ALL deviations were noticeable in the cutter - not the machine. Just my quick observations on trying to get some really precise CNC parts without having to take goods to my local HAAS machine shop.

Gerald D
Thu 17 July 2008, 20:53
You can be sure that actual machine flex also had something to do with the variability in the results. All machines do flex - that is just the nature of metal, or any material for that matter. For the MM design I just got to a point where I said "this is stiff enough, folk will blame their cutters and not the gantry for flex"! :)

chopper
Tue 28 April 2009, 20:50
This post copied from elsewhere:

I am currently running the VEXTA PK296A2A-SG7.2
I was having an issue with a small amount of chatter in my cuts
so I switched from running the 20 tooth gears to a 30 tooth gear
thinking the bigger dia. may smooth out the cut with more gear contacting the rack but what happened was I ended up with more backlash, I wasn't surprised I thought that this might happen, but I didn't think it would be this bad, and I was wondering if anyone else has had the same issue??
the backlash on the Y axis is about 13 thousands of an inch
and on the X it is 7 thousands on one side and 20 thousands on the other
is this a normal amount of backlash?
I was wondering if I am expecting to much from the machine?
thanks
//chopper

Gerald D
Tue 28 April 2009, 23:46
Chopper, I only once measured a new motor and got 5 thousands. Your motors are obviously no longer new, and your results are useful for the rest of us. Thanks.

Belt drives certainly have their place as the economical non-backlash option.

J.R. Hatcher
Wed 29 April 2009, 01:59
Believe it or not ............... I can move my X or Y axis toward the negative .001 then toward the positive .001 and see the laser move back and forth, in other words 0 backlash ............ believe it or not. No brag just fact. Belt drives rock :p.

Gerald D
Wed 29 April 2009, 02:21
Not even bragging about your eyesight? :rolleyes::D

J.R. Hatcher
Wed 29 April 2009, 11:10
I was thinking (wishing) I might need to explain it, but only if someone asked. Thanks Gerald.
I put something like a 2" thick board on my table. Bring the laser to the edge so it projects all the way down the side, then if it moves even .001" you can easily see it move. I think it's because the laser light is round and so the width of the light shinning down the edge of the board changes. Trust me, my eyesight is not that good these days. I'm not nearsighted or farsighted ... just about blind, I've even been thinking about what design I might want on my cane :rolleyes:.
Anyway give it a try and you see what I'm talking about.

Gerald D
Wed 29 April 2009, 11:24
Okay, I can see how you see that.

Just for fun, put your laser back in the same place and then push/pull the router collet with 2 fingers without any motors running . . . . . . how much difference do you see then?

orotemo
Tue 19 January 2010, 12:09
Copied from another thread:

. . . . . . .

I have backlash. no lost steps, always manage to get back to the exact same locations, only that there is a 'play'. I grab the gantry in my right hand (the strong one:)), and I am able to get a 1mm movement when moving by hand back and forth. this is quite a problem. if I always cut the same direction, I'm able to get by with perfect cuts. if I change direction on the same cut in a different cut level, the backlash shows, you see a stair on the material. I also cut decorative panels (when will sales pick up...:mad:), and then it becomes a major problem.

I want to cancel the backlash, and from threads around here I learned that the timing belt and pulley is a solution, and a great one.
I also have now SG7.2 geared oriental motors, want to change to keling 906oz-in with 3.6 ratio belt (20-72 teeth).

. . . . . . . . .

bradm
Tue 19 January 2010, 14:09
Orotemo, if, in fact you have a sloppy set of gearboxes with backlash, then the belt drives should help you to eliminate the backlash.

Will eliminating that backlash solve your cutting issue? That is harder to be sure of. There are two things that I would check before spending a lot of money. Both of them should be checked during your test of pulling on the gantry.

The first is to make sure that your pinions are fully engaging the rack, and stay fully engaged. The second is that the pinion gear isn't sliding on the motor shaft.

Based on your description above of hearing the gears slam around, it does sound quite possible that your gearboxes are bad. There seem to be a lot of people for whom the OM gearboxes aren't a problem, so it might be worth checking with OM; maybe they'll replace them. Have you always had the problem, or did it develop over time?

MattyZee
Tue 19 January 2010, 16:21
I had a similar issue early on with my belt drives, I tried to get away with small grub screws and the pinions came loose and i could move the gantry a millimeter or two. After changing to bigger grubs screw i haven't had the issue again. So i agree with the others, a belt drive won't solve your backlash issue if your motors are in good order.

. . . . . . . .

Gerald D
Tue 19 January 2010, 20:19
We have found that the most common source of "backlash" seen in the x and y direction is caused by a loose z-slide. The z-slide roller eccentrics should be adjusted up to hold a pre-load on the z-slide. The z-slide should not wobble side to side.

lumberjack_jeff
Tue 19 January 2010, 21:20
With the steppers energized, I can get .4mm of movement (entirely gearbox movement - I can feel the pinion rotate) in my y axis. For reasons of which I'm not entirely certain, I have much less in my X axis.

my z-slide is quite rigid.

When doing things like cutting signs, this hasn't proven to be a problem. It only becomes a problem when I try to do inlay-type work.

I'm able to work around this with backlash compensation, but my next machine will have timing belts.

(Jeff has Fulling stepper motors & gearboxes)

orotemo
Tue 19 January 2010, 22:37
lumberjack jeff - I didn't input any backlash into the controller - I just make sure most of my cuts are done same direction, that way the backlash behaves the same for every cut, so say I'd like to cut a rectangle out, I do it with two-three step downs, and I get a cut with zero marks by doing all the cuts in the same direction. my 2 cents.

one of my x axis motors shows no backlash, the other - a millimeter(0.04''), and the Y is around .4mm. I'm sure most people are happy with the OM SG gears, but I'm afraid the company that sold it to me clearly stated that this gear is not for our purposes and won't replace it.
Maybe the gear was damaged while I worked with Mach using exact stop G61 - because I couldn't get Mach to the tolerance I needed with G64. Mach actually didn't gracefully accel and deccel, it stopped with a bang - and maybe that what damaged my gears.
anyhow - thanks a lot. I will now build my gears with more confident.

sailfl
Wed 20 January 2010, 03:59
Jeff,

I don't understand. If you can feel the pinion gears moving, have you checked that the set screws are tight?

Orotemo, have you checked to see that your pinion set screws are tight on all your motors.

One of the problems I had when I first started was that my set screws were not holding. I had to file a flat spot on the motor shaft to help the set screws do their job.

I would check the set screw to insure they are tight. It is easy to blame the motors.

It doesn't matter what the company that sold you the motors says, it is what the manufacturer states. If the motors are bad but I don't believe they are, they have to replace.

Check the set screws.

orotemo
Wed 20 January 2010, 06:31
sailfl, The motors are awesome, its' the gear, and for all I know, the manufacturer is fine with 1 degree of freedom on this gear.

The set screws are also good, tight in place.

thanks for the tip though.

Gerald D
Wed 20 January 2010, 08:46
The couple of reports trickling in where the Oriental Motor PK296A2A-SG7.2 gearboxes have developed excessive backlash is a bit worrying. Seems like they wear, or bed down, when used under heavy duty. (Nobody has broken a tooth yet.)

Under these circumstances I think we need to re-visit the choice of gearbox, and discuss these factors;

- Is 7.2:1 the best ratio, or would 3.6:1 be better?

- A smaller pinion (driving to rack) would reduce the loads on the gearbox, and reduce the effect of backlash.

(Please note, this is not a discussion of belt drives)

sailfl
Wed 20 January 2010, 08:49
Gerald,

If this is the case then because we are using or recommending so many drives, do we want to have a discussion with Oriental to also get their input. Or are also considering a different manufacturer?

Gerald D
Wed 20 January 2010, 09:09
ShopBot uses thousands of geared Oriental Motor drives and have enough buying power to have a "custom" gearbox on their motors. They get the TH (taper hobbed) gearboxes which are not available to us small buyers buying 4 at a time. I don't think that the TH gearboxes are any stronger, or less prone to wear, it is just that they have internal spring-loading which takes up some of the wear. I also think that a MM, being heavier and stiffer than a SB, puts more load on the gearbox.

I am not at all optimistic that Oriental Motor would offer an alternative gearbox in the same price category as now - their official answer is going to be that we are using the wrong gearbox. (They have other gearboxes in their range much more suited to CNC, at a much higher price). Other manufacturers would also shake their heads at using a spur gear type gearbox for CNC.

PEU
Wed 20 January 2010, 14:08
What about a Kelling or similar motor and a planetary gearhead? prices for a nema34 10:1 reducer are around $250 and backslash is around 10 arc min (0.1deg) the only obvious drawback I see is that the motors become longer and that may be a problem on the long rails because the operator can bump into them.

Links:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=120514167413
http://www.homeshopcnc.com/gearheads.html

There is a lot of expensive units $500+ here is one example
http://www.electromate.com/products/series.php?&series_id=100530

lunaj76
Wed 20 January 2010, 19:44
Consider the gear boxes only have one flat spot not two.

lumberjack_jeff
Wed 20 January 2010, 20:07
Jeff,

I don't understand. If you can feel the pinion gears moving, have you checked that the set screws are tight?

I wasn't clear. While holding the pinion gear, I can rock the y-carriage back and forth exactly .015" - no more. The pinion gear rotates, but the motor shaft (my motors have dual shafts) does not. Further, my machine zeroes repeatably.

My steppers are not the OM steppers they are the less expensive fulling units with straight cut gears.

I said in an earlier thread the the fulling steppers have comparable backlash to the OM ones, but I no longer think that is the case.

sailfl
Thu 21 January 2010, 03:09
If the motor shaft does not rotate and the pinion gear does then your set screws are not tight. The motor shaft should turn when the pinion gear turns. Maybe I am not understanding correctly.

Can you list your motors. I have not seen motors with two shafts but I am new to motors.

Thanks

bradm
Thu 21 January 2010, 07:12
Nils, many stepper motors are available with a option to have the shaft extend out of both ends of the motor - this is commonly used to add a sensor to provide feedback on the actual shaft motion / position and detect lost steps (*)

I think that Jeff is observing that at one end of his gearbox and motor combo, he can see the .015" motion occurring at the pinion connected to the gearbox, while at the other end, he's seeing no motion of the motor shaft. That is a good description of backlash in the gearbox.

(*) Whenever this comes up, there's a temptation to think "Cool, I should have that to detect missed steps." However, the reality is, if your machine is missing steps, the workpiece is already ruined. There's nothing useful to do with this information, other than rework your machine / software so that it never misses steps. So it's an expensive boondoggle, really, in our application.

lumberjack_jeff
Thu 21 January 2010, 08:17
Yes Brad, exactly.

The motors were an inexpensive eBay find, I wouldn't have ordinarily purchased dual shaft motors.

I hadn't really intended to put encoders on those shafts, but it is my understanding that Mach and EMC can use that feedback to stop the machine immediately when it detects mismatch before it becomes severe.

pblackburn
Fri 30 November 2012, 18:39
I know this is an older thread but I to have seen the backlash in the PK296A2A-SG7.2. All of mine have between 9.5 and 10.3 thousandths. They probably have less than 90 hours runtime on them with no crashes. I have not noticed any problems with circles or carvings but on complex carvings with 3d models inset in a dish you can tell there is backlash. I measured it, and used Mach's backlash comp to eliminate it. I think the PK296A2A-SG7.2 are a great motor for the x and y but not for the Z. Just my 2 cents.

litemover
Sun 22 September 2013, 00:23
What were the numbers you entered int he Mach3 backlash settings Pete?

pblackburn
Sun 22 September 2013, 07:58
That is best done with a dial indicator and use the jog with step only to find your true backlash. Mine was 0.009" (0.2286mm). I will say it does work but it is not the best solution to the problem. Remember all gearboxes have some backlash, they cannot be absolute. They are not a wave generator. So 9 thousands was my setting but I then converted to belt drive to truly minimize it.