# 4th Axis Backlash

#### BigDon

##### DESi Customs
Silver Member
So backlash on a 4th axis is measured in arc seconds and minutes. I have seen them advertised from a couple seconds to 15 plus minutes. I know anything in the seconds would work just fine but I am not looking to spend thousands on a indexer. So if my math is right something with 8 arc minutes works out to be .00145 at the butt end of a cue? Does this sound right? Anyone have any experience with acceptable backlash on an indexer?
Thanks for any help!

#### RBC

##### Deceased
Well Don, let's figure it out.

If the butt of your cue is 1.250", then the circumference is 3.927" (Diam X Pi)

3.927" divided by 360° = .0109"

Divide that by 60 for 1 arc minute and you have .00018"

Multiply that by 8 and you have .00145" at the surface.

I came up with the same number, but wanted to show the math.

Obviously, the diameter can make a little difference too. Smaller diameter, less distance at the surface.

Royce

#### BigDon

##### DESi Customs
Silver Member
Well Don, let's figure it out.

If the butt of your cue is 1.250", then the circumference is 3.927" (Diam X Pi)

3.927" divided by 360° = .0109"

Divide that by 60 for 1 arc minute and you have .00018"

Multiply that by 8 and you have .00145" at the surface.

I came up with the same number, but wanted to show the math.

Obviously, the diameter can make a little difference too. Smaller diameter, less distance at the surface.

Royce

HaHa, Thanks Royce. Well I passed the math portion of the test but still having a hard time processing it. Would that .00145 be spread out over the 360 degrees or is it saying that if I drilled a hole and spun the cue one full time that I could be off .00145 from lining back up with that hole again?

I am looking for an indexer that I can cut pockets, inlay and then come back around and recut and inlay again. Leaving a maybe 20 thou border that looks the same all the way around. So what backlash would I be looking to achieve?

#### harris cue comp

##### Banned
indexer

if you have backlash in you indexer, you should change your setup.

#### Thomas Wayne

##### AzB Silver Member
Silver Member
if you have backlash in you indexer, you should change your setup.

Even very, VERY expensive indexers have backlash to some degree - it's the nature of the beast. The real question is how much backlash is acceptable.

On a machinist forum I read that a member had tested a relatively inexpensive Enco worm-drive vertical rotary table to have an accuracy of 30 arc-seconds. That's 16 times better resolution than the 8 arc-minutes referenced above, and would result in a potential discrepancy of less than one thousandth on the above described buttsleeve.

But the real secret is not so much in achieving "zero" backlash in your indexer, as in knowing how to deal with whatever backlash you do have. One way is to never have a negative value for your rotary axis in your G-code. In other word, always rotate to home ("zero") in the same direction you did your repeat indexing. You can easily set this up in your Mach3 configuration.

OR you can have a method for establishing your rotary home position accurately before each operation with a witness mark on a large dial attached to your indexer - especially if you're using stepper motors.

OR you can punch a locater hole in your cue where it will not matter later on and re-zero your rotary axis to that hole before each operation.

All three of the above solutions are being used every day in many cue shops around the country with good success. Pick one and perfect it, or come up with another method for achieving the same results.

TW

#### harris cue comp

##### Banned
backlash

.........................................................

Last edited:

#### conetip

##### AzB Silver Member
Silver Member
When I served my apprenticeship, they had a very precise rotary table that had the least amount of backlash I have ever seen. On a 250mm radius was less than 0.01mm . It got sold as it had not been used for a few years. Now it resides in a metrology lab.
Thomas is right, since most setups do have backlash of varing amounts. So it pays to work in such a way that it does not become the problem. Backlash on a rotary table is really like backlash on a lathes cross slide, yet millions of lathes are producing good parts with the backlash.
One way is to apply some drag on the system that is more than the cutting forces. So the cutting will not pull the work into the cutter. You can also figure out the amount that your system has with a particular amount of drag.
Neil

#### Thomas Wayne

##### AzB Silver Member
Silver Member
Thomas: while we all know you are the all seeing eye. If you have a fourth indexer that is a direct drive and rotates in a positive direction on a continuous basis I don't believe you can have backlash. You may have rotational inconsistency due to your setup. By having a consistent positive rotation you are not allowing the fourth axis rotation to have backlash as it is moving only in one direction.
Robert Harris

No need to get snotty, Robert. I provided actual practical methods for overcoming indexer backlash, while you merely decreed that such backlash indicates an incorrect setup. So if anyone was posing like they are the "all seeing eye" it was YOU. Aspiring cuemakers want solutions, not simply confirmation they're doing it "wrong".

As for driving your indexer in a positive direction only, isn't that exactly what I said when I wrote:

[...] One way is to never have a negative value for your rotary axis in your G-code. In other words always rotate to home ("zero") in the same direction you did your repeat indexing. You can easily set this up in your Mach3 configuration.

[...]
?

TW

#### Scratchy

##### AzB Silver Member
Silver Member
What about harmonic drives used for the 4th axis? I have read that their design is zero backlash. Is that so, or not so? Have not measured one, so I can't speak from any experience, only what I've read.

Several options on Fleabay offer the harmonic drive option.

#### RBC

##### Deceased
I think you've got all the correct answers here.

Thomas and Neil have it right. Move in only one direction and keep the resistance or drag high and you won't have any problems.

As for Harmonic drives, I'm sure they do have some amount of backlash, but I'd expect it to be very very small. I think one would make a great rotary axis. I have one on my desk with plans to build an indexer from it, one of these days! lol. If you are looking to get one, just make sure it's not a planetary mechanism. The companies that make them offer both and the planetary systems will have a fair amount of backlash. And, by the way, a new Harmonic Drive can cost into the thousands, but you should be able to get one suitable for around \$1000 new. Of course, the way to go is Ebay and buy a used one.

Royce

#### BigDon

##### DESi Customs
Silver Member
Thanks for the help guys! I was already thinking of rotating in one direction and making a "center" hole to check my position when I spin back around when it allows. I just wanted to make sure I didn't already get myself started with a "sloppy" indexer and have to worry with it all the time.

#### Scratchy

##### AzB Silver Member
Silver Member
Don, I measured the backlash on my Chinese harmonic 4th axis tonight. I bought it a year ago or so for \$420 off Ebay including shipping. It is a 50:1 unit with the 80mm chuck.

With a bar chucked 'sideways' to act like a big indicator hand, at a radius of 6" out from center I measured .001" backlash error. Translated to a nominal .5" radius, such as working on a butt inlay, that would represent an error of only .00008". If I figured it correctly, is an error of around 34 arc seconds.

Hope the info helps in your decision.

mac

#### RBC

##### Deceased
Don, I measured the backlash on my Chinese harmonic 4th axis tonight. I bought it a year ago or so for \$420 off Ebay including shipping. It is a 50:1 unit with the 80mm chuck.

With a bar chucked 'sideways' to act like a big indicator hand, at a radius of 6" out from center I measured .001" backlash error. Translated to a nominal .5" radius, such as working on a butt inlay, that would represent an error of only .00008". If I figured it correctly, is an error of around 34 arc seconds.

Hope the info helps in your decision.

mac

Scratchy

Do you have any info on this unit?

I wasn't aware that the harmonic mechanism was being made in China. I was under the impression that it was pretty well protected with patents.

Is it a true "strain wave" type of mechanism?

Here's a link to those who aren't familiar with these.

Royce

#### Scratchy

##### AzB Silver Member
Silver Member
Hi, Royce -

Here is a link to a virtually identical unit currently on ebay (I have NO financial interest!):
http://www.ebay.com/itm/CNC-4th-Axi...t=LH_DefaultDomain_0&var=&hash=item3aa802e71e

There are several sizes advertised with the harmonic configuration. The seller I bought mine from does not seem to offer any at this time. I don't know anything about IP protection, perhaps they have a workaround?

I disassembled my unit when I received it, and it for sure does have the flex ring configuration of the strain wave concept. (AND, it still worked after I reassembled it - odd, but true...)

I had not taken the time to actually measure the backlash previously, since a quick run early on seemed like it was close enough for my work, but I thought folks on the list would find it interesting so I checked it last night.

best regards,

mac

#### LGSM3

##### Jake<built cues for fun
Silver Member
Interesting, I've seen those on eBay 100 times and just assumed they were direct drive

Last edited:

#### Scratchy

##### AzB Silver Member
Silver Member
Re: ratios - I've seen 50:1 and 100:1, don't know if there are other options offered by any of the sellers. Note carefully, not all of the 4th axis setups are harmonic drives. I don't believe any of the belt-driven units would be.

50 turns of the input shaft to 1 turn of the output shaft, i.e. with a 10 microstep Gecko drive, 2000 steps per revolution on the motor x 50 = 100,000 steps per revolution on the A axis. Works out to 277.778 steps per degree as my setting in Mach 3 .

HIH,

mac

#### BigDon

##### DESi Customs
Silver Member
Hey Mac,
Thanks for checking that. I know TW shared a picture one time of the inlay cnc he built and it had a Chinese 4th on it. He said he had to rebuild it some if I remember correctly. I still have not come to a conclusion yet but this does help a lot. I just remembered the other day that Deepgroove1 had an indexer for sale. It is a direct belt drive 5C set-up with a 6-1. I sent him a message about the numbers on that one but have not heard back yet. Still not sure what route I am going to go yet.

#### conetip

##### AzB Silver Member
Silver Member
Hi, Royce -

Here is a link to a virtually identical unit currently on ebay (I have NO financial interest!):
http://www.ebay.com/itm/CNC-4th-Axi...t=LH_DefaultDomain_0&var=&hash=item3aa802e71e

There are several sizes advertised with the harmonic configuration. The seller I bought mine from does not seem to offer any at this time. I don't know anything about IP protection, perhaps they have a workaround?

I disassembled my unit when I received it, and it for sure does have the flex ring configuration of the strain wave concept. (AND, it still worked after I reassembled it - odd, but true...)

I had not taken the time to actually measure the backlash previously, since a quick run early on seemed like it was close enough for my work, but I thought folks on the list would find it interesting so I checked it last night.

best regards,

mac

Can your drive be used as continuous spinning unit ? If so, what sort of rpm can it run at ?
Neil

#### Scratchy

##### AzB Silver Member
Silver Member
Neil - yes, the unit can run continuously in either direction. I just maxed out the Axis setting and M3 reads as "5400 inches per min".

I get around 15 RPM counting revs watching my second hand.

15 RPM x 360*/Rev = 5400*/min, so I think things jive.

I have a 48V supply and run Gecko 540.

Hope it helps.

mac

#### harris cue comp

##### Banned
4 th axis

if you run a direct drive in mach 3.you can run the 4th axis at max of 168000 and 50000accelerations. this will allow you to do engraving in record time as the a drive moves like the other 3 axis.
any a driver that has a ratio seriously slows down your indexing and your production time.
there have been some good advice here in this thread but a direct drive indexer is by far the best way to go.
thanks
Robert Harris