View Full Version : A fourth axis attachment concept

Thu 01 October 2009, 15:45
Greetings to all,

This is an experimental design I am working on for a 4th axis using the available materials in the shop. It will be attached under the table beams and could be used both as vertical and horizontal rotating axis positions just by 90 degrees indexing the main assembly on the table beams.
This 4th axis on my machine is designed for approx. 1000 mm length x 500 mm diameter in horizontal axis position and 250 mm height x 500 mm diameter in vertical axis position.
It should fit on MechMate with only minor changes made to the last two of the table beams, and to the control panel of course.

Some specs:
The motor is PK296A2A-SG36.

A simple L297 MOSFET driven home made unipolar driver board will be used. Although it is only a half step driver, it proved reliable in several applications up to 40V and 4A supply. Theoretically these two combined should yield a resolution of 1.5 minutes disregarding backlash and everything.
This driver has no idle current reduction which due to its reduction of the stepper motor torque and eventual motor slipping is not a useful feature in the 4th axis anyway.
The downside of this driver is that it rotates the motor maximum at around 250 rpm, which when calculating the 1:36 reducer, will be about 7 rpm max. Horribly low I suppose.

A 4” chuck will be used which along with its inverted gripping jaws will serve many applications and seems adequate. The chuck’s neck bearing is a simple UCF205 which has enough lateral load tolerence.

A simple design for a tailstock is also improvised which consists of a M16 threaded rod, fitted in two 6001 bearings and a simple livecenter. The total displacement is 140 mm which I am not sure if it wobbles or not when fully extended.
The gripping blades on the livecenter are spring steel from a 3 mm scrap leaf spring. Quite hard and robust material it is. The piercing end is ground manually on a wheel grinder.

Attached are the concept and the pictures of the progress will follow in a short while.


Sat 03 October 2009, 17:19
Hello all,

This is the progress made in the last two days.
I wonder how to attach pictures in between the lines of the post. Therefore I had to split it into several posts. The (picture) refers to the logical order of the attached files.

These are the main purchased components:

For ease of attaching/detaching the 4th axis assembly it is better for the first two rails to the front of the machine, i.e. farthest from home, to be arranged face to face, therefore they will be like left and right hands and care should be taken in locating the holes:

12mm diameter holes to be drilled along them:

Two pair of holes to accommodate the motor assembly in its horizontal axis position, 60mm apart:

A set of holes to accommodate the motor assembly in its vertical axis position and also the live center assembly for different lengths of workpieces. This distance and the number of holes are up to your choice. Mine were 140mm apart:


Sat 03 October 2009, 17:25
A piece of 120x12mm angle iron, 150mm length is squared on a sanding disk to an acceptable condition.

This part will accommodate the bearing and the motor housing. One wing cut to 90mm to sit on the base support which is a 90x50x4mm square tube. Since this angle iron is thick enough, it will be drilled and tapped so that no nuts will be used. The spaces are rather tight:

Next part is the motor housing. It is made from a piece of 100x50x3mm square tube. One face will carry the motor and drilled and bored accordingly, the other face for the UCF205 bearing housing to be mount. Please note that the motor face is configured to accept Oriental Motor gear motors, in my case PK296A2A-sg36. Cut 45 degrees on ends to provide access for the bolts and screws:

The assembled parts:

Please don’t mind the shorter or longer than necessary screws and not using washers. This is only a temporary assembly.


Sat 03 October 2009, 17:33
Parts for the motor assembly support:

Motor assembly support welded:

Parts for live center base:

Live center base welded and drilled.


Sat 03 October 2009, 17:40
Parts for the live center. The square tube is 25x25x2.5mm and 150mm length. Nuts and screw rod are M16:

This part was very annoying to make and I am not satisfied with it yet.
The first nut was welded with care and the rod screw was quite parallel to the length of the tube. The problem arises when I weld the second nut. The tack weld tend to shrink and off center the second nut so that the rod screw will get jammed and won’t move a bit. Maybe somebody could help.

The live center casing:

Preliminary setups:

Horizontal position of the 4th axis and the live center:

Vertical position of the 4th axis:

An approximation of height relative to the table board surface. The spindle fitted with a tool bit will be clear off the chuck and therefore not running into it:

Works to be done:
Flanged shaft from motor to the chuck.
Live center inself and assembly into the bearings and the M16 rod screw.
Aligning and welding the fox shaped live assembly.
Mounting the locking screw and the end handwheel.
Mounting the motor.
Checking plays and centering procedures.


Sat 03 October 2009, 19:16

That's a VERY interesting indexer.:) Thanks for sharing the design and progress pictures.

Now, we are anxiously awaiting completion so we can see it in action.

Sun 04 October 2009, 00:08
Did you have the threaded rod through both nuts when you welded the second nut?

You might want to grind it off & try that. Or drill out the second nut so it is just supporting the threaded rod, not engaging it. I don't think you really need those threads.

I see in your 3d model that you are planning a second "lock" nut between the support and the hand-wheel. I think it would work better in front of the first nut, between the support and the live center.

Everything looks good, and you have an amazingly clean shop!

Gerald D
Sun 04 October 2009, 00:27
To "insert" the pictures in the right place in the text, see this post (http://www.mechmate.com/forums/showthread.php?p=8923&postcount=20)and the few just before it.

Sun 04 October 2009, 16:34
Thanks for the help Leko, that's the way it should be. My mistake was that I was actually making a non-backlash screw, backed with a locking nut.

Thanks John. I am redesigning the live center part. I noticed a slight buckling in the 16mm rod screw and am pretty sure it would wobble while at work. maybe I should choose 20 or 25 mm screw to add more rigidity to it. I will reconfigure and remake this part and will post the progress.

Thanks Gerald for all your efforts and care.


Sun 04 October 2009, 18:09

I was wondering about a larger ( 20 or 25 mm ) screw.

Alternatively, you could make the live center support mechanism movable, so you would not have to extend the screw very far from the support...kind of like in a lathe. I realise that this would mean developing some sort of slide for the support mechanism. Or you could have multiple mounting holes. Then you could select a set of mounting holes based on the length of the workpiece being setup.

We anxiously await your progress...good luck.

Tue 06 October 2009, 14:41

I worked on your idea for multiple holes and also a sliding mechanism. It is achievable but I have to broaden the base to more than 170mm. I encountered two problems which we could work it out.
1- the maximum length of the workpiece will be reduced; please bear in mind that for the longest workpiece and with the tailstock fully retracted there is the danger of the X Echain path being cut through.
2- the assembly will be heavier and harder to attach/detach, say when you want to use the assembly in vertical axis position.

There is also this possibility to use the lower wings of the U80 channel as a bed for the slide. I am working on the design and will post it soon.

For the time being I have changed the design so that now:
It is M24 rod screw. The M24 nuts will house well in a 40x40 mm square tube.
The locking nut put at the livecenter side as Leko suggested.
The end nut will be drilled 24 so that it acts only as a sleeve, again as Leko suggested.
The assembly will be bolted with flanged screw on a rolled steel strip which is welded on the base and will be tapped. This arrangement provides some lateral movement as well as the possibility to shim the assembly in order to get minor alignment possibility with the chuck.

Tue 06 October 2009, 19:22

Looks like you have the problem well in hand. I hope it works for you. Please continue to keep us informed about how your testing goes.

On another note, I and some others would like to invite you to show us your MechMate. You've given us just a glimpse, so maybe you would consider showing the whole machine. I also realize that you are reluctant to do this since you've had to make some changes to the design and are concerned that it might detract from the Forum to show us your "hybrid" design. From the little I've seen, I think your MachMate could qualify for a Serial Number. Please give this some thought. I hope you decide to show your MechMate. I further hope you decide to apply the MechMate Logos and seek a Serial Number.

Just my opinion. Whether you decide to show your MechMate or not, please continue to contribute to the Forum. From what I've seen, your contributions are well thought-out and articulate. Keep it up.:)

Wed 07 October 2009, 10:53
Asalamalikum Nader - I would really back John on you showing your machine - having seen a glimpse means that I can say that is is really close to the spirit of the Mechmate!

and Kudos to the 4th axis design

I wonder how and where you would place a proximity switch on the 4th axis, well is there actually a real home???


Wed 07 October 2009, 16:03
Just an update. This is the latest version which seems more versatile to me. It has free moving along the width of the table and the live center displacement is only 42mm. A piece of a U channel wing will be cut to act as matching clutch. The previous holes on the U80 channels are not needed
anymore except for the ones that carry the main assembly in both vertical and horizontal axis positions which will be left to the builder. Of course still the table U channels should be arranged face to face. I just wonder if two M10 screws could harness the tailstock from backing off.

Today I bought the 24mm rod screw and nuts and will continue to complete it. It might take a week or so since I am busy painting the machine.

John and Irfan. Thanks. I will arrange the logs on the machine from the beginning and start a new thread soon. Actually there are two machines and unfortunately neither are blue!

Irfan, please refer to the second picture from the top. The dark green part is a roller type microswitch. It is quite sensitive with PT=0.4mm and MD=0.05mm (which I don't know what these abbreviations stand for) but I can see when activated, it returns in very minute distance. It is also present in the third picture from the top. It is locally made under Italian license and is equivalent to $1.5.

While still too early, I was wondering if I could define two sets of profile in MACH. One with a rotary axis with a homing switch and one without a homing switch. I thought the second set might mimic a wood turning lathe. Is it logical?

Gerald D
Wed 07 October 2009, 23:55
. . . . I was wondering if I could define two sets of profile in MACH. One with a rotary axis with a homing switch and one without a homing switch. I thought the second set might mimic a wood turning lathe. Is it logical?. . . . .

You can create as many profiles as you want - just rename the .xml file for each one.

Thu 08 October 2009, 06:49
Irfan, please refer to the second picture from the top. The dark green part is a roller type microswitch. It is quite sensitive with PT=0.4mm and MD=0.05mm (which I don't know what these abbreviations stand for) but I can see when activated, it returns in very minute distance. It is also present in the third picture from the top. It is locally made under Italian license and is equivalent to $1.5.

Hmm I have seen that switch in the local electrical stores - need to enquire more on this.

Thu 08 October 2009, 07:33

I like your new 4th axis design very much.:)

However, I thought of another problem you may face. Since the U channels are unlikely to be exactly parallel in both dimensions (X and Z), you may need to be able to adjust the support mechanism to re-align the centers of the two ends of the 4th axis as you slide the support mechanism back and forth. In a lathe this is handled by having ways that are machined to be exactly parallel and flat. You may be lucky and it won't be a problem. It will require testing.

I'm anxious to see the logs of your MechMates. As for colour, to earn a Serial Number a MechMate must be painted (MechMate blue is preferred but not required), it must be cutting and it must have MechMate Logo's applied. Of course, the decision to award a Serial Number is made by Gerald.

Thu 08 October 2009, 14:15

That is so true. I will work on possible ways.

My original thought was that you have to:

1- With no workpiece loaded lower the tip of the tool to the center of the chuck.
2- Determine the approximate location of the tailstock and set it there.
3- Jog in the Y direction to the tip of the live center.
4- Find out how much corrections are needed for Y and Z.
5- Adjust the tip of the live center to and forth from its base to correct the Y deviation.
6- Shim the bed for Z correction.
7- Mount the workpiece.

Of course this is a crude procedure and will only apply to a one on one part basis. Maybe a targetting laser from a cheap key ring could speed up the process. Anyway, I have to come with something better.

Thu 08 October 2009, 15:50
I like your alignment procedure...I hadn't thought of it. I had thought of how to know where the end points of your 4th axis are...your procedure would solve that.

Perhaps you could add some adjusting screws with lock nuts to make fine adjustments to your Y and Z positions and then lock them in.

I don't know much about how to use them, but there are Add-Ins for Mach3 that will allow Mach3 to find the center of a cross hair and to zero the Z axis. You use metal touch plates so that when the tool touches the metal surface, a circuit is completed and Mach3 knows that a target has been reached. If you are interested, see Wayne #25's build thread here (http://www.mechmate.com/forums/showthread.php?p=23617&postcount=99)for an example and pointers to the construction details.

Fri 09 October 2009, 17:19

Thanks. I am preparing the machine thread which I suspect will be ready in a couple of days.

I came up with some ideas regarding your comment on adjusting screws which I will work on them and send the post.

Sat 10 October 2009, 10:05
Cool design Atifeh, I can't wait to see how the design works out.

Sun 25 October 2009, 15:10
A colleague’s lathe was down due to its faulty chuck jaws. I fixed his problem and he let me use his lathmachine the required parts for the 4th axis. Here are the pictures:

Parts required for the tailstock:


As seen, the M24 nuts sit pretty good inside a 40x40 square tube. Only a minor sanding was necessary to fit it completely inside the tube.


Here is the semi assembled tailstock. There were not adequate tools to machine inside the live center head piece, so I had to change the design.

The motor shaft was machined to fit tightly on a 100mm diameter steel plate. Welded and machined.

The motor side sub assembly:


Sun 25 October 2009, 15:12
After installing the chuck assembly on the motor, I noticed a considerable amount of play on the shaft of the Oriental Motor PK296A2A-sg36. I think it is about 2-3 degrees. I tied together all the motor wires to confine its movement but the play still existed , which led me to the fact that this must be in the gearbox section. First I thought I could compensate the backlash in Mach, but then it occurred to me that during back and forth movements of the spindle, the workpiece will freely move accordingly and this will totally scrap my selection of this type of motor. I thought there might be some adjustment inside the gearbox so that I can tamper with to eliminate the backlash even with the expense of shortening its life or sustaining more friction between the gears, but I don’t know which part to dismantle. Please help me solve this problem.

Mon 26 October 2009, 04:39
My suggestion is to go to a belt reduction.

I would aim for 10 to 1 ratio. This would necessitate a two stage reduction. Unless you can source or make a large (120 to 150 tooth) pulley.

Any backlash in those stepper gearboxes is not so significant when driving a 30mm pinion. However when driving a 300mm diameter project that same amount of backlash becomes very significant.


Mon 26 October 2009, 13:43

You are right. My original intention was to make a two stage 1:9 timing belt reducer (I only have 24

and 72 teeth timing pulleys available). the problem is that a 1:9 reduction will only provide 12 minutes

(0.2 degrees) of positioning accuracy. On a 800 mm diameter work this will resolve to 1.4 mm which is not good for example if you want to make a PU or fibreglass mould. I read somewhere that the preferred ration of reduction is between 15 and 30. In addition these reducers will be massive and heavy. The Oriental Motor stepper gearboxes are very slick and will not take too much space if implemented in the MechMate, and that was my main objective in selecting them.

My assumption was that the stepper gearboxes with high reduction ratios will have reduced backlash

which turned to be wrong. I wonder if this amount of backlash is normal in them?

Is there any solution to correct these geaarboxes?

Gerald D
Mon 26 October 2009, 13:48
The ratio for rotary work should be about the inches. Eg. 30 inch diameter needs about 30:1 ratio. A very rough guide.

Mon 26 October 2009, 16:27
Gerald, Thanks for your useful note.

Mon 26 October 2009, 16:33
The below picture shows the sliding mechanism simulation which was devised. It will have a clutch locking mechanism. The longer piece will act as the table beams, underneath which is the sliding cross beam which supports the tailstock. A piece of 30x100x10 mm hot rolled strip shall be welded on the square tube which will be tapped M12. The upper parts are a piece of a U80 channel wing, inverted to compensate the draft and welded to one of the most widely used steel, i.e. the remainder of the rail cut. This part will be bored 13mm diameter. The square tube will be drilled 14 mm diameter, just beneath these holes to accommodate the length of the locking screw.

This is the tailstock body welded and complete.

In order to dermine the welding height of the 50x50 angle iron to the tube assembly the following method was used:

The angle iron has one elongated hole which will be used for yaw adjustment of the tip of the livecenter with the end of the workpiece. Besides, the cross beam has about 4 mm room to play sideways. You can see the elongated hole on the angle iron, in front of a M12 tapped 50x100x10 mm hot rolled strip.

Shims will be used for the pitch adjustment. So far I could not figure out a simple way for this adjustment.

These are the parts which make up the sliding tailstock cross beam. Please ignore the holes on the welded tabs on the square tube, as these were done during the first version design.

Mon 26 October 2009, 18:11
On a 800 mm diameter work this will resolve to 1.4 mm which is not good

I get about one minute per microstep, but you are right, 10 to 1 is not enough for 800mm dia work.

I assumed, incorrectly, that given the size of your chuck and bearing support that this 4th axis was for smaller work.


Tue 27 October 2009, 16:41
Today I went after some gearboxes to fit to a stepper motor.

So far I found out that worm gears are not suitable at all. I recall that some 20 years ago they made these gearboxes with vertical adjustment that forced the worm on the worm gear. Now they have only horizontal adjustment.

Found a planetary gearbox with 15-20 minutes backlash.


Assuming that microsteps are not positioning steps, and that you probably set your drivers to 1/10 microstep, therefore I assume that you should get about 10 minutes backlash. What was the biggest diameter that you machined without any problem?

Do you see this gearbox suitable for this purpose?

Tue 27 October 2009, 19:45
Yes that would be about ten minutes per full step.

The reason I like belt reduction is because backlash can be pretty good.

Impractical for large reductions needed to handle large diameters though.

My rotary axis is based around a small worm drive right angle industrial gearbox. 40 to 1. It has about 15 minutes backlash which I compensate for in Mach3.

Again this backlash would be unworkable for large diameter work. I only do small (up to 150mm) jobs. The inherent stiffness in the gearbox helps with the backlash at these small diameters.

I think if I wanted to do large work I would look for a harmonic drive. Theoretically backlash free.

I see on a link I found here on the forum that OM has a harmonic drive model,


50 to 1 and 100 to 1 ratios. They are probably very expensive.


Wed 28 October 2009, 01:19
Greg I have picked up a harmoinc drive gearbox from ebay for $100, that was like a year and a half back its a 1:100 reduction, never got a chance to use it - may be now!

Wed 28 October 2009, 15:49
I bought this gearbox to replace the Oriental Motor's.

Later I will use the oriental motor to make a unidirectional indexer to make big timing pulleys which are extremely expensive over here.

Here you can find the specifications:


Wed 28 October 2009, 18:41
Are you planning on using the stepper part of your oriental motors? I think I remember it being posted here that the gearbox could not be removed from the stepper. Cannot remember why at the moment though.

Thu 29 October 2009, 08:51
Lots of used harmonic steppers available on ebay - take a look here


Mon 02 November 2009, 11:30
Thanks Heath,

I was going to dismantle it to see what could be done inside it. for the new gearbox I bought a new stepper motor.

I am still amazed what good is this motor for? With so much backlash.

I asked for the price of a 30:1 harmonic gearbox. its price is astronomical down here. Equivqlent to $2200! i.e 10 times the gearbox I bought recently.

I think 100:1 ratio will be too slow unless you use it with a servo. I ran my 36:1 with around 750 rpm and found out the speed is frustratingly low for a lot of jobs.

Mon 02 November 2009, 11:35
Here are final pictures of the tailstock:




Continuous traveling distance along the Y axis.

Amount of sway for yaw adjustment.

The following picture shows the top profile of the tailstock relative to two 16 mm mdf boards. I used a 100mm diameter handwheel (which was the smallest size available). If 80 mm was available, the total unit would be clear off from the tool mounted spindle. The chuck is on the home side dead zone of the table and spindle never reaches it.

Future jobs:
Placing the homing microswitch.
Design some kind of brake for the PK motor in order to confine its free movement while machining, along with backlash definition in Mach.
Make arrangements for attaching the new gearbox to a separate stepper motor, using the existing construct.

Robert M
Mon 02 November 2009, 12:45
Nice going Nader…
It’s clean…I must say, I like the way you execute & make things happen.
Congrad….Robert ;)

Mon 02 November 2009, 22:52
Thank for nice Information, it's a great concept :)
With your Picture very clear to understand, what's you want to say. Thank again.:D

Fri 06 November 2009, 16:42
I am still amazed what good is this motor for? With so much backlash

I used pk296sg7.2 . May be you should power up motors to check backlash. I see backlash when no power, but while working or holding, I see no backlash or its effect on work.
I am not familiar with gear box.

Tue 17 November 2009, 17:15
Thanks everybody.

Thanks. You have a point. I will try this too and see how it goes.

Tue 17 November 2009, 17:29
Greeting to all and sorry for the delay.

Following the previous post agenda, here are the new progress.

1- I was thinking on this scenario for the case of the PK296A1A-sg36 motor:
How about gripping the output shaft tightly, with some polymer material, in a form that the clamping pressure on the shaft be adjustable (say, like drum type brakes on a vehicle) which will confine the shaft’s movement under normal cutting forces, but obviously less than the torque exerted by the motor and the gearbox during movement. The idea is to localize the backlash only to directional changes of the motor and assign this amount of backlash to MACH. Please comment.

2- Design for the new gearbox
New Al plates for the motor side and chuck side of the gearbox.

I use an old height gauge for marking up the parts. The blue dye is just a refill for water proof markers.

Center pinning is done under a magnifier, first with a 2.5mm sharpened ejector pin for precise location and then the ordinary center pin.

The finished parts:

A new motor is attached to the gearbox.


The new gearbox needs a new shaft and chuck plate. The procedure is the same as the previous one.
As you have noticed on the new gearbox, it has a keyway shaft. I will try to make the matching guide on the inside of a hole using a method named “Chiseling” (just google “keyway cutting” and you will find the sources). The idea is to make a hole, (in my case 4mm in diameter) and then square it with a 4x4 mm tool steel either on a lathe using the support movement, or a mill or drill press, by minute amounts of feed.

A plug is machined to the exact size of the hole, which has a pulling lug.


The drilling surface should be exactly the height of the hole, otherwise the drill will slip off its position. The shaft diameter of the gearbox is 12mm and the overall diameter and the height of the key is 13.5mm, so I have to drill 0.5 mm inside the perimeter which falls on the pin itself.
The coloured area is to provide a contrast for better viewing.
This is the setup so far. I will cut the keyway by tomorrow and will post the results.

Tue 17 November 2009, 20:10

Could you verify length of these parts?


Sat 21 November 2009, 14:48

I have left the drawings in the workshop but for a visual approximation consider these:
The Chuck and the hand wheel are 100mm diameter.
The screw in the tailstock is 24mm.
The bearing is UCF205, the bolt disntances in which are 70mm.
The angle iron is 120x120x10mm.

If anyone interested, I could post the drawings after completion and debugging.

Sat 21 November 2009, 14:59
2- continued…
This, I had not done before but it was an exciting experience.
The flange and shaft part was put in the freezer and the plug is heated and hammered down into the hole. This should be done, otherwise the tip of the drill acts as two teeth gear and will try to rotate the plug inside the hole. This should be done quickly or else the plug will get loose in the hole. A center drill provides a better positioning way that ordinary drills.

The guide hole drilled:

Lug pulled out. As you have noticed there is a slight draft towards outside of the plug.

The tool is a 4x4 mm HSS, ends sharpened in this configuration. Both ends will be needed to cut opposite walls and the top wall.

The working setup.

I did the job on my press drill. Tightened the pulley as much as I could in order to restrain the chuck rotation. The feed was 0.1 mm and several passes of the quill were required to scrape the metal. The total process was about 20 minutes.

The tool being about 100mm in length, flexes along the height of the hole towards the inside of the hole. After completing the initial phase, a slight inclination should be given to the part so that in the second phase the top wall becomes straight cut.

The finished job:

The cut keyway had about +0.1 mm tolerance sideways. It was satisfactorily eliminated by a 0.1 mm shim and now the assembly has no noticeable backlash.

These are my sources for shim material: Aluminum sheets inside a blown capacitor (0.1mm) and shaving blades which are cold rolled stainless steel, precisely 0.08mm thick.

The new assembly now in place, attached to the previous construct with no changes.

3- The homing microswitch
This is the preliminary setup for the microswitch assembly.


Later a simple bump will be filed and polished from a piece of brass (only for its contrast of colour) and glued to the outer rim of the chuck holder by Loctite.

So far this project is completed except for some minor refinements and adjustments which I will post as progressed.
It will remain till the machine is completed so that we could see how it will actually work.

Robert M
Sat 21 November 2009, 16:49
Nader…..Love this !
Neat the way you thought out your complete 4th axis machining steps & process, with basic tools !
I like your DIY ways & approach.
Congrad, Robert ;)

Sun 22 November 2009, 21:07
Inspiring and nice work.

Mon 23 November 2009, 14:05
Thanks Robert,
I very much hope they are useful.

Thanks Sharma,
A couple of picture showing the approximate dimensions:



Mon 23 November 2009, 20:20
Thanks Nader.

Fri 04 December 2009, 09:07
Nader. Thank You for a nice work

Sat 19 December 2009, 14:14
I have never done 4 axis machining before and all I know is theoritical or even deductive.
I assume:
On a rotating axis (in this case B axis) only Y and Z axis will be engaged. The rotating axis will be a X axis shaped into a cylinder.
There is a case though that I believe will need X,Y and Z axis movements. Suppose you want to carve a shape on four sides of a rectangular shape. Therefore the rotating axis will only do an indexing job. The rest is machining in YZ plane with incrementing X or XZ plane with incrementing Y. so X must have an offset relative to the center of the chuck (like below picture):

Now here is the trouble.
According to my proposed layout to move the B axis to the end of the MechMate, there will be very little space for X travel before hitting the limit switch. In the case of my machine it is about 70 mm, which means a rectangular shape of ony 140 mm could be worked. As you can see I have moved the spindle to the last holes of the carriage square beams. Now, the tool changing will not be clear off the table, as it was before. And the dust collector head should be mounted on the spindle side.

It would be wise to move this assembly further in, to obtain more X axis travel.

Sorry for the upside down ruler. I noticed it right know.

Sat 19 December 2009, 16:39
On a rotating axis (in this case B axis).........
If you are using Mach3 as controller you might want to designate the rotary axis as A.

Yes I know by convention it is a B axis, but Mach can only visually represent an A axis rotary toolpath. Not B or C.

My rotary is set up in the same configuration as yours and strictly speaking is a B axis but I call it A for that reason.

Any 4 axis CAM that I have used allows you to designate it either way.

Sorry for the upside down ruler. I noticed it right know.At least it has proper markings on it and not medieval english. :)


Gerald D
Sun 20 December 2009, 00:04
Why don't you want to use the x-axis to the right side of the rotating axis, or "behind" the rotating axis from where you will be standing?

Sun 20 December 2009, 12:50
Thanks for the valuable notice. I did not know that. I have another question though. In Mach we normally use the A axis, slaved for the X axis. Should I define the B axis to be slave to the X axis, and then define the A axis as rotary?

Do you mean something like this arrangement? I wonder, do I have to reconfigure Mach with a new axis definition, in which Y axis has two motors and X with only one, i.e. exchanging the X and Y axes?


Gerald D
Sun 20 December 2009, 14:22
I was wondering why you wanted to work against the limit switch of the x axis. I suggest you consider to work on the other side:


You can configure the axes any way that you want, in any direction that you want.

Sun 20 December 2009, 15:20
Please take a look at the last two pictures of:

Suppose you want to carve a relief on each side of the cube part.The rotating axis will be fixed at each 90 degrees index and the relief will be cut in ordinary XYZ travel. Now, the maximum width (in this case the X travel) will be limited to 2 times of the distance between the center of the chuck and the ultimate distance that the X axis can go (i.e. to the end limit), in my case this width is limited to 140 mm.
This would possibly be the only case that one might need X travel. In other cases your notion is right.

Mon 21 December 2009, 17:24

The link doesn't work for me.

Mon 21 December 2009, 18:09
Yes, Art did some redesign of his site a little while ago. Some of my old links do not work either. The site is still up though.

Tue 22 December 2009, 01:12
Thanks for the reminder. Heath is right, some pictures are omitted.

Visiting the following link:
The second and third picture from the top shows the problem I previously mentioned.

Studying his design more closely, presuaded me that this should not be worrying. The square part, as big it might be, could be done on a separate piece using XYZ ordinates and mounted to the circular section.

Tue 22 December 2009, 03:59
Thanks for the presentation, its great to see an idea come to life.

Time for me to open Solidworks and get designing

normand blais
Fri 22 January 2010, 10:22
Hi here is a good example of 4axes work.The rough is xyz then flip 180. The second part is where the 4 axes are working ''simultaniously '' .Indexer work and 4axe work are different. With the right software you will be able to do both with .

Gerald D
Fri 22 January 2010, 10:44
In that video, with all 4 axes running simultaneously, notice that the cutting bit is always perpendicular (90 degrees) to the surface of the work. That was a strategy decision by the programmer. I wonder why he decided not to use the side of the bit . . .

normand blais
Fri 22 January 2010, 10:57
He waisted lot of time machining the mass instead of contouring the wood last. Machining 4 axes allow also to do undercut.

Tue 26 January 2010, 15:02
Most likely, if he was using a package like Rhino 4th Axis, cut3d, etc...it's easier to write out a 4 face cut file with indexer being either manually written to index 4 times or as I like to do it, write four files and have Mach play them in sequence.

It looks like the file did the following:
4 - face, spiral roughing, parallel to face with collision detection for bit
1 - 4th axis continuous rotation final pass with a very high overstep for resolution.

On the larger index models, I found that the 3d model software would crash after about 3 million polygon faces. Thus, I had to do each face separate to generate a toolpath before trying to let the software "remove" all that material for a final toolpath. Computer crashes are my driving factor when trying to produce a part file.

Have a great day.

Fri 29 January 2010, 15:10
A piece of copper was cut and filed to make a bump for the rotary axis homing system. A strip of sand paper was fixed on the rim of the plate and the copper piece manually rubbed against its perimeter to get the exact diameter of the rim.
It was then fixed with Loctite. I usually cover parts that I fix with Loctite with adhesive tapes, so that they are isolated from the air and speeding up their curing.

After curing, the top surface is manually finished with grit 600 sandpaper. Next the microswitch is adjusted against this bump to have the least movement.


Fri 29 January 2010, 15:11
Is discussed in post #116 in the following thread:

Wed 26 January 2011, 03:28
Did you get your 4th axis working?

Thu 27 January 2011, 03:12
Hi Brad,

Yes it is working fine and quite precise.

Thu 27 January 2011, 04:12
Excellent. I'm thinking of doing a 4th axis build like yours. I've just picked up a 100:1 harmonic drive. Im a little concerned the 100:1 isnt divisable of 360 so not sure how mach3 will go with the steps.



Thu 27 January 2011, 04:30
Hi Brad, don't know how the harmonic drive will interface with Mach 3 on the indexer.
I have a 130 tooth module -1 gear fitted to my lathe shaft.
The Stepper motor has a 24 tooth pinion fitted.
Mach-3 has a function to calibrate the movement.
All I did was to set the indexer with a spirit level.
Tell Mach-3 to go 360....full revolution.
Eyeball the spirit level after this movement and told Mach that it moved 280 for instance (depending on where it stopped)
Within 3 or 4 calibrations I was going exactly one revolution when mach moved 360.

Thu 27 January 2011, 04:41
Excellent.. I'll try that method. I was doing the maths method and 1 degree movment was calculated at 555.555555556 steps per degree. I'll give the calibration a try and if it doesnt work perhaps use a pulley on the stepper 20:18 HD to reduce the ratio down to 90:1 to give an even 500 steps.


Fri 28 January 2011, 06:30
Maybe you would give a second thought to that. A harmonic drive is virtually a zero backlash gearbox. Interfacing it with another set of gears might introduce some backlash.
You can add as many decimal places as Mach accepts. It will not round it up. And don't worry about 555.555555 and 555.55555. it will be less than 1xE-6%.

there is a note that should be considered when using microsteps with the fourth axis. As long as the fourth axis is rotating, it will provide a fine machining, but if your toolpath should require that the fourth axis to stop, while e.g. the Y and Z axes is working on the workpiece (like making fine splines along the length of a cylinder, then microstepping will be neglected and the fourth axis will stop at full step, or at the best at half steps, depending on your wiring, since positioning will not obey microsteps (or at least this the case with my driver).

Fri 28 January 2011, 07:11
Hi Nader,

So good to hear from you. I think of you and your build from time to time and wonder how you are doing. Perhaps you would consider "showing" us your final (latest) machine and some of your results.

I'm especially interested in your indexer experience, since I'm planning to use a similar design on my next MechMate.

John Higgins

Fri 28 January 2011, 22:44
Nader - I would also need some more inspiration! do post some pics

Fri 28 January 2011, 22:46
Brad get a servo for your harmonic drive - there are lots on ebay - they top upto 3K rpm and would definitely work with the harmonic drive

Sat 29 January 2011, 06:46
Hi John and Irfan,

Thanks. Right now I am working on another new design which is intended for cutting mainly Aluminum. Also I am doing some research and experiments to introduce a 4th and 5th axis on this machine. But I will send some of the workpieces made before.