[LeeWay]

I was kinda kidding about the math part. I don't think many can get by in life without it. At least the basics.
With that pulley in hand, I might have went the way you did. It helps to be able to see what would be left after machining, so may have been too thin. Especially with a servo system.

I have a Tormach lathe and I have been looking for something that may be challenging on it. That looks like a perfect project for it. I have been wanting to try some broaching as well. I will be needing to know how to do that soon enough.

[BMSTECH]

My timing pulley now has a collar that can have a Keyway broached and a grub screw fitted. Collar can be reduced a bit to match the motor shaft length. The 3x3x15mm key length will be reduced to 10mm so as to not undercut the teeth. Will make these modifications when the motor mounting plate is made and belt alignment is finalised.

[BMSTECH]

A kitchen table servo motor test.

[BMSTECH]

This is the Servo mounted on the Mill head operating the Quill for the first time. Seems OK at this early testing stage. Still lots to do.

[BMSTECH]

I have added the 3rd Axis Servo Drive to the Servo Drive Enclosure and made some of the connection cables. With the Mill head rotated back to vertical, the Servo Motor and Quill Ballscrew Actuator is looking OK and may not be as intrusive as first thought. I have drawn up a timing pulley cover while I wait for some new cable for the servo encoder and power to arrive in the post. I've also ordered a couple of cheap limit switches.

[BMSTECH]

I have made the timing belt cover.

[Muzzer]

Wow, nice progress!

This is the first build thread I've seen that uses the same approach as mine for the Z-axis drive. I've built mine with the motor hanging out the other side of the head and my motor is a Leadshine closed loop stepper.

I did most of the design in Solidworks although I was on the Onshape closed beta and imported the compts into OS to see how I got on with assemblies. The CAD views are from OS. As you can see, I had access to a 3D printer and was able to make samples up for test fitting.

The X and Y axes are using DMM Tech servos with belt drives (2.5:1) and jockey tensioner. The biggest job remaining is the machining of the yoke which needs to be bored out a fraction of a mm to accept the ballnuts. I'll have to dismantle the machine to do this and hopefully manage to get the yoke onto my lathe. I have 2505 screws from AliXpress machined to my design.

I'm using LinuxCNC and Mesa 5i25/7i76 boards.

I'm between houses right now, so the build is on hold while I am without a workshop. The wait is killing me....

The really bizarre thing is that I also have a Honda S, although being a Brit, mine is an S800, not an S600. I've seen your build thread on the Honda forums - very good work there too.

Keep it up!

Murray

No idea why the photos are now upside down. Can't see any clever way to rotate them. Sorry.

[BMSTECH]

Murray,
Your Z Axis is looking good. I have also not seen any others done this way. Mine is still experimental at this stage but is looking promising. I am still to test the downward force it can exert on the Quill before stalling the servo. My complete mechanism is attached to the lower Quill stop mount by just 2 screws up through the new fixed bearing block. I got a 12x32x16 dual row angular contact bearing to replace the 2 cheap single row deep grove bearings supplied with my Ballscrew.. Are you using a 25mm diameter Ballscrew? I am still thinking about the upper floating bearing and wondering if it is required at all. The end of my off the shelf 1605 Ballscrew locates in the original depth stop spindle hole and may be OK just like that. Is your Honda on the road? What are the chances? We will surely be the only 2 guys with Honda S cars and CNC Mill conversion projects in the world?
Cheers,
John

[Muzzer]

Like you, I've used a 16mm ballscrew (1604 in my case) because it slides in through the original hole without machining but I've taken the thrust force through the top of the leadscrew, via a 5201 deep groove bearing (12 x 32 x 15.9mm), with another such bearing at the bottom, taking the belt loads. The top bearing sits in a cup that locates in the hole left by the feed trip rod. This way, there's no need to rely on a couple of small threads - simple and robust.

If you do the sums, multiplying the motor torque by the pulley ratio and the screw pitch, there is likely to be more than enough force available for the design limit, which would be drilling most likely. I'd be more bothered about the risk of damaging something rather than failing to produce enough force for the job.

I only have an 18:24 reduction due to the space limitations but with a 2Nm motor and 4mm pitch, that stall torque translates to over 4000N ie almost half a tonne of force. I reckon that should suffice for most jobs....

I also used a sliding tensioner pulley rather than sliding the entire motor. Solidworks has a nice belt mate function that allows you to fit a standard length belt to your pulleys and automatically position the tensioner in its adjusting slot. You can then finesse the position of the motor and ballscrew to get things in the right place. When the belt and pulleys arrived, I was pleased to find they all fitted together perfectly.

My 1967 S800 Coupe is at my mother's house where it's resided for some years. We've just moved back from Canada having lived there for 4 years. I didn't take the car with me but instead built up a new spare engine. I fired it up OK and it's ready to go back in the car. I've also fitted fuel injection from a CBR600RR, along with an Emerald K3 injection controller. Before I moved out there I had it on the road - incredible fun to drive. Changing gear at 10500 rpm, doing 5500rpm in the next gear is quite enjoyable.

About the only thing I didn't change when I first built up the engine was the oil strainer gauze on the oil pump. Literally a couple of days before I was due to take the car off the road, the strainer got sucked into the pump and of course the first thing to suffer in that situation is the inlet cam, being the highest journal bearing in the engine. That seized, causing the chain to slipp and bu99er up the valves. Hence the second rebuild. Lucky it happened just before laying it up, rather than within days of next running it. The original oil strainer was made of brass which slowly dissolves in the acidic oil, becoming weaker with time. The modern replacement is stainless steel.

I have a spare body, with several panels still in original wrapping and a galvanised chassis. Realistically that's a retirement project (it needs to be done properly) but I may run the car as a rough-looking dog in the meantime, using the current body.

[BMSTECH]

My Samsung CSM 200W Servo is rated at 0.6 and can peak at 1.9 Nm. torque. I've used 12:24 T5 Timing pulleys and 5mm pitch ballscrew. I had issues with the mechanism binding that were mostly resolved by straightening the ballscrew, replacing the cheap bearings and squaring up the new bearing block with an epoxy shim. There is still a reasonable effort required to manually wind the motor pulley to overcome friction. I still have the Quill counter balance spring attached. Alignment of the Quill Yoke could be contributing a little but difficult to workout how to improve. I have taken a "suck it and see" approach so I have not done the Torque calculations. Proof will be in the pudding. I think my Encoder and Power cables arrive today so I can fabricate the 3m leads to run back to the Drive enclosure. Mach3 is ready for a real test as it seems to work with a spare servo on short leads around the back of the mill. I ordered some smaller limit switches yesterday and can now draw up a mounting plate. Also thinking about a 75x5x225mm Aluminium tube to machine a ballscrew cover to completely enclose the mechanism.
By the way- "Merry plucking everything!"
Cheers,
John

[Muzzer]

I bought some inductive proximity switches from China. They are mostly NO open collector but in my purchase (20 finally) there were some NC version as well, which is actually quite handy in some places They work on anything metallic, not just magnetic. They are IP67, so immersion proof, repeatable and no moving parts.

My Z-axis worked OK without binding anywhere. Given that it has bearings at every point of contact, there would need to be a fair misalignment / bend for there to be any serious friction. Was yours actually making metal-metal contact somewhere?

I'm going to make up some shields for the Z-axis and also for the spokes of the X and Y hand wheels, using ABS and 3D printer. I can see that they might make rather effective mincing machines otherwise. I'll replace the long handles with short knobs so they don't rearrange my lower body parts but that would still leave a chopping hazard.

Wife's gone to pick up the turkey just now. I suppose you lot will be preparing to flash up the barbie! You'd be doing well to get one working here with all the wind and rain....

Have a great break. Make sure you eat and drink far too much. We will.

Murray

PS - you probably can't tell but the 2 hyperlinks in this reply were actually added by me - not by the crappy forum software. They link to the proximity sensor ebay trader I used and the datasheet. Haha - I see it's added a couple to this post script note.

[BMSTECH]

I have put together a video that shows this Z axis Design and Build to this stage where the mechanical side is complete and basic a test with Mach3 control passed.

[CSX66]

Looks great! You could gain a few mm of extra travel by changing out the allen cap screws retaing the ball nut for countersunk heads.

Cheers

Mike

[BMSTECH]

There is some smaller limit switches coming in the post that will be mounted on the original depth ruler mount.
This will allow a neat ballscrew cover to be made.
Thinking about a 225mm length of 75x5mm Aluminium tube machined out to fit vertically.

[Muzzer]

I chose to fit the ball nut the other way round, with the flange at the top. This means that the cutting loads are naturally passed to the yoke and don't require substantial bolts in tension to hold it in place.

I also turned down the flange so that there was little left apart from a step of a couple of mm. I could then fully contain it within the yoke. It didn't make the yoke / ball nut assembly much shorter but it is more compact. Although the ball nut is hardened, you can simply machine it using standard carbide indexable tooling as long as you turn your speed up to max. In my case that was 2000rpm. You need to make sure there aren't lots of flammable materials around when you do this! I lost a couple of edges when I was going through the bolt holes but in the end it only cost me 2 inserts (CCMT style). An insert grade that is designed for hardened steel and/or interrupted cuts would have lasted better but it really wasn't an issue and hardly broke the bank. Power feed and saddle stop made the process pretty quick and simple.

I split the yoke and fitted a pinch bolt. This means I can simply slacken the pinch bolt and revert to manual operation of the quill.
Attachment 303012
Attachment 303014

[BMSTECH]

Mike,

The quill lower travel was limited by the rack length and is 2 mm below the cap screw heads. Countersinking hardened Ballnut would be difficult.

No loss of travel at top with this set up, and that, after all, is the business end.

With the limit switches operating, a safety margin of about 4mm top and bottom reduces overall quill travel to 115mm.



Murray,

Like your idea of being able to use the Quill manually. I have become quite proficient at typing G-codes into the MDI to operate the Z axis and will get by. I will not be able to Power Tap but I am not keen on that anyway.

I have added some pictures of the bearing block and motor plate using a 5201 dual angular contact ball bearing.

[Muzzer]

That's looking good. Looks as if Mach 3 uses the end limits switches as well as the home switch. LinuxCNC is happy (and robust enough) to work from the home position and implement soft limits. However, I will use hard limit proximity sensors possibly wired into the e-stop circuit to protect against damage, most likely caused by finger trouble.

I'm grievously jealous now. I won't be able to resume work on my machine until I've moved into our new (to us) house this month, demolished the old single garage and put up a proper replacement, so I'll be an armchair engineer for another few months.It's killing me.

One machining job I have still to complete is the saddle yoke for the X and Y ball nuts. Although my Taiwanese-made machine is sort of metric, in practice that means it has a mixture of threads and dimensions. The 2 bores in the saddle yoke seem to be around 38mm dia, which I assume is actually 1.5". Naturally the ball nuts are 40mm, so I have an annoying problem to fix. Although I was boasting earlier about machining the hardened, ballnut flange for the z-axis, for the X and Y I have double ball buts (for alleged zero backlash) which would require me to machine down the ODs of 4 nuts. I don't fancy that but the alternative is boring out the yoke bores on the lathe. You were blessed to find a machine with ballscrews already fitted!

The yoke bores appear to be 4" between centres, so I'll need to be able to swing something like 135mm (radius) on my lathe (a Bantam) that claims to be able to swing 143mm. That sounds lucky, as the alternative would be to bolt it to the cross slide and use a between-centres boring bar. Not easy, not least because the lathe doesn't have tee-slots. You might have noticed from my photo that my z-axis motor bracket was welded in the middle. It was too long to swing in my lathe (165mm) so I had to chop it in 2 and weld it up after turning the bores. You might call that a design error.

[Muzzer]

Another x-axis concept here. The blogger didn't do the conversion himself and the blog has been dead for several years. However, it also fits the ball screw in the feed trip position. Interestingly it feeds into the end of the quill rather than using a bracket mounted on the old trip mount. Probably more robust this way but requires the ballnut to rotate rather than the ball screw. And keeps the motor up and out of the way.

[BMSTECH]

Mach3 and Fusion360
A series of screen shots of Fusion 360 CAD of a Bridgeport Mill Crank Handle Adaptor and CAM Gcode generated to be sent to Mach3.

Autodesk Fusion 360 is Free to download and use as a hobbyist and has most of the functionality of Solid works which has cost a small fortune in the past. The CAM driver for Mach3 alone would cost hundreds from others.

Initial testing with my Mill now equipped with 3 axis control seems to go through the motions correctly. I am stoked!
Just experimenting at this stage with a Video to come when the machine is doing more than cutting air.

See an old guy make one on a manual mill the hard way.

All machinists can benefit from a 3D model before they start to make something.

Take the time to learn the basics of F360 and the reward will save you time, money and disappointment.




Mine uses a 5mm Endmill.

[BMSTECH]

Drinking beers around a campfire over Easter maybe better than watching Fusion 360 tutorials and pissing around a computer screen. Each to their own! Maybe next year.