Weiss VM32 Mill Conversion

[picclock]

Weiss VM32 Mill Conversion, 1.5kW brushless motor belt driven design. Large 840 x 210mm table with 525mm X movement and 220mm Y. Z travel is 370mm assisted with a gas strut.
This mill was purchased from Toolco under the VM32B name or is available from Amadeal. I spent a lot of time at shows and suppliers before settling on this machine, which I think is a very good candidate for conversion, as well as being an outstanding mill. By far the hardest part of this project was getting hold of the mill.
Disclaimer - I have no connection to any of the suppliers mentioned in this thread or anywhere else come to that.

Built the stand as shown in picture 1. silver rails are drawer runners for mill parts. Centre section houses drivers power supplies and electrical stuff. Kitchen worktop in picture behind forms worksurface.
The big day finally arrived and I was left with a large box on my drive (2).
Removing the box revealed the mill and accessories covered with a transparent bag (3)
Bag and bits removed revealed the mill. Inspection revealed no transit damage (4).
Connecting strops under mill to allow a safe and stable lift (5).
Test lift showed all Ok. (6)
Finally onto the stand with no dramas (7) :-).

Next installment soon.

Best regards

picclock

[routercnc]

Nice work. Warco also sell that machine or similar and it has been on my consideration list for some time. Having a CNC router is great but sometime a quick bit of manual machining, especially steel, would be handy.

[picclock]

Hi routercnc
I visited the Warco factory on their open day, primarily to check out the WM18. The main difference is that it is gear driven and has a less powerful brushed motor. I am sure the ones I looked at had a different base casting to the one shown on their site. Having a belt drive is important to me. Its quieter and when it fails you just replace the belt. I also have a smaller WM14 mill from Warco, which has had broken gear issues. It is far noisier than the 1.5Kw belt driven VM32B despite being only 600W. I did snag a fairly decent vfd lathe while I at the show, but for me the WM18 did not meet my requirements.

Once I had tested the mill to make sure it functioned correctly I proceeded to disassembly. This started with removal of the table end cheeks and threaded x screw. The tapered gib was removed and the table removed. The original X Axis bearings were kept as I plan to reuse them. Removing the table was easy enough but it is heavy. This I slid onto a workmate. The saddle was similarly removed. Both castings were of excellent quality.

I am fitting ballscrews to all three axis, but I think I made my first mistake here. I assumed that because the leadscrew diameter is 20mm that replacing it with a similar diameter ballscrew would be the easy option. The snag is the size of the ballscrew nut which is far larger that the original - oops!.

I have read about machining ballscrew nuts with carbide mills. My advice is don't bother. I took a brand new 12mm carbide mill and attempted to make a small test cut 0.25 mm deep and 0.25 mm wide. The end result after a very short distance was a very blunt endmill :sorrow:. So I attacked it with an angle grinder with very good results. It is very important not to let the nut get hot or it will lose its temper, so cooling with wd40 or similar is a must, with long pauses between grinds.
To work on a ballscrew itself you need to have a lathe with a spindle hole large enough for the ballscrew, a 4 jaw chuck, a custom collet, headstock centering bush, and a nut bearing keeper. The ballscrews for my machine are all 20mm x 5mm pitch. They were bought with the ends machined (BK/BF15), so if you do this you must remember the length quoted by the sellers includes this. The threaded end is 60mm and the other is 13mm, so ballscrew thread is length quoted - 73mm for 20mm ballscrews.
As a guide the lengths needed for this machine are X=840 table length+end cheeks@30mm each (depending on what you do with the free end). Y= 450mm (inc ends), using a 10mm plate for the thrust bearings. Z = 650mm inc ends.

The collet is simply a length of aluminium tube to fit the ballscrew. This is cut to the length of the jaws in the chuck plus a 2 cm. Four slots are cut in one end to just over the length of the jaws. When fitted over the screw it can be clamped without marking the thread.
The centering bush is just a piece of aluminium bored to the screw OD and made with a shoulder. Approx 2 cms long should do. This is placed on the end of the screw protruding from the spindle and can be fixed in place with the ballscrew nut or a suitably sized oring. It jut stops the end of the screw whipping about.
The keeper is an aluminium cylinder, drilled through 6mm clear. The ends are drilled 10mm to a depth of 15mm or so. The outside diameter is 18mm.

The conventional method is to use deep groove ball races for the axial thrust at the ballscrew ends. This BK/BF system uses two races in which the outer part of the race is spaced apart and the inners compressed until they rotate with little axial movement. balls used in this way run on the edge of the grooves in the casing, which makes me unsettled. Using a thrust race for axial loads the ball is fully supported by the machined recess in the support washer, so a much larger area of contact is used.

In order to use the thrust races the diameter of the ballscrew end needs to be reduced to 12mm for the 51201 thrust races. These have a load rating of 13.3kN, which should just about do :beer: The end will be tapped M12x1mm for the clamping nut which will preload the bearings. The nuts are simply made out of 1" bar stock, drilled and tapped in the lathe, with the corners rounded to an OD of 28mm. They are 8mm thick and after parting off the corners are drilled and tapped 3mm for a locking grub screw.

Thats all for now

Best Regards

picclock.

[Clive S]

My experience with the conversion of WM18 is that I used 16mm screws with double nuts and used AC bearings.

I am helping someone do another one (purchased about 8 months ago) and was surprised just how much that they have lightened the table and saddle castings.

[magicniner]

I'm converting a similar machine from Chester Machine tools and found it's fairly cheap (around £5 each) to replace the lead screw end bearings with a quality brand from my local industrial bearing & seal supplier.
Don't forget your oil delivery system to the ball nuts, it will significantly increase their working life.

- Nick

[picclock]

Y axis start.
Picture 1 original X Y acme screw nuts
picture 2 Ballscrew mounts on saddle. These are 15mm thick (from scrapbox) but 12mm should work.
picture 3 Saddle with nut fitted. To fit to the mill base the mounts/nut must be removed from the saddle to the keeper(post 3 picture 7). After screw is located into the base, the nut and mounts can be attached. The keeper has a 10mm recess to fit the turned down end of the ballscrew to give alignment. The saddle can then be fitted and attached to the mounts with original 8mm screws.
picture 4 Y axis ballscrew mounts
picture 5 Y axis top view. All y axis parts are 10mm mild steel.
picture 6 Y axis plate dimensions.
picture 7 Y axis sides for motor mount. Length may be altered depending on coupling ans shaft lengths.

Thats it for now

Best Regards

picclock

[m_c]

The conventional method is to use deep groove ball races for the axial thrust at the ballscrew ends. This BK/BF system uses two races in which the outer part of the race is spaced apart and the inners compressed until they rotate with little axial movement. balls used in this way run on the edge of the grooves in the casing, which makes me unsettled. Using a thrust race for axial loads the ball is fully supported by the machined recess in the support washer, so a much larger area of contact is used.


The 'conventional' method should use angular contact bearings, not deep groove bearings, which are far more suited to the purpose than a standard thrust bearing.

A basic thrust bearing gives very little radial support, so unless they're housed and adjusted perfectly, when spun at any speed your shaft/ball screw will likely whip, resulting in the kind of issue you are attempting to avoid. Any play in an angular contact should be noticeable in backlash long before damage from whipping become an issue, unlike a basic thrust bearing.

[Hulksmash]

Weiss VM32 Mill Conversion, 1.5kW brushless motor belt driven design. Large 840 x 210mm table with 525mm X movement and 220mm Y. Z travel is 370mm assisted with a gas strut.
This mill was purchased from Toolco under the VM32B name or is available from Amadeal. I spent a lot of time at shows and suppliers before settling on this machine, which I think is a very good candidate for conversion, as well as being an outstanding mill. By far the hardest part of this project was getting hold of the mill.

Really nice thread Picclock, been looking forward to this for some time.

I'm getting mine through Hugh at Amadeal - he has been really great to deal with. He has ordered it direct from the Weiss Factory, hopefully it will be arriving April. The Vm32L or VM32B (same machine) are not the same as the AMAT30's - these are made from a different manufacturer apparently.

Unfortunately I believe the owner of Toolco lost the contract with Weiss Machines - The owner of Weiss said he went in for an operation and not been heard from since :( I hope he is OK.



I too will be converting to CNC and will be watching your thread closely -

I may have to pick your brains too - I have already been chatting to Clive S loads - he's been great and has loads of patients with my noob questions.


keep up the good work

[picclock]

Slight project deviation today. Glass scales.
Are pretty much the best choice for these classes of small machines. They have excellent repeatability and do not drift with temperature or time. The snag is that you must order precut sizes, which is a problem because the delivery times from china are long. A simpler approach is to order the scales as long as possible and then shorten for your application. This also has the advantage that if one gets broken its easy to make up a spare. I have done this many times and evolved a good method for doing it..

Scale construction :-
The scales are enclosed in an aluminium extrusion with caps attached at either end by 4 crosshead screws, around PZ0 size. These can be removed to access the interior. The long strips of rubber/neoprene are dust seals. They are handed, the lower side having protruding lines of material to help the seal maintain the correct shape. The interior read head straddles the glass. One side is spring loaded to ride on the top and side of the glass scale with small bearings, the other side is automatically spaced on the opposite side, think inverted U. The bottom of the extrusion is asymmetrical and the scale itself is butted to it, and held in place by silicon rubber and nylon spacers placed at intervals. The end of the glass scale finishes approx 1 cm or so (min 4mm max 12mm) before the end of the extrusion. This is to give clearance for the end cap and is not critical.

To shorten you will need a tap and drill for the end screws, a BS1 centre drill and some general workshop tools, clamps battery drill, vice etc. . The end screws are generally m2 or m2.5. I will generally fit M2.5 as they feel more secure to me (we all have our foibles :-)). Firstly remove the end cap and read head. Place the read head and cable somewhere clean for later. Slide out the seals, or just remove the other end and pull them out of the way. Next cut the extrusion/glass to the travel length+80mm(read head)+ min 20mm overtravel. I have found the best way is to use my 6x4 bandsaw with the open end of the extrusion facing away from the motor so that the open edge gets cut last. The lower face of the extrusion should be the one that supports the glass. See pictures. Next take a pair of 1/2” pliers and break the end 4-6mm off of the glass. Its important to bend the glass away from its supporting edge or it will break longways. This glass is really soft and breaks very easily, you can even just crush the end. Remove the debris and place the extrusion in a mill to square up the end.

Next drill and tap the threads. Put a piece of square bar stock onto a bench vice and clamp the extrusion to it gap upwards, leaving 1 cm or so of protruding stock. Once secure the end piece can be fitted and clamped in position. Use a battery drill with a BS1 centre drill to a depth of a 1-2mm or so. Remove the end and proceed to carefully drill and tap as normal. Its only necessary to have a short length of thread, determined by the length of the screws protruding from the end cap. Small taps and drills are easy to break, so be gentle. Use lubrication on the tap, molybdenum grease or similar. Unclamp the extrusion and clean away all debris (I use an airline for best results). Refit the rubber seals making sure they are fully abutted to the other end cap and cut flush to the extrusion end using a razor blade. Refit the read head, and end cap. Job done.

Whilst we’re on scales a quick mention of wiring. There seem to be two standards of connections for the 9 pin D connectors. Seemingly designed to stop you using one read head with a different controller. I have only found two but there may be others. I have successfully altered the interior connections on one of my readouts to accept both types and they are electrically compatable. All my stuff is 5V so I don’t know how this would work for 3V stuff. See drawing.

More stuff soon

Best regards

picclock

[magicniner]

It might be worth copying all that useful info to a new thread with an appropriate title in the appropriate section, that way anyone surfing the forum or searching for that kind of info in the future stands a greater chance of finding it. ;-)

Nick

[picclock]

Continuing with the saddle Y axis and some electrical work.
Limit switches, soft limits and home (zero return, grid 0, origin, etc.) position. These systems are designed to prevent the machine trying to go to areas outside the set envelope. Hard limits are set with switches that are triggered when the axis exceeds its normal working range and are used to set home coordinates. Soft limits are often set to confine the work area to a subset of the total area. With stepper motors if the mechanical limit is exceeded all that happens is the motor stalls with few other consequences. If there is a reduction drive it will be subjected to the maximum force the motor can supply. Forunately no smoke or fires occur. However, if this occurs it means that things have gone seriously awry, and the work may or may not be recoverable. There is little to be done to prevent the Z axis hitting the bed during an error as the tooling used is of indeterminate length, so no mechanical switch can be used. When a CNC controller starts, usually the first thing it does is to go to the axis home and detect the position with the limit switches. From that point on all coordinates are calculated relative to these positions.

As this conversion will use glass scales I have used the Y scale bracket to trigger the microswitches. The rear of the saddle has the X limit microswitches, and the left of the saddle the Y. I will upgrade these at a later date to improve the homing accuracy using a latching optical disc sensor I read about here :-
https://www.cnczone.com/forums/uncategorised-metalworking-machines/11623-newbie-home-switch-limit-works.html. The switches will be wired normally closed to ensure that any break will stop the machine.

Sketch with dimensions is really just my crib sheet so take it with a pinch of salt. Also check dimensions of your machine before assuming my drawing is correct. Initial test install of the ballscrew showed an issue with the hole in the front base of the machine. This was approx 1mm too small for the thrust bearing I was going to use. This was opened out using a 33mm hole saw. The position of hole saws is tricky to maintain without a pilot hole to guide. Positioning was achieved by use of a metal plate turned to the outside diameter of the saw cup. This was a simple and successful way of enlarging and centering the hole. The table and saddle were then marked out using a scriber and engineers blue, then drilled and tapped. As shown on one of the pictures its easy to use a metal piece with a right angle cut out to align the drill/tap perpendicularly. The base was similarly treated. To get good right angle holes and threads tapped into the base the drill needs to be held horizontally. This can be simply achieved with a few books or magazines under the drill body, piled to the correct height. I have used std V5 microswitches, the Y ones are mounted on a carrier plates for simple adjustment. Last pictures show the current state of play.
Best Regards
picclock

[routercnc]

Nice work. Following with interest !

[picclock]

As requested my take and decision making on :-
Controllers,
I have chosen to go with a standalone option, the DDCSV 2V1. This is a 4 axis controller integrated into a single unit which runs Gcode. It incorporates a colour screen and keypad, providing the signals for the stepper drivers, and inputs for the limit/homing switches. The advantage of this approach is cost, simplicity, reliability and mechanical robustness, as its all solid state. The gcode is written to a USB stick and plugged into the controller. There are no software licence fees or internet connection needed for licence verification, no spinning hard disks to fail or corrupt, no fans to fail, no external interface compatability issues . Additionally the performance of the box is outstanding with stepper rates on all axis up to 250KHz and pendant support thrown in. In size terms the box is about the same size as two cd rom drives ~ 160x100x50mm. It requires an 18 - 32V supply which can be derived from a dedicated supply or the stepper psu. Power requirements are very modest at up to 0.5A max (12 watts or so). Inputs are opto isolated, and it incorporates a VSO for motor speed control. I have no financial or other interests in this company, just a user.

Motors and ballscrews
How much force do I need ? How quickly do I want to go ?.
Mill cutting forces* for metal under maximum load measured at ~ 50-60kgs, with smaller cutters around ~20kgs or less.
How fast can the motor turn while producing this force?. Stepper motor specifications generally only quote the holding torque - useful only if the motor is not turning. Getting a torque curve for the motor you are going to use is paramount - if you cant get one don’t buy it !!. Hybrid stepper, bipolar 4 wire versions are generally best. A torque curve for the popular longs motor 23HS9442 indicated I could get 1Nm (140 ozin) at 450 rpm.

Using a 5mm(0.2”) pitch ballscrew with this calculator :-
http://www.cncroutersource.com/linear-force-from-torque-calculator.html
indicates a useable force of 109Kg whilst cutting at a speed of 2.3M/min (90 inches/min), more than enough for my requirements. Smaller pitch = less speed more force, larger pitch more speed less force.
The speed of a stepper motor is limited by the driver voltage and its inductance. The inductance resists the change of current which produces the magnetic force. But its the magnetic force which produces the torque. So low inductance = fast motor, low torque. High inductance = slow motor, high torque. The maximum voltage for a motor is reckoned to be 32*sqrt(inductance in mH)**. For my 3.8mH choice this gives 63V. DM860 drivers meet this and seem to perform well without fan cooling.

https://www.hobbytronics.co.za/Content/external/548/StepperCalc_ALENZ_rev1.xls
https://www.allaboutcircuits.com/tools/stepper-motor-calculator/Ballscrew%20and%20motor%20info/

Tests on my Y axis shown below. 1st picture shows controller, driver, motor and steel cable attached to scales. Bathroom scales indicate 130kgs @ 0 (having gone round once) and a further 25kgs before they hit the stop at 155kgs. Calculations indicate that with 4.2A max force should be over 200kgs.

Best Regards

picclock

* the cutting forces I estimated are simply using the same bathroom scale method and remembering how much force I had to apply to the axis wheel. At the high load figure with a big cutter, deep cut in steel the machine was overloaded - so just my estimate, marginally better than no idea :-).

**Although this formula is widely quoted I cannot find any derivation of it. IMHO the voltage should be limited only by the breakdown voltage of the windings.

[Hulksmash]

Thanks for the great post.

I think choosing your operating system / hardware is so difficult. Everyone has a different opinion as to what their preference is. For a complete noob like myself trying to get my head round the pros and cons is so hard.

I was only recently chatting to one of the forum members re Linuxcnc, and 4 and 8 wire motor wiring.. series connections vs parallel.

There is CNC operator near me who swears on 4 wire stepper motors, however he really only works with servo's. So it's good to see the reasoning behind using a 4 wire motor and who supplies them.

Thanks again for the detailed information, I look forward to your future posts.

[Clive S]

There is CNC operator near me who swears on 4 wire stepper motors, however he really only works with servo's. So it's good to see the reasoning behind using a 4 wire motor and who supplies them.


Well you can make 4 wire steppers out of 8 wire ones but not make 8 wire out of 4 wire ones:devilish:

[takeda]

Lovely, looks absolutely lovely!

In some months I am starting the conversion of my VM32B which is branded as Weiss (R8, 4K RPM)

It's blue in color. Cost 2K (euro) not including VAT and I waited one damn year to get it, direct import from weiss.

I will use nema34 closed loop servos (12Nm for Z, 12Nm for Y and 4.5 to 8Nm for X although not sure yet), 2 external gus struts for Z, SFU2005 for Z and SFU1605 for XY. I am trying to find the SFU2005 with a long nut at reasonable price at this moment.

Except for Nema34 closed loop beeing cheep these days (150$ each including the driver) they don't need a DC supply, a toroidal transformer for each driver is enough. I had really bad experience with open loop steppers in the past (1 failure in 1 year is enough to lose trust).

For supports I will go with a bit longer ballscrews and FK12/FK15 blocks mounted on aluminum plates. Only X will be supported on opposing side. FK12/FK15 have proper angular contact bearing in place. Most chinesium BS supports carry deep groove bearings. Especially for Z is super important.


Also I am extending the Z height by 65mm (the extension length flactuates like bitcoin price in my mind) by adding a spacer to the bottom of the Z column. I have my 3D taster in place and the extra Z seems really necessary.

Apart from that I'm looking in replacing the spindle motor with something more powerful but this for a later time. More power is super required for low RPMs. I will use the machine manually as well. But also I need a FAULT signal wired to my control electronics. I don't really trust this chinese speed controller.

TTS/R8 tooling already in stock :)


I am getting a 12Nm stepper for Y because when the X axis is at edge position, Y gets stiff, ofcourse I will avoid using it at such positions if possible to avoid wear but... Oversized stepper whouldn't hurt.

Just can't wait to start my conversion. Unfortunately so busy that it'll have to wait.
If I am hijacking this thread sorry, I am too excited to share the info!

[takeda]

Unfortunately I believe the owner of Toolco lost the contract with Weiss Machines - The owner of Weiss said he went in for an operation and not been heard from since :( I hope he is OK.



That's terrible news.

Their website is down some time now.

[takeda]

And some comments,

Thrust bearing won't work well for a bs application, as stated, no radial support.
I see you use a couple per axis, maybe you can find a drop-in replacement angular contact bearing instead.

For me it's important to have speedy tool changes, so I went with R8 and TTS tooling.

I was considering DDSCV in the past, but having used PlanetCNC software I can't switch to anything else. Everything else just wastes my time (haven't used DDSCV but beeing chinese.....)

But most important, is DDSCV supporting a tool table with tool offsets? Even if it does, editing a tool table in such small screen using arrow buttons would drive me nuts.

Maybe it's nice and will work OK but, I don't know.

About the stand, you will propably need steady feet or some retractable casters (https://www.ebay.co.uk/itm/Retractable-Leveling-Machine-Casters-Heavy-Duty-Machine-Retractable-Casters/293016320576?hash=item4439223a40:g:7DYAAOSwd9dck1S Z&frcectupt=true).

When you'll start accelerating with the motors the whole thing will start dancing with these wheels (except you use them temporarily only for moving arround).

Also, how will you cycle coolant, is it the mid compartment designed for a tank ?

Nice work with the ballscrew machining :) I had a different experience than you, machining the nut was OK, machining a ballscrew was a pain. So much interrupted cut in the beginning and broke many carbide inserts. Finally I made a jig to grind the ballscrew a bit before machining. Also machining long ballscrews is a pain, needs support on the other side of the lathe spindle (we have the same lathe or close).

For the limit switches, is there any reason you prefer mechanical switces instead of inductive proximity sensors? They will be flooded with coolant and might get sticky due to chips (chips can prevent them from switching). Also they are bulky.


As for the strength of Y axis motor, try to mount the table on the carriage, move the table to one of the positions and try to accelerate the Y axis back and forth. I found that in my case you need multiple times the force than in mid position.

Last, you could consider filling the base epoxy granite. I am most definetely doing so.

[dazp1976]

I'm just finishing a 25 size mill.
Thrust bearings? ooooh no. Never again.

Did thrusts on my X2 and they did not sit right at all on the screw ends. Wobble and excessive wear all over the place.

Looked at thrusts on the 25 to see and then tried AC ones instead and the difference is massive.
Had to machine some mounts myself to suit them but worth the effort in the end.

Thrust bearings NOT recommended imo.