CNC metal machinists (for Stax amp cases) unite?

[chinsettawong]

Very nice!  Congrats, @jamesmking!

 

[Kerry]

Congrats!

[jamesmking]

11 hours ago, Kerry said:

Congrats!

Thank you Kerry, however, yesterday I then tried an uncoated 2 flute 2mm slotting endmill... the results was not good at all.... very poor chip evacuation and lots of chips welding to each other. Almost no chips get thrown clear of the cut. I'm not sure if the end mill is garbage or if something else is wrong.... I was looking online for some kind of guide that had photos of poor results and what to do about them... but there seems to be very little information other than feeds and speed calculators which seem to suggest parameters that I cant get close to.

I tried to extrapolate settings from the great results I got from the single flute 2mm end mill to a single flute 3.175mm end mill but to no avail. The 3.175 end mill could not run at anywhere near the rpm or depth of cut of the 2mm end mill without welding, vibration and large burns on the edges of the cut.... Perhaps the 3.175mm end mill is just crap (its not the same brand as the 2mm)... 

wow cnc aluminium is frustrating...

Edited August 26, 2022 by jamesmking

[audiostar]

James, make sure you use the best possible (carbide coated) cutters and also keep in mind 1.5kw is not that much when you calculate speeds. Make sure cooling the bits and material while cutting is adequate as well.

Looks like it is getting pretty hotm here, cutter is not good enough or speeds are too high.

Edited August 28, 2022 by audiostar

[audiostar]

Which aliminium alloy is this exactly you are working with?

6061-T6 being the standard for CNC machining and excellent all purpose alloy.

[jamesmking]

1 hour ago, audiostar said:

Which aliminium alloy is this exactly you are working with?

6061-T6 being the standard for CNC machining and excellent all purpose alloy.

i have no idea of the type of aluminium, its a disapante case I purchased some time ago from modshop. 

The fact I got such good results with a 2mm end mill, especially on chip removal and surface finish on it makes me suspect the 3mm end mill is just crap (the 3mm comes from a different supplier and is a different brand to the 2mm). So I have ordered some 3mm end mills of the same brand as the 2mm.

The spindle does not seem to be slowing down, vibration was low, cutting noise was low and I did not use any coolant with the 2mm...

as far as I can see if I can go 300mm/min 1.6mm depth of cut on a 2mm end mill with zero issues, low cutting noise and no lubrication,
(I managed to break it at 500mm/min 1.6mm DOC but I was deliberately seeing how fast I could go), I should be able to do similar with a 3mm end mill of similar geometry (both single flute and both DLC coated)... but the best result I got with the 3mm was 0.2mm DOC, poor top edge and much reduced rpm compared to the 2mm. Any time I tried to increase the rpm chip welding became worse and chip evacuation did not improve, any DOC above 0.2mm I got lots of vibration and nasty noises.... looking carefully at the cutting the 3mm was not pushing material up the flute instead material was wrapping itself around the outside of the cutter and then welding itself to the top edges of the slot. I did not seem to be cutting so much as ploughing. Slower feeds resulted in the chips becoming like power and welding themselves all over the cut like fur... I tried cutting oil with the 3mm and this did not stop the welding.

Feeling the edge of the flutes the 2mm feels super sharp all the way up whereas the flute the 3mm only feels sharp right at the bottom tip of the flute and the rest of the flute all the way up does not feel more than moderately sharp. I suspect this is why I cant get any sort of depth of cut. Even plunge cuts at the beginning of the slotting sounded much nicer on the 2mm than the 3mm end mill. (plunges straight down at 25mm/min)

 

 

Edited August 28, 2022 by jamesmking

[audiostar]

Hi James,

from your detailed description, it looks like the cutter being the culprit. It would be the first to swap out.

[jamesmking]

On 8/29/2022 at 2:38 PM, audiostar said:

Hi James,

from your detailed description, it looks like the cutter being the culprit. It would be the first to swap out.

Things are looking up again... I just took delivery of some 6mm single flute DLC coated end mills, selling says they are designed to cut aluminium without liquid lubrication. Initial testing looks very good, absolutely throws chips out of the cut into near earth orbit when doing a full width slot. Zero chip welding. Surface finish as good if not slightly better than the 2mm end mill (So far got the speeds and feeds for the 6mm up to 400mm/min feed rate, 1.5mm DOC and 11K rpm). Cutting is very quiet and almost silent at 1mm DOC... The vfd and control box fans are louder than the cutting noises. I am sure there is much more to be got out of this end mill, but I'm progressing conservatively because these 6mm end mills  are not cheap and I don't want to push it to destruction. 

The biggest take away for me is that buying cheap end mills from the manufacturer of the cncs website is not the best move and that end mill quality absolutely radically effects what you can do. I also think i'm going to invest in a oil/air mist system...

(the hideous grey mess is the 3.175mm POC (piece of crap) end mill WITH cutting fluid dripped onto the cut... .) 

middle slot is 6mm end mill 1.5DOC 3mm total depth 11Krpm 400mm/min no cutting fluid no finishing pass no deburring...

front slot same settings as middle slot but 1mm DOC and 10Krpm.

 

 

Edited September 1, 2022 by jamesmking

[audiostar]

Yes, get the absolute best possible cutters you can get, there is simply no way around if you wang excellent results. 

[audiostar]

James, now you need to add re-anodization to the process 🙂

[jamesmking]

First attempt at a complete front panel. Single flute 6mm end mill and 0.1mm V bit used. Depth of cut 0.02mm on the engraving. I also used tape and CA glue as the hold down method for the first time - I really like it. Much less vibration and far quieter when cutting in the middle of the work piece compared to just clamping the corners. As usual no deburring or finishing pass. The surface finish on the volume pot pocket is silky smooth - the photo does not do it justice. Very happy with the results. Not so happy that I have zero artistic talent for cool looking front panel designs....

 

 

 

Another go at a Stax socket with the better quality end mills. Much better surface finish than my previous attempt cheapo mills and I'm using faster feeds and speeds this time. Just shows that a crap end mill gives inferior results even on soft plastics. There are no burs around the edges of the cuts and the surface is smooth.

Edited September 6, 2022 by jamesmking

[n_maher]

Looks awesome to me.  I mean you could add a logo of some kind if you come up with some artwork?  I’m not 100% clear on where the required blue LED goes.

[jamesmking]

5 minutes ago, n_maher said:

Looks awesome to me.  I mean you could add a logo of some kind if you come up with some artwork?  I’m not 100% clear on where the required blue LED goes.

REQUIRED blue led is the blue ring around the power button on the power supply....

 

Edited September 5, 2022 by jamesmking

[JoaMat]

That panel looks good. Is it for an original DIY T2 or a Modified DIY T2?

 Me to have done some milling.

 

0.38 mm end mill, 18000 rpm and 300 mm/min. Two sides, 121 holes (5 diameters) and cutout took a little less than 2 hours.

[jamesmking]

10 minutes ago, JoaMat said:

That panel looks good. Is it for an original DIY T2 or a Modified DIY T2?

 

mostly modern T2 or possibly fully modern if the tta/ttc substitution for the 79 and 216 proves to be reliable. I don't think I will ever build an original.

Edited September 5, 2022 by jamesmking

[JoaMat]

tta/ttc and kca1220/ksc2690 are, in my opinion, reliable. I've used them for years as substitute for j79/k216.

 Next step is the 37 vias.

 Board, 0.5mm wire and tools. Anyone prepared to pay a fortune for handmade printed circuit board?

 

[Kerry]

Congrats @jamesmking and @JoaMat!

Nice work!

[jamesmking]

T2 psu back panel, design and cut. I wanted to have the pin outs of the sockets for the umbilical chords from the psu to amp to be on the psu back for future reference..

I decided to go for 10degree 0.1mm tip engraving bit for the lettering, 0.025mm depth of cut rather than laser. Screw holes done with 3.175 end mill and main cut outs 6mm end mill. Now I have a good collection of quality bits and little experience now. CNC machining is starting to become FUN 🙂. Its so much nicer and more satisfying watching a CNC cut than using a hand drill and Dremel... Tool changing is a bit of a hassle and then I have to use a Z block and Z probe command get the new Z height reference for the tool sickout. So I try to minimise the number of different tools used.

Tape hold-down and CA glue works perfectly. I'm very happy with the surface finish, perhaps the text for the pinouts and lines could be a little narrower, but its certainly not a disaster. Cut still currently in progress:

 

The nice thing about symmetrical panels is that you can always adjust your settings and have another go with the other side...

I decided to decrease the line width by half, nock 20% off the size of the socket decals, decrease the socket decal font size, change the style of the lines and reduce the engraving depth from 0.025 to 0.02mm... 

I think the new version should look a little less blocky and a bit more refined... The 0.02mm engraving depth is slightly too little, other than that I think its an improvement.

Edited October 9, 2022 by jamesmking

[audiostar]

Phew... couple of days already finishing up the final wiring, setup and calibration of my CNC router. Ended up with a 2.2kW 400Hz water cooled spindle on a 4 axis machine that is connected to a Mach3 system with two LPT cards fully automating spindle start/stop, water pump start/stop for cooling, air cooling for the cutter, etc. Very happy for now how everything goes together. Need to fabricate a heavy duty cabinet for it holding all the electrical stuff and computer below the machine with some drawers as well for the cutting tools, fixtures, vacuum pump and water tank as well. Going to keep me busy for some time.

[jamesmking]

I thought I might post some of the modifications I have been doing to my Flox Alien vasto.

I got fed up with the not so rigid spoil board and so filled in the bottom with 20mmx40mm aluminium extrusion... and replaced the thin t slot and mdf spoilboard with a 30mm thick mdf slab. The result is far more rigit and will take my own weight (60kg) with ease and no detectible flexing.

The plan for the spil board is to have m6 threaded inserts on the underside of the board to allow me to screw in work holding clamps/brackets. A 20mm by 20mm grid engraved and a set of 3.175mm holes around the edge so I can insert small pegs to make lining up work with X and Y axis easier.

 

I also got fed up with the cheapo controller. It has no rs485 support so it could not control my spindle vfd. I could not change spindle speed or direction in software. In fact the original unmodified machine could only switch on and off the cheap dc motor it was supplied with and had a potentiometer for manual speed control with no indication or rpms. It has no wifi support and no user interface, is also only 3 axis and completely undocumented. I opted to rip out the controller and put in a rootcnc controller (https://www.rootcnc.com/root-controller-iso-about/) running fluidnc (http://wiki.fluidnc.com/) control software. It took a while to configure fluidnc and re do all the control box wiring.

I replaced the 2 pin spindle power socket on the back with a 4 pin to carry rs485 from the controller to the spindle vfd. I also removed the pause and resume buttons and replaced them with sockets for a tool height probe and a 3d probe.

While I was at it I replaced the now redundant potentiometer that used to control the voltage on the original 48V spindle output with a fixed voltage divider of 12V so I could run a good quality noctura cooling fan and power diode lasers directly. I also got a cheap minbot labeller and went on a labelling spree.

The result is I can store multiple profiles (one for the laser and one for the vfd spindle) in the controller flash, In fact rootcnc and fluidnc is versatile enough and has enough I/O I could run a vfd and laser at the same time if I could find a sensible way to mount both. I can control the spindle speed and direction in software and by using standard S and G3,4 and 5 gcodes, control the machine via the built in web server, usb or wifi and have much less fan noise, power 80W diode lasers directly from the control box and have 2 more axis. The spindle control and a spare axis makes an automatic tool changer such as the rapid change atc https://rapidchangeatc.com/ a viable option.

I now have enough expandability that I could add control for a rotary axis, run a pendant or even consider automated tool changes... 

I have been experimenting with automated tool height measurement and have written some gsender macros to set the tool length offset using the tool height setter I recently purchased. I still have to manually remove and insert and screw down the tools but the aim is that I will then just press a button and the machine will measure the new tool height and Z axis work coordinate so the tip of the tool is at the same physical Z height as the previous tool. I chose the tool height sensor because it is also compatible with the rapid change atc.. Gsender has poor documentation for its macros so I got the source code and reverse engineered the macro language parser. I love open source software, if you can't find the information online you can always find it in the source code. 

The Z axis mount is too narrow to accept a mounting bracket for an 80mm diameter spindle. So at the moment I am stuck with 65mm spindles which limits me to only er11 collets and a maximum end mill diameter of 6mm. So when I have finished rebuilding the machine I will make a new zaxis mounting plate for 80mm spindle brackets and am looking at buying a 2.2KW er20 spindle with ceramic bearings and better dust sealing. Further in the future I would like to get the rapid change atc which will probably necessitate extending the y or x axis extend the x axis from 420mm useable to around 550mm or so. I'm also thinking about changing to closed loop steppers.

happy cnc'ing

James

Edited February 8 by jamesmking

[jamesmking]

Well, I decided I needed to upgrade my CNC machine (replace almost everything from the original machine I purchased). having already upgraded the controller, spoil board, spoil board mounting and spindle it's time to look at the axes ....

First  I got myself an 80mm diameter er20 ceramic bearing 2.2KW water cooled spindle with labyrinth seals to replace the cheap 65mm 1.5Kw er 11.

To fit the 80mm spindle I had to make a new mounting plate. This included experimenting chamfering and threading. I settled on 0.3mm chamfers at 45 degrees and experimented with thread milling. The result I think is ok: (full discloser I purchased tooling plate so the top and bottom surfaces were already machined flat, but all the other surfaces got machined.

 

I have ordered 4 nema 23 closed loop steppers (3NM - about twice the holding torque of the original open loop steppers the machine came with), which will entail replacing most of the wiring loom, removing the existing open loop motor drivers, removing the motors, some software config of the new closed loop drivers and some changes to the controller wiring. The original divers have no error output pin so the controller has no idea if a stepper has an issue. The new drivers include error output pins so they will need to be connected to the controller. (stock jpeg from the ebay listing, I'm still waiting for delivery)

I have also replaced the standard 'deep' groove (6001 series) bearings inside the bk12 bearing housings with angular contact bearings (7001 series) as per this you tube video:

 

I also noticed that all the bearing blocks were attached to the machine with m5 bolts despite the blocks being designed for m6, So I plan to replace those bolts and find a way to make new threads in the machine housings.

The nema 23 motors are also designed for m5 bolts but the machine uses m4 (on avery axis except Z), again I don't know why fox alien decided to use one size smaller bolts on everything that drives the axes so I will have to machine new back end plates/motor mounts too.

After these upgrades and fixes the only original parts of the machine left will be the frame, axe and the metal case the electronics came in.

It's so nice to have a cnc machine so I can make parts to upgrade my cnc machine...

 

Edited May 7 by jamesmking

[JoaMat]

Impressive!

What kind of tool do you use for threading?

[jamesmking]

junk z axis.

holy crap, I just disassembled the Z axis. The entire spindle weight is held by a 12mm ball screw. On each end of the ball screw is just one f688z bearing - (8mm ID, 17mm OD with a flange 18mm diameter). These tiny bearings have no preload and not designed for axial loads and have axial play. Looks like fox alien are relying on the weight of the spindle alone to push down on the lower bearing and that's the only thing stopping the spindle from dropping. Piece of absolute SHIT! The bearing housing has no recesses, no preload mechanism, no adjustment mechanism, nothing it's just unbelievable. Looks like I'm going to need an entire new Z axis too...

Does anyone know if there is a angular contact version of the f688z??

sigh

 

 

44 minutes ago, JoaMat said:

Impressive!

What kind of tool do you use for threading?

single tooth carbide partial profile (so it will do multiple thread pitches and the tool pressure is a lot less than other threading tools.). It's more versatile and cost effective than buying complete profile thread end mills. With 5 thread mills I have every metric pitch (both fine and course) covered from m3 all the way to m10.

https://www.shop-apt.co.uk/single-tooth-thread-mills-dlc-for-aluminium-internal-60/internal-single-tooth-carbide-thread-mill-partial-profile-05-10mm-pitch-dlc-coated-for-aluminium-39mm-head-diameter.html

The thread mills don't come with all the necessary parameters (e.g. tooth height, distance from the centre of the tooth to the bottom of the end mill etc.) for the software I use so I had to take multiple measurements with a 0.001mm micrometer and average the results and guestimate other parameters. 

 

The threads are m6 I ran 7000rpm, 150mm/min feed, 70mm/min plunge 0.0054mm per tooth chip load, bottom to top cutting a right handed thread. This resulted in the threads being cut in two passes and the chips falling into the bottom of the hole (which I made 1.5mm deeper than needed to hold the chips). To make the thread go all the way through the 10mm mounting plate I found I had to start about 1mm below the bottom of the material, so I painters tapped and superglued a 3mm sacrificial aluminium plate underneath. 

 

Edited May 7 by jamesmking

[audiostar]

By the time the closed loop steppers detect an error it is already too late as it already happened. Always better to have the machine itself running more accurate and perfect cabling and controllers. 

[jamesmking]

1 hour ago, audiostar said:

By the time the closed loop steppers detect an error it is already too late as it already happened. Always better to have the machine itself running more accurate and perfect cabling and controllers. 

true, which is why I have been upgrading the axis bearings etc. I agree prevention is better than detection, which is why I am going for steppers with double torque than the originals, plus the software that comes with the closed loop system provides graphs, including positional error which will allow me to fine tune the pid parameters and see what is actually going on and give me more options to fine tune and correct. I am also learning a lot - which is fun too. I loved metalwork and woodwork at school but when the route of computer science, but have always enjoyed making things - even if it's just putting together flat pack furniture for friends or soldering pcbs.

Edited May 8 by jamesmking