Game on!

[Voicecoil]

After some lurking, quite a few questions, a lot of sums and some FEA I've just started building my first machine. It's not as big as I really need (I want to cut out large loudspeaker cabinets amongst other things), but I'm having a lot more difficulty finding larger premises than I thought I would. So having kind of got the bug and wanting to get on and do something I thought I'd make a small (but capable) one anyway, the experience will do me good, it will be useful for machining up metal parts and a lot of it will be transferable to the big machine design when I get the space.
Sorry I've not shared any of the initial drawings, but I don't really do 3D CAD and without being able to turn layers on and off and have the thing in your head, the 2D drawings are a bit of a nightmare to interpret. The design ethos is:
1) To try to keep all individual deflections of the gantry extrusion/side beams/bed to less than 1um for a 50N cutting force - this has been determined by calculation (thanks @routercnc for the spreadsheet) + FEA
2) Minimise twisting moments by keeping all assemblies as compact as possible - no massive spindle overhang please!
3) Keep the number of bolted joints to a minimum.
4) Try to get a good price performance compromise
It's being made out of aluminium extrusions/plate simply because I'm able to machine that better and more accurately - and you can anodise it to a nice finish :congratulatory:

Not having any CNC machining capability until it's made, I'm having to "bootstrap" the process, apart from one or two pieces that I need to get laser cut, it will all be done using the trusty Elektra Beckum chop saw, a drill press (Axminster ED16B2) and my little lathe. First job was to fettle up the saw and make a decent clamp for it:
25467
With a new blade fitted and a few minutes true-ing it up it goes through 15mm plate like a knife through butter. Squareness is about 0.25mm across a 125mm cut, and with a wee drop of WD40 to help the cutting, the cut ends are smooth as a b******s bottom.
First parts to make were the carriage plates for the Y axis (please note I denote X & Y from the point of view looking down at the front of the machine i.e. the X-Axis runs L to R along the gantry):
here's marking them out using the trusty optical centre punch:
25468

[Voicecoil]

I've found time to do a bit more on this in the last few days, firstly got the gantry extrusion fully end drilled and tapped..... it just fitted under the pillar drill with 15mm to spare :tears_of_joy:
25480
talking of tapping I soon realised that there were going to be A LOT of holes to tap and if I wanted to save a lot of time (and avoid w*****s cramp!) machine tapping was the way to go. Tapping machines looked a tad expensive for the moment, thought about changing the motor of the pillar drill to give a quick reverse function, but some bits were on a couple of weeks lead time.... then I tried one of the Europa tool interrupted-thread spiral point taps:
25481
These seem to require a lot less tapping force than anything I've tried before and seemed hence offer up the option of tapping with a cordless drill (which convieniently has a torque limter to stop breaking taps, and a reverse to extract them). In order to keep things straight I added a circular level to the back of an old drill, using the pillar drill and a length of tool steel rod to get it properly level whilst the glue dried.
25482
With this and a known flat surface I seem to be able to tap holes pretty straight and all for the outlay of a £1.80 level off the Bay of fleas; and you can do it in large pieces that wouldn't fit under a normal pillar drill.

[AndyUK]

In order to keep things straight I added a circular level to the back of an old drill

Nice bit of customisation! I've taken to using a guide for my taps to pass through; find a block of something (wood, aluminium) that is about as thick as your tap is long, minus 1cm and the length of the square drive section. Using the pillar drill, put a guide hole through the block - diameter to be as close fitting as to your tap as possible. Now you can place this block up against the surface you're going to tap, passing the tap through for the first bit of the hole - hey presto, nice and straight tapping every time.

[Voicecoil]

I've got most of the pieces machined up now and waiting for the dozy anodisers to do their stuff on some of the beams - whilst it's obviously a home built machine I wanted it too look vaguely OK and I don't really have the space for some 3ft long anodising batths at the mo. Today I finished off making the spindle clamp, I'd cut the outside profile and done the holes a few weeks back, but I've been waiting for a friend to have some free time so I could use his mighty steam-powered lathe to bore the hole. Here it is after putting a 30mm drill through to start things off: And yes, that's a 300mm chuck and it's a heavy lump!
25634
Then it was something like 3 hours gentle boring out, with a few minor hiccups and several Km of swarf - here's the end result:
25635
The internal finish wasn't quite as good as I'd hoped as he had a rather limited range of boring bars and with 80mm of tool sticking out from the toolpost there was a certain amount of resonance going on; never mind, no-one will ever see it, and as long as it grips the spindle I'll be happy.
The reason for making the spindle clamp was initially because I couldn't easily source an 85mm part OTS, then it kind of led to some lateral thinking about how to implement the Z axis with good strength and minimum width - all will be revealed soon when I start putting it together :encouragement:

[Voicecoil]

I've been cracking on with stuff the last few days, here's a pic of some of the parts:
25691
From LHS clockwise we have Z axis slide with clamp sitting on top of the bed cross beams, front bed end plates, Z axis top plate, gantry end plates and the Z axis back plate. Just hit a bit of a snag though - the big bit of plate for the top of the bed has been cut well off square by the folk at Smiths - thus even with the 2mm tolerance oversize by the time it's trued up it will be 2mm short causing a few major headaches - I will be bending their ears on Tuesday :mad:

[Voicecoil]

Marking out the bed plate:

25731

I need a bigger surface plate!!!

[AndyGuid]

Marking out the bed plate:

25731

I need a bigger surface plate!!!

As a pen-pusher, I'm guessing that your method is a very accurate way of manually Marking Out Using a Surface Plate and Height Gauge ?

[Voicecoil]

Yes, I bought the height gauge and optical centre punch for this project and I've been suprised at how accurately you can do things - I reckon on smaller pieces the max error isn't much more than +/- 0.1mm and maybe twice that on longer bits: quite a novelty not to have to file out holes to get things to line up! Mind you, when you think about it people were making accurate stuff 60 to 70 years ago before CNC was even thought of, so it had to be possible.

[Voicecoil]

Yesterday and today I've been assembling the various parts of the bed/Y-axis to check fit before doing a bit of anodising.
25739
25740
25741
25742
All went reasonably well apart from a couple of screw heads (that need to be moved) fouling stepper motor bodies and a small CAD error on the carriage side plates that I'd transferred really precisely into metal - time to get the file out! :disillusionment:

[mekanik]

Looking Good:thumsup::thumsup:

[Voicecoil]

Having found it very difficult (or very expensive) to get anything done on a reasonable timescale by the local anodisers, in desperation I thought I'd have a go myself. £50 worth of chems and some plant troughs later I managed to get the gantry box done this evening, took a bit of fiddling with the process but not too shabby in the end.
25748

[Gav74]

looking good

[Voicecoil]

With all the beams now anodised I thought it was time to indulge in a bit of damping, as handling the box sections in particular showed that they were ringing a lot at mid-audio frequencies (which match up nicely with the stimulation given by typical spindle speed/cutter flutes). I thought I'd try my old fave Tecsound, but wasn't overly confident of it working as I've never used it on anything thicker than 2mm thick metal sheet before, and the box sections I've used for the side beams and gantry are 7 & 10mm thick wall. However in reality it seems to work pretty well, hanging an untreated section up on a piece of wire and giving a gentle tap with a hammer gives something not unlike a tubular bell, whereas a damped section is just a bit of a thunk. Here's some sound files from the tapping to illustrate the difference
25765
25766
Looking at the decay time in an audio editor, the damping has been increased by something like 6X - a useful improvement I think - now to do some of the other parts!

[Voicecoil]

Thought I'd better check fit on the Z-axis before assembling the gantry as the X carriage is integral to it: all fitted suprisingly well once I'd clocked that the Z slide extrusion was supplied under-width and was causing the assembly to sieze at the end of travel
25817
25818

[Voicecoil]

So, the last few days I've found time to build up the chassis/Y axis. Getting the side rail assemblies properly aligned was a bit of a faff, but it's done now!!
25830
25831
25832

[Lee Roberts]

Looking good, nice to see something diff to the norm :thumsup:

[Voicecoil]

Looking good, nice to see something diff to the norm :thumsup:

Cheers Lee - I thought it was worth going back to basic principles and trying a few slightly different approaches to see if was possible to make an aluminium machine that was a bit more rigid than normal. I've learned a lot as I've gone along and come up with some improved ideas which will be incorporated into this build once I've got it cutting metal and can make more accurate and complicated parts - those ideas will also go into the big machine which is the long term goal.

[kingcreaky]

as lee says its great to see new methods. thanks for posting, keep it coming,

[routercnc]

Somehow I managed to miss this thread and have just caught up. Great work Voicecoil and I heartily approve of the twin Z ballscrews. Glad the stiffness spreadsheet is still helping people out.
Let the over build commence - great machine !

[Voicecoil]

Somehow I managed to miss this thread and have just caught up. Great work Voicecoil and I heartily approve of the twin Z ballscrews. Glad the stiffness spreadsheet is still helping people out.
Let the over build commence - great machine !
The twin ballscrew thing kind of evolved, it came from trying to minimise twisting moments on the gantry - I'll be finding out in about a month's time how well it works I guess! Meanwhile lots of wiring and screwing together to be done :-)

[Voicecoil]

Had to take a break from this for a few weeks due to the pressures of (paying) work, but now I'm back with a vengeance and looking to get it finished ASAP. First I've installed the steppers and wiring into the base:
26239
then it was time to start putting the gantry together:
26240
here's getting the rails levelled:
26241
Just realised I need to get the oiling sorted before I go any further, so there's been some frantic buying of bits on the bay of fleas this afternoon: more pics coming it a few days when they land!

[Voicecoil]

So I've mostly gone with an idea Lee suggested, using 4mm pneumatic connectors/tube to supply lubrication to inaccessible parts:
26376
26377
On the HGH15's I couldn't find any suitable adaptors so had to turn up some barbed brass push on thingies:
26378
26379
Tried the same on the long axis, but then realised they would foul the ballscrew covers:
26380
Tried spinning the ballnuts through 180 degrees but then found everything was seizing up - turned out the ballnuts weren't concentric on the screws - thanks Fred :sour: - so those are going to have to be a manual lube.
On the subject of lubrication, does anyone know where I can buy a sensible quantity of synthetic semi-fluid grease please? Looks like to be the DB's for this sort of setup.
Z-axis ont' gantry:
26381
So onto the wiring... Found that the closed-loop steppers have the connectors on the encoder cables of such a length that they would sit in EXACTLY the wrong place - half way down the drag chain :mad:
26382
....so had to do some inline splicing......
26383
and to finish off added a touch off probe socket into the base :congratulatory:
26384
So onto the wiring

[Wal]

>On the subject of lubrication, does anyone know where I can buy a sensible quantity of synthetic semi-fluid grease please?

My mini-mill has no practical way to get oil to the ball-nuts or the bearings riding on the supported rails, instead I lift the rubber skirts every once in a while and squirt a load of a TF2 Teflon oil onto the shiny bits:

https://www.google.com/search?q=tf2+oil&safe=strict&client=firefox-b-d&tbm=shop&source=lnms&sa=X&ved=0ahUKEwjA3f2VrLLkAhVl5OAKHdX9DVkQ_AUIDigC&biw=2560&bih=1286&dpr=1#spd=17336399006240699812

No doubt this isn't the done thing, but having done it for five or so years, I'll probably keep doing it. There's logic for ya.

Nice build BTW.

Wal.

[Voicecoil]

After a lengthy break due to Bonfire Night commitments I started work on the control box a few weeks back, and it's getting there now, in spite of being much more of a ball-ache than I anticipated* - I've never had to do so many screw terminals up in my life! Anyway here's a pic from a few days ago, everything's in there now apart from the stepper power supply limiter/clamp module which I'm making.26814

[mekanik]

Did you forget something, ventilation ?

[Voicecoil]

Errr, no - I'm a cool dude when it comes to electronics at least :hysterical:. Sorry for the poor pic, due to the direction of the lighting it's not too easy to see, but there's a fan on the bottom LHS, you can just about make out the chrome grille. The exit is in the top RHS, just above and behind the VFD, which I reckon is what's going to make the most heat.

[Muzzer]

Use a simple household thermostat to control a cabinet fan so it comes on at ~30C or so.
https://www.screwfix.com/p/flomasta-wired-thermostat/7593g

Or a process controller
https://www.ebay.co.uk/itm/PID-Digital-Temperature-Controller-REX-C100-with-K-Thermocouple-SSR-Output/323849706141?_trkparms=aid%3D555018%26algo%3DPL.SI M%26ao%3D1%26asc%3D60712%26meid%3D7b6ab0aec9b14dab a608123f5e9b764d%26pid%3D100005%26rk%3D5%26rkt%3D1 2%26mehot%3Dco%26sd%3D163734209554%26itm%3D3238497 06141%26pmt%3D1%26noa%3D0%26pg%3D2047675&_trksid=p2047675.c100005.m1851

CPC are good place to look for fans - cheaper than Farnell (sister company) or RS. Mains or 12V / 24V
https://cpc.farnell.com/w/c/electronic-electrical-components/fans-cooling-devices/axial-fans/prl/results?st=fans&aka_re=1

[Voicecoil]

Finally got the control box wired up late last night. It turned out to be much more of a PITA than anticipated, and wasn't helped by paying work/Christmas/New Year getting in the way. Anyway here it is:
27072
Here's the little clipper/clamp thing I was talking about, seems to work OK on the bench, clips the charging waveform to the HT reservoir cap at 45.5V at high mains/off load conditions and clamps stepper back-EMF at 49V.:Whist making it I realised I could add a cap and make it do soft start as well.
27073

If I find myself doing another machine, I'll likely do things a bit differently, having a small "sort out" box mounted on the machine to minimise the wiring, also I'll put the closed loop stepper drivers on the machine as well - that way the step/direction signals can be easily run on some small multi-twisted pair cable on e.g. M12 connectors.

[ericks]

I just completed a panel. Was going to offer it for sale but decided to use it on my own machine. The way i do the ventilation/cooling is filter on side bottom and fan on side top opposite side. Been doing it this way in control panels for many years...no issues. Separation between "noisy" and low voltage devices and wiring is also very important, that way noise related issues would be greatly eliminated :)
As you can see i used plugs for all the external devices cause this was a plug and play unit for sale. Same as you i used the AXBB-E and although it works good the way they designed it makes wiring a bit complicated....i am currently working on a better solution for that. Another thing i am working on is to make wiring to the sensor plugs easier and faster. Just thought this might give you some ideas for your next one :)

27074

[ericks]

Good job by the way!!

[Voicecoil]

I just completed a panel. Was going to offer it for sale but decided to use it on my own machine. The way i do the ventilation/cooling is filter on side bottom and fan on side top opposite side.
Ahhh, so you're using your fan in sucking mode I guess?? Mine's blowing into the enclosure, but the airflow route must be similar. I always tend to go for blowing into the enclosure in equipment I've designed in the past as it's marginally better dust-wise to have the enclosure slightly pressurised, though with a well sealed box like these control enclosures it won't make any difference at all.


Same as you i used the AXBB-E and although it works good the way they designed it makes wiring a bit complicated....i am currently working on a better solution for that. Another thing i am working on is to make wiring to the sensor plugs easier and faster. Just thought this might give you some ideas for your next one :)

Aye, the enormous number of screw terminals is a PITA :thumbdown: - if I were making even just a few units I'd lay out my own breakout PCB with onboard connectors and interface to something like the UC400ETH via some flat cables.

[ericks]

Yes i basically suck the rising heat out of the box. Busy making a breakout board/terminal board/plc for the axbb-e. Problem is my spare time is limited atm. I found the connections to the stepper drives an issue due to the lack of 5V0 terminals on the AXBB-E. But i still like the unit :)

[Voicecoil]

Good job by the way!!

Cheers! it was quite a job cramming it all into a 380x380x200 box, but having had one kindly donated by Nealieboyee it was worth the effort.

[Voicecoil]

So I've finally got all axes working. Next job was to square up the gantry; first I got it roughly right by eye and some shoving/using the Y motors independently, then fine trimmed it with the eccentric bushings on the carriage plates using a fine DTI and a large square left by my late father. It's true to about 10u now :tickled_pink:
27223

[Neale]

Does the software/motion controller you are using support proper gantry homing/squaring? I'm wondering because you talk about mechanical adjustments but I assume you will still be at the mercy of the homing routine at the start of any cutting session?

[paulus.v]

Does the software/motion controller you are using support proper gantry homing/squaring? I'm wondering because you talk about mechanical adjustments but I assume you will still be at the mercy of the homing routine at the start of any cutting session?

What do you mean by proper gantry squaring?
Both Mach3 and UCCNC have a good working gantry squaring routine. At homing both axes will move together until one axis will reach the home sensor, then that axis will stop and the other will continue until it will also reach its home sensor. After this, both axes will move back off the sensors and the homing/squaring is finished..
I think it is also possible , with a macro, to add an offset distance between the sensors if it is difficult to adjust the sensors distance for the gantry squaring.

[Neale]

That's pretty much what I mean by proper gantry squaring - the software understands the hardware configuration and does the right thing with both ends of the gantry. The main thing that prompted my question was that the previous post described very accurate mechanical squaring but I wasn't sure how that tied into the software-initiated homing which would be used for normal operation.

I've been using Mach3 for a few years now but I wasn't aware that it understood the concept of master-slave axes. I use Mach3 with a CSMIO-IP/M and it's the motion controller firmware that takes over the homing function from Mach3. Unfortunately that firmware does not include the ability to square a gantry during homing - in effect, it just runs both master and slave in synchronism and you have to find another way to square the gantry. In fact (as someone pointed out on this forum a while ago) you can go into the CSMIO config menu and disable slaving. I then use a specific homing macro that homes X and A simultaneously; both ends of the gantry move to their home position but you do need to separately adjust the two homing sensors so that the gantry is square when it is homed. Once homed, you reconfigure slaving and it all works fine from that point on. Assuming that you don't hit e-stop or that anything serious happens, the gantry retains its "squareness" for the rest of the session, even if rehomed the usual way.

However, you have made me think a little more about this, and I suspect that because I separate X and A for homing, I could use the Mach3 homing offset parameter on A to provide the squaring correction more easily than trying to tweak a proximity sensor. I'll look into that.

[Voicecoil]

What do you mean by proper gantry squaring?
Both Mach3 and UCCNC have a good working gantry squaring routine. At homing both axes will move together until one axis will reach the home sensor, then that axis will stop and the other will continue until it will also reach its home sensor. After this, both axes will move back off the sensors and the homing/squaring is finished..
I think it is also possible , with a macro, to add an offset distance between the sensors if it is difficult to adjust the sensors distance for the gantry squaring.

Cheers for that Paulus - you've saved me digging through the UCCNC manual to find out if it does it!

[Neale]

Every time I have to work around a deficiency in my setup I start reading the UCCNC manuals... I reckon that a UC300ETH/UB1 combination would be a drop-in replacement for my IP/M, which is a great piece of hardware let down by the software around it. I spotted that UCCNC does proper gantry squaring, although I haven't bothered to chase down exactly how. Would have to replace the MPG as well as I couldn't live without one of those.

Be interested to hear how you get on with UCCNC. At least it's a supported piece of software, unlike Mach3!

[AndyUK]

Be interested to hear how you get on with UCCNC. At least it's a supported piece of software, unlike Mach3!

This is what pushed me towards UCCNC rather than Mach3 - and the horror reports from Mach4.

I'm slightly worried however by the multiple updates to UCCNC which are released without any accompanying release notes or change history. :( Feels like the developers are starting to get into some bad habits.

[paulus.v]

I still like the CSMIO IP-M with Mach3.

The real benefit of the controller support for axis slaving/homing (e.g IP-S) is homing on index. Homing on index has almost 100% accuracy/repeatability of homing/squaring as it doesn't rely on the sensors accuracy but you need motors with encoders, e.g servo drives with index pulse..

Otherwise I find the software homing/squaring quite reliable and do not see any real benefit for a homing sequence performed by the controller.

It is even possible to home on index with Mach3 with a simple macro. Not sure about gantry squaring on index but it may be doable as well.

[JAZZCNC]

Be interested to hear how you get on with UCCNC. At least it's a supported piece of software, unlike Mach3!

There will come a time when even UCCNC and older controllers will be unsupported as technology/software PC's etc moves on. This is basically what's happened to Mach3 along with Art fennerty selling the company to Brian who took a different approach to development and marketing, a wrong one IMO.
That said Mach3 is still a very capable piece of software and the Cslabs controllers are high quality. Cannot really blame Cslabs for Mach3 or for not developing for mach3 anymore and they do fully support Mach4 which I'm told is much more robust that it was and lot of the bugs have been ironed out.! . . .I'll have to maybe revisit it.? They also have there own software which is still a work in progress.

Also on a side note and in Mach3 defense and to some degree Cslabs regards MPG's. It's no surprise that some are having problems when using MPG's plug-ins from other company's with Cslabs controllers. They do offer an MPG module that works perfectly with the controller and obviously they are not going to help or support another company's products at the cost of not selling there own, who would.? This problem is not a mach3 problem, it's the controllers plug-in who is in charge.!




I'm slightly worried however by the multiple updates to UCCNC which are released without any accompanying release notes or change history. :( Feels like the developers are starting to get into some bad habits.

Why.? I don't get why people feel the need to update software just for updating sakes when they don't have any issues.! . . . . If it's working ok leave it alone.

[AndyUK]

Why.? I don't get why people feel the need to update software just for updating sakes when they don't have any issues.! . . . . If it's working ok leave it alone.

Well that's exactly my point. If I have an issue, I should be able to look through the release notes to see if its been addressed in a newer version. I shouldn't have to try pot-luck installs.

Maybe there's a new feature I really want? How would I know if there are no release notes...

Maybe I know there is a new feature, but they've removed some other functionality I use... again how do you know?

I've worked as a software engineer before, and release notes are a basic requirement for any quality product. It makes me worry what other shortcuts they're taking, which might not lead to problems right away, but a couple of years down the line could make a project unsustainable to work on, and means your code base becomes a buggy mess (which I gather is what happened to later versions of Mach3?)

[JAZZCNC]

Well that's exactly my point. If I have an issue, I should be able to look through the release notes to see if its been addressed in a newer version. I shouldn't have to try pot-luck installs.

Maybe there's a new feature I really want? How would I know if there are no release notes...

Maybe I know there is a new feature, but they've removed some other functionality I use... again how do you know?

Yes, Agree on that side of things your 100% correct regards release notes. I wasn't actually referring so much to that but rather peoples need to update for the sake of updating. Just don't get the need if things are working ok.!

[Voicecoil]

I wasn't actually referring so much to that but rather peoples need to update for the sake of updating. Just don't get the need if things are working ok.!

+1 Perhaps Microsoft (mighty offenders in this area) could do with some Yorkshire wisdom!

[Voicecoil]

After much swearing at the rather arkward connector I've got the spindle installed :thumsup:

27231

[AndyUK]

Looks like a nice tidy job!

*goes back to my spindle and gets some cable braid out*

[Voicecoil]

To be honest, it wasn't just a desire for neatness :blush: Having painted myself into a corner with feasible belt paths on the Z-axis I came up with the idler-with-the-cable-through-the centre idea. Problem was, the cooling pipes tended to have a bit more friction through the bushing than I was comfy with - the cable braid is a lot more slippy and makes everything move a lot more freely .

[Voicecoil]

The spindle of course needs cooling water. After a long and fruitless search on the Bay of fleas for a suitable flat sided metal pot, I made one out of a bit of 100x100x2 ali tube that's been at the back of the workshop for ages. Add a pump + flow switch and everything's cool :cold:
27232

[JAZZCNC]

After much swearing at the rather arkward connector I've got the spindle installed :thumsup:

You might want to add strain relief to that, it's just a short circuit waiting to happen. Also not a good idea to pump heat into the wire which is what the water pipes will do, esp on long jobs.

Strange Z-axis design. Why did you feel the need to use 2 ball-screws and have the linear rails side-mounted.?

[Voicecoil]

I'm not too worried about the strain relief (the supplied one wouldn't fit over the cable!) as I've added a couple of layers of adhesive lined heatshrink over the cable so the clamp is pretty damn tight now, and the last inch or so of the cable has been stiffened up a lot. The heat transfer is an interesting point, I'll keep an eye on that, if necessary up the cooling - easy to add some finned extrusion to the pot and up the pump speed. As mentioned above, having to bundle the water pipes and spindle drive together and take them through the bushing wasn't really intentional!
The Z-axis design was done that way to keep things compact and the spindle axis as close to the gantry as possible (it's saved between 40 and 50mm off a more conventional design), and was partly inspired by Routercnc's machine. Should see how well it works in a few days once I've got the VFD programmed.....

[JAZZCNC]

The Z-axis design was done that way to keep things compact and the spindle axis as close to the gantry as possible (it's saved between 40 and 50mm off a more conventional design), and was partly inspired by Routercnc's machine. Should see how well it works in a few days once I've got the VFD programmed.....

Ermm not sure if that gain alone will be worth it.? It's going to make tramming the spindle an absolute bitch of a job. Accurate Alignment of rails, ball-screws, etc looks much more difficult.
Also, the bearings being sat on thin edge-mounted plates will allow more vibrations into the spindle.

Don't get me wrong I'm sure it will work fine but the conventional approach would have been stronger and easier to setup. I also think it's a bit of a wasteful design.? In that doesn't require 2 screws which cost extra money and also put excess load on the motor robbing it of performance.! . . . I'd rather have it simpler, stronger and more efficient than save 40mm travel if I'm honest.

Sorry to be negative about it,:dejection: . . . But I'd rather say it than think it, then sees someone else copying it thinking it's better because saving travel without seeing the negatives I see. I

You made a nice job of it thou well done..:thumsup:

[Voicecoil]

Ermm not sure if that gain alone will be worth it.? It's going to make tramming the spindle an absolute bitch of a job. Accurate Alignment of rails, ball-screws, etc looks much more difficult.
Also, the bearings being sat on thin edge-mounted plates will allow more vibrations into the spindle.

Don't get me wrong I'm sure it will work fine but the conventional approach would have been stronger and easier to setup. I also think it's a bit of a wasteful design.? In that doesn't require 2 screws which cost extra money and also put excess load on the motor robbing it of performance.! . . . I'd rather have it simpler, stronger and more efficient than save 40mm travel if I'm honest.

Sorry to be negative about it,:dejection: . . . But I'd rather say it than think it, then sees someone else copying it thinking it's better because saving travel without seeing the negatives I see. I

You made a nice job of it thou well done..:thumsup:

[Voicecoil]

Tramming's not such a bad thing as might appear - the vertical blocks that the Z-axis Hiwins are attached to have increasing clearance hole diameters from 6.1mm (bottom) to 6.6 mm bot( top) to allow some side to side movement at the top to get it square. And they make the Z-axis plate seriously chuffing rigid top to bottom-wise. Front to back tramming (which would always seem to be a bugger on a gantry machine to me) was done using the end fastenings on the gantry and a bit of shimming :whistle: As regards the bearings being on thin plate, that was a result of me not having much in the way of machining facilities, the top plate and a few other bits are due for a re-make once I have it making chips, with thicker material and better alignment once I can sink the bearings and the motor into the plate.

[JAZZCNC]

Tramming's not such a bad thing as might appear - the vertical blocks that the Z-axis Hiwins are attached to have increasing clearance hole diameters from 6.1mm (bottom) to 6.6 mm bot( top) to allow some side to side movement at the top to get it square.

In that case, you are bending the ball-screws because they are fixed to the top plate while the bottoms are fixed to the upright plates which your moving causing a twist on the screw. Also, the ball-screws must stay aligned perfectly parallel to the rails which they cannot do this way because you have them fixed to the top plate.

This design fails the golden rule of DIY to me.? Which is build-in Adjustment, adjustment, adjustment and bit more adjustment just for good measure?


And they make the Z-axis plate seriously chuffing rigid top to bottom-wise.

Pointless if they cause vibrations through the spindle, which they will do compared to if they were fixed directly to backplate like on conventional design. You could also use a similar supporting strip on the rear plate if felt it was needed on conventional design but without causing vibrations.

[Voicecoil]

In that case, you are bending the ball-screws because they are fixed to the top plate while the bottoms are fixed to the upright plates which your moving causing a twist on the screw. Also, the ball-screws must stay aligned perfectly parallel to the rails which they cannot do this way because you have them fixed to the top plate.

This design fails the golden rule of DIY to me.? Which is build-in Adjustment, adjustment, adjustment and bit more adjustment just for good measure?
.

Well, the holes that secure the top plate to the back plate are a bit oversized to allow the bit of L to R adjustment whilst the idea is that the top of the side blocks and bearings stay fixed relative to each other - and after all the position where tram is spot-on should be when everything is perfectly aligned at 90 deg. If I were to do it again I would likely split the back plate into 2 stacked thinner plates to make assembly and tramming easier, but right now the important thing is to get this one up and running.And whilst it took a couuple of goes round the loop to adjust it all to get it running smoothly, it does basically work, it goes up and down quite happily apart from a slight wobble on one of the pulleys which I seem to have managed to bore out slightly off-centre, that's a job for tomorrow.

[Voicecoil]

So the last few days have been a "getting it all working" scenario. First off was a power problem: I'd had some issues with the control box and other gear (typically with big SMPS) tripping out the RCCB in the workshop, and as soon as I tried to power up the VFD it did it every time. Investigation showed that the whole workshop was on a 30mA trip, fine for a domestic ring main, but unlikely to be stable with a couple of big EMI filters on the line, I've come across such issues before. Thankfully I have a humungous (1.5KVA) isolation transformer I use when developing switch mode supplies, putting that in the line has at least enabled me to get going. looks like I'm going to have to have a word with the estate to get a more realistically rated breaker fitted :sulkiness: Then came the joys of programming the VFD (a Sunfar 300 - the more usual Huanyang doesn't go high enough for a 4 pole spindle), although the manual (or at least the newer version of it) isn't so bad there's a couple of things that vary between ambiguous and rather misleading. Anyway finally I've managed to get the spindle spinning (and the right way round!!) and sorted out stopping and starting from UCCNC. This showed up and interesting twist in that for survivability's sake you need a high voltage open collector O/P from the AXBB-E to switch the spindle On/Off as it's switching something that seems to vary between 5.5 and 6.5V; OK so you could use one of the Port 2 isolated outputs. BUT the switching input is referred to the 5V/processor ground on the AXBB-E through the 0-10V speed control line, NOT the 24V "machine" ground that the isolated outputs run from, so the choice is to either common all the grounds up, or get creative: I chose the latter and added a wee MOSFET on the output of Port1 p17 to buffer it. Then lastly I had some issues with limits triggering when the spindle started, after some parping about this was solved by tying the 24V ground to earth but floating the 5V ground.

[Kitwn]

Thankfully I have a humungous (1.5KVA) isolation transformer I use when developing switch mode supplies, putting that in the line has at least enabled me to get going.
You know this, but I'm going to say it anyway in case anyone else wants to have a go: You are now completely unprotected from electric shocks. What would previously (thanks to the annoying 30mA RCD) have resulted in nothing worse than a bout of swearing could now lead to a funeral.

[brman]

You know this, but I'm going to say it anyway in case anyone else wants to have a go: You are now completely unprotected from electric shocks. What would previously (thanks to the annoying 30mA RCD) have resulted in nothing worse than a bout of swearing could now lead to a funeral.

Care to explain that? Because, like voicecoil I use an isolation transformer when developing 240V switching devices (in my case 3 phase pmsm drives) specifically to prevent myself getting an electric shock. That and stopping me melting 'scope probe leads anyway.... ;)

[Kitwn]

Care to explain that? Because, like voicecoil I use an isolation transformer when developing 240V switching devices (in my case 3 phase pmsm drives) specifically to prevent myself getting an electric shock. That and stopping me melting 'scope probe leads anyway.... ;)

Yes it does need qualifying.

If there is no earth reference connection anywhere following the transformer and you touch a live connection then that side of the supply will float to your potential and no current will flow. Same with your scope probes, the scope floats to the same potential as that side of the supply (or vice versa) and there is no problem.

The problem arises when one side of the supply is connected to exposed metalwork either deliberately or by accident such as touching a metal bench or via another piece of equipment plugged into a dis board fed from the transformer. This creates the risk of you contacting both sides of the supply at once.

You are quite right that an isolating transformer can protect you in certain circumstances but most electronic workshops have strict guidelines about the use of isolating transformers including only ever plugging in one piece of equipment at a time, covering metal benches with rubber mats etc. My concern was if the whole CNC machine and other gear was going to be isolated by the transformer something somewhere could be providing an earth reference and that creates a danger.

The primary danger I was concerned with is the fact that a transformer negates the protection provided by an RCD. If you connect yourself between a standard mains supply and earth it's the leakage to earth (which creates a small difference in the live and neutral currents at the RCD) that is picked up by the RCD and that trips the supply. Put a transformer downstream of the RCD and connect yourself across it's output and there is no earth leakage back to the mains supply so the RCD won't protect you.

[Voicecoil]

Y

The problem arises when one side of the supply is connected to exposed metalwork either deliberately or by accident such as touching a metal bench or via another piece of equipment plugged into a dis board fed from the transformer. This creates the risk of you contacting both sides of the supply at once.

......That's why SMPS engineers tend to work with one hand behind their back I guess!

Seriously though, the danger isn't that great as to get a shock off an isolated supply you'd basically need to be touching something live with one hand and something neutral with the other, which is likely a smaller risk than touching something live and leakage to earth via feet etc.. And of course as part of the procedure I've always gone through when working in such a way, you do a check that there's mains earth continuity through the iso. transformer, so that in this case the machine chassis is properly earthed.

[brman]

......That's why SMPS engineers tend to work with one hand behind their back I guess!

Seriously though, the danger isn't that great as to get a shock off an isolated supply you'd basically need to be touching something live with one hand and something neutral with the other, which is likely a smaller risk than touching something live and leakage to earth via feet etc.. And of course as part of the procedure I've always gone through when working in such a way, you do a check that there's mains earth continuity through the iso. transformer, so that in this case the machine chassis is properly earthed.

I could have misunderstood but I think kitwin was making the point that, even with a good chassis earth, you will not notice if you have a short from line or neutral to the chassis. So, if you then touch chassis and live you will get a belt the RCD might not protect you from.

[Voicecoil]

.

[Voicecoil]

.

[Voicecoil]

OK, I see what you mean, I must have read it wrong. But I guess in an industrial setup you're not going to have a 30mA RCD, so similarly it's going to be a matter of working carefully.Anyway it's a lot less scary than 370V DC on the primary side of a power supply (or on the braking terminals of a VFD for that matter :upset: )
Interesting to hear you work on PMSM drives, what sort of power? I've only ever come across quite small PMSM's


I could have misunderstood but I think kitwin was making the point that, even with a good chassis earth, you will not notice if you have a short from line or neutral to the chassis.

It's not beyond the wit of man to make a circuit that detects that, if I have a spare moment I might just add one to my isolation transformer. Not that the current arrangement is anything more than temporary until the workshop electrics get sorted out (type F RCCB maybe?), but I can see it might have other uses.

[brman]

OK, I see what you mean, I must have read it wrong. But I guess in an industrial setup you're not going to have a 30mA RCD, so similarly it's going to be a matter of working carefully.Anyway it's a lot less scary than 370V DC on the primary side of a power supply (or on the braking terminals of a VFD for that matter :upset: )
Interesting to hear you work on PMSM drives, what sort of power? I've only ever come across quite small PMSM's



It's not beyond the wit of man to make a circuit that detects that, if I have a spare moment I might just add one to my isolation transformer. Not that the current arrangement is anything more than temporary until the workshop electrics get sorted out (type F RCCB maybe?), but I can see it might have other uses.

up to about 1kW motors. It is not our main business though so I don't claim to be any sort of expert.

edit to add: yes, I agree. I am much more worried about being zapped by me putting my finger somewhere it should not be than worried about the possibility of a secondary fault invalidating the protection. But I think Kitwin is right with his warning. It is all to easy to forget that not everyone understands the consequences of what might happen and so might think that an isolation transformer is a guaranteed safe solution.

[Muzzer]

You may be overthinking or misunderstanding this. Having spent most of my career designing and developing mains powered SMPS products, I would endorse the use of an isolation transformer. It's true that the RCD no longer protects you against accidental contact with an HV circuit - but when it's floating, by definition one such point of contact isn't in itself an issue. On the other hand, if it's non-isolated mains, you can get a lethal (deadly) belt just by touching a live node unless you are fully enclosed in a heavy duty rubber gimp suit. Unfortunately these are expensive, uncomfortable and can get you arrested.

Most professional labs use an mains isolation transformer and a variable source, either a variac (variable autotransformer) or a modern switchmode AC source. There is generally an e-stop to completely kill the power remotely in case you observe anything untoward. And you don't let anyone near the stuff without full and formal training.

There's no protecting against complete idiocy and at some point you need to learn the importance of NOT touching 2 different nodes, particularly with different hands. I once did that across 415Vac and somehow live to tell the tale. It hurt like f*ck and I never did it again but might not have been so lucky. For that reason, only using one hand to attach probes etc, combined with an isolation transformer is an <almost> foolproof way to avoid unnecessary bouts of death. Nowadays, we are trained in CPR and have access to defibrillators but those are the ambulance at the bottom of the cliff rather than the fence at the top.

[Voicecoil]

On the other hand, if it's non-isolated mains, you can get a lethal (deadly) belt just by touching a live node unless you are fully enclosed in a heavy duty rubber gimp suit. Unfortunately these are expensive, uncomfortable and can get you arrested.

PMSL :hysterical: :hysterical:

[Voicecoil]

Got the machine up and running, then realised I needed to get some coolant sorted before cutting anything much in the metal department. Rather than reinvent the wheel from scratch I thought I'd start with one of these off the Bay of fleas:
27426
Trying it out it seemed to deliver coolant without too much fog, but was rather feeble in the chip clearing department running off my rather low pressure air supply..Dis-assembling the thing it is made of 4 parts like so:
27427
the inner coolant nozzle has the same pitch thread as the outer collar, so threads into the collar when it's put on. The lack of puff looked like it was being caused by the very small apertures:
27428
Then I noticed the internal diameter of the yellow plastic end bit looked like it would fit a bit of 6mm AF hex bar, and that the collar and final guide had the same M10 x 1 threads. So I turned and slotted a bit of 6mm AF hex brass to make an alternative inner piece:
27429
This push fits firmly inside the end of the yellow bit thus:
27430
Screw the conical end guide back on and hey presto there was a good blow & coolant, quite a decently narrow stream too, about 10mm diameter at 100mm distance. And the bits cost less than a fiver :beer:

[AndyUK]

Nice one! Looks great!

Don't suppose you'd knock off another insert for those of us without a lathe handy?

(Technically the lathe is just a few hundred miles away covered in a mountain of crap, but I've never used it, and plan to eventually start gently...)

[Voicecoil]

Nice one! Looks great!

Don't suppose you'd knock off another insert for those of us without a lathe handy?

(Technically the lathe is just a few hundred miles away covered in a mountain of crap, but I've never used it, and plan to eventually start gently...)

I'll have a go - means I'll have to measure that one though! Might be a week or two as v.busy at the mo.

[Voicecoil]

This was borne out of a cockup, but might be of interest to someone else. I'd originally intended to have some bellows covering the rails and ballscrew on the gantry (my X axis), but when I came to fit them I found I'd made an error in the drawing and there wasn't quite enough clearance at the bottom so they fouled the carriage plates. It won't be a problem in the long run as I was intending to remake the carriage plates anyway now that I have proper milling capacity, but that's been put back a bit due to work pressure and supply problems with aluminium stock due to the virus. Observing that the X axis ballscrew was frequently getting sprinkled with chips, I thought some other form of protection was in order. I rang a few of the machine guarding people but everything they had on offer was too big and/or too expensive. After a bit of thought and fiddling, it looked like some kind of blind arrangement would be the best option given the limited space. I fiddled around with springs out of an old window blind, but the force was far from constant and it was a pain to put together without getting custom springs made. Then the thought struck me that a DC motor supplied with a constant current gave constant torque..... So I came up with these:
27795

Not a 100% perfect solution, but they seem to be keeping the ballscrew clean, and apart from a couple of £10 motors off the Bay of fleas, were made from scrap bits lying around. I've set them up so they're switched on by the X axis enable signal, though the motors don't even get warm if on continuously.

[JAZZCNC]

:hysterical::hysterical::hysterical: that picture really got me confussed. I looked at it and thought why as he got green caps on the ends of those rails... then i realised the green was the bearing caps...:whistle::stupid:

[dazp1976]

Not a 100% perfect solution, but they seem to be keeping the ballscrew clean, and apart from a couple of £10 motors off the Bay of fleas, were made from scrap bits lying around. I've set them up so they're switched on by the X axis enable signal, though the motors don't even get warm if on continuously.

I did something similar on my mill Z axis. Took some car window blinds apart, stuck some 0.5mm nitrile sheet on the roller and it actually works.
The things you have to do :smile: