Aluminium Profile Type Router - Initial questions

[CNCRY]

Thanks for the encouragement all, looking forward to having a moving machine soon..

Andrewg - I plan to keep the log going until the end, I'm sure there are much better build logs out there than mine, but I'm trying to show some of the smaller details/photos which made no sense to me when I started this process, and if it helps a few people along the way, then great as this forum has helped me loads so far.

As you say without seeing the fruits of someones design/pitfalls you can't be sure if it's a good one. As Andy mentioned though there are certainly more, like his which do continue.

On the dual plate, to be honest I did think of it earlier on, but at the time I was going in circles with the design and if I could physically make it so opted for a simple single plate. Jazz - I see what you are saying, essentially the ballscrew is fixed square with the frame, but the gantry may need slight squaring.

I'm hoping I can get slight adjustment if I have to go there, from shimming either ballnut or the gantry side/ballnut fixing. I have built in a very slight tolerance by using a 1mm thick piece of rubber between gantry side and ballnut fixing, which can compress 0.5mm ish to accomodate a slight twist.
So watch this space whether it works or not! If building new, I'd go for the 2 plate though.
Jazz - looking forward to your machine unveiling, maybe I can swap it with mine..

So, wasteboard, just best to start with MDF and hold down clamps screwed into it to get started?

[Kitwn]

I'm hoping I can get slight adjustment if I have to go there, from shimming either ballnut or the gantry side/ballnut fixing. I have built in a very slight tolerance by using a 1mm thick piece of rubber between gantry side and ballnut fixing, which can compress 0.5mm ish to accomodate a slight twist.

So, wasteboard, just best to start with MDF and hold down clamps screwed into it to get started?

I like the idea of deliberately designing with a built-in gap for a shim in certain places. There's no such thing as a -0.2mm shim however much you might want one when finally assembling the machine.

Having recently got away from just screwing clamps into an MDF spoilboard to using something a little more sophisticated I can honestly say it's worth the effort to do better from day one, even if it is just to get you started. One of the problems is the little dimples that form around the screw holes which must be removed before you can put another workpiece over them and it's a pain having to drill holes for where you want to fit the clamps every time. Re-using the same holes many times over isn't too reliable either. I let my machine drill a grid of holes in the plywood base board and then put M8 T-nuts underneath. A sacrificial layer of MDF goes over that and then drill again. I also put a row of hand-drilled holes at either end of the base board to hold any pieces which cover the whole cutting area of the machine.
You can always upgrade to a slotted aluminium base in the future if you need to.

[JAZZCNC]

On the dual plate, to be honest I did think of it earlier on, but at the time I was going in circles with the design and if I could physically make it so opted for a simple single plate. Jazz - I see what you are saying, essentially the ballscrew is fixed square with the frame, but the gantry may need slight squaring.

Yes and no.
What you are doing when squaring the gantry is setting it perpendicular to the MASTER RAIL and NOT the frame. Infact forget the frame because unless you have built it perfectly square and aligned the Master rail perfectly parallel to it then you CANNOT measure from it. The MASTER rail is the ONLY reference point you work from.

So the ball-screw must also be perfectly parallel to the MASTER rail. Likewise, the Ball-NUT is set parallel to the gantry side.

So if you have set your ball-screws perfectly parallel to the MASTER rail, which you should have done. Then any angular adjustment on the gantry in respect to the MASTER rail means there must be an angular twist between the Ball-NUT and Ball-SCREW and this will cause excess wear and binding.

The goal is to get the gantry perpendicular in respect to the MASTER RAIL and the ball-SCREW perfectly parallel to the MASTER rail. This will then put the Ball-NUT perfectly parallel to the ball-SCREW.

When all this is done then this will show you how much Shimming and spacing is needed at various places around the machine in relation to the Frame. If you have built a perfectly square frame then it should be easy.!!

You may find the best place to shim to get square without impacting the screws/nut relationship too much is between the Gantry plates and profile at each side.
This is where building in lots of adjustment points comes into play and helps with setup and fine tuning, don't be surprised if this takes you many weeks to get right with lots of strip downs.

Edit: I've just done a quick sketch in SW to show how much just 1deg of misalignment translates across the length of a gantry or machine.
The measurement show the numbers except one and that is the twist of the ballnut in relation to the ball screw but the number is 0.70mm of twist which would wear a ballnut out in no time and cause lots of binding.

I'm hoping I can get slight adjustment if I have to go there, from shimming either ballnut or the gantry side/ballnut fixing. I have built in a very slight tolerance by using a 1mm thick piece of rubber between gantry side and ballnut fixing, which can compress 0.5mm ish to accomodate a slight twist.

The Rubber is BAD idea, your basicly building in Back-Lash. The rubber will allow it to float even under compression. Don't underestimate the forces and the amount of inertia that will be applied to ballnut which transfers to the ballscrew which pushes on the Endbearings which will be sat basicly on rubber engine mounts.!!

Jazz - looking forward to your machine unveiling, maybe I can swap it with mine..

Oh don't be getting too excited I've not re-invented the wheel and the design is nothing special or new. It will just be built properly and strong compared to all the Cheap offerings on Ebay etc.

[Andrewg]

That's just because people don't bother to post in the gallery thread..

I was being a little provocative to see if anyone else might think to post their machines. Shoudl be a sticky, as it was a good idea. That some very capable people turn up with some great designs and get huge assistance here but fade away before getting a working machine does seem a pity.

I'm going to answer this Andrew because many won't realize the difference and why it's important.

Thanks, I am beginning to understand how alignment and precision in the build process is hugely affected by these design details. However with a single under-table ballscrew design, this would mean ideally the gantry side plates are not a single part...or are you suggesting the adjustment come at the top of the gantry plate with its connection to the X Gantry beam?

But sorry CNCRY, lets keep focused on your final hurdles to cutting! You are almost there.

[CNCRY]

What you are doing when squaring the gantry is setting it perpendicular to the MASTER RAIL and NOT the frame. Infact forget the frame because unless you have built it perfectly square and aligned the Master rail perfectly parallel to it then you CANNOT measure from it. The MASTER rail is the ONLY reference point you work from.

So the ball-screw must also be perfectly parallel to the MASTER rail. Likewise, the Ball-NUT is set parallel to the gantry side.

So if you have set your ball-screws perfectly parallel to the MASTER rail, which you should have done. Then any angular adjustment on the gantry in respect to the MASTER rail means there must be an angular twist between the Ball-NUT and Ball-SCREW and this will cause excess wear and binding.

The goal is to get the gantry perpendicular in respect to the MASTER RAIL and the ball-SCREW perfectly parallel to the MASTER rail. This will then put the Ball-NUT perfectly parallel to the ball-SCREW.

When all this is done then this will show you how much Shimming and spacing is needed at various places around the machine in relation to the Frame. If you have built a perfectly square frame then it should be easy.!!

You may find the best place to shim to get square without impacting the screws/nut relationship too much is between the Gantry plates and profile at each side.
This is where building in lots of adjustment points comes into play and helps with setup and fine tuning, don't be surprised if this takes you many weeks to get right with lots of strip downs.

Edit: I've just done a quick sketch in SW to show how much just 1deg of misalignment translates across the length of a gantry or machine.
The measurement show the numbers except one and that is the twist of the ballnut in relation to the ball screw but the number is 0.70mm of twist which would wear a ballnut out in no time and cause lots of binding.

The Rubber is BAD idea, your basicly building in Back-Lash. The rubber will allow it to float even under compression. Don't underestimate the forces and the amount of inertia that will be applied to ballnut which transfers to the ballscrew which pushes on the Endbearings which will be sat basicly on rubber engine mounts.!!

OK so full honestly, in terms of the rail and ball-nut alignment, what I've done is to set the master rail using the thickness of the extrusion as a gauge and then use a fixed size block offset from the edge to ensure the rail runs parallel to the edge of the extrusion.
Then the "slave" rail is firstly set it the same way with not fully tight bolts, then run the gantry along it to ensure it runs smoothly along the rail and moves the rail very slightly to a final adjustment. (but in theory its the same offset from extrusion as master rail)
The axis runs smoothly with no ball-nuts attached at least!
Then the ballscrews were attached to extrusion using tooling plate spacer so in theory are fully parallel to the extrusion.
So yes clearly its not fine precision, and a fair amount is riding on the aluminium profile being to good tolerances which I guess is much harder on box steel build, but to see where I'm coming from I didn't know what a DTI even was was a couple of months ago!
So with that said, I'm hoping I can tweak things without any major changes to the machine.

The rubber part is only on the dual Y (formally known as X :) axis) ballnut. As per below in red. The bolts are still coupled into the ballnut mount, but I can easily replace that with 1mm aluminium shim.


I've attached the motor mounts now so am going to run it tomorrow and see how it moves. Is it obvious it its binding , does it judder etc?
Then if shimming is needed use this on the gantry and Ball screw mounts.

How should you align the ballscrews vertically - I tried using a DTI on the rail touching the ballscrew but hard to where on the thread it sat. Or use a small block sat on the screw..?
Do people use the 3, 4, 5 triangle method to check square then adjust and set limits? As there are various ideas out there..

Will report back after running it - wish my luck

[JAZZCNC]

The rubber part is only on the dual Y (formally known as X :) axis) ballnut. As per below in red. The bolts are still coupled into the ballnut mount, but I can easily replace that with 1mm aluminium shim.

Dump the rubber it's a bad idea for the reasons stated before. Shimming with aluminium is the better way to do it.
What you have done so far with rails etc and screws is fine so with shims you'll get it pretty much spot on with a bit of patience. My point in all the other post's was to high light the difference and why the extra plate makes things easier and with less chance of messing things up in other areas when making adjustments to gantry.

I've attached the motor mounts now so am going to run it tomorrow and see how it moves. Is it obvious it its binding , does it judder etc?
Then if shimming is needed use this on the gantry and Ball screw mounts.

Be Very careful when approaching the ends of travel because it's here where most get it wrong and it's very easy to bend the screws. At slow speeds the steppers will quite happily bend those screws like it's liquorish sticks and not complain about it.
At higher feeds any error here will stall the motors but still damage will be done.

When jogging near the ends watch the ball screw carefully and if you see it lift or bend then you have some misalignment. I CANNOT stress this enough move slowly and watch carefully because it only takes a second to bend the screw.!!

If you see any movement then loosen the end bearings and jog slowly towards the bearing, this will give you some idea of what's wrong. Most likely either a Gap will appear between bearing and frame or it will lift up/down. However it could be pushing the bearing into the frame which isn't so easy to see so pay careful attention to what the screw does.

How should you align the ballscrews vertically - I tried using a DTI on the rail touching the ballscrew but hard to where on the thread it sat. Or use a small block sat on the screw..?
Do people use the 3, 4, 5 triangle method to check square then adjust and set limits? As there are various ideas out there..

Will report back after running it - wish my luck

It's difficult on a long machine to use the DTI because the screw sags so you need a combination of straight edge and DTI. However it's often easier to go by feel and just loosen the end bearings, then shim or adjust as needed.
Between adjusting the Ball-nut mount and end bearings you will get it spot on eventually but don't be surprised if takes several setups, but the effort will be worth it because any binding robs power and cause excess wear so the extra time spent here pays off BIG time.

Regards testing for squareness then if you haven't got precision squares etc just initially use a good set square off the master rail to get you close. Then just cut a Large square, nearly the size of your bed and measure the diagonals with a good steel rule, NOT a tape measure. Tweak and adjust as needed then cut another square by trimming the last one.

[CNCRY]

Very mixed day of results - finished most of the Z axis strengthen plates and connected the 2 Y motors up.
I ran each motor individually first with the ballnut uncoupled to check how it runs. The good part
The gantry motion on rails is very smooth and all good, and one ballscrew is running true (until squaring anyway!)
Using a digital square the gantry seems good enough to progress to cutting a square to determine further.
The bad bit - the other ballscrew is definitely bent, not 100% if it arrived that way but much more likely it's a result of drill testing or something else I did! Some swearing involved and am waiting to find out if Fred can amend my spindle order...
Advice for anyone in the research stage - plan some budget for screw ups!
I connected both screws up for a short time just to test the Axis slaving - and it was cool to see my creation moving via the laptop finally!

Now waiting on replacement parts I'm going to fit and wire in the limit switches. This should be simple but the diagram which came with the switch confused me..
Can someone tell me is this right? I thought I understood it until I see the black cable on the switch diagram shows 6-36V while "out" on the AXXB manual shows it at 0v?

AXXB manual image, switch diagram and my understanding of it.

[Kitwn]

The black wire is the output of the sensor. It is pulled up to the supply voltage by an internal 10k Ohm resistor until the output transistor conducts and pulls it to near zero volts. 'Normally Open' (NO) means the transistor is not conducting (output pulled high by the resistor) until the sensor is triggered. 'Normally Closed' (NC) means the opposite. Only NO types can be connected in parallel as shown in your diagram, any single output pulled low will pull the others with it and the input to the controller. NC types (which you have according to the diagram) must be connected in series since, if in parallel, all the outputs will be normally low and the ONE that tries to go high will be held low by the others. Unfortunately your AXXB manual image does not explain that but the text should.

EDIT: Your controller software should include a configuration for selecting whether your sensors are NC or NO. If you get continuous limit alarms then that my need to be changed.

[CNCRY]

Ok so they are NC type to cover the issue of broken wires etc.
So the black wire will normally be at supply voltage until limits are hit.
I'm planning on wiring each to its own limit input on axbbe as there are 6 available, leaving a couple for probes and future use.
So imagining just one switch in my diagram above , l1+ will always be 24v and l1- would be at 24v normally unless switch triggered.
Is that right ?
Thanks !

[Kitwn]

Other way round! NC will normally be at ground voltage (Normally Closed to ground) until triggered. "Normal' means you have NOT driven the machine to it's limit and it is continuing to cut material as intended. This gives you protection against broken wires as an open wire shows up as a high voltage if your controller has it's own pull-up resistors and the machine stops. Failure of a NO connection only shows up when the limit switch fails to make the ground connection required and your gantry crashes through both the limit switch and the end wall of your shed.

My personal preference is for one switch per input as you are planning. NC is best as explained above but NO is more commonly available, presumably because of the ease of parallel connection.