Aluminium Profile Type Router - Initial questions

[Doddy]

The power of the internet! One of the key attributes of toroidal transformers is a very low flux "leakage" which will limit any potential induced current. I understand the argument but would challenge the actual risk - if I had a toroidal going spare I'd set up an experiment (it's interesting, and this is the first that I've heard of it, and the evidence from the internet looks to be more anecdotal than academic). Actually, because of this thread I've ordered a toroidal just for this experiment - I'll report back in a few days. My thoughts are any current is likely to be measurable but insignificant. The risk can be mitigated in any event by isolating (air gap) the top of the bolt from the chassis to avoid the short.

Edit: What Clive said. The more I read the more interesting this experiment might be.

[Doddy]

Wow!, just... Wow! I take back everything.

Following the previous post I'd realised I have a bench PSU with a chunky toroidal transformer that I'd previously hacked to pieces. Sounds like a good experimental platform. Ignore all the crap on the bench - I really need to tidy up.

Pic 1: General topology, next-to-nowt resistance (0.1R) from the top of the mounting bolt to the chassis.


Pic 2: Open circuit Vacrms reading from the top-bolt to chassis - about half a volt. A lot more than I'd expected and enough to ignore the fancy stuff with a shunt resistor and 'scope, and go straight for the "blow-the-hell-out-of-the-meter" probe...


Pic 3: Current short, top bolt through meter and 3 metres of meter leads to ground. Some 3.1A.


So, yeah, isolate that top bolt!

[JAZZCNC]

So now I'm thinking to mount to a phenolic (insulating) sheet so the mounting bolt does not touch the cabinet back plate directly. The transformer also has an earth wire.
Any issues with this, or better suggestion people use?

I'm going to throw a spanner in the works here and point out that I've built dozens of machines that use the bolt thru chassis method with rubber pads and I've never had any issues with shorts, fires or getting zapped with stray volts/current etc.
Like Clives mentions it's a Big No No earthing to the bolt but other than that then you won't have any issues provided the transformer sits on the rubber pads.

So just get on with it and don't worry is what I'm saying. If you want to put a rubber washer on the bolt then why not for what it costs but I wouldn't mess around with Phenolic boards etc.

[CNCRY]

Thanks for all the replies - I wasn't expecting a full on experiment Doddy :) Good to have some real results though as everything I found was purely theory, no-one tried it until now!
So in summary if the bolt is used along with the chassis to form a "loop/partial loop" around the core then there is some serious potential for amps to flow, but its low risk if the bottom of the bolt only contacts the chassis, as Jazz mentions this is commonly done with no apparent issues.
I happened to have a piece of phenolic offcut so ended up using that on some spacers, plus the bolt itself has a rubber washer between the head and the metal transformer plate. Probably overkill - and I'm sure I'll do something else instead !
PS before I power up.. Red and Yellow (Vsec from each secondary) and Orange/Black (0V) are connected together - is that all correct for parallel?




Ryan

[Doddy]

It was an interesting and eye-opening (for me) quick test - I've learnt from it so all is good for me. Don't blame Jazz re. the mounting method - that's how industry does it - single bolt through the grounded chassis. As Clive says the two rubber sheets are a form of mechanical protection to allow you to positively clamp on the transformer windings without deformation of the windings and compromising the insulation coating on those windings. The sequence is chassis (bottom) -> neoprene washer -> toroidal transformer -> neoprene washer -> steel plate -> nut.

Re. colours - that's a less than useful label on the transformer but I'd interpret it the same as you say - positional reference indicating the phase of the windings. I've had a quick look on Airlink's website and there's no further useful information that I could spot. Your fuse will tell you if it's wrong.

I've missed the earlier conversation on this but guess you're using AC to power the stepper drivers (no sign of provisioning for rectifier and caps) - I'd think carefully at this time of routing the secondaries in a tidy(!) route away from any low-level signalling, to the drivers - I'm sure you're looking to route to the top of the picture of the box.

[JAZZCNC]

Don't blame Jazz re. the mounting method - that's how industry does it - single bolt through the grounded chassis.

Ye go on blame me I've got wide shoulders I can take it..

Re. colours - that's a less than useful label on the transformer but I'd interpret it the same as you say - positional reference indicating the phase of the windings. I've had a quick look on Airlink's website and there's no further useful information that I could spot. Your fuse will tell you if it's wrong.

Here you go you blind bugger.

https://airlinktransformers.com/post...echnical-notes

[Doddy]

Admittedly - I just looked at the technical data for the transformers, e.g. https://airlinktransformers.com/prod...ange-cm0750255 ... which you really would think that they'd replicate that most basic of info. Even on the link you provided there's no clear statement as to the phasing of the windings (I've seen better data sheets).

Edit: Actually, I take that back. The link you provide does, in the schematic symbol, give the phase relationship of the tails.

[CNCRY]

So the transformer lives, and nothing exploded/shorted etc! I did a quick reading across the output and its 70V, which is higher than the 65V its rated at. Is that normal with no load applied? The drivers are up to 70VAC and don't wan't to fry them..

Made some progress over the long weekend - if only I had a CNC machine this would be so much easier:)

Reading some of the new threads I've been re-thinking and slightly concerned that the accuracy I'm building too could result in inaccurate final results.
As this is all "by hand" and printing 1:1 scaling, drill holes are perhaps up to 1mm out, I'm using a digital caliper and angle measuring device during the build, but check as much as possible
There is some slight adjustment by the hole sizing on everything - how much can the software make up for sub mm "corrections"
Maybe overthinking it - but also don't want to end up with something not accurate after all the effort.
As I haven't owned a CNC machine before I don't know what accuracy I'd like, but for example if I could produce something like the Aztec calendar I'd be happy..

Some photos of current build.





Thanks

[Doddy]

Re. your 70VAC. The rated 65VAC would be at the rated power of the transformer, and off-load (or lighter loaded) you would expect the terminal voltage to rise. There's normally an attribute "regulation" applied to the transformer that describes the difference from full load to off load - typically 5-10%, which looks to be the sort of zone that you're looking at here. Reading AirLinks's technical data sheet I can't find the regulation value (though I clearly struggle with their site), although they identify the attribute, and strangely relate it to efficiency which is a new one on me (they're kind of right, but the actual attribute of regulation is in the context of change-of-voltage).

Others here will (rightly, in my eyes) present a cautious (but reliable) approach to all this and recommend a 10% headroom on supply voltage to rated voltage. For a drive with 70VAC max input this is likely to result in a recommendation of 60VAC. Even that's a bit simplistic but it's a useful rule-of-thumb. So lets have a think about where you are now (no-man's land).

Firstly, what you know is what your meter is telling you. That's probably +/-1% or so, and realistically on a cheap meter (just making a blind assumption here) which isn't calibrated and on an AC supply that's probably going to give you +/-2V uncertainty on the display.

Next you're going to have an issue with line regulation. The UK grid used to be 240VAC (domestic), it's currently around 230VAC (-6%/+10%) trying to get to a EU standard of 220VAC (-10%/+6%). Your transformer is likely rated with primaries at 220VAC. If the mains is at 230VAC then you have a 5% uplift on the primary voltage - which will translate to a 5% uplift on the secondary voltage. That's also about where you find your measured secondary. Worryingly the mains supply can rise a fair bit above this - if you're sited close to a big transformer (e.g. on an industrial estate) you might find the supply around the 250VAC mark, with an appropriate rise in terminal voltage on the secondaries.

So, in a nutshell, you're likely to be at or slightly above the rated voltage of the drivers. And you have to figure on this before you consider any load from the drivers on the transformer. And, in an e-stop situation the load from the drivers is likely negligible, so you should design with this open-circuit terminal voltage in mind. I am a bit surprised that your build has got to this stage without attracting some concern from this thread.

Now, I don't know where to start with the driver. Quickly looking back I thought you was looking at LCDA86's - but one of your diagrams shows a LCDA86H - I think that 'H' is significant here as it raises the operating voltage and introduces a AC-supply option to what was previously a DC supply-only, and at a higher voltage. Clearly the non-H variant would require you to rectify the supply and you'd be so far beyond the rating I'd expect you to blow the drivers. If you do have the 'H' variants, then I think you're on the ragged edge.

You might, on paper, be in spec, but only just. You might also be conscious of supply line regulation and want to consider a little head room between the transformer output voltage and the max input voltage of the driver. You might (I would) also squint at the max rating of the driver and wonder if the designer built in some spare float in the electrical design - but the accountants usually engineer that out. Clearly I'm uncomfortable at the voltages (and power!) that you're playing with. It will probably work, but you might expect that the driver to be running to the limit of what it can, and you might find that the life of the driver is compromised, or, that it lets the magic smoke out when you first apply power, or that it runs fine until the end of time. The problem is you're essentially on/outside of the specification of the device and all bets are off.

Personally I'd probably risk it if I could afford to replace the drivers if needed. But, that's a personal (and some would say foolish) choice.

Ideally I'd have chosen a transformer closer to 55VAC secondary and give myself plenty of headroom.

Other than swapping components (drivers or transformer) for others, is there anything you can cheaply do?... not easily - you are looking to drop the terminal voltage without dissipating huge power (heat).

You could add a couple of bridge rectifiers on the secondaries, in series - rely on the non-linear behaviour of the diodes to introduce a typical drop of a couple of volts per device, and throw them onto the cabinet to passively cool them. At full chat these would dissipate about 20W each - not insignificant, but they would provide better regulation than a chunky resistor. The first thing that the driver would do internally is rectify and smooth the supply - all you're doing is adding a pre-rectifier rectifier... it'd work but it's not elegant.

In theory, you could remove a number of windings from the transformer. I wouldn't touch this solution myself - if only for the buggeration involved with this.

Beyond that you're talking serious solutions that would add cost, weight and power. Nothing springs to mind that provides an elegant solution.

Are you beyond the point of returning and swapping the transformer? If not, I'd give that a shot. Next I'd look to sell/swap on the transformer as near-new here, you might be lucky. Or, you can risk it. You might be lucky.

EDIT:

Going back to where I came into this thread... you could always short out the mounting bolt on the toroidal... I hear that puts a bit of a load on the transformer

I'll get my coat.


There's another option with transformer design. You could wind maybe 10 - 20 turns of 13A-rated (min) insulated flex in the opposite direction onto the transformer per secondary, and feed this in series with each secondary. If in phase it'll add a few volts, or if you swap the phase it'll reduce by a few volts (you'll not be able to tell the phasing - just try either way to wire them and measure to work it out). I'm not advocating this, nor will I do the experiment to prove it. It'd introduce additional losses (power) into the transformer but if you was in a corner it might offer a solution.

[AndyUK]

Reading some of the new threads I've been re-thinking and slightly concerned that the accuracy I'm building too could result in inaccurate final results.
As this is all "by hand" and printing 1:1 scaling, drill holes are perhaps up to 1mm out, I'm using a digital caliper and angle measuring device during the build, but check as much as possible
There is some slight adjustment by the hole sizing on everything - how much can the software make up for sub mm "corrections"
Maybe overthinking it - but also don't want to end up with something not accurate after all the effort.
As I haven't owned a CNC machine before I don't know what accuracy I'd like, but for example if I could produce something like the Aztec calendar I'd be happy..

First off, looks like some great progress in the photos, well done!

Regards the by hand - quite a few of us have used this method to good results. Yes, sometimes you cock it up, but just be slow and methodical double checking measurements.

No idea of your experience so lets ensure we hit all bases; make sure to centre punch well (optical if possible - the press type I also like), and let the drill bit align to that punch mark (if you clamp too tightly aligned slightly off, the bit won't happily sit in the punch crater). Use the drill press as much as possible (but check its square to the plate in both directions). If you're not using printouts, use marking dye. Set your digital calipers to the value from a reference surface, lock them off, then use them to scribe the line. When using a steel rule, never use the end for final marking (its a bit variable), always align to markings on both sides (e.g. to measure 10cm I'd align the 1cm mark and the 11cm.) - Now check using the end on the first mark!

If you come to fit components together and the holes are a bit out, you can always identify troublesome ones and widen only those. An M5 clearance hole already gives you 0.5mm of wiggle room, a 6mm hole isn't going to be the end of the world. Worst case is you have to make it again (oh no!) or remake it using the CNC later.

Regards accuracy - mostly to do with how you align the linear rails and screws. Use a dial gauge and get everything running smoothly and true to a reference surface. In your case this is mostly on ali extrusion or cast ali plate, so the height of the rail shouldn't be a massive issue - shim slightly if required, and you'll have plenty of left-right wiggle on the extrusion T slot. Your main problem will be aligning the two sides of the long axis so the rails sit at the same height. You won't go too far wrong if things move smoothly. Avoid software corrections as much as possible - the main one is the steps-per-unit, which shouldn't be too far from the calculated value (if it is very different, somethings probably lose).