how to twist large copper strands of wire together for crimping ?

[Kjelt]

Get the strands as straight as possible and seize them together in a bundle with a single strand of thin copper wire, rolling the bundle as you put on the seizing to get it as well compacted and circular as possible.   Done with care, thin enough wire and with the seizing as tight as possible (just short of breaking the fine wire), the increase of diameter due to the seizing wire thickness should be less than the decrease due to better alignment and packing and straight strands.

YES! This was the golden tip.
I untwisted 30cm the copper thin wire worked but slipped badly so I went to tiewraps and the diameter was just 0,3mm below the lug diameter.
First try I only missed two single 0,4mm wires so more than good enough! 

[Ian.M]

Get the strands as straight as possible and seize them together in a bundle with a single strand of thin copper wire, rolling the bundle as you put on the seizing to get it as well compacted and circular as possible.   Done with care, thin enough wire and with the seizing as tight as possible (just short of breaking the fine wire), the increase of diameter due to the seizing wire thickness should be less than the decrease due to better alignment and packing and straight strands.

YES! This was the golden tip.
I untwisted 30cm the copper thin wire worked but slipped badly so I went to tiewraps and the diameter was just 0,3mm below the lug diameter.
First try I only missed two single 0,4mm wires so more than good enough! 

Cable ties tend to make a non-circular bundle so may actually make the job harder, however they are useful to stop the end unlaying any further.

Start by taping the very end of the unlaid strands if they are difficult to keep all together.  Next  form a constrictor knot in the fine wire (which may be tinned wire but must not be enamelled).  Slip it over the unlaid end of the cable with the short end from the constrictor knot furthest from the cable end.  Tuck the short end back under the constrictor knot so it lies parallel to the cable strands, but leave a loop.  Work the knot up tight where you want the start of the seizing.  Working towards the end of the cable, put on about ten turns as a whipping, work them up tight and follow them with a half hitch so it so they cant loosen. After the first group of ten, pull on the tail of the short end to get rid of the loop, then cut the tail flush with the cable end. Repeat putting on ten turns of whipping then a half hitch till near the end of the cable. When you get near the end, remove the tape and, working with a long loop, lay the end of the fine wire parallel to the strands leading away from the cable end, using the loop, put on another ten turns, and a half hitch, then work them up tight and pull the tail through till the long loop is gone.

An alternative if there is very little clearance is to put on a plain whipping (no half hitches after the constrictor knot, and at the very end clip out a couple of wires from a strand to reduce the diameter for the last ten turns, and put them on working with a loop, over the fine wire end so you can pull the tail to lock them in place without increasing the diameter.   

If there is virtually no clearance, put on a whipping using whipping twine, waxed linen thread or waxed dental floss, starting from a constrictor knot but don't bury the tail.  Tape the end of the whipping to hold it, but *DONT* lock it with a half hitch or bury it or do anything to secure it that cant simply be unwound.  After trimming the end of the cable, slide it into the ferrule unwinding the whipping as you do so.


 

[Kjelt]

Cable ties tend to make a non-circular bundle so may actually make the job harder, however they are useful to stop the end unlaying any further.

Yes I was looking for a hose clamp, if that is the right english term, it is a metal ring you can adjust the diameter by rotating a screw. Did not have one.

     Start by taping  "........................   the whipping as you do so.     

Wow thanks  that is a very detailed description, I have to study on it esp with the english "knotting terminology" let me get back on that later

[Ian.M]

Edit above: I left out a sentence about repeating the ten turns then a half hitch.

If you have any problems with the terminology, I can attempt to explain any specifics.

Even a hose clamp is problematic unless you put it on over a thick layer of rubber tape, as the part under the screw tends to dig in, and the band tends to kink as its tightened if you need more than a small reduction in diameter.    A constrictor knot in thin nylon cord, braided fishing line, kite line or similar high strength thin cord is far more likely to result in a circular bundle, especially if you tightly tape the bundle first so the knot cant dig in easily.  Tighten it using two spikes (a pair of stout Philips screwdrivers will do), wrapping each end of the line several times around the shank and holding the tail against the grip, then lever against two sides or ends of a solid block placed under the cable.

[T3sl4co1l]

If this is for EMC, you're not gaining much over a handful of hookup wires.  It's good galvanic grounding, but not much better than spaghetti once you get up to some MHz (give or take length).

For that, you really need metal boxes (grind off the paint where connections are made), metal connectors and conduits.  Treat it like RF transmission lines, because, well, it always is.  Conduits serve a particularly useful purpose: all the currents are gathered together within, and you can insert a line filter there, if you like (and if applicable).  Do this for the mains wiring, and the control signals if they need it (preferably, they're isolated and aren't carrying much noise).

Tim

[Kjelt]

If this is for EMC, you're not gaining much over a handful of hookup wires.  It's good galvanic grounding, but not much better than spaghetti once you get up to some MHz (give or take length).
For that, you really need metal boxes (grind off the paint where connections are made), metal connectors and conduits.  Treat it like RF transmission lines, because, well, it always is.  Conduits serve a particularly useful purpose: all the currents are gathered together within, and you can insert a line filter there, if you like (and if applicable).  Do this for the mains wiring, and the control signals if they need it (preferably, they're isolated and aren't carrying much noise).
Tim

All the machineboxes are Rittal full metal enclosures with mains filtering, I got two: one controlling the milling spindle (very noisy frequency converter) and one for the control and powering the motors. Then there is the target machine the CNC machine from aluminium profile on a metal stand. So the purpose of this cable is to connect the metal spindle box, the cnc machine and metal stand all together to the metal control box in order to get a star grounding and low impedance between all the parts. This would ideally be Litze but this will also be good enough (i hope).
So your concerns are valid but addressed inside the boxes which is a bit offtopic.

[Kjelt]

If you have any problems with the terminology, I can attempt to explain any specifics. 

Thanks I am a vision person so a picture tells me 100 times more than words , so thank you for the knot website that was very helpfull.
The ten turns are those on one spot so over eachother to fix or do you mean over a length of wire so kind of circelling, I am trying to visualize the end result.
If you happen to have a picture of one of your previous cablefixes that would be great.

[Ian.M]

The turns go on flat and close spaced (touching side by side) as-if you were winding a helical tension spring.  DO NOT let them pile up in one spot as that would stop it fitting cleanly into the ferrule.

Unfortunately I rarely document the simple stuff photographically, and I'm fairly sure I don't have the right materials on hand to prepare an example as I don't currently have an offcut of fine stranded heavy cable, or suitable bare fine wire in stock.   Also although I have a good mental 3D model of it, I cant draw in perspective well enough to convey it clearly.

[Kjelt]

Hi Ian,
not a problem I will be experimenting, I got a good idea what needs to happen, almost there yesterday so confident I will get it going.
For now I have some work to do  So it will probably be after the weekend I can have another go at it again.

[T3sl4co1l]

All the machineboxes are Rittal full metal enclosures with mains filtering, I got two: one controlling the milling spindle (very noisy frequency converter) and one for the control and powering the motors. Then there is the target machine the CNC machine from aluminium profile on a metal stand. So the purpose of this cable is to connect the metal spindle box, the cnc machine and metal stand all together to the metal control box in order to get a star grounding and low impedance between all the parts. This would ideally be Litze but this will also be good enough (i hope).
So your concerns are valid but addressed inside the boxes which is a bit offtopic.

Can you diagram this?  Since as you say,

Thanks I am a vision person so a picture tells me 100 times more than words

My first instinct is, if everything already is metal conduits, then what's the cable for?  You still want a ground wire in the conduits and boxes, for safety purposes, but they're not there for RF purposes of course.

Tim

[Kjelt]

Can you diagram this? 

I am at work will be after the weekend probably. But I will try to explain below in words.

My first instinct is, if everything already is metal conduits, then what's the cable for? 
You still want a ground wire in the conduits and boxes, for safety purposes, but they're not there for RF purposes of course.

Before these copper cables will be attached the situation is:
Metal box 1 : grounded through earth (mains) but 2,5mm2 cable to mains junction box or how you call that.
Metal box 2: grounded through earth (mains) but 2,5mm2 cable to mains junction box.
CNC machine not grounded at all.
CNC stand not grounded at all.

The cables from box1 and 2 are shielded but the shield is ONLY connected on the connector on the box side, NOT on the cnc machine side. This is to make sure no equalisation currents are running but I still want to shield the environment from the noisy vdf and motor cables. This is discutable since I know that one side shielding will not protect from some fields so my other option is to run an extra earth cable in the cable (so 4 pole instead of 3 pole) and hook the earth cable to the shield at the motor or spindle connector, but anyway the result is the same there is no ground-loop current running from the box to the cnc machine, they are seperated.

Now the motors and spindle are mounted on the cnc machine so their body is hard connected to the cnc machine chassis, so the chassis should be grounded, for safety but also for emc I guess (correct me if I am wrong since I am not an EMC expert at all).

So to get an equivalent potential on the boxes and the cnc machine in respect to earth/ground the CNC machines and boxes have to be connected together in a star topology.
The Star point is the ControlBox earth bond point, that will be the point the large copper cables will go to the stand and the chassis of the cnc machine and vfd box.
So that is what these cables are for to simulate as if they all three were bolted together on a metal chassis.
Hope that is a bit more clear? If I make wrong assumptions or design error I would like to know that ofcourse. 

[T3sl4co1l]

So there are two mains circuits, from the main junction panel, one to each box?  Are these metallic conduit?  (If the line filters are doing their job, this shouldn't matter.)

You didn't list more connections, how do "Metal box 1" and "Metal box 2" connect to "CNC machine" and "CNC stand"?

Is "CNC machine" just sitting on top of "CNC stand", as the names suggest?  Is the stand incidental, or is there anything electrical on it, too?

You also later mention a ControlBox but didn't say how it connects.

It sounds like you're being very paranoid about ground loops, as in safety ground.  I don't know of any problem with bonding already-grounded appliances together (but, someone better versed in CE code, do chime in?).  It doesn't sound like you're measuring tiny, low-frequency signals between boxes, so ground loop (as might apply to audio systems) is a non-issue.

The reason I'm asking very precisely is, whenever I see "star ground" I am instantly suspicious.  Star grounding is almost exclusively applied badly.  Like slotted ground planes, it's a massive pitfall for newbies.  Yes, it's a thing; and the designer should know about it.  The technician even, say when it comes to pro audio systems, or setting up lab or test equipment.  It is a possible option, another tool in the toolbox.  But as with all tools, one must be keenly aware of its conditions, traps and downsides.  And these gotchas are far more numerous, for star grounds and slotted grounds and such, than the benefits are.  And because there are so many gotchas, it's very likely one or more will always be missed.

It's an electronic design equivalent of Roko's Basilisk.  It's literally a kind of story that is better left untold, so people don't try to use it just because they, I don't know, think it's neat?

So that's where I'm coming from, and just want to sanity-check here.

By far, the biggest sin people commit, regarding star grounding, is the same problem as with slotted grounds: running connections between sides, rather than following the ground path.  Crossing traces over a slot, takes exactly the current you meant to avoid, and puts it into the traces (or vice versa); connecting a cable between two boxes, that are star grounded separately, creates a loop that's altogether worse than, say, routing the ground with the cable itself (or, better yet, running the cable and ground together through a conduit).

So, at this point, it seems like you'd most benefit from using metallic conduit (flexible if needed) between the various boxes and machines, so you can get the drivers' noise currents back from the motors, inside a conduit, rather than with a huge loop around all kinds of metalwork, that will radiate.  Or if conduit is unacceptable, filtering the offending sources (but, that's probably even less practical -- you can't simply hang a line filter on a VFD output, unfortunately).

Tim

[X]

With thick cables, I score the insulation with the strippers, then twist the insulation while I am removing it. No waste, and the twist looks very clean, but you may need to practice to avoid broken strands.
EDIT: I should mention, it works well even for bog-standard 14-strand security cable. This method tends to be difficult with very thin cable (eg. found in chinese USB cords).

[Docara]

NO NO  NO

You're not working on yuor car!

On cables the size you're dealing with - you should NEVER twist cables together or put two cables in one connector - that's it! Its against the wiring regulations.

When crimping is done properly it forms a better joint (than even soldering) with much lower volt drop across the joint. When crimping is done correctly it fuses (practically welds) the copper to the lug without stressing the cores or causing the joint to over heat in use.
The lug size is closely matched to the cable CSA witch should be matched to the (hydraulic) crimper you use. The top companies match their lugs to crimper. If you are using tri-rated or highly multi-stranded cable like you are, then you should also opt for a heavy duty lug which helps to maintain to longevity of the joint.

Now taking the twisting point further.
By twisting the cores inevitably you will have cores crossed over each other - probably multiple cores in practice. This causes local high pressure stress points when crimped together with a cable which is thicker than the lug was designed for. The crossing of cores will result in some cores breaking when crimped causing drop in current carrying capacity of the joint. Having a lug over compressed because the cable is thicker than it was designed for can also be a point of failure because the copper can be compressed enough to crack or cores break again causing a drop in current carrying capacity of the joint. The joint will probably overheat in operation as a consequence of the joint becoming slightly looser in operation.

'Doubling up' cables.

Again as mentioned above lugs are designed for cables od a certain Cross Sectional Area. By forcing cables into a lug which it (they ) are not designed for is increadable bad practice and potentially dangerous for the reason stated above. I am an Electrical Test an Inpection electrician based in the UK and I would fail a joint crimped like this.

The proper way to terminate your cables is to individually crimp them with the correct size lugs and connect them together with a Bus Bar or similar.

[Kjelt]

Well noted and you're right.
If the cables were used for high amperage I would certainly do so, however the cable is for ground potential equivalence, highest current would be perhaps a few mA.

[senso]

If its for grounding and you expect a couple mA, why dont you just use 4mm^2 cable with m8 eyelets and crimp it properly and just run a wire from the cnc base to one of the Rital enclosures?

[Docara]

What does high amperage have to do with what your doing. A crimp joint is a crimp joint.

What is you application / installation - metal work in a field or machinery?

 

[Kjelt]

That has all been answered in posts above, it is about mutual grounding and EMC, bringing the same grounding to three metal enclosures and chassis together that are connected to a cnc machine.
When I have time I will open a new topic with again more details since the questions and challenges from T3sl4co1l are quite relevant.

[BradC]

Get the strands as straight as possible and seize them together in a bundle with a single strand of thin copper wire, rolling the bundle as you put on the seizing to get it as well compacted and circular as possible.   Done with care, thin enough wire and with the seizing as tight as possible (just short of breaking the fine wire), the increase of diameter due to the seizing wire thickness should be less than the decrease due to better alignment and packing and straight strands.

Just a loud vote of thanks for this concept.
This morning I had to re-do an improperly crimped connection (a mangled 50mm² lug on a 35mm² cable) and I was having a pig of a time getting the cable into a correctly sized lug. It's one of those super flexible cables with a gazillion strands and it's a nightmare to work with. I bound it tightly with some 0.25mm diameter (30AWG) tinned copper wire and it slid right in. I've fought these connections for years, so have a virtual beer on me!

On that note, I can't believe the amount of times I've seen way oversized crimps on cable that is obviously too small. Maybe if everyone knew this trick they'd use the right sized connectors.