PDA

View Full Version : Earthing?



manofgresley
03-09-2012, 10:41 AM
Hi again.

I need advice on earthing, generally I am ok, but being new to electronics I thought I would get a second opinion.
My question is this,my PSU, Drivers etc; are enclosed in a metal cabinet. I ordered an On/Off switch for the cabinet, it arrived with 2 poles only. The cabinet is powered from a 3 pin mains socket, can I use the earth cable of a 3 wire circuit, to earth the metal cabinet, by attaching the cable through a hole in the cabinet wall, and attaching with a bolt/screw?

6721 This is what I have now.
Regards

Ray

m_c
03-09-2012, 11:32 AM
Your main earth wire should go to a central (star) ground point within the box, and should be permanently connected.

Web Goblin
03-09-2012, 12:02 PM
A bit more information required. Are you directly connecting your power supply for the machine to the socket circuit or are you connecting it via a 13amp plug?

manofgresley
03-09-2012, 04:07 PM
Hi

Thanks for reply, excuse my ignorance what is a STAR?

Regards

Ray

manofgresley
03-09-2012, 04:08 PM
Hi
Thanks for reply, i am connecting via 13 amp plug.

Regards

Ray

JAZZCNC
03-09-2012, 05:40 PM
Hi

Thanks for reply, excuse my ignorance what is a STAR?

Regards

Ray

Hi Ray,

Star means all AC Earths come from one point "Star point". This is very important so you don't introduce electrical noise into the system thru creating ground loops.
The signals coming and going from control software thru the parallel port to drives or sensing switchs like home switchs etc are very sensitive to electrical noise. This has the affect of tricking the motors, sensors etc into false moves or switching when shouldn't and cause real head scratching problems.

Mostly when people have missed steps or false e-stops etc it often turns out to be bad earthing practices. This is also why it's important to use sheilded cables for stepper motors and spindles ETC and correctly route back to the star earth point. (See quick pic showing)

NOW don't confuse AC Earth with the " - " or negative of the DC side of the system. While they ultimately all go back to the same Earth point they should be treat has separate and any DC negatives are best taken from a DC negative point. IE from the BOB Com or " - " pin or the " - " negative of the DC side of a PSU.
Often I'll run a single wire from the Main PSU DC negative to a distribution rail where I then pick up all the other DC negatives I need for drives ETC. . . . I do the same for DC positive as well.

All shielded cables should only be earthed at one end and that taken back to the main earth star point.

Hope the rough sketch helps.?

Robin Hewitt
03-09-2012, 06:37 PM
Earth loops are always trying to sneak in and act like shorted out transformers. For example, an AC electric motor has an earth connection that goes to it's casing and that casing connects to the machine body which has a seperate ground connection. That is an earth loop with some major magnetic field activity in one corner and current starts to flow. Not only does the motor earth wire heat up but so do any other loops in the system, especially if they pass through the machine body. The Volts may be low but the current can be enormous.

JAZZCNC
03-09-2012, 07:43 PM
Earth loops are always trying to sneak in and act like shorted out transformers. For example, an AC electric motor has an earth connection that goes to it's casing and that casing connects to the machine body which has a seperate ground connection.

Yep it's possible but lets not confuse the poor bloke any more than need be Robin.!! . . . In the main If he sticks to the star grounding then he shouldn't have any issue's.

Web Goblin
03-09-2012, 08:06 PM
The guys going to be confused if he stays around here much longer anyway:confusion:

Jazz got it with his description of earthing. If you follow that you should be ok. Your original drawing made me think that you were connecting directly to the ring final circuit.

manofgresley
03-09-2012, 08:19 PM
To all that have helped me so far, many thanks, it as restored my faith in human nature. It's great to know strangers can help each other.

Best Regards

Ray

motoxy
03-09-2012, 09:17 PM
Trust me they don't come any stranger :witless:

Edward
13-01-2017, 04:03 PM
After reading these post it's all becoming a little clearer, so many thanks.

A few questions that may be a bit naive, but important for me to understand things better:

-I'm in the process of putting all the elements into an enclosure. From a practical point of view, should I buy a plastic or metal enclosure?

-Regarding the star ground connection, if the "star" point is a bolt making contact with the metal enclosure, to which all the grounds are connected, and the 3 drivers metal base are also touching the metal enclosure, does that affect the efficiency of the star connection in any way?

-Also, my 68V linear PSU (Leadshine) the mains input Ac doesn't have an earth connection point so only the Live and Neutral are used, shouldn't the Earth wire be connected to something? When it finally goes into the enclosure the 240V mains enclosure input socket will be earthed to the star point, so I guess that's all that's needed as far as the drivers PSU?

-On my vertical mill spindle motor, there is a ground wire bolted to the back of the motor. Why do you need this if the motor casing itself is already touching the mill metal body, which presumably is in turn also grounded somewhere. Is there a star arrangement within the mill itself?

I can't think of anything else to ask...wait a minute, I notice in the schematic (Jazz) you have a fuse to the PSU, etc. Is this essential or recommended since the PSU is already connected to a fused mains socket?

Edward

Neale
13-01-2017, 04:27 PM
OK, I'll kick off but there will be plenty of opinions given to this list of questions! One point to remember is that sometimes you earth for electrical safety and sometimes for screening/electrical noise reduction. Good earthing practice via star point will generally address both concerns.

Metal case - shielding comes for free that way. You will probably need a bigger box than you thought, and not sure about strength of larger plastic boxes.

Driver enclosure is isolated from all internal electronics so doesn't matter whether it's connected to ground or not. Bolting an anodised enclosure to, say, an epoxy-coated steel box might or might not give a decent electrical connection but I wouldn't worry about it myself. I didn't in my own control box.

I reckon that anything that has mains going in should be earthed, so if your PSU doesn't have a connection for mains earth, then the case should probably be earthed (via a wire to a mounting bolt? See comments about bolting to epoxy-coated steel) anyway, and that would go to the star point, where the mains earth is also connected.

My first router was built from MDF so the spindle motor (usual 2.2KW water-cooled) was isolated from ground. Under certain circumstances, you could feel a slight tingle from it when you touched it while running, which was some kind of electrostatic or electromagnetic leakage from the power feed to it. Not dangerous but mildly inconvenient. Decent earth connection immediately stops this kind of thing. Also, if you are going to use a touchplate for tool height setting, the spindle must be earthed. It might be that your mill builder took a belt-and-braces approach to earthing, but it might be because it's better to avoid relying on a metal-metal contact in sliding surfaces which, theoretically, are held apart by a tiny layer of lubricant. My new all-steel router uses earth wires running through the various cable chains to make sure that there is an earth wire between all pairs of moving parts to avoid relying on contact via the profile rails and bearings. Like that, there is a solid electrical earth run from spindle platform right back to the main router frame which is in turn earthed inside the control box by a wire from a mounting bolt back to the star point.

I use a 10A (I think) MCB in the control box as the mains enters. It's not clear to me what use further fuses are as any average electronic item will blow in a thousandth of the time it takes a fuse to react; fuses do not protect electronics, faulty electronics blow fuses (occasionally). Fuses sit there saying, "I wonder where all that magic smoke is coming from? Good job I'm still working!" Primarily, they might stop a fire if there is a major wiring fault causing a sufficiently high current short.

One man's view...

Clive S
13-01-2017, 04:31 PM
Ok A metal enclosure is best by far just make sure it is big enough 600 x 600 x 200 is popular.

The star point is where you connect all the cable screens to including the mains in earth.

Fuses I personally rely on the MCB in the box for the power supply fuse But it does no harm to fuse the smaller PS ie 5V or 24V desperately (good to get din rail type for these)

Regarding the mill the earth to the mains frame is probably for the electronics in the speed controller that might be on a swivel type mounting. You would be powering (mill) this from a separate 13A plug.

Edward
13-01-2017, 05:14 PM
Thank you for the helpful answers:)

However, could I ask further about the PSU earthing? This is the PSU which I bought from Zapp ready-made for convenience:

http://www.zappautomation.co.uk/electrical-products/power-supplies-43/linear-power-supplies/ps806-12-linear-power-supply.html

As you can see, it has no metal case, it all sits on the pcb and I can't see an earth point. Maybe this one has internal earthing, if such a thing exists?

I see you are both using MCB's that's cool, I'll have to investigate this. At the moment, the limit switches (micro type) IO's are fed by the 3.3V supplied by the Kflop board so I don't need additional low DC supply. However, I am going to use proximity sensors instead, and that will require a 24V supply. So as I expand and start to use relays to trigger this and that, I will need an enclosure that allows for this expansion, even though at this stage it may look huge and rather empty.

Another question regarding proximity sensors, I ordered a couple of the cheapo ebay ones for less than 10 quid each to play with, a bit bulky, but are they any good? I notice that some proximity sensors can cost a small fortune. For our type of usage, what does an expensive sensor offer that the cheapo ones lack?

Edward

JAZZCNC
13-01-2017, 05:48 PM
Hi Edward.

The PSU will need mounting either directly to the back plate using stand offs or on separate backplate(with standoffs) which then mounts to enclosure backplate or Cabinet side. My prefered way is to mount on separate plate then fasten this to cabinet side as it saves space.
In either case if Metal enclosure you won't need Separate earth as the Whole Box will be Earthed. However if you want to bolt the gate so to speak then run separet wire from Star point to one of the holding boltings. !!!!!! WARNING !!!!! DO NOT put Earth wire to the Bolt which holds transformer.!

If Plastic case and PSU isloated then Good idea to run Wire from Star point to the point plate.

The main thing is you don't have Wires Earthing to different points, if running earthing wire then always to Star point.
Personaly Even with metal back plate and the component does have Earth connection I always take it from the Star point with wire and never rely on the back plate for Earth.
Same with Shield cables, while technicly could connect to back plate I don't. Always back to Star point. This way I never have any issues or chance for ground loops to occur.
Proud to say never have any ground faults or false trips on my machines and these little details are mostly why. These and good components.

In general you want to Fuse all AC devices with MCB ie: Main PSU, 24Vpsu, VFD etc. The DC devices like Drives, fans then I use Glass type fuses.

Regarding the cheap Prox switch then depends on what usage and accuracy/repeatabilty your chasing.?
If just limits then yes they are fine. If Router/Plasma etc then they are also fine for Homing. But if it's Mill and your chasing V-High repeatabilty on Home Sw then may pay to buy better quality.

Neale
13-01-2017, 06:16 PM
I am using one of those power supplies myself, and I have just bolted it to the case on stand-offs. Works fine for me. Doesn't need earthing as there is no external metalwork that is likely to come into contact with mains voltages.

I have also just come in from the garage where I have been setting up my new machine. As part of squaring the gantry, I have been checking repeatability of position of the gantry using a dial gauge. I'm using cheap proximity switches (mine cost something like £20 for 10 when I bought them a couple of years back but they are a bit suspect so I've just bought some more at about £10 for 4 although I haven't tested these yet). You can't expect wonders at these sorts of prices, but doing a set of runs where I move the gantry down the X rails a distance, then home it, I'm seeing something like 0.02mm repeatability in the homed position. However, I haven't gone round and checked things like the ballscrew bearing nuts since I first put it together and there might be a bit of play as the bearings settle in after a little bit of use. There's also a bit of uncertainty in the position as my dial gauge isn't ideally positioned and isn't as solidly mounted as it might be. Nevertheless, that's probably acceptable for a machine primarily intended for wood.

Edward
13-01-2017, 06:17 PM
Thank you Dean, for the clear advice and for the warning about the transformer! I've learned so much from this forum.

A metal enclosure it will be, more expensive, but worth it.

I think I will try the cheapo prox sensors first, as I am more interested in them acting as limit switches, at least at this stage. I'll keep you posted.

Oh, a last question unrelated to this. What is the average or expected backlash on bog standard ballscrews C7, chinese. My backlash is between 0.045mm and 0.054mm for each axis, i.e. including the bearings, etc. Is that reasonable?

Edward

Edward
13-01-2017, 06:26 PM
Hi Neale,

Good to know about the PSU. I will use standoffs too.

The prox sensors I ordered are these:

http://www.ebay.co.uk/itm/162016718402?_trksid=p2060353.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

Maybe yours are something similar. I will play with them when I receive them before I commit. If I get 0.02mm repeatability I will be very happy:)

Edward

JAZZCNC
13-01-2017, 07:01 PM
I think I will try the cheapo prox sensors first, as I am more interested in them acting as limit switches, at least at this stage. I'll keep you posted.

They will be fine but bit of advise Fit Home switches ASAP. Infact if I could only choose one it would be Home every time. Far more usful than Limits and can use Softlimits to put you one until Limits are fitted.


Oh, a last question unrelated to this. What is the average or expected backlash on bog standard ballscrews C7, chinese. My backlash is between 0.045mm and 0.054mm for each axis, i.e. including the bearings, etc. Is that reasonable?

Edward

About good as it gets for Chinese C7. Acceptabe or not again depends on usage. For Router yes more than enough. For Mill then might want to chase higher but again depends on what your doing and prepared to accept.!

Didn't see you say what type of machine.?

Edward
13-01-2017, 07:28 PM
Hi Dean,
Yes, I can designate them as Home or limit switches, or both. I think Mach3 works similarly. With Kflop for instance, you assign a Home button to execute a C file for homing that basically reads the sensors as home switches, homes the machine, then pulls off a little, zeros the coordinates and then turns the sensors into limit switches.

The mill is a Sieg2.7 conversion, which is just a little smaller than an X3, but with better Y range and I think a similar brushless motor, 750W. It's massively better than the SX2 in every sense, for those who want to know. I haven't actually cut anything with it yet, though everything moves as expected. I want to get the electronics in a box to get them away from the muck. In terms of accuracy, well, I am not expecting great miracles, but it's as neat as one can do...for the money spent. We shall see what the pieces look like, the time it takes to produce them and so on, it's my first foray into CNC.

I have been playing with the backlash compensation, and it's pretty unbelievable what software can do! The compensation happens in milliseconds, and using a dial gauge, the backlash "removal" is bang on. I don't know how good it will be in real life, cutting a circle, changing direction, for example.

Edward

JAZZCNC
13-01-2017, 07:58 PM
I have been playing with the backlash compensation, and it's pretty unbelievable what software can do! The compensation happens in milliseconds, and using a dial gauge, the backlash "removal" is bang on. I don't know how good it will be in real life, cutting a circle, changing direction, for example.

Backlash Comp is a fudge at best, better if can remove any slop.! . . Not seen one yet that will get of rid the chatter. . .Lol

Robin Hewitt
14-01-2017, 10:17 PM
If you have tight Gibb strips then you have backlash. No Gibb's then what you have is slop :wink:

If you have a circuit going through a hole in a metal box, I suggest you make sure that both wires go through the same hole.

Clive S
14-01-2017, 10:38 PM
If you have a circuit going through a hole in a metal box, I suggest you make sure that both wires go through the same hole.Is that because you will get a ground loop:whistle:

Robin Hewitt
15-01-2017, 10:57 AM
I cannot pretend to understand metal boxes and the strange things that happen when you drill holes and run wires through them. Putting the whole circuit through a single hole sounds like a good idea, balance things out, like an earth leakage trip.

OTOH, sounding like a good idea is not always enough. Recently there was a lovely picture doing the rounds of a plated steel nut and bolt holding an earth loop to a copper bus bar. Should have used copper, they used to use copper nuts and bolts in that golden age of understanding when I were a lad. In the picture dissimilar metals had formed a thermocouple. The earth loop probably only gave a fraction of a Volt but the current was enough to heat the nut red hot :nightmare:

Neale
15-01-2017, 12:15 PM
Electric current through a wire creates a magnetic field around it. Current flowing in opposite direction creates an equal and opposite field. Keep the wires close, ideally twisted together, and the magnetic fields cancel out, don't transmit interference, and are resistant to outside magnetic fields for the same reason (rather simplified but the principle is true). Pass wires through separate holes and not only do you lose that "cancellation" effect because the wires are further apart, but you now create a magnetic field between the two wires that can induce (= create) a current in the metal box between them, which is exactly what you don't want to do. I can bang on about differential signalling and common-mode rejection, but the simple model is pretty sound to understand what's happening.

Edward
15-01-2017, 12:50 PM
I can't see why anyone would want to separate wires of a circuit through different holes or not....

A practical question about wiring the 8 wires of a stepper in parallel:

For short run desk tests I normally join and solder each pair length as supplied to go to the respective A+,A-,B+,B- twisting it together along the way, as Neale mentioned above. But now I have to do it properly using 4 core screened CY cable, 1.5mm2. which I think is rated at a whopping 18A. Should I cut the motor cables as close to the motor source as possible, solder each pair to one of the four cores of the CY cable and use heatshrink on each joint, and then a bigger dia heatshrink to hold together the lot? This bulky point where all the cables join at the motor side, and the screening ends, isn't it a possible source of interference?

If so, is there a recommended best way to do this?

In terms of magnetic fields and interference, is it better to instead use 8 core CY cable all the way to the driver but of a smaller cross sectional area?

Edward

Robin Hewitt
15-01-2017, 01:27 PM
But now I have to do it properly using 4 core screened CY cable, 1.5mm2.


Why is that "properly"? I like big fat wires, as big as will fit in the screw terminals on the driver. I run my big fat wires through flexi-plastic conduit to keep it tidy.

Why would you want screened? Armoured would be more understandable.

If you are worried about interfering with your high speed computer connections, you could screen those perhaps.

Edward
15-01-2017, 01:48 PM
Robin, now you are confusing me:)

As far as I understand, "armoured" hardly offers any electromagnetic containment. All signal cables will be screened. I will use flexi conduit for the motor wires too. Do you mean you run 8 separate fat wires inside the conduit?

Edward

Clive S
15-01-2017, 02:36 PM
Robin, now you are confusing me:)

As far as I understand, "armoured" hardly offers any electromagnetic containment. All signal cables will be screened. I will use flexi conduit for the motor wires too. Do you mean you run 8 separate fat wires inside the conduit?

EdwardEdward just do as you suggested chop the wires off about 250mm from the motor and then join them to the CY cable personally for the motors you are using I would go with 1mm size. If you want to be posh you can get an end cover for the motors and take the cy cable into that.
2037020371

Edward
15-01-2017, 02:51 PM
Thanks Clive.

The 3.1Nm motor would draw a maximum of 4.2A, but we all know that in reality it's probably about half of that anyway.

So a cable of 1mm2 (normally rated for at least 10A) will more than cover this. However, I wonder if there is any benefit in using a thicker cable, other than robustness...or obsession:)


I like the motor covers with the glands. Is this something you printed yourself, or are they readily available for the Nema23 frame? BTW, my Z motor is to the side of the column, just like in your picture.

Edward

Edward
15-01-2017, 02:56 PM
I notice that Robocutters do sell similar motor covers


https://robocutters.co.uk/proddetail.asp?prod=CoverKitNema23

Clive S
15-01-2017, 03:00 PM
Thanks Clive.

The 3.1Nm motor would draw a maximum of 4.2A, but we all know that in reality it's probably about half of that anyway.

So a cable of 1mm2 (normally rated for at least 10A) will more than cover this. However, I wonder if there is any benefit in using a thicker cable, other than robustness...or obsession:)


I like the motor covers with the glands. Is this something you printed yourself, or are they readily available for the Nema23 frame? BTW, my Z motor is to the side of the column, just like in your picture.

Edward

I printed these a few years ago but since then I have milled them from various blocks of plastic including those fencing posts made from car tires. If you do a search for hem I have seen them in Europe somewhere

https://www.hardware-cnc.nl/en/winkel/stappenmotoren/steppermotors/motor-cover-nema-23-detail

http://www.thingiverse.com/thing:243126

Robin Hewitt
15-01-2017, 03:13 PM
I wonder if there is any benefit in using a thicker cable, other than robustness...or obsession:)

Now't wrong with a bit of healthy obsession :02.47-tranquillity:

Thin cable is fine and dandy if you were lighting a bulb or charging your iPhone with a bit of AC but this cable has more to do than that.

On the other end of it is a hefty inductive load, it may have tuned itself a bit to reduce the peak but sooner or later it is going to want to reverse the current in a coil winding PDQ. It will try to control that current using a chopper circuit which depends on measuring the current flowing some way away. Did you fit the optional braking resistors? :tiger:

Fat, low resistance cable takes the strain off more expensive components.

Edward
15-01-2017, 03:42 PM
I don't have braking resistors.

Here is a pic of my conversion, the Sieg SX2.7 with the drivers, etc waiting to go in a suitable box.

Edward

20372

JAZZCNC
15-01-2017, 03:54 PM
Now't wrong with a bit of healthy obsession :02.47-tranquillity:

Thin cable is fine and dandy if you were lighting a bulb or charging your iPhone with a bit of AC but this cable has more to do than that.

On the other end of it is a hefty inductive load, it may have tuned itself a bit to reduce the peak but sooner or later it is going to want to reverse the current in a coil winding PDQ. It will try to control that current using a chopper circuit which depends on measuring the current flowing some way away. Did you fit the optional braking resistors? :tiger:

Fat, low resistance cable takes the strain off more expensive components.

Robin your talking shite again and confusing people with your ridiculous answers.!. . . If your not going to help people with relavant or helpful replys then why don't you STFU.!!

Robin Hewitt
15-01-2017, 06:14 PM
Robin your talking shite again and confusing people with your ridiculous answers.!.

Hi Dean

I knew you would be along to explain why everyone seems hell bent on screening and tin boxes despite the problems they cause. This is something that has bothered me for some time, a trap I think I have fallen in to myself.

Ready when you are, I am all ears... :joyous:

Robin

Clive S
15-01-2017, 06:53 PM
Did you fit the optional braking resistors? Robin would you please explain how to fix a braking resistor to the motor in this case :apthy:

JAZZCNC
15-01-2017, 08:26 PM
Hi Dean

I knew you would be along to explain why everyone seems hell bent on screening and tin boxes despite the problems they cause. This is something that has bothered me for some time, a trap I think I have fallen in to myself.

Ready when you are, I am all ears... :joyous:

Robin

Not even going to get into it with you Robin because it's well documented why shielded cable is required. To say it's not needed is like saying this Gun may or may not have bullets put it to your head pull trigger to find out. . :cower:

Robin Hewitt
15-01-2017, 10:23 PM
Robin would you please explaining how to fix a braking resistor to the motor in this case :apthy:

You can fit a braking resistor if the driver black box has provision for a braking resistor. A braking resistor dissipates unwanted energy and is one of the reasons a body might chose thicker wiring.

JAZZCNC
15-01-2017, 11:00 PM
You can fit a braking resistor if the driver black box has provision for a braking resistor. A braking resistor dissipates unwanted energy and is one of the reasons a body might chose thicker wiring.

Exactly Clives point I think.!!! . . . There isn't a blood option for braking resistor on AM882 and tell the last time you or anyone else seen Nema 23 3Nm stepper motor fitted to Minimill using braking resisitor.!! . . . . . STOP TALKING SHIT.

Edward
15-01-2017, 11:12 PM
Hahaha, this is getting interesting:)
In the EM806 and AM882 don't you have an over-current cut off with the red LED blinking? I think you really have to push them really hard, for instance, with high deceleration to get to that stage. I don't intend to drive this machine too madly:)
Also, I guess the psu will be able to cope with a bit of back EMF if I feel too adventurous...

Edward

Clive S
15-01-2017, 11:29 PM
You can fit a braking resistor if the driver black box has provision for a braking resistor. A braking resistor dissipates unwanted energy and is one of the reasons a body might chose thicker wiring.

Robin Can you please send me some of that shit you are on as it just might help me to understand what the f@*k you a taking about.

Robin Hewitt
15-01-2017, 11:55 PM
Robin Can you please send me some of that shit you are on as it just might help me to understand what the f@*k you a taking about.

I keep taking breaks from PCB design. I am trying to wire up a 168MHz processor without resort to 4 board layers and it is not easy getting sufficient decoupling and copper down to get the power everywhere it needs to be. I am having real problems while you lot are having imaginary problems and I guess I am getting a bit jealous.

To quote Jesus, purely because it is hard to argue with Jesus, "You're far too keen on where and how, but not so hot on why" :nonchalance:

Clive S
15-01-2017, 11:58 PM
To quote Jesus, purely because it is hard to argue with Jesus, "You're far too keen on where and how, but not so hot on why" Can you point me to the document where HE wrote that please.

Robin Hewitt
16-01-2017, 12:20 AM
You don't believe me? Google is your friend :hysterical:

JAZZCNC
16-01-2017, 12:25 AM
To quote Jesus, purely because it is hard to argue with Jesus, "You're far too keen on where and how, but not so hot on why" :nonchalance:

I wonder WHY all the time.? . . . Often wonder WTF your on about and WHY you say the Shite you do.!! . . . . . Go back to the Stone age because your just confusing new users with BUM info that isn't relative to todays modern components.!! . . . . . Sorry to be so harsh but your going to cost somebody lot of wasted time and possibly money with such Rubbish.

Now that's the last I'm saying so OP can get back on track.

Edward:

You won't get near taxing those drives with Mill this size so don't worry about Over current etc. Just fit and enjoy.!

Regards Earthing then just use Star ground and shielded cable and you won't have any troubles. Size wise then 1mm is fine for motors and limits etc so just buy one size as it's easier and often cheaper as Some suppliers have minimum qty. 4-Core will do both

Edward
16-01-2017, 12:33 AM
Brilliant, thanks Dean and everybody, I'll just get on with it.

I think this thread is quite old, but the OP asked questions that I was also interested in.

Edward

Robin Hewitt
16-01-2017, 01:05 AM
Calling me names and telling me I am stupid rather than explaining why I am wrong does rather suggest that you do not have a clue what you are talking about. Your appeal to authority, "because it's well documented why shielded cable is required" is a fallacious argument. Where is this document?

Neale
16-01-2017, 09:15 AM
I like to understand "why", because that helps me know what and how. Trouble with earthing and screening, though, is that you can still get odd problems because you can't actually see or measure what's happening, particularly with intermittent problems.

Been testing my new machine, checking machine limits. X axis kept tripping around 20mm or so before it reached the limit (proximity) switch, although it wasn't clear which switch had actually tripped. Cables are all screened CY but run in same cable chain - limits in one cable, with motor Y and Z cables, but separated as far as I could by spindle cooling tubes. No other tripping problems, just this one point near the upper end of X travel. However, at this point the unrolling cable chain was next to the cable to X proximity switch (the usual unscreened thin stuff); I re-routed the proximity switch cable and the problem went away. Similarly, my earlier problems with home switch triggering went away when I moved the control box from its test position on floor in front of machine to its installed position bolted to frame with earth connection to star point. Moral of story? Use good earthing practice, screening helps, but be alert for the unexpected just in case.

I'm using 1.0mm CY for my Nema 23/EM806 combination, but buggered if I can work out where to put my braking resistors...

Robin Hewitt
16-01-2017, 10:42 AM
My mill either works or doesn't work depending on the route taken by the cable connecting it to the computer.
Everything was fine and dandy until I started putting things in boxes.

Clive S
16-01-2017, 10:53 AM
I'm using 1.0mm CY for my Nema 23/EM806 combination, but buggered if I can work out where to put my braking resistors..
It's obvious Neile in the bin of course:yahoo:

Robin Hewitt
16-01-2017, 11:17 AM
It's obvious Neile in the bin of course:yahoo:

Just because you cannot see the braking resistors does not mean you do not have them. It is all very well you behaving like a sheep on the border collie's day off, but it is only through understanding why things do not work as well as expected that we can improve them. No kind way to say this, you are a Luddite.

Edward
16-01-2017, 11:19 AM
One problem I experienced with one of my motion controlled cranes was that the drivers - 6 of them - were placed too close together in the box (which was open top) and didn't have enough space to "breathe" and there was no fan. To start with, it would all function well, but after some minutes some of the drivers would start to have a life of their own:)

I put it down to the heat, the drivers didn't have the cooling base either. With a 3 axis machine, it is easier to give them space for ventilation plus I will use a small fan as well. I also had some minor earthing problems with another machine, so the lesson learned is to anticipate all this and be on the safe size to avoid later problems.

The Kflop communicates via USB, and I haven't had any problems in that respect, been using it almost daily for three years in my work as a photographer and for motion control you can't afford to lose steps or have a bit of backlash, it's servo motors and harmonic drives for every axis. I use a good quality shielded cable, with ferrite chokes at both ends just as a precaution.

Edward

Robin Hewitt
16-01-2017, 12:24 PM
Hi Edward
Laid out, open plan on the desk your Sieg SX2.7 will be working fine and dandy, I believe they always do. All the data lines are clear, it is not interfering with the telly.
It is only when you start putting things in metal boxes, using screened cables and connecting everything to ground things start to go wrong.
I don't know if anyone sees an improvement. If not our idea of how to do it may be flawed.
Robin

Edward
16-01-2017, 12:38 PM
Well yes, but these electronic elements need to be kept in some sort of box or small cabinet to protect them from swarf and keep it all tidy. I will do it and then let's see how it goes, I will report back to you guys with pictures. I think I am going to make my own box to size as I have plenty of 20 x 20 alu extrusion and with that and a few panels it can easily be made to measure.

However, I am tempted to run the cables all in one length, direct connection so to speak, with as few connectors as possible, just using glands for the entry points on the panel.

Edward

Robin Hewitt
16-01-2017, 01:42 PM
Edward
You do whatever seems best to you, you cannot actually go far wrong. Don't worry about me an Jazz, I like winding him up and he likes screaming obscenities at his computer, it's a match made in heaven, no lasting damage has been done.
"Tempted to run the cables all in one length". Does that mean bring all 8 motor wires back to the driver? I would say bad idea because you still have to join them, the wires do have resistance and there will be heating. If all together in a cable rather than loose in a conduit, there is not much scope for dumping excess heat. Running 4 wires halves the resistance. Also, beware cutting any wires close to the motor, leave yourself plenty of wriggle room.
Robin

Edward
16-01-2017, 02:05 PM
Robin,
I take it these "battles" have been going on for a while:)

I meant running a 4 core cable from motor to driver in one go, without connectors (XLR or similar) on the panel - less work soldering pins. At the motor side, just cut the 8 cables and solder/insulate accordingly to the 4 core cable.

Not that I've ever experienced any problems with connectors, but I am just thinking that once all is in place, I won't be moving the box anywhere, so using XLR or GX16 connectors is not going to be necessary, though I suppose it looks a touch more professional, if one is bothered about that.
If for whatever reason I need to disconnect a motor or something in the future, it's not a big deal to do it from the drive own connector.



Edward

Edward
16-01-2017, 02:26 PM
Changing the subject a little, I read somewhere else in this forum that Dean's sensors are floating - that way it minimises cable runs.I didn't quite know what he means by floating and I'd like to know, maybe from the man himself, if he's reading, please.

I just saw a video by him talking about the repeatability of the cheapo proximity sensors and I think the trigger was the metal of the ballnut flange, or in any case, something that moved past the sensor to trigger it- sideways, rather than front on. So there was no possibility of trigger and sensor crashing into each other. I quite like this and maybe it's common practice to do it this way, although I have seen videos where the machine would crash into the sensor if it overrran a little.
Mind you, if the machine overruns, a crashed sensor is probably the least of the problems:)

Edward

njhussey
16-01-2017, 03:09 PM
He means that he (and I) have one home (or/and) switch for each axis and it travels on the axis with two triggers each end for it to see. Less wiring....

Sent from my HUAWEI VNS-L31 using Tapatalk

Edward
16-01-2017, 03:23 PM
Thank you, Neil, I'm with you. Though in my head I am initially thinking that it may not be so easy to implement on a vertical mill, say, the X axis, sensor somewhere in the middle, hitting targets at either extreme...
I can see the opposite working, i.e. sensor stationary in the middle and triggers at the extreme ends moving.

Edward

JAZZCNC
16-01-2017, 08:15 PM
I can see the opposite working, i.e. sensor stationary in the middle and triggers at the extreme ends moving.

Edward

Neil answered for me but no difference if sensor moves or Target moves. Infact it's common on mills for the Switch to be stationery and the targets or Ramps if mechaincal type Switch to move with table or saddle. See pic.
20385


Edward
You do whatever seems best to you, you cannot actually go far wrong. Don't worry about me an Jazz, I like winding him up and he likes screaming obscenities at his computer, it's a match made in heaven, no lasting damage has been done.


Robin you don't whind me up and believe it or not don't curse you to the PC, like you it wouldn't listen.! . . . You do Piss me off with the one line sarcastic unhelpful replys and excrament which comes from your fingers.

However much as I'd like to go at it with you decided In the interest of not hi-jacking (more) Edwards thread and fact can't be arsed after long day I'm not going to endulge your delusion of whinding me up by replying to your challenge regards Screening and Metal enclosures.

Anyone who want's to follow your Stupid approach is very welcome to go for it.!! . . . . They will soon see who was correct.

Edward
16-01-2017, 08:24 PM
Thank you, Dean. The picture says it all. Not only do you save on wiring, but you save on sensors too! I think I will implement something like this.

Edward

the great waldo
28-03-2018, 09:53 PM
A couple of questions regarding earthing in a control box
I'm going to run the screened 4 core wire from steppers to sockets on the metal case and run the screened wire from the inside of the the case (soldered pins on the stepper connector insulated with heat shrink) is it sensible to cover the inside of the socket with copper tape as a screen and to continue the screen continuity? Would I be right in thinking that as the case is earthed I won't need an extra wire from the connector to the bus bar ? I'm going to use a bus bar to connect the earth wires (star earthing) should I bolt the bus bar to the case or insulate it? The switching power supply has a metal case I assume as it will be bolted to the case it doesn't need an extra earth wire. I assume all metal cased items such as stepper drivers are earthed by the mounting bolts, or is it better to insulate them and run extra earths to the bus bar. One last thing is that the boxes holding the connections from the steppers are plastic. Is it worth screening the inside of the boxes but having no electrical contact to the machine frame so that the earthing is only connected at the busbar. I hope this makes sense.

Cheers

Andrew

JAZZCNC
29-03-2018, 01:44 PM
A couple of questions regarding earthing in a control box
I'm going to run the screened 4 core wire from steppers to sockets on the metal case and run the screened wire from the inside of the the case (soldered pins on the stepper connector insulated with heat shrink) is it sensible to cover the inside of the socket with copper tape as a screen and to continue the screen continuity?

It's not that critical really but won't hurt if you want to go to that much trouble.


Would I be right in thinking that as the case is earthed I won't need an extra wire from the connector to the bus bar ? I'm going to use a bus bar to connect the earth wires (star earthing) should I bolt the bus bar to the case or insulate it?

Bolt the Bus bar to Cabinet and run all screen wires back to this. Because case is Earth via bus bar then won't need wire from connector body if that's what you mean.



The switching power supply has a metal case I assume as it will be bolted to the case it doesn't need an extra earth wire. I assume all metal cased items such as stepper drivers are earthed by the mounting bolts, or is it better to insulate them and run extra earths to the bus bar.

Just bolt them down to the case.


One last thing is that the boxes holding the connections from the steppers are plastic. Is it worth screening the inside of the boxes but having no electrical contact to the machine frame so that the earthing is only connected at the busbar. I hope this makes sense.

No they will be fine. Don't over complicate things because you can easily introduce problems in your attempts to avoid issues. All you need to do is run all Screens and Earth's back to Star point. If you introduce extra Earth wires etc you can easily create ground loops.

Mains things are.
Shields & earths back to Star point.
Stepper sheilds Earthed at star point only not motor end.
Run all signals wires far away from Mains voltage wires as possible.

the great waldo
29-03-2018, 02:28 PM
Thanks Jazz
As the screen for the steppers is connected at the socket via the chassis should I leave both ends of the screen unconnected or is it better to run the driver end of the screen to the bus bar?
Thanks again
Andrew

Clive S
29-03-2018, 04:14 PM
Thanks Jazz
As the screen for the steppers is connected at the socket via the chassis should I leave both ends of the screen unconnected or is it better to run the driver end of the screen to the bus bar?
Thanks again
Andrew

Others might have a different view but I have done it like you and grounded the screens at the sockets and the plugs. But then I connected a ground wire bolted to the panel close to the sockets and ran that to the star point.

Then you don't ground the other ends of the cable ie at the driver or motors

the great waldo
29-03-2018, 04:22 PM
Thanks Clive.
Just for other forumites make sure there is continuity on the chassis holes for the sockets. On my box only one socket made a connection, all the others were insulated by the case coating/finish. This needs to be corrected.
Cheers.
Andrew

Neale
29-03-2018, 10:46 PM
Thanks Clive.
Just for other forumites make sure there is continuity on the chassis holes for the sockets. On my box only one socket made a connection, all the others were insulated by the case coating/finish. This needs to be corrected.
Cheers.
Andrew

I'm also using an epoxy finish steel cabinet, and don't trust any mounting bolts to earth to the cabinet. I've used ground wires from sockets back to the star point, which is a stud welded to the inside of the cabinet.

Clive S
29-03-2018, 10:51 PM
I'm also using an epoxy finish steel cabinet, and don't trust any mounting bolts to earth to the cabinet. I've used ground wires from sockets back to the star point, which is a stud welded to the inside of the cabinet.

Yes I should have said that as I have also done the same thing with the mains earth to the star point.

m_c
29-03-2018, 11:56 PM
The key thing with grounding, is to avoid ground loops, which you potentially create anytime a screen had more than one ground connection.

However, in power wiring, ground loops aren't likely to be a problem. The point of grounding in power wiring is to contain noise. A ground loop there is unlikely to cause any problems. A prime example is if you wire up a VFD as per most manufacturers recommendations, you ultimately create a ground loop in the wiring, as the motor is likely to be grounded through the machine.

It's in signal wiring, especially low voltage i.e. 5V, where ground loops are far more likely to cause problems.