Terminating a shielded cable with a VFD & spindle?

[sploo]

I have a VFD that I'm going to use to run a 3-phase spindle on a CNC machine, and I've been advised to use a shielded cable between the VFD and the spindle.

The VFD (unsurprisingly) takes a 3-core single phase 240VAC main input, and has terminals for L, N, and E.

The VFD's terminals for the spindle are U, V, W, and E. There does not appear to be continuity between the single phase and 3-phase earth 'E' terminals.

The spindle socket has 4 pins, with the pin for the earth core connected to the can of the spindle. The socket on the spindle is isolated from the can (i.e. no continuity between the socket and any of the socket pins).

Without using a shielded cable, my only question would be whether I should wire the earth core on the cable coming from the spindle to the E terminal next to U, V, W, or to the E terminal also used by the mains input?

How should I terminate the cable shield? If I expose the shield and clamp it to the plug that goes into the spindle socket then it won't actually be connected to anything (as the spindle socket is isolated). I could wire it to the earth core in the VFD-to-spindle cable, but I'm unsure if this is the correct procedure. If it were wired to the earth pin at the spindle end then I assume it would definitely need to be connected to the earth terminal at the VFD end too?

I've put together this image to try to illustrate the current connections - with the VFD-to-spindle earth core wired to spindle can, and the U, V, W lines going to the motor inside the can. The spindle socket is isolated from the can and motor, and the (grey) shielding around the VFD-to-spindle cable is currently not connected to anything at either end:

 

[guineafowl21]

An odd one. If the VFD output ‘E’ is not connected to mains earth, it’s not really an earth, more a local circuit ground.

On the other hand, the VFD output is not isolated from mains, and so is considered mains-referenced. In this case, the mains earth should continue right through to output.

My instinct would be to bond mains earth, VFD output earth and cable shielding together. However, I wouldn’t do this without consulting the VFD manual and/or the manufacturer.

 

[minilathe22]

Have you tested with a multimeter that the two earth points are indeed not connected?

On mine there is a distinction between "earth" from the mains, and "ground" which is a much smaller screw, and designed to be the connection for the shield on any signal cables. If you have the PDF manual for the VFD they usually go into alot of detail, perhaps even with an example wiring diagram.

 

[Spectric]

There does not appear to be continuity between the single phase and 3-phase earth 'E' terminals.

This is because it is classed as a non linear load, the 230Vac is rectified to Dc before being electronically switched to generate the 400 volt three phase outputs. These outputs are internally protected by the VFD and are not referenced to the 230 single phase supply, but to ensure full protection that "Earth" is provided so if a short occurs to the motor body there is a path so the device can detect the short. This is also why a lot of cheap VSD's are cheap, they cut back on protection just like a lot of cheap wall chargers and the end result can be a thermal event.

 

[Phill05]

You need 3 core 240v cable to input on VFD and 4 core shielded cable to connect VFD to motor, connect 1,2,3, of the four to U,V,W connect the 4th wire to earth, connect shield to a star point earth near to VFD only, the cable shield at motor end leave unconnected or you will get a lot of interference affecting the running of the system.
If motor turns the wrong way interchange 2 of the U,V,W wires.

 

[sploo]

An odd one. If the VFD output ‘E’ is not connected to mains earth, it’s not really an earth, more a local circuit ground.

On the other hand, the VFD output is not isolated from mains, and so is considered mains-referenced. In this case, the mains earth should continue right through to output.

My instinct would be to bond mains earth, VFD output earth and cable shielding together. However, I wouldn’t do this without consulting the VFD manual and/or the manufacturer.

There appears to be a weak continuity between the output earth and the input earth (my multimeter gives a weak pulse when set to give an audible continuity signal). Perhaps there is some connection (albeit through the VFD "conversion"). I could get an actual resistance reading, but I don't know how useful that'd be.

My initial instinct on bonding/earthing was largely the same, but I'm getting conflicting advice on whether the VFD to spindle cable shield should be earthed, or only bonded to the spindle plug (which is isolated from everything). And obviously if the shield was earthed, does that mean it'd be connected to the earth code in the VFD to spindle cable (at both ends), and would that shield + earth be connected to the earth terminal on the VFD output side or input side. Confusing!

On mine there is a distinction between "earth" from the mains, and "ground" which is a much smaller screw, and designed to be the connection for the shield on any signal cables. If you have the PDF manual for the VFD they usually go into alot of detail, perhaps even with an example wiring diagram.

Chinese Huanyang, so whilst the manual is pretty good for all the configuration settings, it's a bit light on the actual wiring.

This is because it is classed as a non linear load, the 230Vac is rectified to Dc before being electronically switched to generate the 400 volt three phase outputs. These outputs are internally protected by the VFD and are not referenced to the 230 single phase supply, but to ensure full protection that "Earth" is provided so if a short occurs to the motor body there is a path so the device can detect the short. This is also why a lot of cheap VSD's are cheap, they cut back on protection just like a lot of cheap wall chargers and the end result can be a thermal event.

Does it make any difference due to the fact it's not converting the voltage? It's outputting three phase, but it's still at 230Vac.

 

[sploo]

You need 3 core 240v cable to input on VFD and 4 core shielded cable to connect VFD to motor, connect 1,2,3, of the four to U,V,W connect the 4th wire to earth, connect shield to a star point earth near to VFD only, the cable shield at motor end leave unconnected or you will get a lot of interference affecting the running of the system.
If motor turns the wrong way interchange 2 of the U,V,W wires.

3 core from mains to VFD: check, got that.

4 core shielded from VFD to motor: (sort of) check; it's on order.

By "a star point earth near to VFD" do you mean a separate point to which I earth various components (e.g. the VFD-to-spindle cable shield, the CNC machine body etc.)? I assume then that this star point earth would be connected to the mains earth?

Interestingly (and this isn't a criticism) the advice of earthing the shield at one end but not connecting it to anything at the spindle end is exactly the opposite to other advice I've received; hence my confusion!

 

[minilathe22]

How long is the motor cable? I am running a motor on a VFD and the cable has no shielding, I have not had any problems. Have you tried running it?

 

[Phill05]

What we found was If your machine is a cnc connecting the shield creates an earth loup and you get interference on the controls, we know if the motor end shield is not connected it does not interfere with anything the shield acts like an Ariel and collect interference so by connecting one end to earth it dissipates.
You might not be connecting to a cnc but by connecting the shield you might create problems on other control systems running close by on other machines, or as Minilathe22 has pointed out run with no shield at all.
You would only need to create a Star point if you connect a shield and yes a star point is a where a group of all earth wires are connected together at a point away from the VFD.

Edit: just reread your 1st post you do have a CNC so use 4 core shielded and leave the spindle end of shield not connected.

 

[minilathe22]

I should point out I am using a manual wood lathe, so no computer equipment in close proximity to the VFD/motor. I do have a PC in the workshop, I have not noticed any interference issues.

 

[guineafowl21]

If input/output earths are not continuous, as you have found, it’s probably best not to bond them together, on reflection.

Best advice is - install according to manufacturer’s instructions. In the absence of these, I found this:

https://inverterdrive.com/file/Invertek-E3-Easy-Start-Guide-IP20-3ph
In short, mains earth -> earth input. Bond output earth and screen together at both ends, and to motor earth.

My inverter doesn’t have screened cable on the output, but I just followed the installation instructions. It runs just fine and causes no problems. Although the inverter output is choppy and noisy, you are feeding it into a large inductor, which would have some filtering effect.

 

[sploo]

What we found was If your machine is a cnc connecting the shield creates an earth loup and you get interference on the controls, we know if the motor end shield is not connected it does not interfere with anything the shield acts like an Ariel and collect interference so by connecting one end to earth it dissipates.
You might not be connecting to a cnc but by connecting the shield you might create problems on other control systems running close by on other machines, or as Minilathe22 has pointed out run with no shield at all.
You would only need to create a Star point if you connect a shield and yes a star point is a where a group of all earth wires are connected together at a point away from the VFD.

Edit: just reread your 1st post you do have a CNC so use 4 core shielded and leave the spindle end of shield not connected.

You see; this makes sense to me - don't create loops in the grounding by having stuff connected at both ends, but do ensure the shielding is earthed at the end nearest the VFD; I just keep getting different advice from different sources so it does seem that maybe there's an element of "art" to it

 

[sploo]

How long is the motor cable? I am running a motor on a VFD and the cable has no shielding, I have not had any problems. Have you tried running it?

I've actually been running that CNC machine for well over a decade, but with a router. Given that the router was double insulated, and the CNC machine itself is only fed 24VDC (from a control box that takes 240VAC, but is earthed) I'd never bothered with any extra grounding.

Having bought the spindle I have rigged it up with some 3 core flex just to test it, but I'm aware that means there's no earth, the metal can of the spindle has continuity with the whole CNC machine (metal on metal contact), and I'm told that VFDs are quite electrically noisy. I did try a cutting job with the spindle and didn't have an issues, but obviously I want to ensure everything is safe - it's just the confusion of what should be earthed where.

 

[Phill05]

It will help you in the future, if you get a fault on your controls not having to find this one will be a big help.

 

[sploo]

If input/output earths are not continuous, as you have found, it’s probably best not to bond them together, on reflection.

Best advice is - install according to manufacturer’s instructions. In the absence of these, I found this:

https://inverterdrive.com/file/Invertek-E3-Easy-Start-Guide-IP20-3ph
In short, mains earth -> earth input. Bond output earth and screen together at both ends, and to motor earth.

My inverter doesn’t have screened cable on the output, but I just followed the installation instructions. It runs just fine and causes no problems. Although the inverter output is choppy and noisy, you are feeding it into a large inductor, which would have some filtering effect.

Straight up on page 2 of that Invertek manual it shows the earth core of the VFD-to-spindle cable and the shield connected to the earth point on the VFD output; which is what I would assume would be the correct procedure.

However, I see they're also connecting the earth core and the cable shield at the motor end of the cable; so that means I think I've now seen differing advice for just about every combination of shield/earth/one end/no ends/both ends connections

 

[Spectric]

Does it make any difference due to the fact it's not converting the voltage? It's outputting three phase, but it's still at 230Vac.

It may be at the same voltage but it is converting it, once converted to Dc and supplied to the IGFT drivers there is no reference between any of the phase outputs and the Ac input. Another way to view this is that if you were yourself properly earthed and touched the Ac live you would know it, but touch any ONE of the phase outputs and nothing, they are only live with respect to each other.

On the subject of cable shielding it is now common practice to ground both ends, with a properly designed system the issue of ground loop induced currents is no longer considered an issue. There are exceptions and when this is the case the OEM will be explicit in how they deem the shielding should be done. Now I agree with Minilathe in that you should not experience any issues with noise or EMC because you are only running a short distance and do not have any sensitive equipment that may be affected so just try it, but be careful if you have a pacemaker! Again cheap VFD's can be bad for EMC due to the output not being a true sinusoid and with low quality components. So for a little more buy something like Siemens V20 3kW (FSC) 230V 1ph to 3ph AC Inverter Drive, C2 EMC - AC Inverter Drives (230V) which is a quality drive, also ABB are really good.

Technical info Cable Shield Grounded At One End Only - EMC Standards

 

[Woody2Shoes]

I think it's important to be clear on the words "shielded" and "earth".

Shielding may be mechanical protection - e.g. steel wire armour, or may be to reduce incoming (or more likely in this case) outgoing electromagnetic 'noise'. Was the person who advised you clear on which of these they meant (I presume mechanical protection - what with 415V AC).

Earthing similarly, can be used to provide protection from electrocution and the earth wire is usually referred to as the CPC - circuit protective conductor. All bits of metal that you can touch should (almost certainly) be connected to a known good protective earth conductor.

Earth loops occur because all shielding/earthing conductors have resistance - and any current flowing in them can result in unwanted voltage appearing on one end of the conductor (this can have implications for safely and for EM interference). This is the consideration that will help decide whether one end or both ends of the conductor need to be 'earthed'.

 

[guineafowl21]

However, I see they're also connecting the earth core and the cable shield at the motor end of the cable; so that means I think I've now seen differing advice for just about every combination of shield/earth/one end/no ends/both ends connections

In the absence of exact manufacturer’s instructions, you have found an equivalent device with better ones. I would say this is a good start.

there is no reference between any of the phase outputs and the Ac input.

This is my view of an inverter:
AC mains —-(bridge rectifier)—> DC bus —> 6 IGBTs

I can see a path right back to the mains there. At what point are you saying the mains reference is lost?

 

[sploo]

It may be at the same voltage but it is converting it, once converted to Dc and supplied to the IGFT drivers there is no reference between any of the phase outputs and the Ac input. Another way to view this is that if you were yourself properly earthed and touched the Ac live you would know it, but touch any ONE of the phase outputs and nothing, they are only live with respect to each other.

On the subject of cable shielding it is now common practice to ground both ends, with a properly designed system the issue of ground loop induced currents is no longer considered an issue. There are exceptions and when this is the case the OEM will be explicit in how they deem the shielding should be done. Now I agree with Minilathe in that you should not experience any issues with noise or EMC because you are only running a short distance and do not have any sensitive equipment that may be affected so just try it, but be careful if you have a pacemaker! Again cheap VFD's can be bad for EMC due to the output not being a true sinusoid and with low quality components. So for a little more buy something like Siemens V20 3kW (FSC) 230V 1ph to 3ph AC Inverter Drive, C2 EMC - AC Inverter Drives (230V) which is a quality drive, also ABB are really good.

Technical info Cable Shield Grounded At One End Only - EMC Standards

You see now; this is why I'm so confused; because the article above states: "The shields should be “bonded’ to the boxes at the ends of their respective cables, not earth ground, building ground, or anything else". I interpret that to mean there should be clamps at both ends of the cable, that have continuity with the shielding around the cable. However, those clamps should not be connected to any earth wiring (such as the earth core within the 4 core U,V,W,E cable)? That's the opposite of what's shown in the Intertek manual posted earlier by guineafowl21.

 

[Spectric]

Firsty I should say that I used to be involved in the design of switched mode power supplies using various topologies and a good understanding of electromagnetics.

So what is a rectifier, it only passes in ONE direction and nothing in reverse. If this was a linear circuit then the output response would be directly proportional to the input and in the case of an AC circuit that would mean a sinusoidal voltage resulting in a sinusoidal current because the instantanious current will rise and fall in proportion to the voltage, therefore the current waveform is also a sine wave. Once you introduce rectification and filters this no longer applies, the output current is no longer proportional to the input current. This is also evident when performing electrical test and inspection, with motors directly connected to an installation you test loops down to the motor junction box, once a VFD or softstarter is used then you can only test upto the input of the electronic device, there is no continuation of the loop through the device.