Need Help!- My first CNC controller layout - Evaluation needed

[jestah]

Hi guys,

This is my first cnc project and needless to say it has grown somewhat out of hand... I was planning on making a small desktop router and after a few design changes I am now coming to the last stages of a 1.5 tonn steel/ally router that can cut a full sheet of mdf with Z of around 250mm (10inch)!

I have NO formal training other and only a few hundred hours of reading cnc zone with not much luck finding solid info about laying out my cnc controller to give good performance while minimising interference.

I need to house:



Enclosure #1
Pmdx 132 BOB with 4 gecko drives underneath
a small bank of 24vdc relays mostly switching low voltage systems
a 24v dc power supply
2-3 relays switching 240v dc
1 bleed resistor
a computer power supply
a mother board
a hard drive

Enclosure #1
1 1000va trodial transformer, bridge rectifier and a 18800uf bank of caps (supplying around 65vdc to the drives)
1 Huanyang VFD
1 time delay relay for shorting out the soft start resistor
1 soft start resistor

Once things are running I would like to also add a second bob for more input and output options as well as a 5th gecko drive and small power supply for a rotational axis.

Maybe I am crazy but I am going to try and cram all of the above into two enclosures! Now this may be possible but to add the last spanner in the works... I need to make both enclosures fully sealed as this router will be cutting a bit of carbon and the dust is quite conductive as well as being VERY hard to contain.

I had planned to mount the geckos on a heat spreader and 4 cpu heat-sinks (with fans) to pull the heat off the drives and move air around the enclosure. I hope that the steel box has enough surface area to keep the case temps down but if I need to I am willing to water-cool the pc cpu and geckos leaving only the power supply as the major heat producer.

My other big concern was interference between different components. I have done my best to keep the 24vdc wires as far from the 240v lines and plan to place the VFD, drive power supply and soft start items in a second enclosure as I have read these both are rather noisy. I will also use shielded wire for all cables.

I have attached a photo showing a rough display of how I plan to lay things out. Keen to get any advise as I do not have enough experience to feel confident about electrics to start drilling and mounting to my riser.

so a list of questions:
Grounding of the PC, The mother board looks to ground using the mounting screws and will attach to the enclosure in many places. The pc power supply and hard drive will also ground to the riser. Should I insulate both and then lead a single ground wire from each to the star ground point.

Will the pc power supply act as a source of noise that needs to be reduced?

Can the riser act as the star ground?

Do I need to add a conductive cover over the drive transformer and ground to protect it from the vdf noise? If so how do I do this without creating a shorted winding?

Grounding of cable shields with conductive connectors - My connectors have a conductive shell that bites into the cable braid once again grounding to the enclosure/riser. I am unsure what to do in this situation, I am guessing that from the enclosure to the motor (unconnected at the motor end) this gives the correct grounding but what about from the drive to the female socket on the enclosure? I have read that you should always ground a shield at the source end so I could terminate the shield before the connector and lead the shield from the drive end to my star ground.

I have fuses protecting each system of the controller but should I also have a overall fuse that kills both incoming ac lines and ground? (not sure what its called but seen fuses that have 3 independent blocks that are connected at the flip switch)

Any thing I have forgotten to take into account to make a safe controller considering the system has some decent volts and amps floating around it?

I hope this was not to much of a ramble but I am keen to drill the riser once rather than find out that a different layout is needed and having to redrill.

Thanks
Jestah

Ps in the photo the soft start resistor has not been moved

[Al_The_Man]

It is not really necessary to shield a Toroidal Transformer as the EMI tends to be contained within itself.
Here is a paper on general grounding and bonding, which shows that all enclosure backplanes and machine frame etc are bonded to the central star point to which the service ground is fed to.
http://www.automation.siemens.com/do.../emv_r.pdf?p=1
If you are using a VFD, twist the VFD to motor conductors together if fed with 3 single conductors.
Run a ground wire from star point to VFD and on to the motor frame.
If you have a 1 phase supply, do not fuse anything but 'live' conductors, i.e. not the neutral.
I always leave the P.C. grounded as is, but also take a ground wire from PC frame to the star point.
Al.

[brtech]

Originally Posted by jestah Grounding of the PC, The mother board looks to ground using the mounting screws and will attach to the enclosure in many places. The pc power supply and hard drive will also ground to the riser. Should I insulate both and then lead a single ground wire from each to the star ground point.

No. Screw the motherboard down to the case the way it was designed to.

Will the pc power supply act as a source of noise that needs to be reduced?

Well, it has to be kept as it is (in a metal case, well bonded to the case the PC goes in. I'm not sure how you are going to cool it, since it has a fan, needs air, and creates a fair amount of heat. You don't need to screen the power cables. The AC is filtered at the input, and you need that; generally, the AC should be outside the enclosure.

Can the riser act as the star ground?

I would do as Al suggests and put a bolt in the enclosure, well connected to case, and connect everything to that. Any one point CAN be star ground as long as it is well bonded to the case. Worry a bit about the PC power supply. It has it's input AC wired to its case. It's probably okay as is, and you don't want to bring, say, a ground outside the case just to get it to the star ground. Keep the power supply case bonded to the enclosure, and it will be okay.

Do I need to add a conductive cover over the drive transformer and ground to protect it from the vdf noise? If so how do I do this without creating a shorted winding?

I wouldn't think so.

Grounding of cable shields with conductive connectors - My connectors have a conductive shell that bites into the cable braid once again grounding to the enclosure/riser. I am unsure what to do in this situation, I am guessing that from the enclosure to the motor (unconnected at the motor end) this gives the correct grounding but what about from the drive to the female socket on the enclosure? I have read that you should always ground a shield at the source end so I could terminate the shield before the connector and lead the shield from the drive end to my star ground.

Not exactly sure where you are worried about. If you are worried about shielding INSIDE the enclosure, then I would ground at the entrance to the enclosure regardless of where the source is. Mostly, you control noise inside the enclosure with cable routing and not cable shields. On the other hand, if you were coming, from, say the VFD and wiring inside the enclosure to some connector on the enclosure that the motors outside the enclosure would connect to, the VFD has it's own enclosure, which has to be well bonded to the outside enclosure. If I was worried about inside wiring from the VFD to the connector, I would ground the shield at the VFD. I would have the external wire from the connector to the motor grounded at the connector.

I have fuses protecting each system of the controller but should I also have a overall fuse that kills both incoming ac lines and ground? (not sure what its called but seen fuses that have 3 independent blocks that are connected at the flip switch)

I would get an AC input filter (Corcom), and if it had a fuse, great.

Any thing I have forgotten to take into account to make a safe controller considering the system has some decent volts and amps floating around it?

Spacing. You keep high voltage separate from low voltage. The difference between high and low is around 40V (SELV). I'm not recalling the distance, but it should be a couple inches I think. You clearly mark the high voltage wiring. Color codes are good. Having an extra level of enclosure between high and low voltage is good, but not necessary.

[jestah]

It is not really necessary to shield a Toroidal Transformer as the EMI tends to be contained within itself. Here is a paper on general grounding and bonding, which shows that all enclosure backplanes and machine frame etc are bonded to the central star point to which the service ground is fed to. http://www.automation.siemens.com/do.../emv_r.pdf?p=1

Thanks for that reading. I had been under the impression it was EXTREMELY important to make sure that each ground only had a single path back to the star ground point. This had me to thinking that I needed to insulate things like the pc psu and mobo from the riser and lead a single ground lead from each to the star point. I was worried that boning it to the riser and therefore to other parts also grounded to the riser and having individual grounding wires leading to the start ground could form a ground loop.

If you are using a VFD, twist the VFD to motor conductors together if fed with 3 single conductors.

I will be using shielded wire for any VFD to spindle connections. Should the inner conductors be twisted or was this just a precaution with unshielded wire as to keep the wire paths crossing to reduce radiated noise?

Run a ground wire from star point to VFD and on to the motor frame.

The VFD has a plastic case but also a earth wire tag. I have just checked and the cable sheild has continuity onto the body of the spindle. Should I still run a second grounding lead to the motor?

If you have a 1 phase supply, do not fuse anything but 'live' conductors, i.e. not the neutral.

Thanks for that, the fuses I have currently are all only on the live line (a single fuse for each sub system on the active line right after the AC line enters my enclosure). Should I have a master fuse that is before the sub fuses to protect my house hold wiring? Down in NZ we have mostly 15A wiring but my spindle and transformer could draw far more than this. If so I just need the one breaker on the live in line?

Thanks Al for the fast reply and further readings. Sorry for the many questions but I feel that I am at that place where I have just enough knowledge to think I know what I am doing but not enough to be safe!

[jestah]

No. Screw the motherboard down to the case the way it was designed to.

will do

Well, it has to be kept as it is (in a metal case, well bonded to the case the PC goes in. I'm not sure how you are going to cool it, since it has a fan, needs air, and creates a fair amount of heat. You don't need to screen the power cables. The AC is filtered at the input, and you need that; generally, the AC should be outside the enclosure.

This is going to be tricky... I am very reluctant to try and have a filtered air input into the case as carbon dust will manage to work its way past even the most robust dust suits we wear (rated for very fine aerosol spray). The enclosure is rather large so I am hoping that it has enough surface area to deal with the heat the psu produces but failing that I may have to make a water cooled system (not for the psu!!) that has a few heat sinks that are bolted to a water block inside the case to help capture some of this heat.

Worry a bit about the PC power supply. It has it's input AC wired to its case. It's probably okay as is, and you don't want to bring, say, a ground outside the case just to get it to the star ground. Keep the power supply case bonded to the enclosure, and it will be okay.

The riser that the other parts of my cnc controller are mounted to is the computer case. The AC input into the computer PSU has an earth that I was planning on wiring to the star ground. I still see a ground loop forming from this and bolting it to the riser that will also be connected to the start ground but as Al and the pdf he suggested I read said this should be ok (or maybe I misinterpreted the information)

Not exactly sure where you are worried about. If you are worried about shielding INSIDE the enclosure, then I would ground at the entrance to the enclosure regardless of where the source is. Mostly, you control noise inside the enclosure with cable routing and not cable shields. On the other hand, if you were coming, from, say the VFD and wiring inside the enclosure to some connector on the enclosure that the motors outside the enclosure would connect to, the VFD has it's own enclosure, which has to be well bonded to the outside enclosure. If I was worried about inside wiring from the VFD to the connector, I would ground the shield at the VFD. I would have the external wire from the connector to the motor grounded at the connector.

My confusion comes with the principal about only having a single path back to the star ground. If the shield links to the connector body and this to the enclosure and this to the riser I can see a loop forming should I lead the braid as well to the star point. OR should I just use the path from braid to connector to enclosure and forget about wiring the shields to the star ground.

I would get an AC input filter (Corcom), and if it had a fuse, great.

Will do

Spacing. You keep high voltage separate from low voltage. The difference between high and low is around 40V (SELV). I'm not recalling the distance, but it should be a couple inches I think. You clearly mark the high voltage wiring. Color codes are good. Having an extra level of enclosure between high and low voltage is good, but not necessary.

I have done my best with my limited skills to lay out the controller. I have placed all of the 24vdc relay wiring together with no other voltages entering the same cable tidy box thingies, placed the VFD and transformers in an other enclosure as my understanding was they are the biggest noise produces, done what I can to keep all 240vac to the bottom right corner of the riser in enclosure #1 and have a strong colour coding system in the case (I plan to replace all of the orange wires with blue and brown)


Big thank brtech for answering my questions, fell a bit closer to getting this right but to be honest feel that I may have the wrong understanding of this grounding issue. I will have some lunch and go over the threads regarding grounding I have saved in my bookmarks.

[brtech]

Don't worry about ground loops from green wire ground and cases. It is better if all the green wire grounds go to the star ground, but opening up the computer PS and removing the connection between it's input and its case is silly. You don't even have to run the green wire to the computer PS as long as the PS case is well bonded to the case it is inside of.

Most cheapo computer PS are not all that efficient. You are going to have to figure out how to cool the components in it that the forced air usually does.

You don't connect the shields to the star ground. You connect them to the case as close to the end of the wire as practical. Sometimes that means going through a connector pin, with the pin connected to case ground. That will change a lot of your thinking.

[jestah]

Thanks a lot for answering many of my questions... After gaining a better grasp of grounding I decided to completely change the layout of my controller.

I have decided to keep the computer in an enclosure of its own and move the vfd and drive psu into the same enclosure as the drives and my relays. I hope that the VFD will not cause too much noise inside my controller. Fingers crossed but just in case I may have goon a little overboard with my cable shielding I will post photos as soon as I am done but would like to ask a few more questions so I can tie up the last few details.

I have around 8 cable shields to ground and would like to know if it is ok to route the shield pigtails within the same cable tidy or near any signal wires?

Can I also route my 12vdc 2a supply for feeding fans, water pumps etc in the same tidy as my signal wires? My guy says yes but I have no idea why.

thanks in advance
Jestah

[brtech]

You want to get the shield tied to case ground as close to where the shield is cut off as possible. In many layouts, where there isn't much noise inside the enclosure, the shield is grounded right as it enters the enclosure. If you have noise inside the enclosure, you ground as close to the termination of the signals as possible, leaving as little unshielded wire as possible. The length of the wire from the shield to the ground is much more important than exactly where the ground is. Shorter is better. An inch is long. You wouldn't make the wire from the shield to ground longer just to get it to a star ground point or a "tidy", which is not a term I know. Ground it to the case right where the signals terminate.

Think of the shield as an extension of the metal enclosure.

[jestah]

Hi Brtech,

Thanks for the fast reply, I was thinking that sort pigtails were the winner but then got confused while reading one of the many grounding posts suggesting it would be best to tie them all to star ground. Thinking of it as an extension of the enclosure is a far better way to look at it.

Just a quick bump about routing my 12vdc fan supply in the same tidy (what I am calling the blue and grey plastic slotted tracks that my wires are tidied into) as my signal wires?

[brtech]

Generally, I think there is no problem in running DC fan wires in the same wireguide as the signal wires for this kind of application. Fans are not sensitive to anything, and usually the fans are on all the time. If you had some kind of speed control, I might worry a bit.

[jestah]

Thanks for that, Ok I am going down to the workshop and grounding every thing. I will post a few photos as to be honest it has turned out FAR better than I expected.

[RomanLini]

I'm more concerned about the lack of heat extraction.

Waste heat (guesses);
PC 150W
Toroid 15W
VFD 30W
Geckos 20W

It's going to be over 200W idling, rising to maybe 250W or more when running. I've built incubators and low temp drying ovens using 200W to heat a volume of 6 cubic feet to well over the safe temperature for running electronics and well over the fail point of your PC micro.

How big is your cabinet and how are you going to extract all that heat so it doesn't cook??

What I would do is use sensible air cooling of your cabinet with a decent filter mat on the air intake. That should catch most stray particles and you can use an automotive air filter oil on it if needed to get really good filtering.

Then deal with the carbon fibre dust at source so you are not eating and breathing it! Maybe a cyclonic extractor and good shop vac will get 99% of it out of the air.