VFD inside or outside the control box?

I am looking to start to put together my control box for my cnc, I have decided to with the DDCSV controller, which I will mount on the front on the enclosure, what I would like to know is, is it best to put the VFD inside or outside of the enclosure? If it should go inside the enclosure, where is the best place to mount it, I have read that it should be mounted at the bottom of the enclosure? Should it be kept away from any of the electrical components within the box? I am looking getting a metal enclosure, I do have an old pc case, but it won't be deep enough to mount the VFD. I do not have any electrical experience, just a little basic knowledge, so any help or advice would be very much appreciated. Thank you everyone

You haven't said which VFD you have, but if it is anything like mine it has a digital display of speed and a manual speed control on the front, neither of which will be accessible if it's inside the box. More importantly you must ensure good air flow through the VFD to avoid overheating. Correct air management is critical for power electronics.

Mine is on the wall behind one end of the machine near where the drag chain ends, partly because I can see it properly and set the speed and partly because I should have bought six metres of screened 4-core cable instead of five.

Kit

Quote:

---

Originally Posted by Kitwn

You haven't said which VFD you have, but if it is anything like mine it has a digital display of speed and a manual speed control on the front, neither of which will be accessible if it's inside the box. More importantly you must ensure good air flow through the VFD to avoid overheating. Correct air management is critical for power electronics.

Mine is on the wall behind one end of the machine near where the drag chain ends, partly because I can see it properly and set the speed and partly because I should have bought six metres of screened 4-core cable instead of five.

Kit

---

Thank you for your reply, it is a Huanyang VFD, with a digital display and manual speed control, luckily I have plenty of cable, so can mount it pretty much anywhere within reason 😁

It's got to be better having the VFD inside a steel cabinet from an EMC (electrical noise) POV, as well as making a neater installation. In both my machines, I have the VFDs mounted inside and have a fan to ensure they can stay cool. Rather than have the fan powered all the time, I have a simple central heating thermostat to turn it on when / if the air temperature exceeds a threshold which is something like 40C. There's an exit grille with a foam filter above the VFD and the fan is mounted near the bottom of the cabinet.

Attachment 26027

Muzzer,
Aren't you shutting one of the biggest sources of electrical noise inside the metal shielding meant to keep it out of the other electronics? However I can see the point of keeping all the noise from the stepper drivers and VFD contained and away from the controlling computer.

Paul,
Perhaps this is an opportunity to broaden the discussion to noise in general, it's sources, consequences and how to deal with it.

My own controller is an old PC case with the PSUs and stepper drivers in it with the parallel breakout board in it's own metal box inside there as well (I use a cheap Chinese breakout and LinuxCNC). I left the PC PSU in there to provide 5v for the breakout and 12v for the fans.1uF capacitors are connected across all the inputs on the breakout board. Cables to/from the steppers, spindle and limit switches are screened with the screens only connected to ground at the controller or VFD end. It's been working OK for now.
Your mains supply can spread noise and there are recommendations to feed the VFD from a socket as far away from everything else as possible. My entire machine, computer, dust extractor and all, is fed by one dual socket and still works so don't spend any money on getting new sockets installed at the other end of your shed unless you have to.

Kit

Inside an enclosure is good, check the manufacturer's requirements though as adequate cooling is essential.
Unless your VFD is really shoddy the majority of EMI is carried through wiring, not the air, although proximity will mean short cable runs and may require a filter on each mains equipment input in the box and shielded cable from the VFD to the motor is advisable.
Outside is fine though, I've run a benchtop CNC mill for about 7 years with the VFD mounted to a bracket on the spindle motor without any issues and my lathe VFD is mounted next to the DRO display with a tacho on top of the VFD, that's worked fine for 20 years or so :D

Most VFDs don't have filtering incorporated, as it's expect you will fit whatever is appropriate for the final installation. Consequently, many hobby installations don't bother unless there is an issue due to the electrical noise.

I fitted Yaskawa V1000 drives in both of mine. They didn't actually cost a lot more than the Chinesium equivalents and you can buy EMC filters that are tailored for those specific drives. They may not guarantee 100% compliance with EMC regs as such but are going to get you in the right ballpark. Certainly, I have not had any issues. This is how professional installations are constructed, rather than simply locate the VFD on the other side of the shed(!).

Getting everything into one cabinet makes for a tidier installation and allows you to construct it on the bench where access is a lot easier. It's a personal choice but I like the end result to look reasonably professional.

One of the main ways in which noise will get coupled between parts of the system is by running cables next to each other. I wanted to make everything neat by having all the cables for the spindle, motors and limit switches running through one set of drag chains but this is obviously the worst case for this. At one point I had the spindle feed looped up with the dust extraction hose to keep it separate but wasn't happy with it. This is why you want to screen all the cables and ground them to a single point at the control box/VFD end. Capacitors in parallel with the limit switch inputs to the breakout board are a great help as well. My understanding is that it's the high frequency, high amplitude noise from the stepper and spindle drives couling back into the limit switch inputs and possibly then into the controlling computer that are the main risk.

Kit

Yes, screening (and grounding the screens) is just good practice. Most drag cables have sets of screened cores within a screened bundle, so you can run signals and steppers etc in the same cable. Good idea to run the VFD>>motor separately though. Decent quality CNC controllers are fitted with noise filtering on the IO, so additional measures shouldn't be necessary if you have a good construction and follow good practice.

Well, It looks like I have started an interesting discussion.

I have watched countless Youtube videos on wiring/connections etc, and most of the builds have the VFD inside the cabinet, now whether that is the right or wrong way, I wouldn't have a clue, but what I will say that a majority have an EMI filter on the mains feed from the VFD, I was also going have the VFD/spindle on it's own power supply, as I said, this is all new to me and it is a very steep learning curve, so expect loads of questions and advice while I put together all the components, none of you live in Hampshire by any chance? Would make my life so much easier just pop round for some one to one advice:smile::smile:

I have just picked up a metal box/enclosure, so I will very shortly be starting the electrical build, wish me luck!

Thank you for all the input and advice.

Quote:

---

Originally Posted by Kitwn

This is why you want to screen all the cables and ground them to a single point at the control box/VFD end.

---

For emissive cables (steppers, spindle) you should bond the screen at both ends to the respective enclosures.

<Opens popcorn>

Haha, nice try. Some (most) VFD manufacturers say one end, others say both ends. Probably doesn't really matter for the sort of work we do but this is an old chestnut. Let's just say there is no right or wrong here...

Quote:

---

Originally Posted by Paul53

none of you live in Hampshire by any chance?

---

Sorry no, I live in Exmouth. The one in Western Australia. I'm not sure my advice is worth a cup of tea anyway.

Earthing everything to a single point is supposed to prevent earth loops, the cause of much mains hum in audio equipment. This is probably the thinking behind only connecting screens at one end. As you'd be earthing the body of your machine back to the same point with fairly short wiring anyway it probably doesn't matter. Belt and braces is to make the cable so you CAN connect at both ends but be able to disconnect if you have problems.

On a safety note: The output of your VFD is at potentially lethal voltages and is isolated from the mains input so you get no protection from an RCD or other 'safety switch' on your mains supply. Make sure the spindle body is well earthed and use insulating sleeving over each connection inside the plug (this will help prevent you detonating your VFD as well). Recent spindles from Huan Yang have the body earthed via a fourth pin on the socket, but earlier versions had only 3 pins or the fourth pin was not connected.

Kit

Quote:

---

Originally Posted by Kitwn

On a safety note: The output of your VFD is at potentially lethal voltages and is isolated from the mains input so you get no protection from an RCD or other 'safety switch' on your mains supply. Make sure the spindle body is well earthed and use insulating sleeving over each connection inside the plug (this will help prevent you detonating your VFD as well). Recent spindles from Huan Yang have the body earthed via a fourth pin on the socket, but earlier versions had only 3 pins or the fourth pin was not connected.

Kit

---

I'm afraid that's incorrect. VFDs do NOT isolate the motor (output) side from the mains. So the RCD would provide some degree of protection. So, you should consider the motor phase connections to be at mains potential (which they are) and use appropriate insulation / creepage and clearance distances. A good quality VFD should tolerate a phase to ground short circuit without "detonating". I speak as an engineer who has developed VFDs throughout my career.

There are some high end industrial installations where the entire VFD high voltage circuit (AC and output) is isolated from ground by means of a dedicated power transformer ("normal" mains is grounded on the neutral side). This is to allow a single fault (phase to ground) to exist without the whole system crashing immediately. These are used in factories such as paper and steel mills where loss of one VFD in a series of interconnected VFDs could be catastrophic. I doubt any of us will ever see such a system.

Muzzer,

You obviously have more experience of this than I do. I'll have to do some rectifier drawings and follow the currents.

Kit

Generally speaking you will find a bridge rectifier, then DC bus caps, then a 3-phase / 6-switch (IGBT) bridge. Nothing much else needed on the power circuit.

With single phase 240V input, only 4 of the 6 rectifier diodes get used but as I said, most VFDs don't check for a missing phase. The manual may define a slight derating when you use single phase instead of three phase input but the rectifiers are pretty robust and unlikely to be an issue for intermittent / light hobby use. The DC bus is going to be around 340Vdc and obviously you can't generate an AC voltage that exceeds 240Vac from that, hence you normally need to connect your 3-phase motor in delta if it's a 240/415 star/delta machine.

On higher power VFDs (>6kW or so), the bridge rectifier will become an active PFC ie another 6 IGBTs instead of rectifiers. And they may then be bidirectional ie can regenerate back into the mains during overrun.

Some VFDs come with built-in EMC filters but that's the exception rather than the rule. It's a good idea to fit one if you can.

Thanks for those details. I've been doing some scribbling (it's been a quiet night-shift so far for me here in Australia) and realised that my long held belief that (when considering the operation of an RCD) a rectifier output is isolated from the input just like that of a transformer is so much kangaroo poo. On reflection I can't think where I got that idea which is all the more worrying, I obviously didn't draw any diagrams to prove it.

Well I did say that my advice wasn't worth the cup of tea Paul would have offered me if I'd visited him in Hampshire:witless:

Kit