Inverter with motor needed

This all seems a little murky!

I'm pretty sure that you can't buy an ELCB any more. Someone is bound to prove me wrong, but broadly I would declare the ELCB technology obsolete. They work differently from RCDs although they do a similar job. An ELCB actually measures earth current. In situations like "TT" supplies where an actual ground spike is the only method of grounding, the ground impedance can vary wildly. If the ground potential does vary, it can render the ELCB useless, and no one will know.

RCD technology compares supply and return current. It calculates the difference and infers the earth current, even if that current is not passing through the earth conductor. It requires no earth connection and helps to make "TT" supplies much safer.

All modern switch mode supplies (of which inverters are a kind) usually require some kind of EMC filter to allow then to pass conducted emissions tests and gain CE and other approvals.

EMC filters work by providing a high frequency path from the live and neutral conductor down to ground. Generally the impedance of an EMC filter from LN to E is fairly high. Indeed there is normally some leakage even in a simple electrical appliance with a line transformer. Typically the coupling is capacitive from the primary into the frame of the of the transformer.

It from the perspective of the RCD these leakages can accrue. If you have a single RCD on a whole premises, then it is possible that all of the leakages can accrue to the point where nuisance tripping occurs.

In general, nuisance tripping of RCD's is much less pronounced than it is for the old, less safe, ELCB technology.

Most appliances leak much less than their allowable class limit, but it is a particular problem with motor inverters, because we all want nice (sonically) quiet motor drives. In order that the switching noise cannot be heard in the windings of the motors we all wind up the switching frequency so that it is outside of the audible range.

The downside to this is that it pushes electrical energy out into the network at frequencies which are quite tightly controlled. The regulatory authorities consequently demand EMC filters to prevent this from happening.

The EMC filters for motor inverters are usually quite big, because AC induction motors are a heavy load. The electrical noise is powerful and at high frequency. These signals can interfere more easily with radio systems. Appropriate EMC filters can leak quite a lot.

Most loads are not big motors.

In your workshop, you probably only need one inverter running at a time. If they have a proper isolator, then they probably don't all need to be connected at the same time. If they are not all connected at the same time, then their EMC filters cannot accrue leakage.

Decent inverters usually have selectable EMC filters. I bought a Schneider, and it says that if you disable the EMC filter then you should run it at a low switching frequency, to remain compliant. Sure you can hear the motor hum but so what. Mine actually has a dynamic mode. If the load increases, the switching frequency reduces. This maintains compliance, and is quiet most of the time. I actually quite like it. I get feedback about how I am loading the motor, by the sound it makes.

I commissioned a medical robot a while back which had several powerful (>1HP) axes, all of which had their own inverter. Obviously the axes all had to be on at the same time to achieve co-ordinated motion. Being a medical robot it had to be compliant. The answer was to provide external EMC filters. A single combined filter does not have to have the same leakage as the sum of each filter together.

Conducted emissions are only high when the motor is loaded. It is not likely that all axes will be maximally loaded at the same time. Even if they are, it is only a temporary situation.

Increasingly electrical switchgear manufacturers are promoting RCBO's. These provide discrete RCDs combined with traditional thermal-magnetic breakers. Instead of having a single RCD each "fuse" has its own. By using these RCBO devices, accrual of leakage can be managed, and nuisance tripping reduced.

In high load situations perhaps greater than 50A it might be the case that it is hard to find an RCD capable of the job. It should always be remembered that inverter systems can often be isolating.

The isolating nature prevents the controlled AC voltage generated from being directly hazardous. Clearly, if a circuit with no RCD protection is used and used in a fixed wiring scenario (supply side trunking/conduit) then safety hazards are inherently constrained.

Obviously no RCD with a non-armoured trailing cable has its hazards.

Andy.

Edited By Andy Ash on 21/10/2016 23:09:31

Never had any trouble with RCDs and inverters, and I've had 4 of them, but the first one used to blow the MCB on startup because of inrush current to the capacitors. I made up a double switch arrangement which put a resistor in series for a secondor so to charge the caps, then shorted out the resistor. Can't remember any values ( probably sized to give 15A or so with full mains voltage) , and that machine went on its way years ago. You can get 'slow blow' MCBs I've since found, must get one for the welder circuit.

Edited By duncan webster on 21/10/2016 23:10:17

Edited By duncan webster on 21/10/2016 23:12:32

Well done, Andy Ash, some sensible rmarks at last! Many of the comments regarding RCD's show a complete misunderstanding of their function and, indeed, where they are required under the current wiring regs.

If you are running a sub-main to a garage / workshop via, for example, a SWA cable then there is no requirement to RCD-protect that cable. Modern consumer units allow for unprotected ways and the submain should be taken from this section, the regs require that the cable is suitably protected and the armouring, earthed at the source, provides this protection. The RCD should be at the next port of call i.e. the garage / workshop consumer unit.

30mA RCD protection is for "person protection", in other words to try to save your life if you receive an electric shock. They compare the current flowing in the line and the neutral conductors which should be perfectly in balance. If there is less current returning in the neutral than went out in the line (or live, as we used to call it) then it is going somewhere else - probably through you to earth. This imbalance trips the RCD before your heart goes into fibrillation.

Another thing to note is that RCD's do not discriminate. If there are two 30mA RCD's in series and a >30mA leakage occur, then both RCD's will trip. Electricity moves, to all intents and purposes, at the speed of light and the mechanics of an RCD cannot act quickly enough to ensure that the unit closest to the fault switches off before the other(s) follow suit. Another reason to ensure that the supply cable to your workshop is NOT protected at the house end.

Caveats: if the workshop supply cable is just some T&E then extra protection, such as conduit may be advisable. If the supply is TT (you have one or more earth rods) then there may be a front-end 100mA RCD. This is not the same as a 30mA RCD and has a different function, it is there for fire protection. The advice is that the heart goes into fibrillation at about 40mA and a 100mA RCd is, obviously, not going to be much protection to you if you are unfortunate enough to receive an electic shock. A 30mA RCD and a 100mA RCD WILL discriminate if the fault current is less than 100mA but, otherwise, both will trip as will two or more of the same rating.

If contemplating setting up a new workshop, it is well worth talking to your local spark. It could actually save you money, and it may save your life if things go pear-shaped. But, hey-ho, we're engineers and know it all, dont we?

Steve

Not on the evidence presented here.

Andrew

Regarding my comments about screened cables.

The screening of the cable between the invertor and the motor is required to reduce radio frequency interference. An SY type cable with braid is probably sufficient for most invertor installations. The reason for earthing screening at one end only is to prevent the chance of a circulating current being set up in the screening, probably not a problem in an invertor circuit but it can be a problem in cables being used with encoders etc because they can pick up pulses that have been induced into the cable (if its not screened or if both ends of the screening are earthed). If an controller starts to count extra pulses, and not just the ones from an encoder then it will lose its position. On a CNC router or similar machine this can be serious. Hence the convention for earthing screening at one end only.

Its probably worth mentioning that, in my opinion, swa cable armouring and SY cable braid (where SY is used as an incoming power cable) should, obviously, be earthed but additional earthing should also be provided. In the course of my work I regularly find machines that are not earthed because the wire armour or the braid has been pulled out of the cable gland.

Martin

Nicely put. Clearly not one to make undue friends or think before opening your mouth. Great to have such a modest expert in our midst!

Teco inverters that I have serviced have a thermistor preceding the caps to prevent inrush current surges. It is not until the under voltage detection circuitry has cleared then the main power relay is energised bypassing the thermistor....

As for mcb rating, there are different trip curves, A, B, C and D.....these are what you are referring to.....

PMSL

We armchair experts are renowned for our thoughtful, informative and sensitive replies when responding to some of the drivel misconceptions written (about) on this august website.

However, it is difficult to endear oneself to all and sundry whilst retaining one's position in the saddle of such a lofty mount.

Regards, Steve

Not followed this thread, but have now looked back at replies, and editing times in particular. It seems that (perhaps) we can guess that some information or other has been deleted/changed since some original post(s)? If that is the case, it would be better all round if mistakes were highlighted with an edit to make the correction clear.

We had an electrical check recently as our wiring is a few decades old. Found a few things and the comment was 'the best bit is the wiring in the workshop' as its the only bit completely up to modern standards. Guess who wired up the workshop

N.

We had to get a new washing machine a couple of weeks ago. Today was the first time I've used the mill, which has a Newton Tesla 3ph package on it, since the arrival of the Bosch washing machine. Lights out in the workshop again. The washing machine needs to be switched off at the wall if I'm to use the mill, so no milling on Sundays 'cos that's the day we get our free electricity from British Gas and the washing gets done.

Sheesh

Rod

Martin, the screen on the motor/inverter cable definitely should be connected to ground at both ends. You will find that this is the instruction from all the manufacturers.

Mark

Edited By Hywel Evans on 17/01/2017 08:24:52

Well put Andy Ash.

I would add the observation that whilst the topology of mains filters is generally symetrical Live to Earth and Neutral to Earth, the single phase supply is assymetric in as far as the neutral and earth potentials are usually within a volt or two. The 50Hz leakage component in the filter is predominantly on the Live side which is what causes the imbalance and as Andy has pointed out is additive when more filters (devices) are connected. Hook up enough filters and sooner or later you will exceed the trip current (30mA).

In our old building (Lab) we had no end of issue with earth current running through both the pipework and the reinforcing to the extent of many amps. I spent a week selectively earth strapping a water pipe in one of our electron microscope rooms in order to reduce the field generated by it. I built an active field cancelling unit in the end. We were aware that the main culprit were the many hundreds of computer mains filters each dumping current into the safety earth.

regards Martin

That is only the case when the screen is being used to provide the safety earth bonding. For reducing radio frequency interference it should be connected at one end only and the safety earth provided by a fourth conductor. Otherwise high frequency currents in the screen will give rise to radiation in accordance with Maxwell's laws.

Russell.

Russell, you might want to take that up with the inverter manufacturers.

Mark

No need Mark. I have Mitsubishi and ABB inverter drives and neither manufacturer makes the statement you claim that all manufacturers make.

Russell.

What confuses many, and isn't always clearly set out, is that screen should be grounded at both ends for the power cable between inverter and motor but at one end only for control signal cables (e.g. pendant to inverter or PLC to inverter).

Russell,

No, you are creating a Faraday cage arrangement. Here is the guidance note produced by ABB. You might want to read your installation guides again...

**LINK**

Mark

We can learn a little from RF engineering here. If you had two screened boxes connected by a coax cable carrying RF energy from one to the other, but only connected the outer screen at one end, then interrupting the coax screen means that currents that otherwise would flow on the inside of the screen will have to flow on the outside when they reach the break. As long as they flow on the inside they can't radiate RFI, but once they start flowing on the outside of the cable it starts to radiate. So to stop RFI generated by the switching current waveforms carried on the inner conductors being radiated you have to ensure that the outer screen is connected right through.