Inverter Tripping RCD

[Barnabas TaylorParticipant @barnabastaylor89961]

Posted by Steve Skelton 1 on 08/04/2020 13:02:37:

Hi Barnabas,

Yes you are right , it is the DC component that causes the problem. You can use a basic AC type RCD which will trip on a pure AC fault but if the DC component is too high it may not trip. I am not an expert on VFD systems but the same problem can occur with charging systems for electric car charging systems and solar power installations.

It is no consolation for you but the price of Type B RCD's will drop as the demand increases.

I imagine that most people do not fit Type B units – probably from a lack of understanding they fit AC types.

My advice would be to do exactly as the manufacturer recommends as they are the experts but I would have thought that it is only the supply to the VFD that needs the Type B RCD and not the rest of the house.

Steve

Cheers Steve, I shall probably just get a sing RCB then. I shall think about the relative merits of the RCDs you sent, then I shall stop lying to myself and the bank manager and buy the cheap one! I'll fit it just to the one DB, that way I am protected and I shall just have to see if the VFD trips the 30mA earth fault. If so, it is back to the drawing board!

[Clive Brown 1Participant @clivebrown1]

I purchased a Schneider VFD for a mill about 6 months ago. Tripped the domestic RCD straight away. Not something i'd experienced with the Teco VFD on my lathe After a brief investigation into type B RCDs, I read the Schneider manual carefully and found a warning about just this problem and its solution. The EMI filters can be disconnected, an internal, removable bridging link is provided for the purpose. Problem solved with no knock-on effects.

Then looked into Teco manual, there is similar provision, which I've not needed to apply. Fingers crossed!

My other VFD, a cheapo Ebay jobbie, gives no trouble. I doubt it has EMI filters.

[Barnabas TaylorParticipant @barnabastaylor89961]

Posted by Clive Brown 1 on 08/04/2020 13:39:12:

I purchased a Schneider VFD for a mill about 6 months ago. Tripped the domestic RCD straight away. Not something i'd experienced with the Teco VFD on my lathe After a brief investigation into type B RCDs, I read the Schneider manual carefully and found a warning about just this problem and its solution. The EMI filters can be disconnected, an internal, removable bridging link is provided for the purpose. Problem solved with no knock-on effects.

Then looked into Teco manual, there is similar provision, which I've not needed to apply. Fingers crossed!

My other VFD, a cheapo Ebay jobbie, gives no trouble. I doubt it has EMI filters.

Correct me if I'm wrong, but if you disconnect/bridge the filters, isn't the whole thing still unsafe unless you fit a type B RCD? (Not to mention the potential 'noise' in the electrical circuits) I am going to wire things up today (probably, if I can find some suitable wires in the box!) and see if it causes trips, if it doesn't, I shall get a type B RCD and fit that. if it does trip, then back to the drawing board! Either I shall have to remove the filters, or I shall have to run a dedicated line just for it!

[Phil WhitleyParticipant @philwhitley94135]

I am taking on board what is being said here, but a VFD without an EMI filter would not last long before a mains spike would take out its electronics. they are there to protect the VFD from the mains, and the mains from any EMI The VFD priduces. EMI filters are capacitive, and therefore BLOCK DC current An RCD is tripped by current flow imbalance, and therefore "you would think" that the AC/DC thing should make no difference, as it is current imbalamce between line and neutral which causes the trip, unless there is some part of the RCD circuitry that reads a DC component in the line as more current than it actually is? If this is the manufacturers explanation, and it works, then the only thing we can argue about is the price, which does seem to be an awful lot for what you actually get. I am lucky, I have "real" three phase, and only have RCD protection on the single phase installation.

Phil

[Clive Brown 1Participant @clivebrown1]

For disconnecting the Schneider filters, the caveats in the manual are that to safeguard the VFD the drive switching frequency should not be set above 4kHz and also EMC compliance is not guaranteed.

[Barnabas TaylorParticipant @barnabastaylor89961]

A quick update, after some fiddling around with wires, I have got the VFD running, as yet I cannot 'load' the motor as I am waiting on a drive wheel but it hasn't tripped the house yet! This bodes well for just having to purchase an RCD for safety and being able to keep the filters intact on the VFD. Fingers crossed I am only out of pocket by another £100 and change…

[LynneParticipant @lynne]

An interesting discussion. Having read the F_____ manual, it tells me that all vfd's have a leakage current to earth, the levels of which are affected by motor cable length and type, effective switching frequency, earth connection used, and type of RFI filter installed. If an ELCB is used, it must be fitted with a type B device, The vfd I have has an external screw which can be 'removed' if tripping is an issue. Now logic suggests to me, that no mfg. would tell one to do this if transient spikes from whatever source are going to destroy the vfd. There has to be some form of protection built into the drive. But hey ho, I am no electronics buffa. Regards, Lynne

[Phil WhitleyParticipant @philwhitley94135]

Thats interesting Lynne! I wonder what the screw actually does? If not dissconecting the filter entirely, it must perhaps change the way the filter works? The leakage curret to earth is the filter dumping what charges it has filtered out of the supply, and also any harmonics that the vfd emits back to the supply Maybe it is unwise to ponder on anything beyond finding a solution to the immediate problem?

Phil

[Anonymous]

Some of the confusion arises from the fact that both RCD and overload functions use letters to determine the details, even though the functions are different and may be different parts.

If we look first at the overload function there are three common classes B, C and D. Class B will trip at 3 to 5 times rated current and is normally used in domestic properties. Class C trips between 5 and 10 times the rated current and is normally used in commercial properties. Class D trips between 10 and 20 times the rated current and is normally used in industrial properties. Class C and D are used where larger inrush currents are expected. All the breakers in my fairly new consumer unit are type B overload curve.

Confusingly the RCD function also uses letters to determine characteristics. The three most common are AC, A and B! Type AC only responds to AC residual currents. Type A also responds to pulsating DC currents and type B further responds to smooth DC currents. So in theory a VFD should be running from type B. But this makes it more likely the RCD will trip as it responds to more types of inbalance. Of course that might be seen as a good thing. The RCDs in my consumer unit are type AC and I run two VFDs on my CNC mill with no problems. Never had an RCD or overload trip.

Both VFDs run via EMI filters which contain a common mode choke and common mode and differential mode capacitors, but do not contain MOVs or transzorbs to catch mains spikes. Of course the filter components will attenuate mains spikes to some extent, but fast spike attenuation is left to the user, at least on the VFD I installed. I live in a rural area with overhead 3-phase feeders and fingers crossed I haven't had a spike related blow up yet.

Andrew

[Steve Dunthorne 1Participant @stevedunthorne1]

Hi Andrew, I trust you are well in Cambridgeshire.

Your third paragraph states you never suffered an RCD trip. I know you are extremely well versed in electrical engineering matters, so I bow to your greater knowledge, but may I 'throw in' a teaser that a VFD, as well as potentially causing nuisance RCD trips due to its EMI filter action, can conversely prevent them from tripping by what is known as 'blinding' them due to the stray DC current saturation effect. And it is explicitly, is it not, the type AC RCD which is susceptible to being blinded, the type you say you have fitted (the normal type most people have).

If you chose to fit type B or type F RCDs specifically because of the VFD drives, and they did not trip, then you would still not really know 'by how much' they were not tripping, if you see what I mean.

Steve

[old martParticipant @oldmart]

When I was researching the possibility of having a VFD for the museum's Tom Senior light vertical, the earth leakage was a worry, especially as I wasn't sure if the RCD's were 30 or 10 milliamp. The Schneider Altivar 750 watt has never tripped the power, fortunately. This model ATV 12 has EMC filters for C1 and C2, which were switchable and the earth leakage would have been less if not used. I did not need to switch them off.

**LINK**

[Steve Dunthorne 1Participant @stevedunthorne1]

But, again, you don't actually know 'by how much' you don't appear to have a problem. That's why, I believe in industry they don't use RCDs on VFD drives. Please correct me if I am wrong.

[old martParticipant @oldmart]

In industry, they don't often use 240V single phase from a wall socket to power machinery.

[Steve Dunthorne 1Participant @stevedunthorne1]

True, but there are 3-ph RCDs

[martin perman 1Participant @martinperman1]

My last job before retiring was installing industrial washing machines in Universities and Research facilities, the wash pump and drain pumps were run by VFD's and we had to change the 3 phase RCD's to B or C class because of tripping.

Martin P

[SillyOldDufferModerator @sillyoldduffer]

Posted by Steve Dunthorne 1 on 02/04/2022 16:34:31:

True, but there are 3-ph RCDs

There are, but I don't understand how they work or what the value is!

In a Live/Neutral/Earth single-phase set-up an RCD trips when it detects a current imbalance between Live (in), and Neutral(Return). The imbalance can only be caused by Live leaking to Earth, which often signals a fault. Straightforward and so sensitive it can detect when the earth leak is through a human and pop before serious damage is done. Unfortunately also sensitive enough to detect high-frequency electric muck dumped to earth by EMC filters.

Presumably, because there's no neutral, a 3-phase RCD works by comparing the current in each phase with the sum of the currents in the other two? In which case, maybe, a 3-phase RCD is only worth fitting to a balanced load?

Possibly circuit breakers are more appropriate to a 3-phase system. My impression is industrial electrics favour locked terminal boxes and procedural protections rather than domestic style consumer units. Perhaps a different approach is taken because properly installed industrial systems don't have numpties and children fiddling with them!

I'm way out of my depth though!

Dave

[Steve Skelton 1Participant @steveskelton1]

SOD,

Three-phase RCD's work in exactly the same way as a single-phase RCD.

It does, however, have four terminals one for each of the phases and the neutral.

Due to the 120-degree phase rotation if all the phases are balanced then there is no current flowing in the neutral but if any phase is pulling more than the other two phases then there is an equal current (to the imbalance) in the neutral. The three phases are wound around the sensing coil as if they were one phase, but obviously insulated from each other.

Steve

[Anonymous]

A residual current detector (RCD) is basically a current transformer. Normally the output of the secondary is proportional to the current in the single turn primary winding. However, if two wires are used then the output will be proportional to the algebraic sum of the currents in both wires.

For a single phase system both live and neutral wires are used. If there is no leakage the currents in the live and neutral should be the same magnitude but opposite sign. So the magnetic fields will cancel out, and the output of the current transformer will be zero. If there is leakage to earth then the currents will not be equal, the magnetic fields will not completely cancel, and there will be an output from the secondary.

The same principle applies to a 3-phase RCD. There are two cases, with and without a neutral wire. Consider a 3-phase system connected in star but with no neutral. Applying Kirchoff's current law to the star node the algebraic sum of the three currents is zero. Hence the algebraic sum of the three phase currents is also zero. Whether the system is balanced or not is irrelevant. In this case a 3-phase RCD simply has all three phases running through the current transformer. Intuitively the same is true for a system connected in delta, but I don't have the inclination to prove it mathematically.

If we now look at a 3-phase system in star, with a neutral line, the same applies. The algebraic sum of the the currents at the star point, including neutral, is zero. In this case four wires (three phases and neutral) are run through the current transformer.

So a 3-phase RCD may have three or four wires, and they will work irrespective of a balanced or unbalanced system.

Andrew

[SillyOldDufferModerator @sillyoldduffer]

Posted by Andrew Johnston on 03/04/2022 10:50:30:…

So a 3-phase RCD may have three or four wires, and they will work irrespective of a balanced or unbalanced system.

Andrew

Many thanks for the explanation and conclusion: it was the balance/unbalance part I couldn't get my head round.

Brought back failing to understand Kirchoff in a Physics lesson. One of several similar moments where I had to admit my maths wasn't anywhere good enough for the sort of career I aspired to. Being an idle youth didn't help either.

Dave

[Anonymous]

Posted by Steve Dunthorne 1 on 02/04/2022 13:46:53:

…'throw in' a teaser…

It is fairly simple to calculate leakage currents due to filter capacitors. It would be more difficult to calculate DC leakage as this would normally result from a fault condition. I agree that a type F is ideal for a VFD, but in practice with the relatively small VFDs used in the home workshop it probably doesn't matter. The only surefire way to know how close one is to the trip limit is to actually measure the leakage currents.

I take the pragmatic view that if something is working then leave it alone.

Andrew

[Anthony KnightsParticipant @anthonyknights16741]

Posted by Andrew Johnston on 03/04/2022 20:47:26:

I take the pragmatic view that if something is working then leave it alone.

Andrew

If it ain't broke………………………………………………………………………..

I wish someone would tell Microsoft.

[Author]

Posts