VFD and RCD compatibility

[Jieve]

Hello,
We are working on a conveyor system to be used for educational purposes. The system consists of 7 control cabinets: 1 main for power distribution and e-stop circuits, 5 sub-cabinets, each with Siemens PLCs and 1 motor control center. The first component in the main cabinet is a 3-phase type A 30mA RCD. There are 5x 0.4hp 3-phase (400V, 50Hz) motors controlling the conveyors, and each will be controlled by a Siemens G120 VFD. All motor cables will be shielded and grounded at both motor and VFD. The VFD’s do not have built in filters, and we are not using any externally.

My question is regarding using RCDs with the VFDs. The G120 manual recommends using a single 300mA type B RCD before each VFD. I have done a lot of web research, and there seems to be mixed opinions about this – there is a large number of people who say to leave the RCD’s out because they constantly nuisance trip due to earth leakage currents. To simulate our system, I ran a bench test running a single type A 30mA RCD upstream of 5 of these VFDs running 5x 1hp motors. I varied motor speeds and run a number of tests with no problems.

As I understand it, type A RCD’s will trip on sinusoidal currents with DC offset, and type B RCD’s will also trip on pure DC currents. I want this system to be as safe as possible as students will be using it for PLC programming practice. We are not using braking resistors on any of the VFDs. Here in Germany, 30mA RCDs are required for direct touch protection. I have read that 300mA are allowed when using VFD’s because the harmonics in the rectified signal results in higher frequency fault currents, which are not as dangerous to the cardiovascular system at 30mA as 50Hz AC currents.

My question is then: Would it make sense to add these type B 300mA RCDs before each drive or are they redundant and can we leave them out? Is it correct that the type A 30mA RCD that is feeding the rest of the system would not catch DC earth leakage currents that may potentially cause injury, therefore requiring the installation of the 300mA RCDs, or is this not the case? In one Schneider document I found on VFD’s and RCD compatibility, they recommend 30mA type A RCDs on each drive in a TN-S system (our grounding system), which is different than the Siemens recommendation, and seems redundant if our main RCD doesn’t nuisance trip.

I’d really appreciate any info from anyone with more experience than I in this area. Thanks!!

 

[jraef]

The problem is nuisance tripping. If you are not experiencing that with a 30mA RCD, then go for it.

Just be aware that the RCD is not necessarilly protecting anyone from anything down stream of the VFD, because a ground fault on me output has no effect on the input current other than the increase. Most VFDs have equipment ground fault (earth leakage) protection for the output side, but not at a personnel protection level.


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[ScottyUK]

That's not strictly true unless the drive incorporates an isolation transformer. The DC bus is connected to the most positive and most negative phases at any instant in time, so there is a definite earth fault path between the input and output although the current which flows on the input side as a result of an earth fault on the output will possibly confuse the RCD.

 

[Parchie]

@Jieve (Mechanical)
Cascading type A and type B RCDs do not ensure those breakers discriminate well. Even if you employ a 300mA, then a 30ma in cascade fashion, there is no certainty that the 30mA RCD will trip ahead of the 300mA if the leakage current will be 300mA--> both will sense that leakage at the same moment and respond accordingly. The reason is that both types do not offer a bit of delay when sensing leakage current. You can only achieve discrimination if you use a type S or a type G RCDs(20-40ms delay min) ahead of the type A's. To be sure, it would be best if you contact the VFD manufacturer for advice on the most fitting RCD for your equipment.

You also need to solve the nuisance tripping problem (tripping on leakage current due to common mode noise or insulation capacitance leak, not on actual ground faults). Measure the leakage current of your loads using a good milliamp clamp meter (I used Fluke 360) when you energize them. If you get a value very near the RCD sensitivity, split your circuits to avoid nuissance trips. Since you have the time to experiment, try different brands of RCDs and see which of them gives you the discrimination you need while not sacrificing personal safety of your students.
I hope that helps a bit.

 

[PersonalProfile]

The following may be of interest:

http://www.eng-tips.com/viewthread.cfm?qid=316664

Regards,
Lyle

 

[Jieve]

Thanks everyone for the responses.

We are not currently using drive isolation transformer(s), and weren't planning on it as long as we don't have nuisance tripping issues.

Jraef: As I understand, any capacitive coupling to grounded metallic objects/conductors would result in an increased current draw by the VFD with the corresponding current travelling back to the mains transformer through the PE connection instead of the phases, which should trip the RCD. A ground fault would similarly result in a high current back through PE and should similarly trip the RCD. And any inductive effects shouldn't result in high enough currents to injure anyone I wouldn't think. Could you explain what you mean with the RCD not protecting anyone downstream of the VFD's?

Parchie: thanks for this input. There is actually a main cabinet on the wall ahead of this system (since it is located in a classroom) and there is a 63A type A RCD with 30mA trip ahead of our main cabinet which feeds 4x 3-phase outlets in the room. While it would be nice to be able to have the RCDs discriminate, at the very least for educational purposes, we ultimately don't know what will be plugged into those other outlets in years to come. So we'd prefer not to swap out the wall cabinet RCD, and if both RCDs trip at the same time, it's not too big of a deal for us. We've already had some difficulty getting our breakers in the system to discriminate, overload is ok but under short circuit conditions it's pretty much not happening. But that's something we can live with, as long as the lowest branch breaker trips we have an idea where the issue is coming from, even if it does cut power upstream.

Once I've got the MCC built and wired up, we'll do as you suggested and measure the earth leakage current, probably with a scope. That should be informative.

So would it be advisable then, to replace the type A RCD in our main system cabinet with a type B also rated at 30mA, that would detect DC currents as well, or is this unnecessary? While I am not seeing a direct need for this, I would prefer to defer to people who have more experience and understand the phenomena better than I do. Again, students will be programming this system, and I want to prevent any potential safety hazards.

 

[itsmoked]

Could you explain what you mean with the RCD not protecting anyone downstream of the VFD's?

The RCD looks for an imbalance in the power supplying wires to the VFD. It won't see ground currents originating from the output (motor side) of the VFD.

Realize that the AC source power to a VFD is simply charging and keeping charged a capacitor bank the the rest of the VFD uses to synthesize the motor's power. All the line-side RCD can see is that capacitor charging function.

Keith Cress
kcress -

http://www.flaminsystems.com