Deye PV microinverters lack a safety relay (EU)

[Siwastaja]

But how is that relevant in this particular situation? It's zero current switching under normal conditions. Apart from the temperature, it's ideal condition.

It is well known that relays which switch rarely under load develop significantly higher contact resistance. Just right amount of sparking cleans the contacts. Too much and contacts wear, too little and they oxidize on their own.

[Marco]

Which goes back to making them hermetic, where there is a wallet there is a way.

[f4eru]

again: sealing does not work for power relays, due to self-contamination of the fumes released by the sparks (ozone, metal particles, etc...)

Relays are like ICE engines:
a complex system that is about fit for purpose since 160 years, but we did never kink out all the random failure modes (we still find new ones, high end or not), and is slowly going towards widespread obsolescence.
Every time we make improvements, there are new modes of failure around the corner.

[Uunoctium]

Deye yesterday released an official statement on their homepage on that issue:
In accordance with regulator (BNetzA) Deye urgent recommends users to disconnect these converters until a solution is certified.

https://de.deyeinverter.com/news/company-news/aktuelle-information-zum-mikrowechselrichter-deye-sun600g3eu230-vor-bergehende-beh-rdlich-angeordnete-netztrennung.html

[Uunoctium]

At that page the second image shows internal of the affected converter(s).
At upper right corner an current compensated choke was inserted instead of two 1pole relay.

https://www.pv-magazine.de/2023/07/17/normenverstoss-bei-balkonsolar-mikrowechselrichtern-von-deye-fehlt-schutz-relais/

Relay type is Hongfa HF115 acc to certificate

https://www.hongfa.com/Product/Item/355

[f4eru]

It is really good to see them taking responsibility and getting towards a solution 

This affects all the legislative improvements that usergroups push for the light schuko-fed-in "balkonsolar" type PV installations.

Now next step for the industry would be to update the norm to allow relaisless disabling.

[NiHaoMike]

The high current relays used in EV batteries are sealed with some special gas mixture, most obvious downside is the relatively high cost.

The main reason is to make sure the mains part is safely isolated from the energy stored in the intermediate voltage (e.g. some 400 V) capacitors.

It looks like in that inverter design, the big capacitors are on the other side of the transformer and so the energy stored at high voltage would be relatively small.

[Siwastaja]

Now next step for the industry would be to update the norm to allow relaisless disabling.

Without reading that standard (but having some general idea about how standards work), I'm 95% sure it does not mandate the use of mechanical relay. Additionally, even if standard mandates it, remember that standards are not legal requirements; their existence is only so that if manufacturer demonstrates their product adheres to standard, then it will adhere to the law (and in case of harmonized standards, the laws of every EU member state), simplifying the process. Law itself is more generic, while standards describe the usual, tried and tested ways of doing something. This means, if a given standard seemingly prevents a superior way of doing something, you can always deviate, that just requires even more rigorous proof. This is actually pretty common in innovative technology, because obviously a standard exists only after a certain way of doing something becomes mainstream. Before that happens, a designer has to find similar enough standards for some parts, and just prove their compliance to law without citing any standard for some novel parts.

[Uunoctium]

It seems, there is no transformer in that design. The four big cores are storage chokes. Each 3,5µH and in my view they are belonging to primary step up converters to gerenate the intermediate voltage. H-bridge at output - done. So, how do they isolate panel inputs from mains?

[Siwastaja]

So, how do they isolate panel inputs from mains?

Solar inverters (string or micro) almost never isolate the panels from mains. That is not required, adds a lot of cost and reduces efficiency.

[Jeroen3]

Yeah, it was funny viewing the german from balkonkraftwerk fuming from rage in his video.


Though, the regulations are more complicated. For compliance the inverter doesn't need to have a relay. The installation must have. For the grid code.
This is in basically all of Europe. Here they mandate a "scheider" or separator. Transistors are not a separator. Thus big inverters use a contactor for that. Space enough.
Now I'm not sure if the regulation specific for inverters specifies any different, but that isn't a VDE, it would be an IEC.
Microinverters cannot use any mechanical relays due to their IP rating and wide temperature range. It's not going to happen, there are no off the shelf relays for this use case.

The worst part of this is Deye intentionally putting such relay inside the inverter, completing VDE 4105 certification, then removing the relay.
They could just as well have validated the inverter without the relay with a stipulation that external automatic separator would be required.
But they would have lost their edge to their competitor...

Enphase for example sells the q-relay in europe which provides isolation in abnormal grid states.
The q-relay also provides DC leakage current protection mandatory for PV systems. Even though Enphase uses isolation transformers, turning the changes of Dc fault to negligible..
Most big inverter include DC leakage protection inside. (if they don't do this right you get the Easee situation)
I'm not sure if Deye has DC-leakage protection. If they don't users may need to install a pricey type B rcd.
This isn't VDE 4105 I think, but at least in our installation code NEN 1010 derived from IEC 60364.

Bottom line, they shouldn't have done this to Germans. It probably wouldn't have been a big issue in any other country. But Germans... they love their VDE.

[bdunham7]

Law itself is more generic, while standards describe the usual, tried and tested ways of doing something. This means, if a given standard seemingly prevents a superior way of doing something, you can always deviate, that just requires even more rigorous proof. This is actually pretty common in innovative technology, because obviously a standard exists only after a certain way of doing something becomes mainstream. Before that happens, a designer has to find similar enough standards for some parts, and just prove their compliance to law without citing any standard for some novel parts.

I'm not intimately familiar with EU regulations and law, but I doubt legislators are going to write laws with detailed technical requirements.  In jurisdictions I am initimately familiar with, this usually involves simply adopting a standard.  Additional requirements and details relating to inspection and enforcement may be added, often by administrative authorities empowered by an enabling law.  So most of the time you cannot deviate unless the relevant authority has some legal leeway to do so.  Construction and fire protection are areas where you aren't going to get much wiggle room.

Solar inverters (string or micro) almost never isolate the panels from mains. That is not required, adds a lot of cost and reduces efficiency.

I'd prefer mine isolated, thank you!  I can't speak for all manufacturers, but Enphase inverters are all isolated by design and are 97%+ efficient.  I'm not sure if this exempts them from some other requirements like the disconnect relay, but the standards that seem to be implicated are mentioned in this letter.  I think proper isolation probably solves a few problems, one being that you really don't need a disconnect relay--even if they are mandated in certain areas.

https://enphase.com/download/rcd-manufacturer-statement

[f4eru]

Without reading that standard (but having some general idea about how standards work), I'm 95% sure it does not mandate the use of mechanical relay.

something something something ... physical disconnection.
try to do that without a relay 

The standard is outdated, and makes the inverters respecting it unnecessarily uncompetitive ....

[bdunham7]

something something something ... physical disconnection.
try to do that without a relay 

The standard is outdated, and makes the inverters respecting it unnecessarily uncompetitive ....

It's especially stupid when applied to inverters that have a HF transformer for isolation and I'd argue that the transformer is a physical disconnection.  Sadly that view has not prevailed and EU customers are stuck with a less-reliable system that protects them against nothing.

[Phoenix]

I no longer have access to the standards as I've been out of the grid connect converter area for a little while now, so I could be out of date... but I will add a few things.

VDE 4105 requires that either the device itself or installation has an "Interface Switch" that is triggered by the "Network and System protection (NS protection)".

For the integrated switch case, all the inverter standards I am familiar with (IEC62109 series, AS4777 series, EN 50438, VDE4105), have essentially the same requirement for redundant disconnects:
  1. For isolated inverters - single mechanical disconnect and the semiconductor bridge
  2. For non-isolated inverters - dual redundant mechanical disconnects
In both cases the integrity of the relays needs to be monitored, if they are stuck closed the converter can not operate.

You can read some test reports to get an idea of what everyone is doing, and what the standards require without having them on hand:

VDE4105
  An isolated microinverter with integrated NS protection (single 2 pole relay): https://www.photovoltaikforum.com/core/attachment/93544-envertech-vde-ar-n-4105-testreport-pdf/
  A non-isolated string inverter with dual relays operated by dual microcontrollers: https://solar-distribution.baywa-re.de/out/media/9-SUN2000_-_M2_SUN2000-8_10_12_15_17_20_M0_M2__VDE-AR-N_4105_Report_BV_Eng_2020.04.22.pdf

EN 50438
  A non-isolated string inverter with dual relays https://voltaconsolar.com/DOWNLOAD-FILES/Hybrid%20Inverters/VOLTASOLAR%20SUPER%204KW%20PV170207N030_4K_EN50438_2013%20report_0.pdf see section 4.6.3 regarding testing single relay defect and "shutdown device".

An example of a stand alone network switch:
https://www.dold.com/media/pdf/95/3d/77/RP9810_en.pdf

It's especially stupid when applied to inverters that have a HF transformer for isolation and I'd argue that the transformer is a physical disconnection.  Sadly that view has not prevailed and EU customers are stuck with a less-reliable system that protects them against nothing.

The HF transformer is considered in the switch requirements - with a transformer only a single mechanical disconnect is required. It's not just EU; the requirements are basically equivalent in every place I am familiar. They key missing here is redundancy... the transformer has no redundancy as a physical disconnection.

Solar inverters (string or micro) almost never isolate the panels from mains. That is not required, adds a lot of cost and reduces efficiency.

Every commercial micro-inverter I am aware of is isolated. Maybe I missed the evidence that the Deye was not isolated, but the construction looks 100% isolated to me (based on the linked youtube video - probably flyback). Thus they would only need a single relay per pole. I am not sure if microinverters MUST be isolated to meet IEC62109 as the PV side could be considered DVC-A (like SELV) and require isolation to the mains at DVC-C (hazardous) - I would need to read the standard again to confirm. It would be interesting if anyone can link a non-isolated microinverter.

String inverters are largely non-isolated now - for obvious efficiency and cost reasons.

[Jeroen3]

I don't know, but VDE is quite clear about what it should have.

The structure of the section switch is designed taking into account the single-fault safety (see explanation in A.6)
to do.
The tie switch (e.g. power relay, contactor, mechanical circuit breaker, etc.) provides single-fault protection
an all-pole galvanic disconnection.
In the case of generating plants with converters, the section switch must be provided on the grid side of the converter.
A short circuit in the converter must not impair the switching function of the section switch.

Dutch grid code (actual law) also uses the word "separator", imho mosfets are not a "separator".

I'm not intimately familiar with EU regulations and law, but I doubt legislators are going to write laws with detailed technical requirements.

You'd be surprised how much technical details, limits and maths is defined in energy laws. A harmonised european synchronous grid must require this I think.

[nctnico]

RelayGate: Deye Solar Microinverters Lack Essential Part (https://www.heise.de/news/RelayGate-Deye-Solar-Microinverters-Lack-Essential-Part-9220307.html)

TL;DR:
- they got their microinverters certified with the required safety relay
- sold tons of microinverters without the relay (more profit, 400k+ inverters sold in Germany)
- someone opened an inverter and noticed, others checked too
- big oopsie

AFAIK this is pretty common. For the Enphase microinverters you'll also need an external relay box between the microinverters and the grid connection in order to comply with regulations.

[nctnico]

Now next step for the industry would be to update the norm to allow relaisless disabling.

Without reading that standard (but having some general idea about how standards work), I'm 95% sure it does not mandate the use of mechanical relay. Additionally, even if standard mandates it, remember that standards are not legal requirements;

In most cases they actually are legal requirements. Just look at laws that define building codes which then refer to electrical and other safety standards as being mandatory for compliance.

[Phoenix]

AFAIK this is pretty common. For the Enphase microinverters you'll also need an external relay box between the microinverters and the grid connection in order to comply with regulations.

The difference here is that Enphase got their system certified with the external switch unit which communicates with each microinverter. Deye would need to under-go recompliance and prove the arrangement with external switch is acceptable.