Dangerous voltage between positive and ground (structure)

Hi All

I havent posted anything for ages, basically all has been going fine and I have had no real technical issues with any installation.

However, last week I got a shock.... literally!! We were testing a system of 5x315Wp panels connected  in a single string (38x5=190Voc) to a 5kW Voltronic 48V inverter. Now these inverters are now coming with MPPTs of higher and higher voltages- a concern in itself when amateurs are involved. Anyway, to get to the point I had to remove a badly crimped MC4 in order to replace it at the +ve end of the string:

  • The string breaker (only +ve line) was in the disconnected position and the inverter was switched off. However, despite this when I inadvertently touched the bare +ve wire at the end the string with my finger and my elbow touched the structure I got an electric shock.
  • I tested the voltage between the +ve and the structure and it gave the full open-circuit voltage of 190Vdc. I checked the negative wire for any chaffing or anywhere that could cause a ground fault and found nothing.
  • I then disconnected the -ve from the inverter and the voltage between the +ve end of the string and the structure dropped to around 60Vdc (which I presume is a phantom voltage)

Therefore my conclusion is that the inverter itself has a negative to ground fault, because surely I should neither get the full Voc nor a shock from an isolated, floating PV string..... I assume! The inverter also seems to be charging erratically, which is why we were asked to check in the first place.

I would greatly appreciate your input on this, to hear if anyone else has come across similar issues at some time and as I wish correctly understand the isolation of floating systems and take necessary precautions in the future

Thanks
Larry


Comments

  • BB.
    BB. Super Moderators, Administrators Posts: 33,431 admin
    Welcome back Larry--Very happy to hear you are alive and kicking!

    My guess... For standard PWM/MPPT type charge controllers--They usually do not have any "isolation" (no transformers, etc.) between Vpanel input and Vbatt output.

    So, that would mean if your battery bank is negative grounded (to ground rod, electrical safety ground, metal water/gas piping, etc.). then the ground is carried all the way through the inverer-charger-solar charger and back out to the solar array.

    And the only way to "float" the solar array is to open both + & - circuits from the Inverter-Charger.

    If you did not tie battery bank return/negative to earth or safety ground, then it is possible that you have some leakage current in any your DC loads, or you have a DC Ground Fault detection system (NEC's brilliant idea from a decade ago)--Basically they tie a fuse or breaker between the negative lead of the solar/DC side and to the chassis/safety/green wire/building side ground side of the house wiring.

    If there is a short to ground from + or - on the solar/battery side (aka "ground fault"), that pops a ganged breaker. 1 amp breaker between DC negative and chassis/safety ground opens, and is ganged with two other breakers to open the +/- wiring to the array. Or they use a fuse--And simply turn off the solar charger if the 1 amp fuse blows. (you have to check the manual to see if they have the fuse/breaker for ground fault or also known as "Arc Fault" protection circuit). (note the can also be a DC "Arc Fault" detection circuit--It listens for the electrical hash from an actual Arc--Then shuts down solar charger. The issue NEC was trying to solve is the Arc Fault delema in DC solar panel wiring--How to stop poor connections/short circuits from arcing and starting fires).

    Anyway, in summary: Two possible locations/causes for "ground referencing" the array--The "ground/arc" fault detector inside the inverter-charger, or simply tying battery negative bus to building safety ground.

    Only solar arrays for transformer based Grid Tied inverters may be floating (the NEC DC Ground fault can ground reference array). And transformer-less GT inverter systems reference the array to 240 VAC power lines (no isolation).

    In a "modern" solar power system these days--Do not assume the array is not ground referenced through the attached electronics (solar charger/inverter/GT inverter)... Very good chance that they are ground referenced through the device or even your own DC negative to ground safety connection. And if the DGFI fuse/breaker opens, you could see the array + line shorted to ground, and the negative side showing -190 VDC (I hate fuse/breaker between DC return and Safety Ground--A total violation of safety design).

    Take care,
    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • Dave Angelini
    Dave Angelini Solar Expert Posts: 6,728 ✭✭✭✭✭✭
    edited April 2021 #3
    Hmm not knowing how these electronics are designed is an issue. Never heard of the company and the store here does not sell them. What does the other string do if there is one?

    Even if one feels that the solar DC ground always should not involve fuses or breakers, they can and do protect a large well designed inverter/charger from destroying itself in lightning or grid surges. Not always ! I attended training from John Wiles retired in April 2013 as a Senior Research Engineer at the Southwest Technology. He backed it up with the testing and I have seen it in practice. Prayer also is valuable when the bolts start flying.   https://iaeimagazine.org/author/jwiles/

    The wire run from the solar and the solar itself are often the cause of these problems. Does the Inverter have it's own battery charger? If so how does it charge. Need to eliminate different blocks! This is why all of these new all in one power systems are flawed to me. I like having spares and easy ways to eliminate problems Offgrid.


    "we go where power lines don't" Sierra Nevada mountain area
       htps://offgridsolar1.com/
    E-mail offgridsolar@sti.net

  • jonr
    jonr Solar Expert Posts: 1,386 ✭✭✭✭
    I encourage you to entirely drop the "there won't be dangerous voltage between points A and B because I opened switch C" thinking.   Test before touching.   

    I am available for custom hardware/firmware development

  • lazza
    lazza Solar Expert Posts: 336 ✭✭✭
    Thanks Bill, and others who have responded.

    Yes alive and kicking! Although I will need to take special care from now on in order to stay so :smile:.

    We are using pylontech batteries. They do required an earth connection unlike lead acid banks which I presumed this was just in case of an internal fault, and as far as I know the negative is not ground referenced. But maybe the negative is ground referenced either in the pylontech batteries, or, as you say, through some of the internal electronics.

    As extra protection I will make sure double-pole breakers are used to break both positive and negative lines of the string/s... and additionally as Jonr said- test before touching!

    Cheers
    Larry


  • Solar Guppy
    Solar Guppy Solar Expert Posts: 1,989 ✭✭✭
    One thing to watch out for in newer generation of combined with HV mppt inverters is the transformer less designs. The PV connections are NOT ground referenced but are 1/2 Vout typically. PV frames are still grounded, but grounding the PV minus input on one of these will result is magic smoke or worse.
  • Dave Angelini
    Dave Angelini Solar Expert Posts: 6,728 ✭✭✭✭✭✭
    Definitely be careful. We sometimes assume that the readers here know more than they really do. I always have second person who knows what I am doing and where I am. Mistakes can happen and so a second set of eyes or at least the ability to call the Paramedics is super important.

    I am not a fan of any of these all in one transformer less battery based power systems. They are cheap to make!

    The old school design of the low frequency XW with an 80 pound Toroid is peace of mind for many reasons. I still know of one from 2006 that is running fine down in the Arizona desert. Back in the days when you were testing them hey!
    "we go where power lines don't" Sierra Nevada mountain area
       htps://offgridsolar1.com/
    E-mail offgridsolar@sti.net