On 2/1/2023 12:25 PM, Alex MeVay via RE-wrenches wrote:
Although our controllers are probably smaller than what you would be
considering in this discussion, surviving a load dump (suddenly
disconnected battery) is an engineering requirement for us, and likely
would be for other responsible MFG's on this list (boB?). The charge
controller can either handle full input voltage on the output, or
there is a comparator that will shut the controller down instantly
when the output voltage gets too high.
So that's part of the picture.
The other part of the picture is the rest of the system. The discussion
and what NEC is mandating is that the battery be disconnected. As far as
I can tell, there isn't an explicit requirement in all cases that all of
the power inputs into the system are all going to go away at exactly the
same time .... or ever. This leaves things like charge controllers
getting input power potentially from PV or wind or grid or generator or
something else. And nothing in the NEC, as far as I can tell, is
mandating that all of the DC loads be disconnected. So the loads are
online. Now the regulation circuit (switching power supply in the case
of a MPPT controller) is regulating variable loads and nothing is
providing substantial resistance to change. Normally the battery is
acting like a very very very big capacitor. But without the battery,
there isn't enough damping in the control loops and voltage stability
will suffer. This is where you get 250 volts on your normally 48 volt
battery bus. Or 3 volts. Or -80 volts. Or all of those in a fraction of
second. This sort of thing is hard on electronics and will cause failures.
Alex's comment about Genasun's controllers handling full input voltage
on the output is probably unique to their niche product. Looking at
their biggest controller, it appears to support VOC of 34 volts. At 34
volts, he can use 50V or 100V rated components on his output. For a 150V
input controller, you probably can. But you wouldn't because it would be
too expensive. But on a 600V or 1000V controller, there's just no way
that you do that because 1000V rated components are big and expensive
and their spacing requirements are huge compared 100V level sort of
stuff. So the way bigger MPPT controllers deal with load dump is with
transient voltage supression or other diodes to handle the voltage spike
caused by the inductor when the load goes away. When operated in
parameters, these parts don't wear out. But it can be interesting to
size them adequately to account for inductance elsewhere in the system
adding to voltage and energy that has to be absorbed.
There is a very simple experiment that anybody can perform to see how
equipment handles a load dump: Simply wait for a sunny day and turn off
the battery breaker. If no magic smoke was released, turn back on
battery breaker. If still no magic smoke was released, then great, your
system survived a load dump.
If you are at all uncomfortable doing this and/or your system gets
destroyed in the process, think back to my earlier security comments
about having a self destruct switch on the outside of your building
allowing anyone walking by to do this experiment for you.
I have personally destroyed Outback, Midnite, and Morningstar
controllers inadvertently or intentionally doing load dumps by shutting
off their output breaker. Typically the TVS diodes short out and
secondary over current protection (circuit breaker) trips before things
catch on fire. Usuaully the UL94V0 rating on the circuit board and the
box the circuit board is in prevents fire from spreading when things do
get wild.
But I've also seen all of those brands survive a load dump.
I have multiple customer who have fielded lithium battery systems to
cold locations and have had battery BMS disconnect the battery from the
rest of the system. Ten's of thousands of dollars of equipment has been
destroyed in these islanding events.
Some thoughts, for what they are worth.
-James Jarvis
APRS World, LLC
_______________________________________________
List sponsored by Redwood Alliance
Pay optional member dues here: http://re-wrenches.org
List Address: RE-wrenches@lists.re-wrenches.org
Change listserver email address & settings:
http://lists.re-wrenches.org/options.cgi/re-wrenches-re-wrenches.org
There are two list archives for searching. When one doesn't work, try the other:
https://www.mail-archive.com/re-wrenches@lists.re-wrenches.org/
http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org
List rules & etiquette:
http://www.re-wrenches.org/etiquette.htm
Check out or update participant bios:
http://www.members.re-wrenches.org
