Thank you for the correction. I was going off memory, and had them backwards. I've corrected the original post above.
The ATS does not have a 5 second "off" time when it switches -- I'd say it's about a 500ms switch time from grid to generator, and a 250ms switch time from generator back to grid. Slow enough to kill power to computers, or to reset the oven clock. But it also has a 5-minute delay for grid stabilization before switching from generator back to grid, and when this process is occurring the Sunny Island has independently seen the return of grid power on AC2 and presumably gone through some sort of process to resync itself with that frequency such that when the 5-minute switchover occurs, you generally don't notice it happened.
Only if breaker feeding SI AC2 is on.
If it is off, would not be sync'd and could blow up SB when input phase jumps. Its FETs could be fighting the grid. At least in theory, based on SMA's warnings.
One guy did put SI AC1 output across the grid and it simply tripped breakers. That was at the end of a long wire.
I apologize, but I am not sure. My primary residence is not at the house, and I cannot get the specific model suffix from the portal on the device. I bought it in 2016, and I have a hunch it is the -40 since it is running (and maxes out at) an older firmware version. It looks like it is set to Island Mode 60Hz. The other options available in the picklist are Default, HECO2017, UL1741/2016, and WorstCase.
I think newer firmware allows Rule 21.
You're switching this direct onto grid? Island mode would not have anti-islanding. It would have wider voltage and frequency settings.
This defeats some of the lineman safety requirements. (conditions needed for it to backfeed a dead line aren't likely to be achieved, but UL-1741 is what is approved for safety.)
Got it. So I think what you're saying is that the SB 8000-US *requires* RS485 to properly work with the Sunny Island in "grid-backup" mode. In other words, I would not have been able to make this older inverter work without RS485. Did I understand that correctly?
Correct, regarding backup mode. Without RS-485 it would remain UL-1741.
It would work in offgrid/island mode. But would not perform anti-islanding. SI would perform anti-islanding. I avoid relying on that, based on what SMA Germany engineers documented (and reports of relays welding shut.)
Fascinating. And scary thought about the relay welding shut, but it makes sense that the possibility exists.
So I figure SMA Germany came up with this during FMEA, deal with it by SI recognizing relay is stuck and telling SB via RS-485 do its own UL-1741 anti-islanding.
I don't trust a different recommendation (from SMA America) if they don't acknowledge the previous written instructions, explain why not applicable, and put that in writing.
It shows "Disturbance VacL1-Bfr". It's been running alongside the SB 5.0 and the Sunny Islands, in the configuration described above, for ~8 years with no issues. It's hard to imagine that it is suddenly being disturbed by the SB 5.0, now.
I think Bfr is frequency; does it quote the frequency observed?
I think mine do. Could occur during dropped power, not just frequency shift (knock offline by going above 64 Hz prior to synchronizing (only if excess production, was above 60 Hz before trying to reconnect.)
Try measuring voltage and frequency.
Perhaps line impedance is different and more noise from SB 5.0 is seen. I haven't checked waveforms from my -41 yet.
I did see poor PF from a VFD upset a SB. Documents say if that error continues, it becomes locked. I was playing with SBS during a grid failure, connected to SI, and discovered my SunPower SB 8000US locked that error, is now bricked. I suspect but don't know SBS current waveform did that.
Yes, I tried this route. Unfortunately the guys who do this don't have SB 8000-US anymore. And I'm just not super interested in riding the SMA train here given the high cost and what I'd argue is a questionable product management strategy.
A bunch of SB show up on Craigslist and eBay. Of course they do eventually wear out, maybe after 20 +/- years, so I prefer low mileage or NOS.
I like to pay $0.10/W, preferably for unused ones.
Be aware some models are 208V only, SB 10000TL-US-10 I think. -12 is 240V (also has AFCI) 10000TL-US-12 does not support backup (I informed SMA, they confirmed and changed compatibility list.)
Interesting. So you're a fellow SunPower user? Did your modules claim to not require positive grounding like mine? Tigo's web site talks about a new device they are working on incorporating which goes between the array strings and the inverter and helps mitigate PID. It looked promising, but not yet available for sale.
I learned about PID some time after getting the SunPower 327W panels. Their data sheet, also the REC and Qcells I plan to use at a new place, claim zero or low PID.
For a year I ran SunPower 327W, AstroPower 120W, Sharp 165W polycrystaline, all in parallel and ungrounded transformerless. Later discovered some Sharp were degraded, don't know if due to that. That model is known for it.
SMA for a while had a bias box to bias string at night to partially reverse PID.
Panel chemistry has been changed to be resistant.
I plan to use SB 5000US series to bias positive or negative as desired, even if low PID.
It appears to be their latest comm bus that interconnects all their competing communication protocols. I am honestly not super impressed with the number of variants they've created in the first place. Hence, combined with the high price tag, my lack of interest in supporting their latest variant.
Monitoring SB isn't so important. Drop in system output, check the displays, use a clamp DC ammeter to check individual strings.
SB -41 with WiFi or SpeedWire, you should be able to hook up and read.
I'm sticking mostly with old stuff, but trying some newer products when they show up cheap on eBay just because. And need recent grid-support features for new system installs and permission to operate.
Thanks again for your detailed responses here.
I have concerns about SB -41 together with SB 5000US on SI, that bad waveforms could upset or brick the other. Don't know if that is a problem.
Rule-21 would let you switch between direct to grid and being on SI.
Backup mode, if I switch, I figure I need to switch RS-485 so when direct on grid it isn't told by SI to be offgrid.
If I was you I would consider getting multiple SB of the 5000 ... 8000 series and use just those. 8kW + 5kW = 13kW is OK going through 2x SI so would put both on protected loads panel. (I considered switching additional SB to support twice as much PV.) It may be better not to have anti-islanding by SB when SI forms island (not sure if a problem or not.)
I certainly recommend not using offgrid/island mode for an inverter direct on grid. Not compliant and my be a lineman safety hazard.
If you use Rule-21, consider adjusting frequency limits for disconnect and/or frequency/watts settings of SB 5000US so they work together better without anyone getting knocked offline. This is 15+ years of products being made to work together. SMA invented frequency-watts long before the grid asked for Rule-21 and grid support.
Got it, thanks. This seems to line up with even SMA's own products more recently, where they claim to enable "grid-backup" behavior by default now, right out of the box. I suspect they got tired of everyone asking for "grid guard" codes?
You would think.
But more than that, if off-grid and frequency of SI is sitting outside UL-1741 range, the SB 5000US series wouldn't connect and feed in, during which time it wouldn't talk on RS-485 so you couldn't change its mode.
I saw that, threw the grid switch, then it connected and I could put it in backup mode.
No harm in having backup enabled anyway, so better that they made the change.