I have been searching for an inverter to suit my house requirements (especially since some of the older products on the market appear to be approaching EOL), and The One looks... well... like the one.
Some questions about the product:
- How does The One manage both an AC coupled PV array and a generator simultaneously (to prevent backfeeding the latter)?
- Does that solution work if the PV in attached to the panel on AC out, not the AC coupling relay?
- Can cloud connectivity be prevented while still allowing LAN access, monitoring and management?
- Will the management interface be opened to allow open software solutions? I want to integrate it with my bespoke home automation.
- Do the load shedding breakers turn off if The One enters an overloaded state?
- If I have to bypass The One (due to it having a fault that needs repair, for example), I assume that the built-in breakers will not receive any power?
- When the generator is operating is the max AC out The One's normal max? The generator's? The sum of the two? Something else?
- Is it possible to inhibit the generator from night time operation, and set an evening operation time if the grid is down?
- Am I correct in understanding that The One can feed power out of its grid connection in order to reduce the power drawn by the main panel from the grid? And it uses CTs to limit the net sell back to the grid if the main panel isn't consuming?
- How loud can The One get during operation?
I will take a stab at these:
1- The AC coupled input will be opened and disabled when the generator is running to prevent damage
2- At this time no the ac coupled in the load center would need a black out relay to disconnect it when generator is running, Use the NC and Common on a relay and power the coil from the generators 240VAC
3- I will ask Rock to chime in but I believe the only access is cloud based not lan based so not likely in this case
4- Open source, I doubt it will be truly open but a published modbus protocol is available to talk to the AIO
5- Yes all AC out would go away if the AIO is overloaded
6- Bypass, depends on how you wire the bypass but yes if the system is bypassed the programmable breakers would be dead
7- Max AC out is what the AIO is allowed to feed out to loads. Battery charging will be additional up to the 60A max
8- Quite time, At this time no the AGS is very basic. In the future it will be expanded
9- Yes the AIO can backfeed into the main panel and not export out to the meter using the CTs
10- The inverter is almost silent the fans however will be heard as with any fans when it is running hard
@FNG, can you private message me the modbus spec for the Mn15?
Shoot me an email ryan@midnitesolar.com
For #2, wouldn't the MN15 use frequency shifting to lower ac coupled production? I know the ac couple in relay is the best but power/frequency shifting should be implemented as well?
Thanks for the answers!
Quote from: FNG on December 27, 2024, 07:38:59 AM2- At this time no the ac coupled in the load center would need a black out relay to disconnect it when generator is running, Use the NC and Common on a relay and power the coil from the generators 240VAC
Hmmm... would that be safe? Wouldn't that make back feeding possible while the generator is spinning up and not yet providing any voltage and power?
Quote3- I will ask Rock to chime in but I believe the only access is cloud based not lan based so not likely in this case
That would be unfortunate. I despise essential systems which have cloud exposure. I very much hope they offer an option to have LAN access but no cloud. I set my LAN up to prevent such appliances from having any internet access.
Quote4- Open source, I doubt it will be truly open but a published modbus protocol is available to talk to the AIO
I'd prefer not to have to run a modbus to Ethernet bridge, but if that's the only option.
Quote5- Yes all AC out would go away if the AIO is overloaded
I was hoping it would do something more incremental, like turning off the relays in a priority order until the overload was resolved.
Quote7- Max AC out is what the AIO is allowed to feed out to loads. Battery charging will be additional up to the 60A max
So the gen will run when the battery is below a threshold, but the inverter will use battery power to assist the gen if the loads are beyond what the gen supports, and otherwise will use spare gen capacity to charge the battery?
Thanks again.
Quote from: goldserve on December 27, 2024, 12:44:19 PMFor #2, wouldn't the MN15 use frequency shifting to lower ac coupled production? I know the ac couple in relay is the best but power/frequency shifting should be implemented as well?
Yes t will frequency shift but only when there is no ac in. If it is connected to the grid or a generator it can not frequency shift
Quote from: ASword on December 27, 2024, 10:38:29 PMThanks for the answers!
Quote from: FNG on December 27, 2024, 07:38:59 AM2- At this time no the ac coupled in the load center would need a black out relay to disconnect it when generator is running, Use the NC and Common on a relay and power the coil from the generators 240VAC
Hmmm... would that be safe? Wouldn't that make back feeding possible while the generator is spinning up and not yet providing any voltage and power?
No as the AIO will not close its relay and expose the Generator to the grid tie inverter until it is qualified
Quote3- I will ask Rock to chime in but I believe the only access is cloud based not lan based so not likely in this case
That would be unfortunate. I despise essential systems which have cloud exposure. I very much hope they offer an option to have LAN access but no cloud. I set my LAN up to prevent such appliances from having any internet access.
Quote4- Open source, I doubt it will be truly open but a published modbus protocol is available to talk to the AIO
I'd prefer not to have to run a modbus to Ethernet bridge, but if that's the only option.
Quote5- Yes all AC out would go away if the AIO is overloaded
I was hoping it would do something more incremental, like turning off the relays in a priority order until the overload was resolved.
Quote7- Max AC out is what the AIO is allowed to feed out to loads. Battery charging will be additional up to the 60A max
So the gen will run when the battery is below a threshold, but the inverter will use battery power to assist the gen if the loads are beyond what the gen supports, and otherwise will use spare gen capacity to charge the battery?
Correct
Thanks again.
A couple of additional questions:
1) In my situation I will want to charge the batteries from grid during night time lower pricing, but leave enough uncharged capacity so that AC coupled PV will be able to charge during the day. This requires a prediction that includes weather, time of year, etc. That is clearly beyond what the inverter itself is capable of, so I would want to be able to do that in my own system and then tell the inverter when and how much to charge... and do that programmatically, not via a manual interface. To do that ideally there would be an HTTP LAN-based interface to the inverter to control most of its user-facing settings. Is that something on your roadmap?
2) Similarly, I want to control my generator run times directly. To do so I need access to inverter information (e.g. SoC, grid up/down state). I could try to take direct control of the generator via its 2-wire interface (although I'd rather have the inverter do it directly under my control via the aforementioned HTTP interface), in which case I would have the inverter set to expect manual control of the generator. Will the inverter accept generator AC input at any time the grid is down, or are there other caveats to generator source power?
3) I came across competing product that is able to accept AC coupled PV input at the same time as a generator input, as long as there is battery capacity available to absorb excess power. If the battery is full, their inverter shuts down the AC coupled PV or the generator (via one of the relays, as freq shifting w/ gen isn't possible). Can The One do this? The answer above at the top of the thread says no, but it seems like the hardware ought to be capable of it.
4) The programming/configuration of The One doesn't seem to have more than one configuration state. It seems like it should have two configurations, one for when the grid is up and one for when the grid is down. Alternatively, if I had the HTTP interface I mention above, I could detect these conditions and change settings as needed.
Do you use HASS for your automation?
Something I've been working on using solarman custom integration.
https://imgur.com/a/ztrQXn4
No, mine is a bespoke system. If there is existing open source code to do something though, I can leverage that to replicate its functionality.
solarman integration and the chint inverter definition is very close to the MN15 (senergy based).
https://github.com/davidrapan/ha-solarman/blob/7f1904d34dc2e9168dbff4f6cb55dc4676f4e700/custom_components/solarman/inverter_definitions/chint_cps-scetl.yaml
Thanks, that will be really useful. Unless MidNite takes my suggestion of adding an HTTP protocol that their app (and any other software) can use to control the inverter more safely. The modbus interface contains a lot of elements which are potentially dangerous, so having a more sanitized interface over the network is safer and more robust. It would also let them change their more detail oriented modbus interface while maintaining the simpler HTTP interface.
Agreed! Development seems to be a bit slow though so don't hold your breath!
Yeah, I'm in software dev myself so have some appreciation of how quickly such things can happen (or not). That said, not asking for rocket science here. I'm probably six months away from buying an inverter and batteries, and could wait a few months for a solution. Or longer if I first went with direct modbus control then switched to a safer interface later. I've got 3-4 alternatives... a good software control option is probably the feature driving my selection.
As much as I love Midnite products (Clipper, 2 x Classic 250's), I was in the market for an inverter in 2015, but their first offering was about 4 iyears later.
Either you wait patiently, or go with another product until MS comes out with their promised things (and they always do)
Fair point, although I'm talking about a software feature added to an existing product. Hardware development is a lot slower and more complicated than adding software features, especially a new interface like described here. Four years for a hardware product like the AIO is a reasonable timeline... maybe even pretty quick. I've delivered software features like this in weeks or months, from start to finish, including QA/testing/documentation. A big factor will be how much computing hardware they've got available in their box, of course... but these days it's surprising to not have enough. Five years ago when they would have been starting component selection, a typical single board computer had 1-4GB RAM and 1-6 cores running at 1-2.5GHz. For such a board to support running a simple webserver is trivial, and plenty of open source solutions exist to do most of the hard work. All they really need to add is the details of the higher level interface they wish to create, and the modbus code they must already have.
All just informed speculation, of course, it's entirely up to Midnite. My guess is that their bottleneck is the manpower to implement such a thing, and whether it is a priority for them. I really hope they see the potential in the idea as it would set them apart from the competition, and I'd guess that their competition would then have to scramble to catch up. Being first mover means they don't have to scramble to be the one catching up!
Is there more detailed data about The One's surge capability available? The quoted 20,000VA & 13,000VA per leg is for 1 second. Presumably there is a ramp from the rated 10kW continuous (i.e. lifetime seconds) down to 20kVA for 1 second. Is there a plot available, or at least another data point or two? I'm contemplating my backup panel loads, and the smart load capability I'm considering and how quickly I would need to detect an overload. I have an energy monitor that samples every 5 seconds, so it would be interesting (for example) to know how big a surge The One could handle for 8-10 seconds.
Also, why is the surge capability quoted as VA and not watts (which is volts*amps)? There must be some subtlety behind that choice?
Does Midnite monitor its own forum for questions like these?
Intimately. :P
In the 0 post of this thre4ad ASword asks: "How loud can The One get during operation?"
And FNG responds in post 1: "10- The inverter is almost silent the fans however will be heard as with any fans when it is running hard."
The spec sheet for the MN15-12KW-AIO does state <40 dB for no load noise level. Very good and in agreement with FNG.
Can we also get a numerical value of sound level for "full load" and for perhaps some more typical operating level, perhaps 30% of full load?
Thanks in advance.
https://www.youtube.com/watch?v=EJSo5EwCrjQ
Thanks, ASword Good video.