New 10kw NHX AIO From Watts247
- Thread starter Kornbread
- Start date
Can the clamp meter give you the AC current out GRID and LOAD connections?
The power has to be going SOMEWHERE so it is flowing out the AC connections or is becoming waste heat, but 600 watts waste heat would be a huge problem.
Mike C.
Grid is not hooked up. Yes, post#470 shows the measured and indicated loads.
Now that I think of it, IIRC, the old growatt would show over 1k when the shop ac compressor kicked on and this one only shows ~600w. Would also make sense that when the grid went down last week and we had to use this inverter to power our well, it showed way lower than I was thinking it should.
Now that I think of it, IIRC, the old growatt would show over 1k when the shop ac compressor kicked on and this one only shows ~600w. Would also make sense that when the grid went down last week and we had to use this inverter to power our well, it showed way lower than I was thinking it should.
When things are weird, you have to get very detailed about what the numbers mean.
I presume the above are AC current measurements taken with an AC clamp meter, properly zeroed prior to measurement, on the L1 and L2 outputs of the LOAD1 connection on the inverter. I further presume there is NO PV connection and no GRID connection, and no LOAD2 connection.
I also presume the above data sets represent two moments in time where the load was different. This is consistent with your prior graphs with picket fence current profiles.
What was the battery current draw for the upper data set?
I see a battery current draw of 32.8 amps at 53.5 volts for the lower set of AC measurements. I presume this was with a DC current measurement, and also properly zeroed prior to measurement.
Input power: 1755 watts. AC output should be about 1650 watts (94% efficient),
You didn't say if you measured the voltage and just indicated 120 VAC. It seems these devices often run a bit hot, around 125 VAC per leg. But assuming 120 VAC for the second set, 2325 watts coming out.
If you can put in 1755 wats and get 2325 watts out, and that is real, then there is a Nobel price in your future.
Otherwise, something is wrong with your data provided by the meter, you are measuring more output than input, which suggests you sampling of data is either wrong, or your timing of collecting them missed a changing load.
Something else to consider is that your loads might be spiky in current, that is, not have a power factor of 1. In this case, your meter might be reading the PEAK AC current and applying the 0.707 factor to get your the presumed RMS current. That only works with PF = 1 loads. A "true RMS" meter does an actual integration of the current flow to give you the AC current. Your meter doesn't seem to be of that class.
Did you zero calibrate? Put nothing through the meter loop, hold it in the orientation and near the circuit to be measured, note the reading, subtract it later from the reading with current in the loop. Some meters have an offset button to do this. This is more important with the DC measurement since static magnetic fields are an offset for that.
Another suggestion might be to get a static stable PF =1 load, like a space heater at max power, and only power that during some measurements so we can eliminate the load characteristics as part of the issue.
Mike C.
Zwy
Emperor Of Solar
On my system, I had to use a Smart Shunt to report battery charge/discharge to SA. It was very inaccurate using SA to determine the charge/discharge from the inverters.
I would not trust any battery charge/discharge on SA without the Smart Shunt.
One more thing about SA is the time sampling is not fast. It takes a longer time frame for Loads to update than for Battery to update.
I would not trust any battery charge/discharge on SA without the Smart Shunt.
One more thing about SA is the time sampling is not fast. It takes a longer time frame for Loads to update than for Battery to update.
My "bug" list so far with the Amensolar N3H-X10-US (aka Watts247 NHX-10kW). I am currently using it only as a backup panel UPS tied to an EG4 Indoor Wallmount battery.
1. Battery comm error drops power
I successfully establish battery comms between the EG4 and the inverter using the Megarevo protocol over CANbus. It appeared to work perfectly as far as information flow. After operating for about 48 hours, the inverter suddenly dropped power to the backup panel. At the same time, a "BMS COM Abnormal" alert was posted on the logging app. It appears that if there is a battery comm failure of any kind, the inverter drops output power. This makes output power unreliable. What should happen is that alert is posted, but the inverter should continue to provide power and revert to the voltage based levels (aka lead acid mode) until battery comm is established again.
My response to this was to disable battery comm and put the inverter into lead acid voltage mode and it has not dropped power since. it is unfortunate that battery comm doesn't work on the inverter since I would prefer that.
I was using a premade short network cable, so I doubt this is an electrical problem.
2. With grid available, the inverter draws battery power
I have the unit in "battery priority" mode which should, based on the manual, use the battery only if the grid fails. But I am noting that the unit seems to draw ~2 amps from the battery even when grid is present. Why it does this is not clear to me.
3. Alternating discharge and charge of the battery
Related to the above, the inverter is alternating between ~2 amp discharge and then charging the battery back up. It alternates every few seconds. This behavior will reduce battery life. This behavior has been verified with a clamp meter, so it isn't an indication problem. Here is a graph of the battery current during these events.
This goes on indefinitely and the battery is at 99-100% SoC. My expectation is that at 100% SoC with grid available, the battery should see almost no current in or out.
4. Grid charge enable ignored
One treatment I attempted for the charge/discharge cycling was to disable charging from the grid. The selection "Grid Chg Enable" was unchecked in the settings with the expectation that would stop grid charging altogether. I made that change and the behavior didn't change at all. It appears the grid charge enable setting isn't implemented properly.
5. Touch screen calibration
A cosmetic issue, but the touchscreen alignment between where the icon is displayed and where you press is off, particularly to the lower right. It feels like the software has screwed up the transform from touchscreen XY to screen XY. Minor issue but can be a little frustrating at times. I'm not the only one to notice this and touchscreen calibration has been mentioned in at least one YouTube review video.
6. Runs pretty hot
You would think that with grid available and simply passing through to the backup panel, there would be little power lost in the unit. Not so. The unit runs about 50-55 C internal temperatures (in 30-35 C ambient conditions) and feels quite toasty to the touch. This has three negative effects, it shortens the life of the electronics inside, it wastes energy, and it puts heat into the space (garage in my case). I'm kind of surprised it runs that hot doing "nothing". The fans rarely run in this mode, BTW, so the internal heat is not tripping the fan control algorithm yet. I estimate the waste heat is about 70-100 watts, so about 2 KWH per day which isn't trivial.
7. Solarman has no remote control
The included Wifi adapter works with Solarman which has decent enough logging capability. There appears to be zero ability to remotely control the inverter, however. I was disappointed with this. Perhaps Solar Assistant can provide this capability, but I have not tried it yet.
8. Shipping damage
The units are not packaged sufficiently to be shipped without damage. This has also been noted by other reviewers. Mine had dents in various places as if it was a car in a wreck and the PV disconnect switch was damaged where it had been pushed into the case. Watts247 sent a replacement PV switch which I installed, and I carefully pulled out the dents. The packaging leaves a lot to be desired and Watts247 said that will be improved.
I have yet to hook up PV to it, or use any grid interactive features, so I have no experience to judge those aspects yet.
In good news, when backup enable is on, the unit switches from grid to battery power fast enough that no devices reboot in the house. This was a key feature for my install and it works.
There is a new firmware version available, but Watts247 has not tested it and I'm unwilling to do so until they do. I have no idea what issues the new firmware fixes.
Mike C.
1. Battery comm error drops power
I successfully establish battery comms between the EG4 and the inverter using the Megarevo protocol over CANbus. It appeared to work perfectly as far as information flow. After operating for about 48 hours, the inverter suddenly dropped power to the backup panel. At the same time, a "BMS COM Abnormal" alert was posted on the logging app. It appears that if there is a battery comm failure of any kind, the inverter drops output power. This makes output power unreliable. What should happen is that alert is posted, but the inverter should continue to provide power and revert to the voltage based levels (aka lead acid mode) until battery comm is established again.
My response to this was to disable battery comm and put the inverter into lead acid voltage mode and it has not dropped power since. it is unfortunate that battery comm doesn't work on the inverter since I would prefer that.
I was using a premade short network cable, so I doubt this is an electrical problem.
2. With grid available, the inverter draws battery power
I have the unit in "battery priority" mode which should, based on the manual, use the battery only if the grid fails. But I am noting that the unit seems to draw ~2 amps from the battery even when grid is present. Why it does this is not clear to me.
3. Alternating discharge and charge of the battery
Related to the above, the inverter is alternating between ~2 amp discharge and then charging the battery back up. It alternates every few seconds. This behavior will reduce battery life. This behavior has been verified with a clamp meter, so it isn't an indication problem. Here is a graph of the battery current during these events.
This goes on indefinitely and the battery is at 99-100% SoC. My expectation is that at 100% SoC with grid available, the battery should see almost no current in or out.
4. Grid charge enable ignored
One treatment I attempted for the charge/discharge cycling was to disable charging from the grid. The selection "Grid Chg Enable" was unchecked in the settings with the expectation that would stop grid charging altogether. I made that change and the behavior didn't change at all. It appears the grid charge enable setting isn't implemented properly.
5. Touch screen calibration
A cosmetic issue, but the touchscreen alignment between where the icon is displayed and where you press is off, particularly to the lower right. It feels like the software has screwed up the transform from touchscreen XY to screen XY. Minor issue but can be a little frustrating at times. I'm not the only one to notice this and touchscreen calibration has been mentioned in at least one YouTube review video.
6. Runs pretty hot
You would think that with grid available and simply passing through to the backup panel, there would be little power lost in the unit. Not so. The unit runs about 50-55 C internal temperatures (in 30-35 C ambient conditions) and feels quite toasty to the touch. This has three negative effects, it shortens the life of the electronics inside, it wastes energy, and it puts heat into the space (garage in my case). I'm kind of surprised it runs that hot doing "nothing". The fans rarely run in this mode, BTW, so the internal heat is not tripping the fan control algorithm yet. I estimate the waste heat is about 70-100 watts, so about 2 KWH per day which isn't trivial.
7. Solarman has no remote control
The included Wifi adapter works with Solarman which has decent enough logging capability. There appears to be zero ability to remotely control the inverter, however. I was disappointed with this. Perhaps Solar Assistant can provide this capability, but I have not tried it yet.
8. Shipping damage
The units are not packaged sufficiently to be shipped without damage. This has also been noted by other reviewers. Mine had dents in various places as if it was a car in a wreck and the PV disconnect switch was damaged where it had been pushed into the case. Watts247 sent a replacement PV switch which I installed, and I carefully pulled out the dents. The packaging leaves a lot to be desired and Watts247 said that will be improved.
I have yet to hook up PV to it, or use any grid interactive features, so I have no experience to judge those aspects yet.
In good news, when backup enable is on, the unit switches from grid to battery power fast enough that no devices reboot in the house. This was a key feature for my install and it works.
There is a new firmware version available, but Watts247 has not tested it and I'm unwilling to do so until they do. I have no idea what issues the new firmware fixes.
Mike C.
SolarAssistant calculates charge/discharge from the aio? I was under the impression, that everything under the battery tab and the battery gas gauge was whatever the jk bms's send SA.
Remeasured today and found two mistakes, 1) measured one ac line and the neutral, 2) didn't zero the meter. Rerunning the second set of numbers in post#470 knowing about the neutral brings the measured watts at the load1 panel and the measured dc battery draw within a 100w difference. Still way off from what the nhx reports.
From today measured at the load1 panel:
2.36amps@120vac=~283w
13amps@120vac=~1560w
measured battery draw was 34.8amps@53.2vdc =~1851w
concurrently, the nhx display shows:
1.2amps@120vac@126w
6.3amps@120vac@731w
reported battery draw of 34.6amps@53.0vdc=~1834w
No matter how you look at it, the numbers the nhx report are way off.
`
cs1234
Solar Wizard
- Joined
- May 9, 2022
- Messages
- 2,790
Sounds like it is powering it's internal electronics from the DC bus instead of taking it from the grid.
Sure does seem like it's powering itself from the DC bus, then it recharges the battery to make up for what it used up.. ad infinitum.
70-100watts is no laughing matter when it comes to waste heat to remove. That will heat it up quite a bit over many hours. I'm guessing it's not running the fans to cool itself, so it's passively radiating?
Last edited:
That was my thought as well. During charge, though, it has to be taking the power from the grid.
With grid available, it should not be using the battery input, of course.
When I turned off Grid Chg Enable, I was expecting one of two outcomes: the battery discharge current would be about ~2 amps and keep going until battery low volts is reached (the unit continued to draw internal power from the battery), or the battery current draw would stop. I got neither response, it basically ignored the Grid Chg Enable setting.
It is not running the fans except on rare occasions. Here is last week's temperature logs:
The dips in the blue trace are when the external fans run, so 5 times last week in total, and seemingly only very briefly. The fans don't seem tripped by temperature only since the blue dips are not when the inverter is hottest. I'm at a loss to explain the external fan behavior.
The temps do get pretty warm, sometimes above 55 C. It is uncomfortable to hold your hand on the sides of the inverter enclosure.
Mike C.
AC output power total: 1843 watts
Battery input power: 1851 watts
Indicated efficiency: 99.6 %
This fails sanity check. It realistically should be about 90-95%.
It sure feels like a math error by a factor of 2. If you double the current readings, then:
AC output power total: 1800 watts
DC input power total: 1834 wats
Indicated efficiency: 98.1 %
Still outside sane bounds, but a lot closer.
Are you measuring the battery voltage *at the inverter terminals*? The battery current can cause a meaningful drop in voltage.
Mike C.
cs1234
Solar Wizard
- Joined
- May 9, 2022
- Messages
- 2,790
That sucks. It would be nice if it could pull it's self consumption from the grid if we asked it nicely, or to not pull it from the grid, if we wanted it that way.
It would be nice if they allowed us to set a temperature we are comfortable with (as long as it meets their minimum for warranty terms). That way worry warts could have the fans run more often for a cooler inverter, and people who aren't concerned about it could let it run as hot as the manufacturer allows.
Another issue I have noticed is insane readings on the Solarman logger. Here is an example:
This is the AC output current.
The blue spike is 46 amps that is almost entirely on one phase. There is nothing in the house that draws that much power (5.5 KW) from a single phase.
These spikes don't happen often, maybe once every 3 days. And they can happen to either phase:
The above had a 76 amp spike on blue, 56 amp on red.
I am fairly certain these are indication anomalies and not real since nothing in the house works this way.
Mike C.
This is the AC output current.
The blue spike is 46 amps that is almost entirely on one phase. There is nothing in the house that draws that much power (5.5 KW) from a single phase.
These spikes don't happen often, maybe once every 3 days. And they can happen to either phase:
The above had a 76 amp spike on blue, 56 amp on red.
I am fairly certain these are indication anomalies and not real since nothing in the house works this way.
Mike C.
Adam De Lay
Solar Wizard
It's possible. I'm wonder if the grid-charging issue is because the "battery priority" mode is overriding the Grid charge setting. Why? Not sure because nothing should override a "global" setting.
Those output spikes look very odd. Has to be some kind of glitch because looking at a 15 hour window, they would have to be sustained surges to show up like that.
Are you only running SolarMan and not SolarAssistant?
Maybe.
I could find no better mode (based on the manual) for no PV setup (UPS mode). Self Consume mode sounded like it would drain the battery back to the grid (I can't have export now, no permission to operate). I'm open to trying something else but that's not all that easy since I am remote and I have to walk a non technical person through front panel menus to do that.
They are strange. I had about 5 or 6 of them in 10 days. Maybe it is some sort of comm glitch which might be related to battery comm problem I had earlier. Some random bit flips and the data is garbage.
I will note that no other parameter I have found has similar glitches. it seems to be only the grid current, which maybe suggests something weird with the CTs?
Solarman.
I didn't have time to setup SA on this trip (I'm remote from the site). I am curious if SA will allow remote config or not of this inverter. Watts247 claims cheerfully "100% Compatible with Solar-Assistant". Not clear this means you get remote control.
Anybody using these inverters in SA and have remote control of the config?
Mike C.
Adam De Lay
Solar Wizard
There's some remote control ability. I haven't gone through and tested all of it. Seems like most of the items you would want to change are there.
I know the "Advanced Modes" seem to be missing as well as any TOU settings in the inverter.
I know the "Advanced Modes" seem to be missing as well as any TOU settings in the inverter.
Adam De Lay
Solar Wizard
Yeah, probably best to come up with some ideas and test while on-site.
You mentioned switching to voltage instead of battery comms. Have you seen the same spikes since switching? That might help rule out if it's related to the battery comms issue you experienced.
I know I've seen once or twice where it seems like a battery comms issue but there's nothing in the log reflecting that. Seems very sporadic.
That's kind of limiting since those settings are material to controlling the device.
The remote config ability is kind of important since I will be remote monitoring this install.
Mike C.
Good thought, but the spikes have occurred when I was on battery comms and when I was off it. It does seem to be less occurrence when off it, though.
I had 3 events in 2 days on battery comms. I had 2 events in 7 days off comms. Maybe that's a clue, maybe not.
The battery comm event is in the log. It is the only event in 10 days of operating. It did bring down the backup panel, though, and that's not cool.
Mike C.
Well, it happened again. Almost exactly 192 hours since the last time, the inverter shut down the backup panel. The backup panel went dark for about 30 seconds, then came back. Last time, a battery comm error was reported, but this time no error was logged by Solarman. I'll have to check the alert logs on the unit itself to be sure, but it appears no error was flagged.
Given the inverter has twice failed power to the backup panel, with no apparent external cause, I being to worry there is a bug in the firmware that causes a problem, and then a watchdog reboots the inverter to bring it back.
One of the reasons to do this project was to stop nuisance outages that were resetting clocks and computers, but the inverter seems to be causing them now, so I need to get this fixed.
Has anyone else with this inverter noted similar problems?
Maybe battery priority mode is not a good mode and I need to reconfigure the inverter to something else?
Mike C.
Given the inverter has twice failed power to the backup panel, with no apparent external cause, I being to worry there is a bug in the firmware that causes a problem, and then a watchdog reboots the inverter to bring it back.
One of the reasons to do this project was to stop nuisance outages that were resetting clocks and computers, but the inverter seems to be causing them now, so I need to get this fixed.
Has anyone else with this inverter noted similar problems?
Maybe battery priority mode is not a good mode and I need to reconfigure the inverter to something else?
Mike C.
Zwy
Emperor Of Solar
Most likely it uses the AIO for the charge/discharge data. With the inverter you have there might be some different settings but with my system I had to add the Smart Shunt for any accurate charge/discharge data in SA. There is a JK BMS option under Battery settings but I wouldn't trust those values. The Batrium uses an actual shunt and is accurate.
One way to find out, add a Smart Shunt. I added one because SA didn't interface with Batrium.
Request:
Can someone detail how you review current and historical alerts on the inverter front panel and take pictures of this? I am not at the unit and would like to have a person at the site do this to check what alerts have been posted. I do have Solarman app running, but I want to make sure there are no alerts posted at the machine that Solarman is failing to report. The device manual is of no use for this since it doesn't have any description of how to see past faults.
A reason I ask is that I did a number of grid down tests to see if the unit would transfer properly to battery backup. It did, and it did so fast enough to prevent reset of devices in the house. But I have just realized that none of those tests produced any alerts. Specifically, I didn't get a "Grid Low Volts" alert which one should expect for that situation. The lack of that alert suggests the unit isn't doing what I would expect. Maybe the faults are logged in the unit but not reported over Solarman app?
Mike C.
Can someone detail how you review current and historical alerts on the inverter front panel and take pictures of this? I am not at the unit and would like to have a person at the site do this to check what alerts have been posted. I do have Solarman app running, but I want to make sure there are no alerts posted at the machine that Solarman is failing to report. The device manual is of no use for this since it doesn't have any description of how to see past faults.
A reason I ask is that I did a number of grid down tests to see if the unit would transfer properly to battery backup. It did, and it did so fast enough to prevent reset of devices in the house. But I have just realized that none of those tests produced any alerts. Specifically, I didn't get a "Grid Low Volts" alert which one should expect for that situation. The lack of that alert suggests the unit isn't doing what I would expect. Maybe the faults are logged in the unit but not reported over Solarman app?
Mike C.
Solar Guppy
Red Cobra Guppy
This is standard requirement for UL9540 compliant inverters.
The inverter <--> battery communications must always respond to requests, retries are allowed, but there are timeouts in the 1 to 5 minute range that will trigger the shutdown.
I suspect with painful data logging of the canbus coms you will find there is a request that isn't being responded too and the inverter considers the correctly it is a fault.
One would need the actual document for the protocol you are using and log the communications 24/7 to track down what actual address is the problem and then be able to inform the manufacture what they missed. If you can, you should try and use a more common protocol like the Pylontech High Voltage if possible with your inverter and bms if the support this.
Adam De Lay
Solar Wizard
