Adding Schneider XW Pro

[400bird]

Well, I just did my taxes. Which means I finally had to add up my costs for the additional panels I added last fall.
It total, I added 3,380 watts for $1.30/watt. Using all premium products, no used panels or wooden racking.
Well crap, I can't exactly say that because I haven't officially mounted those last 2 480 watt panels. Ok so 6 of 8 are racked on the real deal up on the roof. Those last two are just leaning on the fence. There goes my math ?‍

That is:

• 4 - 365 watt panels with optimizers added to the Solar Edge inverter
• 4 - 480 watt panels on a new Midnite Solar Classic 150 charge controller
• Permitting costs, everything involved. I even included the portable band saw I purchased to cut the racking. I definitely didn't need it, but it was sooo much nicer than using the Sawzall like I did last time.

 

[400bird]

I added 1460 watts (4*365) to my Solar Edge inverter, in addition to the original 6,500 watts.

Apparently, I never took pictures or posted that here. Takes to total on that 6,000 watt inverter to 7,960 watts.

I never really saw clipping on the old array, but today was the first time where I really saw clipping, but wow, it hits that limit like a wall, err, ceiling?



Anyways, you'll see my energy meter reports about 5,750 watts. The Solar Edge portal reports about 5,980. The biggest difference is the voltage. My energy meter is measuring just one phase at about 122 volts and doubling it to 244 volts. The Solar Edge inverter reports 250 volts.

I don't spend enough time working on AC to know if a 6-10 volt drop between production (at the inverter) and load (the energy meter gets a reading after a breaker)
But there's plenty of wire, an AC knife switch, 5 ish breakers, and an outlet in between the two measurement points. The inverter needs to raise the voltage to pump current onto the grid and into the house, so I'd expect some increase to get enough potential to push 6kw out.

 

[400bird]

Well, 2 odd and unrelated things in the past couple days.

I received my first NEM2 paired storage bills. One was Dec-Jan, second was Jan-Feb. Then 2 days later, I received updated versions. 40+ pages of gibberish in total. I'm struggling to understand any of it. PGE's online bill doesn't match (in form, $ is similar best I can tell)

First bill was $172.25 on 2/22 (mostly gas) and due for payment by 3/15

Second bill was $20.53 on 2/28 and they had the balls to say the bill sent 6 days earlier was past due.

I paid the total on 3/12 (yesterday)
Today, I've got a $91 credit on account?!

God, PGE is such a mess.



Second thing has to do with the BMS and SOC tracking.
When I was charging solely with the XW, my SOC would slowly gain some SOC over days of not resetting. So, the SOC might hit 102% and rest down to 100% a few times before the cell voltage said the pack was really at my 100% (4.11 volts/cell)

Now that I've added the Midnite charge controller and DC coupled solar, the SOC loses SOC accuracy in the opposite direction. So, today it hit 4.1 volts but the SOC shows 93%

 

[pvdude]

PGE is such a mess.

Mine too...(Duke energy).
Duke seems to have deleted the "compare to last year" feature from their web site.
I suspect because of the rate increases, they don't want us to see the dramatic difference
that would be obvious on the histogram.
(note to self - buy more PV panels.......)

 

[GXMnow]

Second thing has to do with the BMS and SOC tracking.
When I was charging solely with the XW, my SOC would slowly gain some SOC over days of not resetting. So, the SOC might hit 102% and rest down to 100% a few times before the cell voltage said the pack was really at my 100% (4.11 volts/cell)

Now that I've added the Midnite charge controller and DC coupled solar, the SOC loses SOC accuracy in the opposite direction. So, today it hit 4.1 volts but the SOC shows 93%

My JK-BMS on my original battery bank is having the same issue.

When it was just the XW-Pro supplying the charge, it would reset to 100% pretty much every day it completed an absorb charge cycle. But ever since I added the BougeRV DC charge controller, it has not reset since last November. The charge current is never falling to zero. The DC panels keep pushing some charge, even if it falls to float, but the XW then starts pulling current, so it is never sitting at the full charge voltage to trigger the reset. It has been creeping down in SoC every few days. It now reports SoC at 34% when the battery voltage is at 57.7 volts (4.12 per cell). That is over 90% on any Li NMC chart.

Next time we have a few days of clear weather, I may try just turning off the panels on the input of BougeRV charge controller for an hour after the batteries hit absorb and see if it will reset the SoC to 100% again.

 

[Hedges]

Anyways, you'll see my energy meter reports about 5,750 watts. The Solar Edge portal reports about 5,980. The biggest difference is the voltage. My energy meter is measuring just one phase at about 122 volts and doubling it to 244 volts. The Solar Edge inverter reports 250 volts.

I don't spend enough time working on AC to know if a 6-10 volt drop between production (at the inverter) and load (the energy meter gets a reading after a breaker)

Or unmeasured phase was closer to 128V?

If really 2% to 4% drop in wires, within the realm of possibility for a long run but not what you expect for a dedicated, proper size wire just within the house.
You can measure voltage L1/2 to N at each end.
You can plug in an extension cord fed by main panel (to a circuit with no other loads on) and run it over to the inverter. Then measure L1 at inverter to L1 remote "Kelvin" sensed (second wire with zero current), to see drop in L1 between meter and inverter.

God, PGE is such a mess.

That's good to hear
(Some of us hope no one knows when we should be moved to NEM 3.0)

When it was just the XW-Pro supplying the charge, it would reset to 100% pretty much every day it completed an absorb charge cycle. But ever since I added the BougeRV DC charge controller, it has not reset since last November. The charge current is never falling to zero. The DC panels keep pushing some charge, even if it falls to float, but the XW then starts pulling current, so it is never sitting at the full charge voltage to trigger the reset.

Is XW acting as AC to DC battery charger, so current delivered comes from rectified sine wave, drops to zero ever 1/120th of a second?
vs. DC charge controller delivering constant current from HF PWM through an inductor?

There may be some differences in how BMS measures and reacts to current & voltage. Would have expected better behavior with DC source, but maybe peaks from pulsed current cause it to recognize higher state of charge.

However, does BMS have anything to do with charging behavior (closed loop), or is it just reporting? Do the AC and DC chargers each independently recognize need to go to float based on voltage they see? If each has separate wires to battery, they see voltage at terminals plus IR drop of their own current.

 

[400bird]

Wow, that's a huge error compared to my BMS. I need to look into what it takes to trigger the Batrium to reset the SOC in this direction.

 

[400bird]

Or unmeasured phase was closer to 128V?

Possible, I was assuming measurement error between the two devices. The run is pretty short (10 feet) but has lots of connections. It's actually 4 foot to the MSP, so I don't even need to do the extension cord trick.

That's good to hear
(Some of us hope no one knows when we should be moved to NEM 3.0)

?

Is XW acting as AC to DC battery charger, so current delivered comes from rectified sine wave, drops to zero ever 1/120th of a second?

It is acting as an AC charger yes. I'll try and get the oscilloscope out and measure. There's plenty of capacitors in the XW, I'd expect it to smooth out the current pretty well.

vs. DC charge controller delivering constant current from HF PWM through an inductor?

Unknown, I'll graph that on the oscilloscope too.

There may be some differences in how BMS measures and reacts to current & voltage. Would have expected better behavior with DC source, but maybe peaks from pulsed current cause it to recognize higher state of charge.

However, does BMS have anything to do with charging behavior (closed loop), or is it just reporting? Do the AC and DC chargers each independently recognize need to go to float based on voltage they see? If each has separate wires to battery, they see voltage at terminals plus IR drop of their own current.

No the BMS is not controlling the charging. Just monitoring/reporting.

Both AC and DC charging is bussed together and shares battery cables.

 

[Hedges]

Possible, I was assuming measurement error between the two devices. The run is pretty short (10 feet) but has lots of connections. It's actually 4 foot to the MSP, so I don't even need to do the extension cord trick.

First check with DMM, is voltage from grid at main panel balanced or not (this is what would cause people with auto-transformer to attempt rebalancing of grid.)

It is acting as an AC charger yes. I'll try and get the oscilloscope out and measure. There's plenty of capacitors in the XW, I'd expect it to smooth out the current pretty well.

Not very well, at 60 Hz with LF inverter. Calculate what ripple voltage an 0.1F cap would have if supplying current for half a 60 Hz cycle.

Scroll down to 5th screenshot. 16Arms x 4 = 65Arms @ 48V battery current supplying 10kW

Battery ripple current

After reading of someone tripping battery breakers at well below rating, I realized that could be due to pulses of current drawn by the inverter. Ideally capacitors would smooth out current draw so high frequency switching pulses and 60 Hz draw all came from the caps, and current from batteries...

diysolarforum.com

When I measure PV input voltage of GT PV inverter, that has RMS ripple 1% of DC. Unlike battery, voltage can move up and down (with steady current from PV) to deliver 60 Hz power.

Unknown, I'll graph that on the oscilloscope too.

Will be interesting, but I expect inductor at high frequency to smoot it pretty well. DC in, DC out.

 

[400bird]

First check with DMM, is voltage from grid at main panel balanced or not (this is what would cause people with auto-transformer to attempt rebalancing of grid.)

I just realized that if I've got voltage drop, it is possibly on the energy meter side too, not just the PV production side. That increases the possibilities for voltage drop being a contributing factor.

Measurement is taken after the MSP, XW distribution panel, sub panel, breaker, back to an outlet near the XW and other solar equipment.

I'll have to wait for a good sunny day, so the production and loads remain more consistent for testing. Clouds were rolling though all day causing production to jump up and down all day.

 

[GXMnow]

Is XW acting as AC to DC battery charger, so current delivered comes from rectified sine wave, drops to zero ever 1/120th of a second?
vs. DC charge controller delivering constant current from HF PWM through an inductor?

There may be some differences in how BMS measures and reacts to current & voltage. Would have expected better behavior with DC source, but maybe peaks from pulsed current cause it to recognize higher state of charge.

However, does BMS have anything to do with charging behavior (closed loop), or is it just reporting? Do the AC and DC chargers each independently recognize need to go to float based on voltage they see? If each has separate wires to battery, they see voltage at terminals plus IR drop of their own current.

Yes. The DC charge controller is putting out a pretty clean constant DC while the XW-Pro has a lot of 120 Hz current ripple following the AC wave. I thought like you that if anything the BMS would have a better time with the clean DC. The current readout is much more stable, only changing a few tenths of an amps as the MPPT does it's searching. When the XW is charging, it bounces a few amps.

When the XW was my only charge source, I had the CC charge go to 57 volts. It would drop to absorb (constant voltage) mode and the current would ramp down and shut off after 15 minutes. Then the system would sit idle for up to an hour before the Enphase solar production fell to where it would start using battery power.

Now with the added DC charging, the XW is now programmed to go into Absorb CV mode at 56.7 volts. It still makes the obvious transition to CV and the current falls off in about 10 minutes now, because the DC charge controller is still supplying current. On many days, this was happening as early as 11:30 am to noon. So the DC charge controller was still pushing 24 amps into the battery bank. The JK-BMS is only on half of my battery bank. The other half has a pair of dumb Daly BMSs. So when the XW drops to standby, the JK-BMS is showing 12 amps of charge current still. The other 12 amps is going into the Dalys.

Looking at the graph from yesterday, when it was very sunny, the XW absorb phase only lasted about 6 minutes. When the DC charge controller pulls the voltage a little higher, the XW shuts off charge current, as it should in CV mode. On that day, the voltage kept climbing for 4 more hours. At that point, the batteries did reach the DC charge controllers CV set point. They call it "Boost" mode. It is CV at 58 volts now, but it is measuring a the charge controller terminals, AND it reads about 0.2 volts high. So the true battery voltage at this point was 57.8 volts. The XW reports the voltage as staying dead flat from 3:30 to 7 pm. That is 3.5 hours of very little battery current, but it still didn't trigger the reset. The XW battery graph shows the voltage at 57.86 from 4:23 pm to 7:03 pm. That should trigger the BMS to see it as a full battery.

The sun was going down and at 5:51 pm, the Enphase panels were no longer making enough to completely cover all of the loads in the house. So my controller did command the XW to start exporting a tiny bit of power. At 6 pm, the power from the battery bank was up to 243 watts. But that is probably pretty close to what the Charge controller was putting into the batteries, so the true battery bank current was likely very close to zero. But by 7 pm, with the sun really falling, the discharge current went up to 10 amps or more than 570 watts, and the DC panels were likely dropping to zero. I really think the issue is that the current smoothly goes from charging to discharging, and never is actually sitting at no current. The XW would go to absorb for the 15 minutes, and then go to "No Float" where the current just stopped. The DC charge controller never does that. It keeps trying to charge until the sun is completely down, and before then, my house is starting to pull from the XW inverter.

So like I said before, when I have at least 2 predicted full sunny days in a row, I will wait until the DC charge controller is in it's absorb (Boost) mode for 10-15 minutes, then I will shut it off and let the system sit at no battery current and see if/when the JK-BMS finally resets to 100%.

At least with these being Li NMC cells, we can use battery voltage as a good state of charge indication. If these were LFP cells, this wild state of charge reading would be a big mess.

My BMS units have no communication back to the Schneider XW-Pro or the BougeRV DC charge controller. They are just working on reading the battery voltage and monitoring their own current. The XW does not see the DC charge current, and the CD charge controller does not see the XW charge current. And the JK-BMS only sees the current going to/from it's half of the cells. The two Daly BMS units each see the current in and out of another 1/4 of the cells each. The cells are 60 amp hours each with the JK seeing 6P groups and the Dalys seeing 3P each. The total is 12P and they are holding balance quite well. The batteries are performing perfectly, it is just the SoC number is currently useless. It showed 33% charged again with the cells at 4.12 volts each.

 

[400bird]

Now that we're into the spring, which is the easy time for solar. I'm routinely having the XW finish it's portion of the battery charging at 11:00, then DC charging continues for an hour or two past that, leaving and afternoon of DC solar production lost.

I need to do some math, but I've got an idea on how to maximize DC production.
If I stop the XW from charging based on both voltage at time of day, I should be able to utilize more of the possible DC solar production.

For example:
If the battery hits 80% by 10:00 stop the XW and allow the DC to finish the charge.
If the battery is at 90% by 12:00 stop the XW and let the DC solar finish out the battery charging.

This way, I don't need to try and predict the weather a day ahead of time and even if clouds roll in after lunch, the battery will still be mostly charged and the DC will continue.

 

[400bird]

I was looking at something else and this came up on the inverter status.
The grid frequency was mostly 60hz, occasionally it blips down to 59.9 hz.
I'm heading out to the garage to get my own measurement.

 

[400bird]

59.99 hz per the DVOM, but by the time I returned to the desk, the Grid Frequency Stabilization was gone.

 

[GXMnow]

That is odd. Even on the page showing the error, it shows 60 Hz even.

When did the DC over voltage happen? On my system, I saw that once when my BMS shut off.

 

[400bird]

That is odd. Even on the page showing the error, it shows 60 Hz even.

Yeah that's why I pulled the screenshot so low, to show the 60 hz. I dunno why it happened, but it wasn't a one time event, it was coming and going over about 15 minutes.

When did the DC over voltage happen? On my system, I saw that once when my BMS shut off.

It happens nearly daily. There's a hard coded "warning" 3(?) volts below the DC over volt threshold. That's what is shown in the screenshot above.
The text about 68 volts is total nonsense. I emailed Schneider tech support, their suggestion was to just put the over volt disconnect higher. Dumb suggestion.

There was a time that I knew the offset for the warning fault, but it's been nearly 2 years since I did the testing.

 

[GXMnow]

I had to raise the over volt shut off on mine as well. I think I put it at 64 volts to stop that error.

 

[400bird]

I had to raise the over volt shut off on mine as well. I think I put it at 64 volts to stop that error.

Yup, this is sounding familiar. I decided having a functional high voltage cut off was more important and I'd live with the false alarms.

 

[Hedges]

It seems like a power quality monitor could be useful.

Fluke has high priced ones

Fluke 1732 and 1734 Three Phase Power Measurement Logger

The Fluke 1732 and 1734 delivers easy to use three phase power measurements for identifying energy waste, overloaded circuits and routine load studies.

www.fluke.com

Voltage Recorder | Fluke VR1710 Voltage Quality Recorder

Detect and record parameters including RMS average, transients, flicker, and harmonics with the Fluke VR1710, a single phase voltage recorder.

www.fluke.com

We tried a cheapie, but although its documents defined "dropout" it couldn't detect them. I could yank the plug from one socket, move to another, and it wouldn't blink (unless AC voltage had been near minimum.) Implementation was probably diode capacitor resistor.

 

[400bird]

Yikes, $3600
If I pulled out the oscilloscope, do you think we could spot the instability?