Adding Schneider XW Pro

[GXMnow]

Edit, any chance the Victon charge controller was still going and the increased voltage level was causing an issue for the XW?

I am also thinking that it might have something to do with the Victron charge controller. This issue has only happened since that was added. But the BougeRV controller also did keep pushing the voltage higher without an issue. The XW would do a very short absorb cycle and as soon as the voltage went above 56.7 volts, then no more current from the XW at all. Is the Victron CC putting out current/voltage spikes that might be confusing the voltage reading in the XW?? I could possibly se that, but the voltage readings in the XW on Insight Local all show the voltage climbing 0.3 volts above the set point and it was still cranking out nearly 30 amps. The Victron was also pushing 20 amps at the same time.

I got a field job today, so I won't be here to monitor it today, but the weather forecast is calling for "partly cloudy" all day. Battery voltage at 6 am is down to 52.5 volts, (still a bit over 50%) it will drop a bit more before any sun comes out. The system would need to produce over 15 KWHs beyond the house draw to top it out. I'll check the graphs when I get home.

 

[400bird]

I am also thinking that it might have something to do with the Victron charge controller. This issue has only happened since that was added. But the BougeRV controller also did keep pushing the voltage higher without an issue. The XW would do a very short absorb cycle and as soon as the voltage went above 56.7 volts, then no more current from the XW at all. Is the Victron CC putting out current/voltage spikes that might be confusing the voltage reading in the XW?? I could possibly se that, but the voltage readings in the XW on Insight Local all show the voltage climbing 0.3 volts above the set point and it was still cranking out nearly 30 amps. The Victron was also pushing 20 amps at the same time.

My system acts the same (not thr ignoring the charge voltage set points, but the rest of your description)

I got a field job today, so I won't be here to monitor it today, but the weather forecast is calling for "partly cloudy" all day. Battery voltage at 6 am is down to 52.5 volts, (still a bit over 50%) it will drop a bit more before any sun comes out. The system would need to produce over 15 KWHs beyond the house draw to top it out. I'll check the graphs when I get home.

Fingers crossed, how'd I behave? I didn't get enough sun here to come close to showing what mine does. My peak SOC this afternoon was 51%

 

[400bird]

Well, if I wasn't already sure, I'm now convinced, this is an addiction. I don't need it, but I picked up 48 used EVE cells from Ampster. Huge thanks, they were a great deal!

Anyways, was going to put them together into 3 parallel 16s packs.

However, I can't lift a 16s pack, so I was going to clamp them in 8s strings with 2 in series on a single BMS to make each 16s battery.

Then, I picked up my enclosure and discovered I couldn't fit them in place at 2 8s strings. So that's each 16s battery will be clamped as a 2 groups of 5 and one group of 6 (modules of 5 or 6). Still on a single BMS.

That means I'll have 9 clamped together modules and 3 BMSs. Likely Batrium, I prefer the external shunt and I already purchased a shunt trip breaker.

Anyways, I did a test assembly on the first module of 5. I'm glad pulled back to 5&6! This is much easier to move around than a module of 8 would have been.

The end plates are 3/16" steel, the clamp rods are 1/4"-20, there's 1/4" plastic insulation on the bottom and sides. Then a thin sheet of HDPE between the cells.
Yes, I messed up for this test fit and put the last cell on backwards, good thing I'm not doing bus bars on this test fit.

The plastic sheets between the cells is fiddly and was annoying to get in right. I think I'll pick up some katpon tape to both cover the tears in the blue shrink wrap and attach the HDPE insulation sheets.

 

[GXMnow]

Are you going to swap from the NMC cells to LFP? I could not find a voltage range to make them work well together.

 

[400bird]

Seems like I can limit the voltage on each and parallel 14s NMC and 16s LFP

Until you mean mentioned it, I forgot I had this thread discussing exactly that:

Paralleling LifePO4 banks with...

Thanks to a gracious member here, I just picked up 48 LifePO4 cells that I plan to build into three 16s batteries. I had I initially considered retiring my current battery bank (14s NMC) But, now I'm reconsidering... Please, let's stay on track. Can we please not discuss everyone's fear in...

diysolarforum.com

 

[GXMnow]

I looked at the other thread.

A while back I did the same math and figured it would not fail in a bad way, but the batteries from the 2 banks will not share the current.

It was noted in the other thread, so I will just keep my comment short here. During discharge, the NMC cells will provide the bulk of the current while the voltage is above about 54 volts. Then when the voltage goes below 53.5 volts, the LFP cells will take over and handle most of the current. The NMC cells will provide very little of the current until the voltage goes into the lower knee of the LFP cells ate about 51 volts.

The situation during charge will be very similar. While the LFP cells are in their flat plateau, they will be handling most of the current. Above or below that range, the NMC cells take over.

It could make for an interesting experiment to watch the current over time from both batteries as the system discharges. I tried a small experiment with my 2 E-Bike batteries. They are essentially identical, but I can run them separate, or in a parallel pair for more run time. One was only at 75% SoC while the other was close to 90%. When I connected them, there was about 6 amps flowing from the higher charged pack. But as I pulled a load from it, the lower pack just stopped being charged. Under a heavy load, the higher state pack voltage was pulled down close to the lower state pack.

 

[400bird]

Yeah, the load/charge current will pass back and forth between the two. I don't think I need to be concerned about that; either system should be able to support the full load.

My concern would be how LPF wants to have the cell voltage relax down after a full charge. As you know, the NMC pack doesn't do that.

 

[400bird]

And I just realized how glad I am that I assembled this test pack. Odd numbers of series cells have the positive and on opposite sides. Had I drawn or modeled this out more fully I would have seen it

Ok, new plan each 16s string will be two 6s modules and one 4s module all in series.

That will make the connections easier and these are going on a shallow cabinet, where the only layout that makes sense is to have each 16s lined up straight left to right with the back face nearly touching the enclosure wall.

 

[400bird]

On a different subject, I had another one of these events, where the XW charges and discharges from the grid on a wild back and forth full blast in both directions. At 9:30, while I was in the garage to realize what was happening.
I tried a few different things before I realized I should measure the frequency, 59.97 Hz on one meter and 59.99 on the other.
Look what the grid codes calls for, response at 59.96 hz. That seems like it could be just a measurement error away from what I measured.


So, I added 2 second delays to both activation and deactivation along with spreading out the center dead band around 60 Hz to 59.9-60.1

 

[GXMnow]

I think you are onto something here.

It sure would be nice if the system would log when these actions happen. Then we would know which one caused it.

The other thing on this page "Freq-Watt P(f) Droop" that is a bit scary to me is the "Ramp Rate Increment" at 33% per second. I slowed that to just 10% per second.

On the two "Freq-Watt P(f)" entries, I also moved the threshold out to 59.9 and 60.1 Hz. I want it to just ride through between those limits.

I am also looking at the voltage and power reactions. Most of them have the ramps set slow, to just 1% per second. I changed the others that were at 33% down to 10%.

The other one that looks funky is "Ramp Rate"
The "Normal Power Ramp Rate" is at 100% per second.

I will leave that for now and see if it does the wild power swinging again with the other changes. If it does, I may slow these rates as well.

The "Power Settings" page seems to be disabled. So I don't think those settings will do anything. But since my breaker back to the main panel is only 20 amps, should I lower those power limits?? All three are set to 6,800 watts or over 28 amps at 240 volts.

 

[400bird]

I think you are onto something here.

It sure would be nice if the system would log when WHY these actions happen. Then we would know which one caused it.

I'd love to know why, that way we can more easily adjust the parameters.

The "Power Settings" page seems to be disabled. So I don't think those settings will do anything. But since my breaker back to the main panel is only 20 amps, should I lower those power limits?? All three are set to 6,800 watts or over 28 amps at 240 volts.

I'm not sitting in front of the settings page right now, got a screenshot?

 

[GXMnow]

Here is a screen cap of the page

I did it as a thumbnail, I think when you click it you will then get full res.

 

[400bird]

Oh, it was a different drop down under the grid codes section. I hope all the disabled ones are actually disabled, I ignored any that said disabled.

 

[JustJoe]

...
It sure would be nice if the system would log when these actions happen. Then we would know which one caused it.
...

Just something to add here about a quirk I have seen ... Because after experiencing other (not related to your problem) system issues, I have often gone into our Conext Gateway / Events / log page and found nothing there ... But ... Wait a minute ... I would find log messages pointing back to other events a couple of weeks prior ... where those were missing back then ???

Is it possible their crazy firmware is storing the current log messages until after a system reboot ??? I usually don't want to reboot if I am trying to capture symptom evidence. But I'm starting to think I have to come up with a test case which should trigger a log event to confirm it. I'll let you know if I do. For now, just consider adding that to your bag of tricks.

 

[dougbert]

doh, in the GRID CODES, duh

 

[400bird]

Well, I've got some homework to do to figure out the settings and what change (likely adding the WattNode) made this happen.
The XW charge current, now follows the available/excess PV production. Due to some network issues at my house, I've been running without Node-Red commanding the charge current and just randomly, manually adjusting the charge rate when I remember to check. With the additional DC coupled panels, I don't need as much charging from the AC coupled PV so I could run pretty conservative and still have plenty to make it through the peak evening hours.

Yesterday, as the sun was setting, I noticed I had been lazy and not turned down the charge rate, but the XW never pulled current from the grid as the math said it should have.
Today, I decided to test the theory and set the EPC charge rate to 5000 watts. The grid load has held around 0-15 watts all day!

The key is at the bottom, but most importantly, blue is grid load (as measured at the utility meter)
Yellow is solar
Purple is battery current
Green is loads in the critical loads panel.

You can see I started the charger (manually clicking the button) at about 9:45 and at about 1:00 the XW entered CV/absorbtion and started tapering current and grid back feed increased (as it should)

It might take me a few weeks to figure out the settings and what controls the grid load. It still won't command the start of a charge cycle, so Node-Red (or something to command the start of a charge cycle) will be needed, but it now automatically Zeros out grid load.
I should say that it, on average zero's out grid load. I'm sure I was pushing some current out one leg and sucking it back in on the other. But a sum of zero is what counts for my power company.

 

[GXMnow]

Do you have the third phase CT of the Watt-Node monitoring the Grid Tied PV current?

 

[400bird]

Do you have the third phase CT of the Watt-Node monitoring the Grid Tied PV current?

I do, that's likely part of what has enabled this feature..
The yellow line in the graph is hooked on the same circuit.

 

[Hedges]

Polarity -

Blue positive is power drawn from grid.
Yellow positive is power drawn from PV.
Purple positive is power delivered to (not drawn from) battery
Green (isn't positive because power never drawn from loads,) negative is power delivered to loads.

If that is correct, I think purple battery ought to be displayed opposite polarity.

But doesn't seem to agree with 9:45 AM, when power drawn from grid suddenly drops to zero as battery charging kicks up higher.

 

[400bird]

I've struggled with polarity in all of those. I'll take any input.

Yellow (solar) positive is production
Purple (battery) positive is charging the battery
Green (home loads) negative is consuming grid power, yup it will never go positive because it only monitors loads
Blue (grid) positive is pushing power to the grid, negative is consuming power from the grid.

I think you had grid polarity swapped in your description.

If you look at 7:30-8:00 grid and home loads match because the battery was low enough to hit the reserve and not support loads (when grid tied)