DIYrich
Solar Wizard
How do you get 120v from 208v 3-Phase?
Frequency shifting ports for EG4/Luxpower/Sol-Ark - not identical?
- Thread starter godawgs
- Start date
timselectric
If I can do it, you can do it.
- Joined
- Feb 5, 2022
- Messages
- 19,769
208v 3-phase is a 3-phase WYE system.
It's 120v/ 208v 3-phase.
Basically it's three 120v/ 208v split-phase systems in one.
Lt.Dan
Solar Wizard
This is typically because power supplies in most appliances have a wide 100v-240v range, correct?
zanydroid
Solar Wizard
Isn’t that a partnered system where Victron and Fronius advertise some secret sauce? I would put that in the same class as Enphase PLC, except maybe with less theoretical capability for on grid operation since you can’t shift any parameters in that situation while you can still modulate PLC
zanydroid
Solar Wizard
Major appliances are more like 180-240V, to cover shitty grids, Japan’s lower voltage, and U.S. + EU
Resistive elements will output less heat at 208 vs 240 and this usually spelled out (and can be derived yourself my ohms law).
I have seen some appliance spec sheets that list only 240v
Lt.Dan
Solar Wizard
Makes me want to get a 30k or 60k at my next house, just for fun. Lol
Has anybody seen one installed?
Has anybody seen one installed?
wpns
Solar Joules are catch and release
F-W should work if the main inverter and AC coupled addon have matched response curves, and I _thought_ that’s what all the parameters in the F-W settings were for.
However, that feels too much like balancing on a knife edge for me, so I’m decomissioning my grid-tie inverters and going with a known-compatible (EG4) system. Yeah, it’s more money, but life’s too short to chase parameter tuning for months.
However, that feels too much like balancing on a knife edge for me, so I’m decomissioning my grid-tie inverters and going with a known-compatible (EG4) system. Yeah, it’s more money, but life’s too short to chase parameter tuning for months.
No. Victron can communicate via LAN with the Fronius to do zero export. But in an off-grid setup that function is not used. The Victron and the Fronius systems are really just individually configured, and there is no "communication" other than frequency shift.
Now, I will add this. The Victron system will "see" and log the PV production numbers if the PV inverter is connected to the internet via the same LAN and/or WiFi as the Victron system. This works with Fronius, SolarEdge, SMA, ABB and maybe a few others. Fronius is the only brand that the Victrons can control via actual communication to the point of zero export.
So no, there is no secret sauce. Just quick and accurate frequency shift reaction.
Before I had my Fronius, I temporarily had an SMA Sunny Boy (it was an older unit) that did not have new enough firmware for a gradual curtailment, and even that worked ok. But having the gradual curtailment works way better! It will curtail well enough to even allow the batteries to drop to a Float voltage, but maintain that voltage!
Disclaimer* I would personally recommend to go with DC coupled in off-grid, unless someone is using >50% of their PV directly as it is produced. My PV>battery>loads efficiency is not great!
timselectric
If I can do it, you can do it.
- Joined
- Feb 5, 2022
- Messages
- 19,769
Just in the SunSink videos.
timselectric
If I can do it, you can do it.
- Joined
- Feb 5, 2022
- Messages
- 19,769
AC coupling works best if there's also DC PV connected.
It's a much smoother control.
It's a much smoother control.
Ampster
Renewable Energy Hobbyist
Yes, had I known that I was going to end up with a SolArk when I designed my initial system in 2021 I might have planned for more DC coupled solar to give me more flexibility. As it turns out my initial install of 8kW of micros was a lot easier because RSD was built in. I also encountered fewer long term grid down situations and those have lasted only a few hours so not having the benefit of all my AC coupled micros in my current scenerio is not an issue. If I do have a long grid down situation I have a 14-30 receptacle which can put 5.7 kWs of load on my system and when I testedd that the AC coupled micros were able to produce even though my battery was at 95% SOC. I do have 3kW of DC coupled solar and the smoothness of that control allows my system to function well during situations where there is little load on just the DC coupled solar.
zanydroid
Solar Wizard
The secret sauce I meant is, the two companies worked together to ensure that it worked, and it’s a supported combination. To me this means they tested it and were willing to release it together. And there is commitment on Fronius side to guarantee that the specified Microgrid profile will either never be changed in behavior, change is within agreed upon limits, or change is tested for regressions.
Victron has numerous writeups that are easy to find on Google for how to set this up.
That said, I can sort of believe that AC couples retrofit companies are willing to take on the testing load themselves for the most common customer situations (EG IQ7 and IQ8, or SEDG), in the absence of cooperation from the GTI vendor
DIYrich
Solar Wizard
100v-240v are power supplies that rectify to low voltage dc.
240v motors are a different. UL 1741 SA requires the inverter to remain connected indefinitely at 88%, or 212v, and for 20 seconds at 70%, or 168v. Motors should be able to run in those ranges. Although 208 is just below the minimum for indefinite.
Ryushin
Solar Enthusiast
I have two Growatt MIN 10Ks hooked to the AC Coupled port on my Sol-Ark 15K. I know the Sol-Ark techs do say there is no modulation but there is. I even had to prove to the Sol-Ark tech that I was talking to that their own video says it varies the frequency to curtail output:
I have a mix of AC and DC PV. I have 17.6 kW of PV on the AC coupled ports and 16.8 kW of PV on the DC side. I have watched the Sol-Ark 15K vary the frequency between 60-62Hz to curtail the production and the Growatt MIN 10Ks did reduce their output based on the frequency in 34% output steps.
There can only be one source of frequency going into the Sol-Ark. So either the frequency is controlled by the grid and the AC Coupled can't be curtailed and will run full out, controlled by a generator and you can't have the AC Coupled running at all, or off grid and the Sol-Ark will self island and generate it's own frequency (BTW, not exactly 60.00 Hz. It's like 60.03 Hz and we had to modify the Growatt settings because it was not exactly 60.00 Hz) and it will be able to curtail the AC Coupled 1741SA inverters.
BTW, the Sol-Ark can shift it's frequency all the way to 65hz to shutdown non 1741SA inverters. But this works more as a off/on kind of thing.
Actually, the biggest thing is to have the frequency range where the GT inverter goes from 100% to 0% output correct. In the Victron inverters you set those parameters (with a little room for error, e.g. 0% Frequency is set slightly higher than where 0% is on the GT inverter) and you tell the Victrons how many kW of PV there is connected, then they work off of that. And I'm pretty sure it would work quite well with any GT inverter that does proper frequency shift curtailment that isn't just on/off. The SMA that I had used at first shut off at 60.5Hz IIRC. The Fronius starts throttling around 60.5Hz or maybe 60.8Hz is at 0% somewhere around 62Hz, which I believe is a pretty standard range on new GT inverters.
You don't even tell the Victron what brand of inverter you are AC coupling to it when configuring it!
zanydroid
Solar Wizard
Ok, but upthread in post #18 i shared the grid response speed required by my POCO:
1547-2018 defines open loop response time as
“The duration between a control signal input step change (reference value or system parameter) until the controlled output changes by 90% of its final change, before any overshoot.”
Ampster
Renewable Energy Hobbyist
And of course the GEN port relay is the ultimate failsafe off/on mechanism.
Whenever you are in a grid-interactive (I.e. net metering/selling to the PoCo) type of scenario, you are of course, subject to the grid-interactive requirements of said PoCo. The one thing that I'm not sure on in regards to that 5 second response time requirement is if that is "the" time frame, or if it is the maximum response time allowable. I would expect that when curtailing "down" (dropping output power level) you would be allowed to respond instantaneously, as an instant drop to 0% would be desirable in a grid-down scenario. I would expect that the 5 second requirement is to avoid PV production to go from 0% to 100% instantly, as that could cause grid instability, especially right after coming back online from a power outage scenario. Now, there is a 5 minute delay on re-connection when grid comes back on, but imagine 20 homes with 10kW of solar connected and grid comes back on, then 5 minutes later the full 200kW of PV starts dumping onto the grid lines at the same time.
Ramp-up rate of AC coupled PV isn't a huge deal when off-grid or when running on battery backup (grid-down). Yes, it does mean the battery gets slightly cycled whenever a load kicks in until PV ramps up to take over again, but from what I have seen at home, it really isn't a big deal. Now if someone were to have a very small battery bank (<10kWh), that could change the picture. But honestly, if someone wants to AC couple to a battery based inverter, then I would argue that they should have at least 15kWh or more of battery anyway. My argument on that stems from the whole picture of 1.) how much loads someone (likely) has if AC coupled, and 2.) Inverter size for a unit that is capable of supporting AC coupled PV, should then also have +10kWh of battery.
For reference: discharging a 10kWh battery bank at a rate of 10kW (pretty high discharge rate there...) for 5 seconds = 13.9WattHours which is 0.14%.
So, discharge any size bank at a 1C rate for 5 seconds (until GT inverter ramps back up power output) and you have lost only 0.14% of that banks storage! 5 minutes of 1C discharge would drop 8.3%. So as you can see, unless you are going to have massive load swings, to the tune of nearly a 1C discharge, having a few seconds, or even a few minutes, of discharge until PV GT inverter ramps back up won't be anything terrible at all.
zanydroid
Solar Wizard
What is the response parameter on the off-grid code that you posted earlier?
There's also ride-through provisions in 1741SA to avoid destabilizing the grid when there is a blip. 1741 classic was too trigger happy with disconnecting. You don't want the grid to get wrecked if, during a brownout or other blip, the grid tie inverters decide to peace out and withdraw their power.
To clarify, you are only talking about grid tied with the 5 min delay? AC coupling is supposed to avoid that (and that's what I'm happy with, as I've beaten to death a few times already)
These violate standard design rules. You are supposed to have grid former inverter AC-out = GTI AC out (sustained, not surge), so that it can always fill in. And they are supposed to be able to ramp up quickly enough.
What are your thoughts on the ramp-down constraints? I think they're determined by the charge current available at a particular SoC %. Depending on how good the BMS communications is, this may be set more or less conservatively. Also, the AC coupling may not be willing to use the surge charge current offered by the BMS to push it further.
If you have a really large battery to inverter ratio, you might even be able to charge at very high SoC %, if the CC/CV taper still allows enough charge current. Now, I'm not sure if the charger logic is actually smart enough to predict the available charge current at that part of the curve.
I'm actually not sure. I'll see if I can find that.
Do you know if the ride-through provisions are both V and Hz based? Or are they only voltage based? (I think frequency is used as well, if I remember correctly.) If frequency is involved as well, do you know what the upper limits are for Hz? (I mean specifically for ride through.)
I'm pretty sure 5 min delay is required any time you are grid-tied. But in an off-grid setup you could bump that time down to 5 or 10 seconds.
What I am referring to is when the grid former/battery based inverter has frequency shifted up to throttle back GTI because batteries are full or nearly full, with low loads. Then a large load kicks in (let's say electric dryer). When that happens, the battery based inverter drops the frequency, in order to allow GTI to ramp up again on production in order to have PV directly power that load.
Battery based inverter will "play" with the frequency to get GTI to exact power output to match loads+battery charge "demand". Once battery is full, that battery charge "demand" (or allowable charge amps) goes to pretty well zero, so then the frequency will be adjusted to limit GTI to only direct feed loads.
The battery based inverter will use either closed loop comms info or measured battery volts and amps to adjust frequency shift off of. My personal system does not have closed loop comms. Therefore the Victrons are looking at battery amps (measured by Victron SmartShunt) and battery volts, and also looking at charge voltage for current charge state. When battery volts rises above absorb voltage setpoint, the Victrons begin to shift frequency up. I believe they have a PID loop of some sort in their algorithm, and how fast the voltage is rising will play into how quickly the frequency shifts up. Once Absorb voltage has been reached, they will "hold" it there until Absorb time has been reached (I have mine set to 1hr to allow top balancing of cells to happen.) then frequency shifts all the way to 62.4-62.6Hz until battery volts drops to Float setpoint, then PID loops resumes "playing" with frequency to maintain Float volts.
At any time when a voltage is being "held", if negative current happens at the battery (in my case measured by SmartShunt) or if voltage drops below target, the frequency will be dropped to allow GTI to ramp up again to offset whatever load, etc. caused this.
In a closed loop setup, the battery based inverters would use targets set by the battery comms for the charge volts and charge amps, and they would also use the battery's data (live/realtime volts and amps) to run the frequency shift calculations PID loop.
So if the battery's comms adjusts the max charge amps depending on SOC%, that would be reflected in how the frequency shift is handled/controlled.
On my personal system (I am of course running "openloop"), I only have AC coupled PV and I am able to get my batteries completely full. (E.g. 100% SOC.) I think the biggest issue that people might run into is with batteries that do not have active balancing, and only passive balancing in the BMS. I have seen so many batteries with only passive balancing run into cell imbalance issues after 4 or 5 years of operation in off-grid setups! What happens then is that battery volts comes up close to your Absorb voltage target, then one cell (or cell pack) suddenly has voltage spike causing BMS shutdown. If there are a few batteries in parallel, the other units receive unit A's charge current when it shuts down. This causes higher C rate, which causes imbalanced cell voltage spikes in another one...... which dumps ore charge current.... until all units have shut down. That in turn causes battery based inverter to see battery overvoltage (when last unit shuts off, charge current spikes DC voltage) and shut down.
As I had mentioned in comment #43, 5-10 seconds of even 10kW is a pretty small amount of kWh relative to a battery bank that is 10kWh or bigger! So, as long as you are not getting cell imbalance related cell spikes, there should be no issue having a power dump feed into the batteries for a few seconds until the frequency gets shifted up and the GTI inverter responds and throttles back power production.
Personal scenario at home: (not exact numbers, but rough idea of normal scenario)
-Batteries are 100%, frequency is ~61.8Hz, GTI is producing ~300w to offset small loads and trickle charge to batteries
-electric dryer gets started, battery amps go negative/volts drop, frequency drops to 60.0Hz to allow more PV "flow"
-a few seconds later.... GTI ramps up to 5,000w, pretty much playing even on power for next 45 min to 1 hr
-dryer cycle ends, dumping PV to the batteries, Victrons "see" voltage rise and ramp up frequency, GTI "sees" frequency shift and throttles back
Key thing here is that even when full (100% SOC), the few seconds of 5kW dumping into my 20kWh battery bank until GTI curtails doesn't cause any excessive battery voltage spike, because it's really a very minor amount of power relative to battery size.
wpns
Solar Joules are catch and release
So classic cascade failure. I get that one solution appears to be active balancing to keep the batteries from getting out of balance and starting the chain reaction above, but wouldn't closed-loop comms also solve/avoid the issue? [I get that you don't want/have closed loop, just wondering if it's another solution.]
Ampster
Renewable Energy Hobbyist
It would depending on how the BMSs communicate. I have several rack batteries which communicate through a master. I don't know how an inverter would respond in the above situation. If it reduced the charging current the cascade could be stalled.
Reducing charging current can help, but honestly, active balancing is by far better long term! I have seen/worked with batteries that were so far out of balance (after 5 years of use with only passive balancing) that in order to balance cells AFTER ADDING AN ACTIVE BALANCER WITH 1A BALANCE CURRENT that I had to set all charge voltages to ~3.375V/cell (27V/54V for 24V/48V nominal respectively) and hold the voltage there for a whole week before bumping up the voltage gradually until I got it up to ~3.55V/cell. Anything higher than that resulted in cell overvoltage spikes!
A few notes to consider:
-A grid-connected system that is not cycling the battery bank, and the batteries are always held at a voltage of +/-56V (48V nominal battery bank) may never have those issues, as the passive balancing may keep up.
-Personally, I have active balancers. I have zero cell imbalance issues!
-The cascading issue that I had mentioned has become a very common thing for me to see happen with batteries that do not have active balancing.
-While closed loop comms MAY help avoid the cascading issue. I would argue that it is not the BEST option. (As far as fixing cell imbalance.) The reason being that you are basically bandaging the symptoms (cell imbalance>cell overvoltage>BMS shutdown) by reducing charge rate and possibly reducing charge voltage. IF this voltage is help for extended periods, the passive balancing MAY catch up. (Likely not though, from what I have seen.) But very likely this cascading effect is a result of imbalances coming from cycling the batteries, in which case that balance voltage will NOT be held. (After dark the sun goes down, as does the battery SOC% and voltage.) Therefore there will be no improvement of cell balance.
-Cycling of a battery that has cell imbalances results in some cells hitting overvoltage and other cells hitting undervoltage. Both of these conditions adversely effect CELL life. Notice I said CELL life. If one or a few cells start losing capacity due to this, your overall usable capacity ends up being drastically reduced. I worked on a 450AH 24V battery that was ~5 yrs old and had drastic cell imbalances. Once we got the cells top balanced we had a usable capacity of ~200AH left. I feel a lot of the reduced capacity was related to cell imbalances.
Side note on closed loop comms. Personally, I'm not a huge fan, despite working with closed loop quite a bit these days. I feel like closed loop gets touted as being the cure-all/fix-all, while the "target" charge voltages and reduced charge current parameters that I see on many name brand batteries tells me that there have been cell balance issues that are being bandaged by lower current at upper voltage levels, as well as simply reduced target voltages. On a battery that has good quality cells with no imbalance, you can charge at a high charge rate and take the voltage all the way up to ~56V (on 48v nominal battery) and the charge rate will drop from full out to nothing in literally 2 or 3 minutes time and the battery is full. That is without any overvoltage spikes, shutdowns etc., and no taper charge at the top end either! With that scenario I see people getting more out of solar because of no taper charge (documented by charge controller daily solar logs), and also a backup gen will get the batteries full quicker due to no taper charge. That reduces runtime and saves $. (That is if you want to charge to 100%, of course. Though to be fair, most people don't charge to full these days, with LifePo4.)
Similar threads
