It happens sometimes, not every day. Will try to catch the moment.
And seems it does not correlate with price, there were no zero or negative prices on May 27. 13:00-16:00 price was very low, but not negative.
@alexpescaru Iâm NOT in favor of these all in one systems, as they often lack the flexibility required to optimally match a clients requirement.
Re the debate over Current or Voltage control, it IS possible for a BMS to correctly implement BOTH, including allowing for DV PV to send power to the grid. However, getting this algorithm to work correctly requires some thought and a lot of testing with a real system. This expense is mostly avoided by some BMS code writers, taking the far easier option of just blocking charge. As far as Iâm aware, Victron is the ONLY DC MPPT manufacturer that allows for remote control of either the charging current or Voltage limit, In my view this is an essential requirement for charging âsensitiveâ batteries. (Lithium arenât the only batteries sensitive to overcharge). As a result, I use Victron equipment exclusively for my Clients.
I also appreciate the Victron thinking in some aspects.
If you see in one of the previous messages, Iâve said âitâs better with CVL and also suits your hardware betterâ.
As per AIO, lets look at the schematic below.
Because of that high voltage bus of an AIO, the flexibility is maximum.
The HV DC bus is like a nexus, from where all things happens.
Just take a look at those grey arrows. All is bidirectional !!
Anyone can âhelpâ (read PowerAssist) anyone on this topology.
If we take in consideration the initial OP message, imagine that when the SOC is 100, then the HV DC bus to battery DC/DC system can be disabled and then the inverter can take all the energy put on the HV DC bus by the MPPTs and use it to feed-in. All of it or only a part, according with the feed-in limit.
Not to mention that the battery is galvanically isolated from the high voltages of the PV panels and MPPT malfunctionings.
A slight off-topic, but why Victron RS PV input is limited to 8x battery voltage, and Deye is not?
For me it is a show stopper. I have several 560V strings, and it is quite complicated to reconfigure them to < 450V. Thought about swapping Huaweis and Multiplusâes to 3x RS 48/6000 Victrons, but PV voltage requirement is a limiter.
Which ESS mode ist set? âOptimized with BatteryLifeâ or âOptimized without BatteryLifeâ
I had the same phenomenon on my system a few weeks ago⌠Out of nowhere⌠Only after I switched the ESS mode from âOptimized without BatteryLifeâ to âOptimized with BatteryLifeâ as a test and then back again did the system run as usual.
Even restarting the CerboGX and changing the firmware did not help!
That was a very, very strange errorâŚ
In theory it shouldnât be limited to that, as in the Deyeâs case.
But, starting from the conditions that:
- the float voltage is generally the lowest battery voltage
- the battery to HV DC bus DC/DC converter is bidirectional
- the battery to HV DC bus DC/DC converter should act fast and switch from buck to boost and vice-versa almost instantly
probably it has to do with the hardware design and the maximum/minimum (safe) pulse width of the DC/DC converter.
And probably with the fact that the MPPTs are working in a boost configuration for most of the time, considering that the HV DC bus is about 400V.
But for such high PV voltages, Victron has no solution anyway⌠YetâŚ
But this âyetâ is more than 2 years old, if we take into consideration the HS range.
Deye PV max voltage is 650V (on newer models 800V), which is reasonable.
Here we must take into consideration that also the AC/DC bidirectional converter must charge the HV DC bus to the same 400V.
Probably the Deyeâs HV DC bus is 650V or more, not 400V like on Victronâs case.
But this implies different HV capacitors, in different topology (series - parallel) which complicates things.
Also the AC/DC bidirectional converter should be a little bit different.
If you take a look at the Victronâs RS range, you can see that they even used a little bit different topology than the usual double H bridge topology for the battery bidirectional DC/DC converterâŚ
Money vise Deye looks better for household installation. Victron is, of course, much more flexible. But⌠Deye wins. Typical household installation in Estonia is limited to 15kw. 3x 5kw Deye (6 MPPT trackers!) is ~3.5K euro.
System DC voltage is a different issue to ALL in ONE.
Many years ago I did the engineering for a 2MWh system where the batteries were Samsung, the battery inverters were from Leonics and the PV inverters were also Leonics.
Point is that inverter and battery sizing physical layout and system voltages all needed to be engineered.
With AIO systems, a lot of that is taken care of for you, BUT there will be limits on the battery to inverter sizing that may not be compatible with a clientâs needs.BMS systems will be engineered in, but NOT open source - so you are stuck with whatever user interface the supplier deems.
I would love Victron to up the max system volts from 48 to 500 ish for the real large systems, but for <20kVA, 48V is ok.
for DC/DC converters, A three phase bride FET design - going transformerless - for 240V works from +/- 380V relative to Neutral voltage. This gives you a DC bus voltage if 760V dc, with isolation OPTIONAL. This also means that the internal DC bus has a Large AC component, so batteries need to be isolated from ground, as does the PV input. Ideally there could be some High frequency dc/dc isolation transformers included in the design.
Victron currently uses low frequency transformers, these are large and heavy, but the design step from LF transformers to HF transformers is massive, needing a total redesign of the entire system.
Iâm talking about PV voltage, not battery.
@kbb
This is why Iâve said above that some must adapt⌠And quickly.
Of course, different companies of different sizes and budgets.
Like Iâve said above, I am still waiting, with the money in a stash hidden from the wife, for the RS19 range or even HS19, but I feel like a lover seduced and abandoned for 2 years now⌠
And 2 years is an eternity for today marketâŚ
Product cycle time is a little bit disappointing, but lucky for Victron that this is compensated by the usersâ âfanaticismâ in the brand. 
And the fact that they are taking the road of letting the user to beta test their software and the existence of this community.
This is a smart thing and a win-win situation for them: (1) they draw free of charge from the collective knowledge and/or experience pool and therefore are quickly aware of the bugs and (2) the users feel more cared and this raise the above mentioned âfanaticismâ in the brand.
Without this community they would have a BIG problem.
@MikeD
They did a pretty good job on the RS range, probably because they have their way paved from the existing knowledge pool.
Too bad that theyâve took some shortcuts, but it seems that theyâve managed to make it work without problems.
Indeed, battery to inverter sizing could be a problem because of that ability to switch FAST from charging to discharging, but once that problem solved all is home free.
Lucky that these days 20ms (50Hz) is an eternity for proper fast DSPs and FETs.
Industrial PV voltages are now up to 1500V, 1000V for residential applications. For Home size systems, the standard limit used to be 600V.
These PV systems are typically much larger than needed for the market that Victron users would normally require - I operate at about 120V dc for the pv - and PV system fusing switching and safety design is a whole new ball game, requiring specialised components. I think that Victrons largest charger is 200A, at 48V nom, this gives 9.6kW of pv on 4 trackers. If you are limited to 8 x battery voltage, this would give you a total PV current of ~25A, or 6.25A on each of 4 inputs. At this level there is not really need for a higher input voltage.
several manufacturers can do full 4 quadrant inverters up to over 500kW. Size âdoesnât matterâ too much for this. higher switching frequencies mean smaller transformers, but ferrites have limits.
Indeed⌠What a wonderful world nowadays with so many possibilities⌠
But like you said, the home market is having/needing different expectations/needs.
Sorry Konstantin @kbb for my rambling and off-topic. 
Sometimes I get frustrated by these little quirks of the system and one could get surprised counting the number of times the word âfixâ is occurring in the Venus OSâs changelogâŚ
But hey, we are still hereâŚ
Lucky for us that we can personalize, tune and sometimes fix our systems through Venus OS scripts and/or Node-Red.
And some through modifying the firmware inside inverters and batteries to suit their needs.
If I see this picture I strongly believe that these low prices was/is your issue.
Between 13:00 and 16:00 they were negative right?
Compared with the middle picture of this:
What are your DESS battery cycle costs? These are also part of the equation.
Price was positive, but very low (like 0.01 cent/kwh).
My battery cycle cost is 0.01.
I strongly believe that DESS was throttling your solarchargers.
Next time simply increase (temporary) the sell price formula with a few cents in the VRM and see if charging will start (feed in grid)
Unless you got your batteries for free, that cycle cost is way too low and will end up in you losing money (high battery wear) instead of gaining money.
It is tempting to set it very low so you can âseeâ your battery working, but it is far from an optimal setting.
The formula for correct battery cycle cost calculation is explained on the DESS settings page and the reasons to set a correct value have been discussed many times.
Of course, you are free to refuse the logic and run your system in a not-optimal way 