Best implementation of diversion

[Mattia Zamboni]

Hi!

I'm posting this question here since I'm sure it can be helpful for several people.

I'm planning a solar system with a few high power solar panels (space constraints).

Because of the off-grid remote location I want to implement diversion (Or dump load) for water heating.

I do have an SBMS0 and I'm looking for the best way to implement it.

Panels: 5x Trina solar 385W

Battery: 8x 3.2V 280Ah EVE LiFePo4 cells (24V)

BMS: SBMS0

I did use the DSSR20 and DEX16 before on a simpler system with a single panel and it worked great.

Now, I will most likely need to wire the panels in 2 separate strings (in order to capture the sun at best during the day/season considering also shading)

I was wondering what is the best way to divert the power. Doing it after the MPPT(s) would make it easier to combine the power, but can this be done using a contactor?

I was trying to understand how the dump load feature works on the Victrons MPPTs but it is hardly documented.

Has somebody already gone through this?

Thank you very much in advance,

 

[Dacian Todea]

Mattia,

I do not think Victron MPPT has anything related to dump load.

So you have the 5x Trina panel (guess they are 72 cell panels) all in parallel connected to a Victron MPPT ?

DSSR20 can only handle two of those 385W panels in parallel if that is what you were thinking about and then you can connect the diversion to a 2Ohm or so heating element (maybe half of a 36V 1200W element as those are made out of 2x 600W elements in parallel) then the other half can be connected to another DSSR20 connected to another set of two panels.

But I do not have enough information about your current system.

How are the panels currently connected ? parallel / series how many wires ?

After MPPT you can not do diversion unless you have the battery connected meaning you get power from panels but also difference from battery.  If MPPT is disconnected from battery then it will just not work if you connect that instead to a heating element.

[Mattia Zamboni]

Thank you Dacian, sorry for not providing more details in the first place.

Solar panels: 5x TRINA SOLAR VERTEX S TSM-DE09.05 385W BLACK (120 half cut cells)

Panels configuration/wiring: still to be defined.

Panels needs to go on the roof of a mountain cabin. Because none of the roof sides are facing south I was considering placing on both sides.

Because of the mountains the sun shows up only at 11AM, at 2PM it is right between the 2 roof sides and sunset occurs in summer at around 8:30PM.

I was considering making at least 2 strings, and I initially thought to wire the panels in series within each string to keep current low and reduce losses.

However I understand using the DSSR20 for diversion this would not work.

But what are the exact implication to adoption your wiring scheme?

Can a DSSR20 handle 2x 385W panels? In practice this would essentially require panels to be connected in couples on the roof and then 2 wires for each couple to be pulled to the electrical cabinet inside the house? Would 10AWG be enough for this configuration?

I assume that with these specific panels I would then need an MPPT charge controller for each panels pair and combine their outputs to go to the battery, right?

About the Victron energy dump load feature I was referring to this video:

https://www.youtube.com/watch?v=dT89QtG2pS4

Unfortunately I haven't found documentation nor anything built-in in their products, which to me is somewhat a shame considering their expertise and wide products portfolio.

In my wildest dreams I was hoping to eventually be able to integrate a Raspberry PI with Victron Venus OS for remote monitoring reasons, but this would require getting all Victron branded MPPTs raising the cost significantly and diversion would not be computed, so I'm still not sure.

Thank you for your precious inputs.

[Mattia Zamboni]

By looking at the detailed DSSR20 specs it sounds like 2 panels (385W) connected to a single DSSR20 will be too much since potentially exceeding the max 20A current. So I would need 1x for each panel, and every panel would need to be wired separately making the whole installation somewhat overcomplex.

However, I did read that you are working on a DSSR50, right? Sounds like the limit there will be 50A there. I am not in a rush for the diversion part, so I can plan for it, wire panels pair in parallel, and feed my MPPTs. I will then plan for the upgrade later.

Am I correct in my thinking? Do you have a schedule for the release of the new DSSR50?

Alternatively, I assume I could replace the DSSR20 with a traditional contactor, connected through an optoisolated interface board to the DEXT16 or SBMS0 directly.

Generally speaking, I think diversion is way underestimated at this time, and so few products support it. And this is what makes Electrodacus's devices so brilliant and forward-thinking.

Just now people are starting to wake up and come up with ideas to use the access energy for green cryptocurrency mining...

Thank you!

[Peter]

Hi Mattia,

my solar setup is also with Victron MPPT.

I intent to implement solar diversion on higher logic level: I am going to read values from SBMS0 via MQQT: battery shunt value, PV shunt value, battery SOC. And with SOC over some limit I will "divert the power" -> automatically turn on some applience, in my case DC boiler or DC floor heating with power level based on shunts values (so that battery shunt will show zero). Power setting is done via mosfet PWM.

My implementation is in Home Assistant but I guess there are many options to do that on this higher level.

Cheers!

Dne úterý 1. června 2021 v 10:15:53 UTC+2 uživatel Mattia Zamboni napsal:

[Dacian Todea]

Mattia,

The DSSR20 can handle two of those 385W panels in parallel ideally you will have one panel on one side of the roof and the other one on the other side if is not to complicated but the DSSR20 should handle two of those panels even if they are both on the same side directly facing the sun and yes 10AWG wire is excellent for a group of two parallel panels.

You do not need any MPPT as te DSSR20 will already work at max power point with those 120 half cut cell panels (two 60 cell in parallel).

That Victron future will not work with LiFePO4 only with Lead Acid and what happens there is the battery will still be involved in powering the Load as Load will need to be larger power than what you can max get from the panels.

With Lead Acid battery voltage quickly gets to absorption level 14.4 or 28.8V (battery may be just at 60 or 70% SOC) and from there the batteries will take less current than available from panels so you can use the Load output of the MPPT solar controller to connect a Load but that Load will be supplied by the PV panels with different coming from battery or if smaller the battery will continue to charge tho with Lead Acid battery will never get to full charge (not with solar).

With LiFePO4 this strategy will not work as battery will always be around 26 to 27V at almost any SOC and battery will use all the current PV panels can provide so nothing is lost until the battery is fully charged when the solar charger is ordered by the SBMS0 to stop.

You can use the SBMS0 signal that battery is fully charged to divert the PV energy directly (can not be done through MPPT).

The DSSR50 will handle much better your panels and can handle 4 of those panels but the diversion for DSSR50 will be available as a separate device much later probably only early next year.

The DSSR50 I hope to have available by the end of summer.

Peter,

You should leave the battery to fully charged when the day starts then you can maintain a lower SOC. If you never fully charge then SOC will decrease in accuracy as correction is done at full charge and without that correction all the small measurement errors will add up over time and SOC value will be less and less accurate.

There is no need to get to this level of complexity with PWM trying to match the exact PV production as the LiFePO4 battery can cycle at the top hundreds of thousand of times so all that cycling will have no effect on battery life is like a super capacitor in that sort of use case scenario.

Also your PV array can not be that large in this case just 0.2 to 0.3C but for house heating you need a much larger PV array thus you should be able to use the PV energy directly and separately from battery.

 

 

[Peter]

Dacian,

I admire how you take care of all the topics that appear here, respect! I am so grateful for any information from you, this group is a real blessing for me. Thank you!! :-)

I know that if I dont fully charge, SOC of SBMS0 will be less and less accurate.

My use case is fully electric camper van with HUGE battery 24 kWh. It will be 100% SOC thanks to solar at the beginning of trip. And during the ride I just want to use that power. If my solar 750W with only max 0.03C charging will be able to maintain the battery at 100% SOC every day, it will be perfect. But I still want to be able to park for 10 days in deep woods without any solar and use my fully electric solution (including induction cooking, microwave, oven, electric hot water, electric heating, electric freeze prevention of water tanks, charging eBikes, ...). In such conditions its impossible to get to 100% SOC every day. If I reach for example 15% SOC during the trip, I just use my 2kW AC charger at almost any electric car charging station. Although the solar could keep up with consumption, I don't want to rely on it, it's just a bonus for me.

Almost every device is controlled by MOSFET via ESP so I have PWM for free. Everything runs on Home Assistant so that PV energy diversion for heating is only easy geeky cherry on the top of entire system.

If LiFePo4 will have such long lifespan, awesome!

Dne úterý 1. června 2021 v 18:51:25 UTC+2 uživatel electr...@gmail.com napsal:

[Mattia Zamboni]

Dacian, thanks for the hints!

In order to finalize my decision I need to understand one more thing: whether for the efficiency of the system it is better to connect two panels on the same side of the roof or opposite side.

Option 1, opposite sides: In this case the panels they cover a longer period of sun, but at a lower wattage, since when one is at its peak, the other one is not and vice versa. If this in diversion mode is connected to a heater resistor it implies I will never reach the full resistor wattage. (I actually don't know its graph temp/watt)

Option 2, same side: in this case when the sun is around its peak on the panel I get full production, most likely even too much at around maybe 700W. Here I just need to understand how safe really it is to connect the 2 panels at the DSSR20. According to the specs indeed at max this panel can produce 11.46A. Since you say it is ok for the DSSR20 even with panels on the same side does it mean the DSSR20 will limit this current to 20A? No risk of frying the DSSR20? How about running it at its peak? Will it not overheat or have its lifespan reduced? I really need to understand this in order to take a wise decision.

https://static.trinasolar.com/sites/default/files/EN_Datasheet_VertexS_DE09.05_202102A.pdf

The above cases of course assuming the battery reached 100% SOC.

When it comes to charging the cells without MPPT, but only from the DSSR20 directly can you confidently say that it will not affect the battery health/lifespan? Is the DSSR20 taking care of not exceeding the 3.65V*8 =29,8V during recharge? I of course understand this works, and I would be more than happy to adopt this solution, but I just want to understand whether this implies a trade-off and sacrificing something.

In the meantime I will start to prepare my order for the additional DSSR20 with diversion ;-)

Thank you so much. You are such an inspiration!

[Mattia Zamboni]

Hey Peter,

Thanks a lot for your hints, really interesting solution.

From what you describe it sounds you put quite some effort and expertise in developing your solution.

Based on my knowledge I understand the following:

1) you do have some HW platform (Raspberry PI I assume) within you network running an MQQT client monitoring via Wi-Fi the SBMS0.

2) On the same HW a software is running which based on the values read from the SBMS0 activates appliances DC floor, DC heating, etc. This activation is not just either ON or OFF, but modulated via PWM in order to have more accurate control.

3) the implementation of this control is performed via the Home automation OS for Raspberry PI.

Based on your answer to Dacian however it sounds to me like the diversion you are talking about is actually more like enabling a load which is connected also to the battery, right?

Anyway I will look into this solution as well, to see whether it sparks an idea.

In my future plans I would indeed also like to have remote control/monitoring as well as the ability to remotely control water heating and floor heating while away. (to prepare the environment before my arrival).

Anyway, if you are willing to share more on your HW, it could be interesting.

Take care!

[Dacian Todea]

Mattia,

It makes no difference in terms of efficiency of panels are on the same side or different side of the roof. The advantage on being on different side of the roof is that the total current will always be lower since one panel will see full sun while the other less or not at all thus there will be less stress on the DSSR20.

The heating element can be chosen to work best with the both side of the roof setup.

DSSR20 will do no current limitation but is OK with 24 even 25A ocasional from those panels when charging the battery as for diversion the current is limited a bit by the resistor value so there current will likely be in the 16 to 19A depening on resistance value you select.

The DSSR20 is controlled by the SBMS0 so as soon as any cell gets to 3.55V meaning fully charged cell the DSSR20 will stop the charging. It will be the exact same thing with an MPPT as the MPPT will need to be controlled by the SBMS0

[Peter]

Mattia,

1) yes, RPI 4B 8GB,

2) yes, that key software on RPI is Home Assistant. It controls several ESP32s with ESPHome or with Tasmota,

3) not Home automation OS, my RPI has much more tasks to do, not only management and automations. I have Home Assistant Docker module installed on RPI.

Yes, my "diversion" is done by PWM precise control of load in a way, that connected battery is not affected. I want to bring battery shunt value close to zero not to use any cycles on the battery.

I do remote control via mobile phone in the van and Home Assistant.

I have shared some info here: https://docs.google.com/drawings/d/1qYhP9BSZgqCqVNMfGxng1hFc_rXr4rki_F04bAA9R2w/edit

I also recommend this mosfet board with little additional cooling: https://www.aliexpress.com/item/4001360170091.html, it has only 10mili ohm resistance (my real measurements) and can be directly controlled from ESP. Each channel can do up to 10A, if you need even more, you just use more channels in parallel. For example for water heating with https://www.aliexpress.com/item/4000045965784.html I need up to 60A so I use 6 parallel channels.

For AIR heating I use PTC elements: https://www.aliexpress.com/item/1005002197577418.html and for battery and floor heating I have heating wire: https://www.aliexpress.com/item/1000001203553.html.

The water boiler also has heat exchanger so I can get heat from the engine.

Well, it took me "a while" to do the research and pick these components.. :-)

I am glad if it helps! :-)

Dne středa 2. června 2021 v 2:44:44 UTC+2 uživatel electr...@gmail.com napsal:

[Mattia Zamboni]

Dacian,

thanks for the crystal clear explanations: now I know how to plan best my system!

Peter,

That's a lot of info to digest at once ;-) Thank you! I will look through it to see whether I understand the implementation.

I actually already bumped into Home Assistant before, but never dared to try it. Now you are giving me great inspiration by seeing what you achieved.

Very useful to share the used components as well. I personally already own some of them, but it is always interesting hearing about other people's experiences. In fact, I think everyone should try to share as much as possible his own finding to promote all the magics which can be done with renewables, in the interest of our own planet :-) And in this sense, Dacian deserves the highest credit!

Thank you,