SBMS0 / DSSR20 energy diversion to heating.

[Dacian Todea]

Just made an update about progress and some other details here http://electrodacus.com/solarforum.html?place=msg%2Felectrodacus%2F3BRwaLva3qQ%2FGv0zqrRbBwAJ

Here I just want to talk about SBMS0 and DSSR20 diversion options.

The simple option is to have a set of two DSSR20 both connected to the same max 2x 60 or 72 cell PV panels say 2x 300W total 600W of PV

Then one of the DSSR20 will connect to battery and the other to a resistive heating element matching the PV array in this example say panels are 33V max power point 9A then both in parallel will be 33V 18A and so you will have a heating element of 1.833Ohms (33V/18A).

This DSSR20 will be controlled by the same EXT IOx but the DSSR20 connected to the heating element will have the logic reversed by a transistor and a optoisolator or some similar circuit so that when battery charging is not needed (battery is full) the heating element will be ON.

The disadvantage of direct restive element is that at best this is 80% efficient (a full sunny day) and worst day may be less than 20% but in average it will be around maybe 60% depending on the type of weather you have and while not great it may be good enough for simple applications where you do not need all the available energy anyway and you only heat some water not full house heating.

The other option is a bit more complex but still simple and can give you max power point and that is by using diodes. Just got today some parts for the DSSR20 and also ordered a few full bridge rectifiers the GBJ2506-6 to be exact https://www.mouser.ca/datasheet/2/80/GBJ25005-G_Thru248362._GBJ2510-G_RevF-1481313.pdf  this is rated at 25A 600V (voltage rating is irrelevant) but I will also be ordering the GBJ5010 that is the 50A version as that can be had for $0.5USD each here https://lcsc.com/product-detail/Bridge-Rectifiers_MDD-Microdiode-Electronics-GBJ5010_C98735.html

Now inside this are 4 individual diodes and I can use just the + and - terminals to have access to two groups of two diodes in series so energy dissipation will be equal in all diodes.

I measured the voltage drop on one of this connected on a small aluminum plate and this are the results

0.1A  -  1.25V

0.5A  -  1.40V

1A     -  1.47V

2A     -  1.52V

5A     -  1.56V

10A   -  1.60V and slightly dropping as it got hotter.

20A   -  1.65V  (probably not measured due to my constant current limited supply that can provide 12A max).

So you can see that no matther the current voltage drop is fairly consistent around 1.5 to 1.6V for the two series connected diodes.

Now you will want around 20 of this connected in series for 60 cell panels as that will give you the 30 to 32V max power point voltage in winter with cold panels can drop to 18 in series if heat is needed in summer also.

The bridge rectifiers seems to be the best value as is just $10 for 20 of this and they can do 600 to 700W to support two parallel panels.

Now each of this bridge rectifiers will drop 30 to 35W and that requires good cooling likely water cooling as that way you can interface that directly with your thermal storage if that is water.

I talked about this a few years ago so it is not a new idea but then I decided for my use to have the DMPPT450 build and used.

But this thing using diodes is a good solution for the SBMS0 so that SBMS0 + DSSR20's can now have all the functionality of the DMPPT450 even if slightly differently implemented and if you read the update (first link here) you will read also about the 18kW max PV array supported by the SBMS0 and how a larger array can be used for battery charging same way as DMPPT450 even if it will be just 3 power levels.

To make sure the idea is understood I will provide an example.

PV panels are constant current sources so they will provide current directly proportional with the amount of light and say there are 3 conditions with sunny 18A provided by the two 300W 60 cell panels then 10A and 1A (cloudy).

at 18A from PV:

With diode's                  1.65V x 20 diode bridges in series  33V x 18A = 594W heat from 594W available so 100% efficient

With resistive heating   18A x 1.833Ohm = 33V and so       33V x 18A = 594W heat so same 100% efficient.

at 10A from PV:

With diode's                  1.6V x 20 diode bridge in series  32V x 10A = 320W heat from about max that so again around 100% efficient

With resistive heating   10A x 1.833Ohm = 18.33V  so     18.33V x 10A = 183W about 57% efficient only.

at 1A from PV

With diode's                  1.47V x 20 diode bridge in series 29.4V x 1A = 29.4W of heat so again around 100% efficient.

With resistive heating   1A x 1.833Ohm = 1.83V  so         1.83V x 1A  = 1.83W of heat just around 6% efficient thus almost nothing.

 

 

 

[Plamen Yordanov]

      Hi Dacian,

  So to get that efficiency You're replacing the resistance heaters with those bridge rectifiers right?. That's for the SBMS0. What about DMPPT450, will it make any difference or the efficiency there is hi enough and no need? What will be the life expectancy of those diodes compared to a resistance?

[Dacian Todea]

Yes the diodes have a fairly constant voltage drop no matther the current and this makes them naturally work at the max power point.  Each set of diodes needs to be matched with a fixed PV array. It will not work with DMPPT450 as DMPPT450 parallels internally the entire PV array that may be 200 or 300A and so on the outputs you need restive elements so current is limited by resistance and if you connect something like the diode bridge in my example that limits the voltage all those 200 to 300A available will go in to that and damaged both the diodes and possibly the DMPPT450 output.

With SBMS0 it works as you pair one of those diode heaters with just two panels so current is limited to around 20A that panels can put out and will be in the range of what the diodes with proper cooling can handle.

The Diodes will be maybe 4 to 5% more efficient in average than the DMPPT450 but that is not a significant difference.  

[Plamen Yordanov]

   That's good news for me :) because i've already bought the resistance heaters for my DMPPT.

[Julien]

Hi Dacian

How do you manage to invert the EXTIO3 logic for heating devices ? Let's say I found the right SSR to drive the diodes or resistors, how to easily control it using extio3 state ?

PS: I found a  video showing low cost chinese SSR  DD 40A disassembly with inside 15A  rated  mosfet .... (so 10A use in fact) 

[Dacian Todea]

Julien,

You just use a transistor and two optoisolator's the transistor is to invert the logic on one of the optoisolators so that way you have two separate outputs one the negative of the other so when battery charging is ON the heat element is OFF and the other way around.

The fake SSR DD are all made differently. I ordered from a few sellers a few to investigate. Some are made with mosfets and some with IGBT but they are all bad and you need a significant active cooled heatsink to use them even with 10 to 20A as the IGBT ones have about 1V drop so 10 to 20W of heat for 10 to 20A

I always got sets of two from same seller to see the consistency and almost always one is defective or semi defective so is just a gamble on what you get and that is way I was not able to recommend those as you do not know what you will get.

Mosfet current rating is not relevant at all. For example the 4 Toshiba mosfets used inside the DSSR20 are all rated at 150A two are 75V and tow 100V rated but the more important aspect is the internal resistance as that will give you the amount of heat you will need to dissipate at a certain current.

The two parallel 75V mosfets used for the ideal diode functionality are in  parallel each with typical 2.1mOhm so equivalent for the two parallel 1.05mOhm then the other two 100V are 3.1mOhm and since they are in parallel is 1.55mOhm and the two groups are in series so 1.05 + 1.55 = 2.6mOhm but to this there is also the PCB trace resistance and connectors resistance so typical DSSR20 including connectors is 3.5mOhm

Then at 20A x 0.0035Ohm = 0.07V drop x 20A = 1.4W will be dissipated as heat and that is why the DSSR20 needs no cooling as it is easy to dissipate 1.4W passively.

Even 25A is not a problem but it is already 2.18W so it gets warmer and at 30A is 3.15W already a bit much for such a small device without any airflow and maybe in high ambient temperature.

You will need a very expensive SSR to get even close to DSSR20 and they will not have the ideal diode functionality just the SSR so half of the transistors.

This for example will have the same sort of performance as DSSR20 (just the SSR half of DSSR) as it 60V rated 20A and has 3.3mOhm  https://www.digikey.ca/product-detail/en/sensata-crydom/CMX60D20/CC1666-ND/1771455

Another one will be this https://www.digikey.ca/product-detail/en/sensata-crydom/D06D100/CC1520-ND/353618 it is rated 100A and has 5mOhm but that is installed on a large heatsink with adequate cooling as 100A x 100A x 0.005Ohm = 50W so very significant.

This is the reason I needed to make the DSSR20 as there is nothing that is at that level and for this cost and the diode functionality is important so that the two parallel panels do not need individual fusing (extra cost).

So I will say stay away from those fake SSR's as they are a waste of money unless you just want to see what you get for fun.

This are maybe at similar cost with those fake SSR's and you know what you get https://www.digikey.ca/product-detail/en/ixys-integrated-circuits-division/CPC1709J/CLA280-ND/1277135  it is 9A and about 4W of heat as is 50mOhm but you know for sure that is the case and with a good heatsink you can get to 20A (20W)

[Julien]

Thanks for the explanation Dacian. I understand the principle but not yet how it is wired. But don't waste your time to explain me in details for now, I know you are busy these days :).    

I was already convinced by the DSSR20 on the panel side (with the integrated ideal diode). This product specifications are awesome, no doubt on that.

I was just looking on some alternatives  on the load side for heating device (like these cheap 25A board in // with some tweaking : https://reprap.org/forum/read.php?13,850744,850744 ) as it is just a basic function needed here but with huge current (so many DSSR20 needed eventually and it begins to be non negligible in term of cost : ( 1 DSSR20 by panel so for me a ratio of around 25€ for one 100€ of panel (270Wc) )

I think a DSSR20 well cooled on a aluminium plate should already be capable of dealing with 4 panels on the load side. But I'm wondering if it does not make problems as all 4 panels will be in // all the time and then you will lose some benefits of DSSR20 like no need of individual fuse protection and maybe more partial shadow losses . So if I guess it should be very interesting to keep 2 panels loops and not go further. 

Then a dedicated SSR20 without ideal diode part (may just be a shunt to keep same board) or even better a [ DSSR20 + SSR20 ] integrated board would be great. Then some of the control components should be common to be more cost effective (and 1 connector only on the panel side for 2 on the output side for a very easy installation ! )

Best regards,

[Dacian Todea]

Julien,

I may design a lower cost switch for the diverting to heating it will just need to be next year as I'm to busy at this time.

There will be no extra shadow losses with multiple parallel panels but they will need individual fusing and that will cost more than an extra DSSR20.

Maybe some panels accept the 3 parallel panel configuration as I recently seen one panel with 30A rated fuse and DSSR20 will deal with 3 panels if you add a very small amount of air flow like a small 5V fan or something similar.

I was thinking about the integrated DSSR20 + SSR20 but cost will get around 49CAD ~33Euro and I will lose the reverse polarity protection that DSSR20 currently has. So if PV panel is installed with reverse polarity the DSSR20 can survive but if I add the SSR20 this future will go away. Also I will need to deal with two different models as some (maybe most) will not need the diversion future.  I think I will prefer a separate unit for diversion even is slightly more than a combined unit.

[Julien]

For me, the choice between a reverse polarity PV protection (which may be useful once time if we have lost our glasses (should not be the most complicated part of the PV installation) and a useful and economic feature for everyday use, it will be obvious !! :)

I think the diversion of SBMS0 should be successful, as it is not so expensive as DMPPT450 which is dedicated to big PV area, and very easy to understand. And almost free to use. So why not at least try it on 2 panels if it is a "given" feature ?

49 CAD seems a good price for functionality and easy installation. (cabling of 2 DSSR20 with 2 panels means 4 x 6 mm² cable to deal with on one connector... Will not work so we need to do the common somewhere else many times... ) Why not but use space, more components and boring...

Maybe only the integrated one with eventually the SSR20 components not installed, as the SBMS0 wifi board option) ? It's possible too. But I think there should be room for the 2 separated models, as people should order many of them each, and the only parts not common of both are PCB. 

With a average area of 12 panels (3 kW ), SBMS0 will be almost given for free with your integrated option costs. Not bad.

As it is only 33% more expensive than one DSSR20, people may also order it right now and don't report it in a hypothetical future ( how many do I need ? I don't know so I will look later with some search of common SSR on the market to do the job for less money...)

Maybe you can make a small survey ?

[Dacian Todea]

You can always add more futures to any product and each will add extra cost.  The DSSR20 has 4 PCB's all of them will be 50% larger if I add another connector the the diversion plus the extra components of course.

Quite a few people will not care about the diversion future so I can not have just the more expensive and larger version. It will look quite ugly with a larger PCB and no extra connector installed :) and I do care a bit about design.
In the past I was thinking at the possibility of a DMPPT0 so that the diversion can take advantage of all the energy available. DMPPT0 will have been just the logic inside the DMPPT450 with external ideal diodes and SSR's so sort of the same concept of modularity as SBMS0.

There are many solution to this problem and I want to find the most cost effective an elegant one for now only DSSR20 will be available until I can decide on something and I appreciate all feedback.

[Julien]

You are right regarding the option to have only integrated PCB, it will be a big loss in term of design and cost for people who want only DSSR20 or PV reverse protection. It was just a proposal for PCB stock but I'm not at all convinced myself :).

Possibilities of different designs and solutions with a DMPPT0 seems huge and may be a very good and interesting concept too. Nevertheless, I don't think you will do concrete research with prototypes boards on this DMPPT0 before a long time. (maybe 1 or 2 year). And you may forgive this idea too after some research. 

At the opposite, a integrated DSSR20+SSR20 for SBMS0 board can offer very quickly a good solution with a very good refund of invest (Invest seems very small here) and I think this can improve the actual SBMS0 global solution a lot for a consequent part of SBMS0 customers (easier installation, more elegant solution, modularity of concept seen in one eye shot on a beautiful schematic and most important a lot more cost effective solution for every PV area size). Just a bit more of PCB pieces stock but as quantities should be quickly sold (compared to a SBMS or a DMPPT), it should be cost effective even with "small" ordering quantities in a first place. The only disadvantage I can see is a too big success of integrated board vs DSSR20 board, as you already ordered a lot of PCB for these last ones :) and may want to sell them quickly.

You may also keep the concept very simple in a first view in your presentation if you are afraid some people will see it as too complicated with more options. Once the product exist, that's just a marketing choice on how to talk about it:)

[Dacian Todea]

Julien,

I only ordered enough DSSR20 PCB's to have an average of two per SBMS0 that is already sold out so not even sure I will have enough and will likely be sold out by the end of this month when I hope to have everything that is ordered shipped. I had not build more as I want to be shipping as soon as possible since people are impatient.

I have ordered more connectors and mosfets as price is better in higher volume but those can be used no matther what design changes I decide on. Tho I will sure build another batch of DSSR20 next.

I do not have all the data but people that order the SBMS0 and similar with those that order the other models are very varied in the scope of the application but I will mention a few.  Small portable battery storage, RV, power wall, small shed and experimenting with it and not as much offgrid houses.

In most of this cases diversion is not needed as they do not have a use for the excess energy. Most people that get the SBMS do not really need it but just want to play/experiment with solar and batteries.

Offgrid as in powering a house is a very small market and 99% of those that are there already have a Lead Acid based system and Lithium seems to complex / complicated for most of them.

Most of the design is done in winter and winter is almost here :) Already had 10cm of snow two days ago.

[Julien]

Hi Dacian,

I was thinking on DSSR20 design and I have a question regarding the negative side which seems not mandatory with 2x big power connectors. A reference ground is necessary but it should be achieved by a much smaller connector connected by a smaller wire to a common ground somewhere or only one big connector for easier installation (may be important too). As 6 mm² cable should be used here, only one big connector for both input and output negative wires may be OK. It could save some space on board and 25% time for soldering these over sized connectors :). A simple effective and "easy to understand" design may be more difficult to get in this case, this is the inconvenient I see as, but it may be cost effective to think about it :)

[Dacian Todea]

:) Of course I was thinking about this but this connectors are pair of two or a single connector with 3 contacts that will be a bit inconvenient as all connections will be on the same side (inconvenient if you have multiple DSSR20 next to each other).

I have installed two DSSR20 on my test setup one with just a single 18AWG wire to GND (is a short 60cm wire) and all works as expected. The connector I have for remote ON/OFF is to small for a secure GND connection so I will not want to use that.

This connectors are 51 and 76A rated thus the large soldering terminals but I'm looking for smaller maybe 40A rated connectors with similar clamp mechanism where I can have just one for GND one for Remote (overkill) then an input, an output for battery and an output for heating element thus 5 contacts in total with 7.5mm pin pitch instead of 4 with 10mm pin pitch and get rid of the small connector that also take some time to install.  Those 5 contacts should reduce the soldering time to half compared to this 4 higher current connectors and the small 2 pin one. I will have one with 3 and one with 2 in opposite directions like now for better cable management.

But the above is just an idea. I will need to think more about it and maybe get some sample connectors to test the idea.

I was looking in to $1000+ soldering irons 250W and 300W form JBC and Hakko that may also reduce the soldering time compared to current 120W PACE I even looked at soldering robots but those seems a bit out of my range at 6000CAD+ for a china version.

I also looked in to geting back to SMD type connectors but I can not find high current SMD other than those that I used before on SBMS40 and they are a bit to large thermal mass hard to deal with and are not spring type contacts (I like this as is less of a chance for user error since contact will be always perfect).

Best regards,

Dacian.

[Julien]

It's quite surprising you care about time of soldering this small remote connector which looks so insignificant on the pictures :). Seriously, I like these small independent connector. It may be difficult to deal with 2 big 6 mm² ground panel cable and 1 (too small compared)  remote on off  "ground "cable on the same connector. (on DSSR20 or somewhere  else). Which will need redesign as remote on off is separated of ground today.

 If you buy the soldering robot, I don't want to imagine  the amortization costs of it :)

[Dacian Todea]

I just finalize 400pcs of DSSR20 and this may not be even enough for this batch (is for about 200pcs SBMS0 but some ordered quite a bit more DSSR20 that that).  I have this power connectors for another 1600pcs (price was better for this volume and they look great).

When it takes you about one hour just to solder connectors on 10 DSSR20 you are willing to consider a robot.

Even that 1500CAD for that JBC soldering iron seems huge and I will not normally consider that but if that reduces the soldering time to half just for the remaining 1600 DSSR20 is 160h and you can not work more than 6h per day doing this thus about 27 days just soldering connectors. If I can reduce that to half I can already consider that soldering iron amortized saving about 14 working days. Is like paying someone 10CAD/h  to solder those with my lower cost PACE soldering iron and none will do that work here for those sort of money.

Robots even if they look expensive initially are way cheaper than humans.  Is not the sort of work I want to do and prefer the design phase.

With SMD connectors it takes a few seconds per connector to place on the board and then all is done by the reflow oven but if they have large thermal mass is just painful to find a proper reflow profile to have those properly soldered while not damaging the more sensitive components.

I have no idea why there is not more development in high current SMD connectors that have low thermal mass. 

  

[Julien]

Hi
I didn't imagine these quantities. With small indications from your side I imagined something like 100 sbms0,300 dssr20 (before you said that it was only 2x sbms0 qties and something like 1000 connectors max once you said you have a lot of them. So not very far for boards but completely out for connectors :)

With this big connector box, I understand very well your searches for reducing soldering time at any price... Even the robot becomes not so expensive :) my remark above was only for the remote one which seems soldered quickly.

So if you don't buy a robot, I think a more powerful iron will be mandatory in a short delay before going mad !:).
Then even with a design optimization regarding soldering time, you will need to put these big connectors somewhere and not keep them... So if design change before box becomes empty, I only see for next batch using only one of them per dssr20 board and smaller connectors for ground and remote ( why not smd type here ?) But then the box will last 2* longer before another more optimized design. So I think you are going to sold these 3200 connectors keeping actual design.

Best regards,

[Dacian Todea]

Yes this connectors will be used for the current DSSR20 design even if I will come up with a different design in the meantime they will still be available at the same time.

[Julien]

Hi Dacian,

I found this about diversion heating via resistor application.

It is a fairly simple system which charge/discharge a big cap between 2 voltages around the panel  MPP and put it in our power appliance via a mosfet at each discharge cycle.

With 2 panels loops  it may good to pulse power in 36V 500/600 W resistors  with some gain above direct resistive element.

What do you think about this design ? 

http://www.thebackshed.com/forum/ViewTopic.php?FID=10&TID=11217

[Dave Festing]

Jullien,

Inspired by this article http://www.thebackshed.com/forum/ViewTopic.php?FID=10&TID=10086

davef

[Dacian Todea]

Julien, Dave,

I know about that method for quite a few years but that is not my style :)  Way to much electromagnetic noise and should be illegal.  I'm fairly sure it is illegal.

[Philippe Damour]

Hello Dacian.
Happy New Year.

Below, a diversion device with two output chanels toward water heating or ...
System is on 230V, because France do not allow current injection to thé grid in case of "Do it Yourself" installation.

Sorry in french
http://forum-photovoltaique.fr/viewtopic.php?f=18&t=38146&start=60

Kit sold almost finished for 85€

Below the assembly of the parts kit
https://forum-photovoltaique.fr/viewtopic.php?t=40512

Regards

[Dacian Todea]

Philippe,

As far as I understand that is just a device working with AC voltage and trying to keep the house Load large enough so that energy from something like a grid tie PV array is not exported to grid.  To me this seems just a way to go around the legislation but this will still be considered a DIY grid connected generator and probably still not legal.

There will still be current injection in to the grid is just that in average that is canceled by your load assuming the load can always work (temperature disconnect in case of water boiler). So the energy meter may be tricked in not registering energy exported to grid but in reality there will be pulses of energy exported to the grid at all times. 

[Philippe Damour]

Dacian
You are correct, it is a way to Lure the grid meter.
However France is about to complete the conversion of all analogue meters into the new digital LINKY meter.

This meter is much more sofisticated and detects at high frequency the power flow.

Any customer can follow on his PC or smartphone houry, daily, monthly, yearly consumption.

The related router (diverter) is taking into account the build-in electronic

Due to the strong control of the grid company, most installations are based on micro inverters connected directly to each PV pannel. Such all cabling is in 230V. Through this connection the frequency of the micro inverter is synchronised with the grid.
Such way potential disturbances due small injection to grid are negligible.

[Plamen]

http://electrodacus.com/solarforum.html?place=msg%2Felectrodacus%2F3BRwaLva3qQ%2FGv0zqrRbBwAJ - that link from the top post is no longer active. Could it be found elsewhere?

[Dacian Todea]

Not sure what was in that post. That link was when the forum was linked to my webpage and google decided not to allow that anymore.

Maybe it was this https://groups.google.com/g/electrodacus/c/ncACW0N63Sk/m/FGv0zqrRbBwAJ

[Plamen]

Thanks, mailed the discussion about diode heating to a friend that's curious and he told me about that link not working.