DMPPT450 to DMPPT0

[Plamen]

   Hi Dacian

Mounted in place the third string (14 PV each) of my array and before connecting them to the DMPPT450 have a few questions. The new DSSR50 is coming and since those PV will get DMPPT to it's limit it will be nice to take the load off it.

  I split the DMPPT in two parts heating and battery. Although i have purchased resistive heating elements years ago (but still not installed) i like much more the idea of analog DMPPT with matching diodes/peltier to PV. So for the moment i'll leave the heating.

  Battery charging is what i'd like to 'improve'.  At the moment most of the time i see between 35-50 amps on each of PVOUTs. I don't complain battery gets charged and even better at lower C. A few times i've manually connected the TPV channel/s that DMPPT decided to be off and the amps get over 60 or very close so was right to be off.

 If connected 1PV on ch1(TPV1), 2PVs on 2, 4PVs on 3, 8PVs on 4, 16PVs on 5 and 32PVs on 6, and say each PV is about 10A it will give 63 steps of 10A to DMPPT and up to 630A control.

 Hopefully, with your blessing i could do that by replacing the power MOSFETs on the 6 TPVs with DSSR50s and taking out the two power shunts for PVOUT1 and 2. I think its actually what you've done to SBMS120 converting to SBMS0. And if there was bigger demand you probably would've done that to DMPPT450 too.

  Let see if that's just a theory.

 

1 Could the power MOSFETs be replaced by DSSD50s and DMPPT450 just to control them?

 

2  Could the two power shunts be replaced with outside ones? Or even better just one shunt.

 

3  Could the 60A limit per TPV be removed or at least increased?

[Dacian Todea]

Plamen,

16PV PV panels and 32PV panels per input are way to much. I will say about 10 panels depending on model is max as those input are good for 100A input max.

1. No because the turn ON/OFF times are in the order of hundreds of ms for DSSR50 so no more than absolute max 4 ON/OFF per second are possible and DMPPT450 works at much faster speed.

2. Are you refereeing to the two outputs of the DMPPT450 ? Not sure what advantage will that have as most of the heat is from the mosfets so outputs are not capable of more than 60A each.  Also if you do not intend to modify the firmware shunts need to be the same resistance value and if I remember correctly they are 0.25mOhm

3. It can not be removed or current increase as mosfets will get to hot and get damaged.

But also considering absolute max current is 450A the 120A output for battery is proportionally correct since you want at least around 3:1 installed PV capacity vs charge current same as it is the case on SBMS0 with dual PV array.

The idea is to save on max battery capacity needed by over-sizing the PV array meaning a lot of PV energy will remain unused. If you have a larger battery capacity and you will preffer a large PV array size then maybe SBMS0 + DSSR50 are a better option for you as in dual PV array you can have up to a 30kW PV array and charge current up to 600A

I posted an update about the DSSR50 yesterday  https://groups.google.com/g/electrodacus/c/eeOpKlyjnag/m/sfmX0VYqBAAJ so it is in almost final form but there may be problems sourcing components. I ordered Mosfets from mouser but earliest possible delivery will be in spring next year. I can probably build a few by getting some of the mosfets from DSSR20 but it will not be much then finalize those left DSSR20 when I can get the transistors.

Many other active components are hard to source like the ideal diode controller from TI is not available and I got a few at very inflated price from China just so I can build at least a small batch earlier but that will not be much.

Same with microcontroller for SBMS0 as I only have a few microcontroller's left and then I will need to wait for Mouser to deliver the order I made in February this year and there is no estimate on when that will happen sure not this year.

I took everithing out of my website for now as stock was sold out and I will build a few more in October with the parts I have but after that there may be a longer delay until I can source components. Even Toyota needed recently to reduce production by 40% because of inability to source components so what chance do I have :)

[Plamen]

  
- DMPPT450 is with TPW1R306PL. Is it the same in DSSR20/50 too? 

- What makes DSSR50 so slow compared to DMPPT450? The transistors controller.

         Looks like it's not a good time for mistakes that will need repairs. And Toyota can wait..... :)

[Dacian Todea]

No it is not the same in DSSR20 and DSSR50 as those needed to handle also 72 cell panels.  I use the TPH2R608 for ideal diode this are 75V rated and the TPH3R70A for the switching part and this is a 100V mosfet. There are multiple of each in parallel.

The DSSR20 and DSSR50 have a reset type circuit with requires about 250ms to re-enable after voltage dropped below 8V or if remote ON/OFF is open circuit meaning you want to turn OFF there is a 250ms delay when you try to turn back ON

This under voltage protection is needed to protect the ideal diode mosfets since ideal diode controllers are powered from PV directly and if voltage drops below about 6V they will no longer work as ideal diode and as normally diodes they will fail at those currents.  The delay is needed else it can enter in osculation in case of a short on battery side and then mosfets will get damaged. With the delay they are still in oscilation but just about 4 ON/OFF cycles per second with fast disconnect they have enough time to cool down.

In the DMPPT450 there is a microcontroller that takes care in software for this conditions so normal operation can switch outputs super fast and then in case of short circuit or under-voltage it can have a longer delay.

The 250ms is absolutely no problem as delay for starting to charge the battery and should also work fine for DMPPT type diversion when you only have 2 or 3 outputs so 3 to 7 steps but when you have 32 steps like on DMPPT450 it will take to long to find the MPPT if it was as slow as 250ms about 8 seconds and that will be huge as solar output can change much faster than 8 seconds but 1 or 2 seconds to find MPPT is fine so DSSR50 can be used for that sort of setup with just 2 or 3 channels it just need a different software and controller.

The DSSR50 is designed in mind with the possibility of a DMPPT100 equivalent by combining two DSSR50 (that is what the middle connector is to parallel two DSSR50) and then another similar device with DSSR50 will be needed that can be used for simple diversion or in combination with the two DSSR50 can be used for a 3 level DMPPT100 and on top of this a controller board that controlls the 2x DSSR50 and the one diversion probably called something like SSR25x2.

I will build the diversion part but not sure I will build the DMPPT100 controller as that will have some cost to that you need to add two DSSR50 each around 90CAD and the SSR25x2 not sure how much will that be but likely around 60CAD and the controller maybe 60 to 80CAD so total for a DMPPT100 used just for heating will be around 300CAD and this is about half of what the DMPPT450 was and that could do both battery charging and diversion to heating at the same time not just heating and 4x higher power for just 2x the cost. If that was not popular I can not expect people will like the DMPPT100 for half that cost.  The DMPPT450 while it looked expensive it was highly cost optimized and a modular version like this DMPPT100 can not compete on price.

There is just not enough demand for solar PV heating.

The advantage of DMPPT100 and one of the reasons is more expensive is that it uses higher voltage mosfets (75 and 100V vs 60V) and on top of that they are passive cooled so they do not need to be installed on a heat sink like DMPPT450.  

I will like the idea to have a few DMPPT100 to heat my house and replace the DMPPT450 but I will for now stay with DMPPT450 and will see how I feel about this by the end of next year.  If I build the DMPPT100 for me then I will also offer a few (small batch) for sale to see if interest has changed over time.

  

[Plamen]

    You have the best overview of the way SBMSxxx are used for but seeing the posts here i will say that maybe 95% are RV/Boat installs and the rest or even less off grid. So for me it's the main reason for DMPPT450 not being popular. You've said that the last two years in your area are worst then the average and with the lower price and higher power PVs it's normal to think of even bigger array. I consider myself one of lucky ones to have DMPPT450. Just wish i have bought one more. So for the moment i stick with it too, not that i have other choice either at least till next year. 

  Maybe you remember that i have to remove a shorted mosfet on the ideal diodes of one of the 100A TPV. It was not priority and i didn't buy replacement and now they are all gone. I really like DMPPT450 so sorry if i'm asking nonsense things that i have no idea of. 

    LT1161CSW is controlling the PVOUT* and TLD* mosfets. Its of control of 4 mosfets on each TLD but only 2 of the PVOUT . So i asume it can control at least 2 more. 

    - Das it need to be reprogrammed to be able to control 4 mosfets on the PVOUT?

    - The passive cooling on DSSR 20/50 is possible because mosfets are different or because they are spread on bigger surface not so close to each other.

    - If external power is supplied to the DSSR 20/50 ideal diodes driver could they be used/controlled by DMPPT450?

    - When calculating the TPVx combination to supply the power requested SBMS it fast scan each of the 6 TPVs and then calculates the output needed or fast scan all 32 combinations and chooses the closest-under combination?

[Dacian Todea]

Here is a simplified diagram of the DMPPT450 as is easier to see what happens.

There are a total of 3x LT1161 and they control the 6x TLDx and 6x PVxOUT notice with green there are 3 separate mosfet switches each connecting to one of the TPVx inputs so all hardware is fully utilized.

The passive cooling is possible as much more transistors are used and there is more space. So for example on DMPPT450 for 50A there are two parallel TPW1R306PL and so 50A continues 60A peak same as DSSR50 will be

But on DSSR50 there will be 6x TPH3R70A all in parallel so each will see peak of 10A vs a peak of 30A for the ones on DMPPT450 plus the density is much higher on DMPPT450 as there are so much more transistors close to each other and all of them may be with 30A peak trough them.

While the TPW1R306 is just 1.3mOhm and 60V at 30A you have 30A * 30A * 0.0013Ohm = 1.17W for each transistor while on DSSR50 you have 3.7mOhm at 100V mosfets but there is just 10A peak trough them so 10 * 10 * 0.0037 = 0.37W thus quite a huge difference in power loss per mosfet and so no heatsink and active cooling needed but it costs more to have larger PCB area and more mosfets for same current.

The thing is that I could not have made the DMPPT450 larger since then the copper traces will have had to high of a resistance and they will heat up more than the mosfets. There is a reason you do not see 400 or 450A devices where current is transmitted trough PCB. The DMPPT450 was done at the limit of PCB capabilities with 6 layers each layer is 2Oz (75um) including the internal layers. Plus the PCB is in contact with an aluminium PCB than then is mounted to a heatsink.

When supplying current to SBMS the DMPPT450 scans fast each of the TPVx and then calculates the best combination to get as close as possible to requested current but not above and it will also need to react fast in a fraction of second if things change in the meantime as it will not want to trip the SBMS over-current protection.

The TLDx are what needs to be scanned with up to 32 steps if all TLDx are connected and so that part also need to be fairly fast.

and for comparison this the the simplified diagram of the DSSR50 as drawn directly on front panel. So there are two ideal diodes with input at PV A+ and PV B+ and then they are added together at PV AB the connector that may be used in the future to parallel two of them. And then there is a single output Batt+ the one with 6 parallel 100V mosfets and that can connect with about a 250ms delay from the time the signal is sent to RC-

This works great and it is simple but due to delay switching (needed for short circuit protection) it can not be used by the DMPPT450.

The DMPPT450 could mostly relay on software protection as it can react fast and it has as a backup the SBMS that can stop the charging independently if something goes wrong with the DMPPT450.

The DSSR50 needs to be more secure so the use of 100V mosfet and 70V TVS plus hardware protection in case of short circuit by detecting not the current but the voltage drop. 

[Plamen]

- but when you have 32 steps like on DMPPT450 it will take to long to find the MPPT 

- The TLDx are what needs to be scanned with up to 32 steps if all TLDx are connected and so that part also need to be fairly fast.

With 6 TPVs and 6 TLDs why 32 steps but not 63?

MOSFETs on PV1OUT and PV2OUT are TPW1R306PL too?

[Dacian Todea]

Because you can not exceed 100A so likely the last two or three outputs will have the same resistance value connected.

In manual I used as example 0.375Ohm as resistance for TLD1, 4 and 6 so DMPPT will just consider this as two separate levels say TLD1 and next one double of that will be TLD4 + TLD6  thus only 31 + off steps.

Yes same mosfets are used groups of two in parallel for each TPVx so there are a total of 6 mosfets on PV1OUT and another 6 mosfets for PV2OUT but when sunny with a large array only two mosfets will be connected to PV1OUT and two of them to PV2OUT and more of them will be connected only in a cloudy day or very early / late in the day before sunrise / sunset

[Plamen]

- From the value set in DMPPT/Max charge current menu of SBMSx, DMPPT450 knows what power to supply to the Batt. How it's measured, with it's own current shunts on PV1OUT and PV2OUT or takes the current from the SBMS shunts?

- SBMS120 Load+ is used and 50A load is connected there and 60A max current is set in DMPPT menu. DMPPT450 will try to supply close to 110A or just 60A so battery will charge with only 10A?

[Dacian Todea]

The DMPPT450 measures the current trough the PV1OUT and PV2OUT and that is what it uses to limit the current.

DMPPT450 will supply 60A or less if the limit is set at 60A

[Plamen]

Will that work?

TPV2 and TPV5 - One PV connected to each one.

TPV3 and TPV4 - Two PVs to each one.

TPV1 and TPV6 - Four PVs to each one.

DMPPT limit set at 40A. 14 PVs in total.

Two PVs connected to DSSR20/50 which is controlled by BPV2. Same for BPV5.

Four PVs connected to DSSR20/50 which is controlled by BPV3. Same for BPV4.

Eight PVs connected to DSSR20/50 which is controlled by BPV1. Same for BPV6.

28 PVs in total. Outputs of those DSSRs controlled by BPV1, BPV2 and BPV3 connected to PV1 on SBMS120 and BPV4, BPV5 and BPV6 to PV2. 

Only 1/3 of the power passed through DMPPT450. All PVs controlled by DSSRs are identical with those connected to DSSR450 on the same channel.

Could a PC817 optocoupler be connected to the Blue led pin so it can control the corresponding channel?

[Dacian Todea]

Not quite sure how will this help.  You are still limited to 120A on the PV1+PV2

And if some of those panels connected to DMPPT450 is shaded for some reason (can be snow or shade of some sort) then you will exceed the 120A for the SBMS120 and charging will be disconnected requiring manual re-connection.

Maybe instead of connecting the outputs of DSSR20 / DSSR50 to PV1 and PV2 you are connecting them directly to battery trough the External current shunt so same place where you connected your inverter.

This way there is no risk of over current on SBMS120 so not shunting down the charging and you can do a 1 to 1 equivalent in therms of panels so that whatever you set on DMPPT450 will be double due to DSSR20/50 to battery unless there is some shade on some of the panels then you can have a bit less than double or a bit more the double the current but if your battery can handle 200 to 300A of charge current it should be fine for those rare occasions.

If connecting in parallel with blue LED the blue LED will no longer light up as the PC817 infrared LED is just about 1.5V instead of almost 3V for blue LED.

[Plamen]

I don't want to exceed the 120A limit. Just to be able to connect all 42 PVs i've installed. With one MOSFET missing i have that TPV from 100A to 75A. And even if i add it in the future and split the array in two as you suggested it will work at 50% load so no need of fan to cool it. Battery is 400A and if i leave the limit at 40A for split in two array 80A is just fine. And there is no snow in Valencia :) so shade is even less possible.
Do i need to add some resistor for PC817 or just paralel it to the blue LED?

[Dacian Todea]

You can add a 470Ohm resistor as there will be around 3V at BPV1 directly from a 3.3V micro-controller pin and the 1.5V drop on the resistor will result in about 3mA that should be perfectly adequate for the PC817.

[Plamen]

Thanks Dacian, will see how long it takes me to do that.

[Plamen]

     Hi Dacian, The 250 ms delay for DSSR 20/50 is only when switching ON. When disconnecting there is no delay?

[Dacian Todea]

Yes but the DMPPT450 will measure current one or two ms after switch ON command was given and if it measures no current it will consider that open circuit / non connected.

[Plamen]

Sorry, I should have started a new tread. It was not DMPPT450 related.