Living on mostly 12v and reducing power

https://www.mcelectrical.com.au/tigo-energy-solar-panel-optimisers/

I don't think they are really appropriate for feeding into something that expects some form of regulated input, including actual battery charging.  That's not their intent.  Sure some types attempt to provide a constant voltage from the panels for loads that aren't MPPT based, but others attempt to be a better bypass than a diode (not that hard really!) to avoid restricting current through the entire array and don't make any attempt to control the panel output voltage.

Part of the reason these things exist at all is down to really bad MPPT algos in inverters and by bad I mean they don't really MPPT.  If you want an example, Dave of eevblog youtube fame did a story about his older system that suffered quite badly from minor shading,  The ultimate problem there was that the tracker in his inverter couldn't cope with Vmp reduction from the array when the bypass diodes came into effect so when the minor shading happened his inverter kept the array voltage well above Vmp and output power suffered significantly.

This wasn't a cheap junk inverter either, it's a SMA.  A firmware update would address the problem.  SMA announced the fix as a special shadefix algo rather than simply a proper MPPT algo.  <img alt=":classic_rolleyes:" data-emoticon="" height="20" src="/emoticons/rolleyes.gif" srcset="/emoticons/rolleyes@2x.png 2x" title=":classic_rolleyes:" width="20" />

Use a decent MPPT charger with a non-dumb algo that can support no-battery operation (a lot can, makeskyblue can't) and you'll be most of the way home but if you do have shading issues and they are causing problem fix them with Mr Chain Saw if possible.

The used panels I am considering are frameless 145w panels: 61v mppt with 80v open circuit voltage. 
https://sunnycalsolar.com/wp-content/uploads/2018/06/Stion-Frameless-STL-Datasheet.pdf

I have not made the purchase yet, since I am still trying to identify all the pieces I need. 😞 

The used panels I am considering are frameless 145w panels: 61v mppt with 80v open circuit voltage

There are lot  of panel over 300 watts  but few  over 60 VOC .   Sunpower make good panels .

<a href="/monthly_2021_11/561423680_Screenshot(781866).png.274e7795bf5cd2a82dbac251e0707155.png" class="ipsAttachLink ipsAttachLink_image"><img data-fileid="941" src="//forums.genetrysolar.com/applications/core/interface/js/spacer.png" data-src="/monthly_2021_11/677575125_Screenshot(781866).thumb.png.c27e61a7b154733a53b2f4c69e142e42.png" data-ratio="56.1" width="1000" class="ipsImage ipsImage_thumbnailed" alt="Screenshot (781866).png">

<a href="/monthly_2021_11/40595092_Screenshot(781867).png.0004831dfa3f4e4677d38c7ebeda1c6f.png" class="ipsAttachLink ipsAttachLink_image"><img data-fileid="942" src="//forums.genetrysolar.com/applications/core/interface/js/spacer.png" data-src="/monthly_2021_11/1498100222_Screenshot(781867).thumb.png.5aa51bfcd6e67ff9aaa5959be59b2b21.png" data-ratio="56.1" width="1000" class="ipsImage ipsImage_thumbnailed" alt="Screenshot (781867).png">

Well done with your efforts!

The more I read on this topic of Grid Tie & a main Inverter, the clearer it becomes that the main inverter needs a big battery to sync the power from the Grid Ties!

The only other options seem to be: 1) shut off Grid Ties; 2) add some heat & fan & incandescent light bulbs to sink excess Grid Tie power; 3) shut off the inverter; 4) allow main inverter to feed electrical grid: 5) allow main inverter to blow up 

A simple synchronous buck-boost converter on some panel in parallel to high DC voltage to a gateway feeding a main inverter seems like an elegant design, but finding the components for 80Voc panels with 61Vmp seems nearly impossible. Almost like I have to design it, myself. (If I owned the panels already, I would likely have started designing.)

The racking is looking like a massive cost, as well, since I don’t want to pierce my roof & don’t have the land to build a whole solar structure far from my home.

Batteries drop every day, it seems. I have some favorites, which I may execute on in the next couple days, not only because of cost, but because of massive current output. The problem is BMS limited output current, which brings me to maybe just using a balancer above 50v on previously matched & balanced cells.

To use GS without UL / ETL, when not officially off-grid, I guess I basically need an external auto transfer switch? Or maybe off-grid parallel circuits around the house?

yes, the 145W panels @80Voc are smaller & odd, but I can get them with mounting hardware for $40 each, free shipping, but wow have they cost me time in figuring out how to use them! LOL!

 I can get them with mounting hardware for $40 each, free shipping, but wow have they cost me time in figuring out how to use them! LOL!

I  started with  2 solar panels  to  4 battery to 48v powerjack to find  the cost .   The cost for  wires  and  MPPT  solar controller  and  fuse and breaker box  is way more than I  first  design  the system .   Once I  started  I  find that the cost  keep going up  with  building  the system  big and  bigger  and  now no end in  sight  .   I   can only at the most run offgrid  for 16 hours  in the winter  and 6 hours in the summer .   It is  almost  impossible  to plan  when  DIY  with  used parts .  

Well, it seems I managed to acquire a Delta H6 as a grid tie inverter, it is now hanging on my wall… ready for adding panels & battery & a whole house inverter!

The source on the 145w frameless panels are gone. Now, need to search for more panel options, for a 330w panel that looks interesting, now transport is the problem, plus they are more expensive.

Looking at a pair of lightly used EV car batteries… I ran the numbers and a GS12k will charge them to about 4v a piece. At ~70% capacity, these should last a lot of cycles!

https://batteryhookup.com/products/l-lg-28-8v-7-9kwh-8s-lithium-ion-ev-module

16S @4v should be 64v, which is the top end of a GS12K

if the 12K is ready for preorder or order, I may execute on these batteries…

Looking at a pair of lightly used EV car batteries… I ran the numbers and a GS12k will charge them to about 4v a piece. At ~70% capacity, these should last a lot of cycles!

The LG  is  33 v from battery hookup  and you need 2  in series  for  the 48 v GS 12 kw .   My  ev car battery  is charge  to max 66 v  with MSB  60amp charge controller  120 dollars  and  will probably last 20 years or more  but  if charge to 67.2 v  SOC 100 percent then maybe last only 10 years .    I  connect (total 2 chargers )  one   30 volt e-bike charger   to each of the LG battery when the two  LG battery is connected is series  with the  MSB  charging at  66 v  at the same time  ( very important  since I  do not use a BMS  with battery  in  series ) .   

2 hours ago, DavidH said:

Well, it seems I managed to acquire a Delta H6 as a grid tie inverter, it is now hanging on my wall… ready for adding panels & battery & a whole house inverter!

Does said inverter support linear frequency-shift throttling?  I can find practically nothing on it online.

On 5/6/2022 at 4:31 PM, Sid Genetry Solar said:

Does said inverter support linear frequency-shift throttling?  I can find practically nothing on it online.

Hey Sid,

I don’t know to what extent it supports frequency shifting throttling. I read on-line that it will throttle, but that is all I practically know.

I executed on the 8.5k battery modules, bought 4x of them, so it is now official… you got me moving from my 12v systems!

Those batteries are nearly 90lbs a piece! LOL! I need to mount them to the top of a big strong moving dolly! ROTFL!

I have mostly backward engineered the battery connectors, with cheesy breadboard wires, now I will need to formalize the connections to go to a BMS.

I am getting ready to preorder a GS 12kx but I will need 2x BMS’s… I am looking at a pair of JK-BMS’s, rated at 200Ah with 350Ah peak. They should supply all the power a GS 12k will ever want during a peak!

Any thoughts on a serial bus wires or hardwired Ethernet, or CAN bus so I won’t have to communicate over WiFi?

The JK BMS’s seem to offer CAN bus, but I never built anything to communicate over CAN bus. Seems like it would be useful for inter-UPS communications 

Thanks, Dave

19 hours ago, DavidH said:

Any thoughts on a serial bus wires or hardwired Ethernet, or CAN bus so I won’t have to communicate over WiFi?

The preorder inverters (both 6kw and 12kw) will support hardwired RS-232 / RS-485 or CANbus. (I will have to make Sean aware of this...you get to pick which one you want before they ship...it changes the isolation chip!)  All use the same basic RS-232 premise on the WiFi board--so it doesn't make a difference from the WiFi board's hardware perspective.

I don't have any firmware or protocols written at the moment for the physical comm stuff--but that won't be too difficult to sort out, and firmware updates can easily solve that.  Thinking it will need to be a multi-choice option list of comm methods in the WiFi board, including MODBUS.  Not sure what else at the moment.

19 hours ago, DavidH said:

I am getting ready to preorder a GS 12kx but I will need 2x BMS’s… I am looking at a pair of JK-BMS’s, rated at 200Ah with 350Ah peak. They should supply all the power a GS 12k will ever want during a peak!

Ha, nice try.  Current limit in the GS12 is going to be between 1,000 and 1,200A.  Yay!

On 7/21/2022 at 3:44 PM, Sid Genetry Solar said:

The preorder inverters (both 6kw and 12kw) will support hardwired RS-232 / RS-485 or CANbus. (I will have to make Sean aware of this...you get to pick which one you want before they ship...it changes the isolation chip!)

…

I don't have any firmware or protocols written at the moment for the physical comm stuff--but that won't be too difficult to sort out, and firmware updates can easily solve that.  Thinking it will need to be a multi-choice option list of comm methods in the WiFi board, including MODBUS.  Not sure what else at the moment.

 

Ha, nice try.  Current limit in the GS12 is going to be between 1,000 and 1,200A.  Yay!

Ok - I just read that RS485 is multidrop, with up to 32 devices on a drop, and I would have to order a special JK BMS for CAN which will not support a display, so I am going RS485 on everything!

I would love to have the 1x GS12k, 2x JK BMS, 2x BMS displays, 1x grid-tie inverter, and 1x Raspberry Pi on the same twisted pair rs485 bus… so I can see it all with a single dashboard.

Future firmware will be interesting. Figuring out how all of these things can co-exist on the same twisted pair set should be interesting!

2 minutes ago, DavidH said:

Future firmware will be interesting. Figuring out how all of these things can co-exist on the same twisted pair set should be interesting!

Best part?  If there's issues, I can easily fix them / add features, etc. as necessary via firmware update!  The port will be electrically isolated; comm speed, ID, etc. will all be customizable.

Worth noting that RS-485 is just a transport method, it does not define the communication protocol.  I am assuming/presuming that the actual comm protocol will be MODBUS...?

7 minutes ago, Sid Genetry Solar said:

Best part?  If there's issues, I can easily fix them / add features, etc. as necessary via firmware update!

…

Worth noting that RS-485 is just a transport method, it does not define the communication protocol.  I am assuming/presuming that the actual comm protocol will be MODBUS...?

Hi Sid,

I have done work in RS232 & Ethernet (UDP & TCP) but never RS485 or CAN bus.

With RS485 limited to 30 on a multidrop, MODBUS limited to 247, I am not sure how they interact.

Are the RS232 drivers written for getting health information out of the GS? (i.e. in/out V, A, W… temperature, uptime, etc.)

if so, maybe it is just best for me to go RS232, out the gate, if I want this data?

20 minutes ago, DavidH said:

Hi Sid,

I have done work in RS232 & Ethernet (UDP & TCP) but never RS485 or CAN bus.

With RS485 limited to 30 on a multidrop, MODBUS limited to 247, I am not sure how they interact.

Are the RS232 drivers written for getting health information out of the GS? (i.e. in/out V, A, W… temperature, uptime, etc.)

if so, maybe it is just best for me to go RS232, out the gate, if I want this data?

OK, from the beginning....

RS232, RS485, CANbus, Ethernet, etc., etc. are data transport methodologies, equivalent to a hose or a pipe.

They do not specify the underlying protocol used to communicate data (i.e. the actual stuff being transferred)

MODBUS is commonly used as a protocol on RS232, RS485, and even Ethernet.  (Consider it like orange juice)

UDP & TCP are protocol layers that generally go on top of Ethernet.  Or WiFi, etc., etc.  (Maybe Gatorade or something.)

In other words, you can put almost any juice (or even motor oil) down the transport method (i.e. the hose or pipe).  The hose will transfer the material from point A to point B...but doesn't determine what material is being transferred.  It just determines how material is transferred.

RS-485 as a transport methodology is a half-duplex, differential signal method.  It's very robust, and supports multiple devices on one bus (RS-232 doesn't).

But simply because something uses RS-485 is no guarantee they all speak the same language on RS-485.  You would have to ensure that all of the devices on the RS-485 bus speak the same "language" (and run at the same baud rate) so they don't cause bus collisions.  If all of the devices support MODBUS, then that's perfect--as each of them could have a separate MODBUS device ID, and the "bus master" simply requests data from each one in turn (by ID).

...but that's highly dependent on the specifications of each and every device that you'd like to have share the RS-485 bus.