Grid Tie, modular and flexible

Seriously, i hesitate to publish my starting project of a smaller type Powerwall.

Seeing the perfection and/or the size of some of the builds mine will be a shame.

So I will at least start to point out the different (to what i mostly see here ) concept i have.

I come from having a totally simple 600 Wp Grid Tie solution running since 3 months, still in experiment. Taking used panels and a cheap Inverter, i have a system providing power against basic continuous household use, having a payback time below 3 years. The obvious problem is, if I increase the Solar Power, i have too much electricity during sunny day which i cannot use, during night (and muddy wheather) i have no electricity to input into my house to achieve saving.

Every standard Powerwall solution is able to fix that, with one "minor" disadvantage: the costs for the battery.
And now i have found secondlifestorage.com.

Basic idea is to feed power to the standard Solar MPPT Grid Tie Inverter at times, where solar panels don't deliver, or at any time i like.
I have a nights electricity usage of perhaps 1,5 kWh, that gives an idea about the size of my battery needed. Since i will use nothing additional eccept the inverter I have already, I am limited to 500 Watts max input. Thus, i can only supply general basic loads as refrigerators, no peak loads.

This has many advantages for the battery concept, because the max currents are not that high compared to what i have seen in other powerwall projects. Preferrable i would like max 1/10 C, short time 1/5 C discharge current, preferable less.

I can deal with all the technical problems of faking an solar like MPPT source, with the necessary measurement in the house and the regulation of the battery power feeding, using components freely available. But i want to have that all tested and brought to function, before I dive into the relative adventure of spending full costs of a SL Battery of appropriate size.

As a resource I have a small stock of some 150pcs 18650 cells, and some Lipos, which have been "collected" since awhile for the latter reason, but without enough trust to start such a project solely on my own.

I made up my concept, as a flexible and modular one as follows:
- i chose 6S /24 Volts as a working voltage, because i can deal with any packs of that with my several B6ACs very easily.
- 6s4p build a basic pack. After combining them, balancing them and checking equal capacity i have a "submodule".
- i combine 2 submodules into a "module", putting them in a small wooden "rack".
- Every Rack gets a cheap 6S balancing BMS, not mainly because of the balancing aspect, but because of the cell-exact under- and Overvoltage cut off. This gives me a 250 Wh module of 6S/24 Volts.
- these racks shall be arranged in a cabinet. All outputs in parallel.

I start with 2-3 Racks, giving me about 6 Amps for quite awhile for testing all circuitry, and of course the modular concept as well.
I can then upgrade to many more racks, in parallel, which will simply share the load. I can take out a rack at any time, even under load, for inspection or whatever. I lack an automatic surveillance of the cells, or if a rack has cut-off, i will deal with this later, if necessary at all.
I can serial 2 Racks to a 12S8P, in case I would later prefer to work with 48 Volts, without having to rework everything.

That concept leaves a lot of loose ends, where i have made no final decision until now. I am not sure about the BMS, i have no general control monitor, i am not sure about the concept at all. But it looks interesting as a start to get everything in order without having to spend too many bucks before i know everything works as planned, and especially how many control/surveillance/Info Board is necessary at all.

I know about no discussion here in this threads, but any comments are welcome.

Sorry for my bad english.

Why are you saying 6S for 24V and 12S for 48V? The traditional numbers used are 7S and 14S. Unless your equipment has a hard limit of ~25 or ~50 volts, 6S is going to be slightly under voltage for most of its operation. (3.9-4.2V is 10-20% of capacity for most Li-Ion / LiPo type batteries)

We all gotta start somewhere Welcome to the power wall building team

Cherry67 said:

I made up my concept, as a flexible and modular one as follows:
- i chose 6S /24 Volts as a working voltage, because i can deal with any packs of that with my several B6ACs very easily.
- 6s4p build a basic pack. After combining them, balancing them and checking equal capacity i have a "submodule".
- i combine 2 submodules into a "module", putting them in a small wooden "rack".
- Every Rack gets a cheap 6S balancing BMS, not mainly because of the balancing aspect, but because of the cell-exact under- and Overvoltage cut off. This gives me a 250 Wh module of 6S/24 Volts.
- these racks shall be arranged in a cabinet. All outputs in parallel.

I start with 2-3 Racks, giving me about 6 Amps for quite awhile for testing all circuitry, and of course the modular concept as well.
I can then upgrade to many more racks, in parallel, which will simply share the load. I can take out a rack at any time, even under load, for inspection or whatever. I lack an automatic surveillance of the cells, or if a rack has cut-off, i will deal with this later, if necessary at all.
I can serial 2 Racks to a 12S8P, in case I would later prefer to work with 48 Volts, without having to rework everything.

That concept leaves a lot of loose ends, where i have made no final decision until now. I am not sure about the BMS, i have no general control monitor, i am not sure about the concept at all. But it looks interesting as a start to get everything in order without having to spend too many bucks before i know everything works as planned, and especially how many control/surveillance/Info Board is necessary at all.

Sorry for my bad english.

Click to expand...

Why 6s over 7s? What are the specs on your inverter? You state later on planning on 48V. 12s is not suitable for this, but 14s is. It would be better to go with 7s4p for your 24V to start off with, then serialize those to get 48V later on. Since you don't have the 48V inverter, the 14s is preferable so something you should look into if you want to keep them modular, going with 7s over 6s if you can.
I like the idea of modularizing your packs. Makes modify, adding, removing, maintenance a lot easier to work with.
At this stage, yes, each pack would need a BMS, mostly. Depends on the connection of the packs. If you plan on Paralleling the packs only at their absolute Pos/Neg, then each pack needs a BMS in this case. But, if you are connecting each 4p set of cells together (in other words, you have 7 connections for every 2 packs connected in parallel), you would need effectively have a 6s8p (for 2 packs; or 6s16p for 4 packs), then you could use 1 BMS to handle all of them. Though, you would need to get a beefy enough one that can handle higher currents.

Keep amp usage low is really nice. You've worked on that a lot I see, especially with the 1.5kWh/night usage. That's awesome!

Your english is just fine. Probably better than most American's english

brwainer said:

Why are you saying 6S for 24V and 12S for 48V? The traditional numbers used are 7S and 14S. Unless your equipment has a hard limit of ~25 or ~50 volts, 6S is going to be slightly under voltage for most of its operation. (3.9-4.2V is 10-20% of capacity for most Li-Ion / LiPo type batteries)

Click to expand...

You are totally correct, and i am aware of all. This does just not apply to my planned hardware. My fault, i should have gven more details about that.
I'll explain together with the answer to Korishan.

Korishan said:

Why 6s over 7s? What are the specs on your inverter? You state later on planning on 48V. 12s is not suitable for this, but 14s is. It would be better to go with 7s4p for your 24V to start off with, then serialize those to get 48V later on. Since you don't have the 48V inverter, the 14s is preferable so something you should look into if you want to keep them modular, going with 7s over 6s if you can.
I like the idea of modularizing your packs. Makes modify, adding, removing, maintenance a lot easier to work with.
At this stage, yes, each pack would need a BMS, mostly. Depends on the connection of the packs. If you plan on Paralleling the packs only at their absolute Pos/Neg, then each pack needs a BMS in this case. But, if you are connecting each 4p set of cells together (in other words, you have 7 connections for every 2 packs connected in parallel), you would need effectively have a 6s8p (for 2 packs; or 6s16p for 4 packs), then you could use 1 BMS to handle all of them. Though, you would need to get a beefy enough one that can handle higher currents.

Keep amp usage low is really nice. You've worked on that a lot I see, especially with the 1.5kWh/night usage. That's awesome!

Click to expand...

My inverter is, as i mentioned, the one and only grid tie solar inverter with one MPPT input i have. No external control at all. If a battery is connected, it runs up to its (power or current-) limit, which is more than i need (in terms of Power) and less than i need (in terms of analog control).
Using this very inverter was the only choice of getting the project to go, since i do not know any cheap (!) battery operating inverter which has my requirements.
In trying to fake the battery as MPPT supplier it is obvious that this is more or less a current limiting function with a nice smoothe shape. There is little chance to get any shunt principle going (because of power loss) and so any type of dc dc inverter has to do the trick.
That means, a step up or a step down will have to provide the current shape, and thta defines the input voltage range.
6S 22 Volt is a good start for a step up giving 36 volt MPPT.
Thus, 6s is no bug, its a feature !
The 2 sub packs are completely connected, 7 rails. And one BMS.

An MPPT doesn't care about the powers of 2. It will take what ever voltage input and convert it to what ever the output needs to be. This is done using pulses through an inductor and a capacitor and diodes on the output to make sure the power is smooth and doesn't feed back.
So, going with 6s or 7s to fake the input, makes no different to the MPPT. It's just the very nature of an MPPT makes this irrelevant. Now, if you were using a device that used a transformer to step up or step down, then yes, that makes a difference. But transformers are only used with AC, not DC.

The next question is, what MPPT are you looking at getting. And what inverter do you currently have? Make/Model would help out a lot here.

Korishan said:

An MPPT doesn't care about the powers of 2. It will take what ever voltage input and convert it to what ever the output needs to be. This is done using pulses through an inductor and a capacitor and diodes on the output to make sure the power is smooth and doesn't feed back.
So, going with 6s or 7s to fake the input, makes no different to the MPPT. It's just the very nature of an MPPT makes this irrelevant. Now, if you were using a device that used a transformer to step up or step down, then yes, that makes a difference. But transformers are only used with AC, not DC.

The next question is, what MPPT are you looking at getting. And what inverter do you currently have? Make/Model would help out a lot here.

Click to expand...

I am not sure how to explain.

The Inverter is a chinese brand, Input range 22-60 V, Mppt Range 24-48V.
The typical 72 cell panel provides about 41 V unloaded. The Tracker starts to load it more and more, in 0,2A steps,a few 100ms per step, until the panel voltage starts to drop. at abot the Panels Mppt Voltage, nominal 33V for mine, the product voltage times amps (the el. power) reaches a maximun. The tracker loads even more current, afew steps, the voltage drops more, the procuct Volts Times Amps now drops, so the tracker unloads again, Voltage rises. This is an iterative process done by algorithm in the MPPT Trackers firmware. after a few cautious up-down-up-down moves the tracker locks in and stops changing the current. this stays stable until the Panels procuction changes, means Voltage rises (by same current) or Voltages drops, the Software starts again to find the V/A point of max power. Its a try and error process.

(This is done to squeeze the max power out of the panels, independently of the incoming radiation level. and with fading radiation the max power Voltage changes/drops a bit)

Now, if you simply connect a battery, the Tracker loads current endlessly higher, but the voltage never drops steep enough to reduce the V*A product, so the current finally reaches max limit of the inverter. To get the MPPT algorithm find a voltage drop, you need a "current limiting curve", like any protected power supply has. An i use a step up converter, wher i have added, by hardware, this current limiting function.

I use a step up converter for this task. The current limiter is added. The disadvantage of a step up is usually, that it can provide no voltage lower than the input. this is inherent in the function of a step up. It just feeds through then, completely out of control, the current limiter does not work anymore. If the mppt tracker would find that working point, unlucily, it would not find a voltage drop anymore, ending up in loading its maximum current.
But, with a 6s, 23 Volt, minus voltage losses, i am just out of the mppt range.... just a little bit, and the tracker will not reach current lock in at that voltage. It unloads, the voltage rises (to the level defined by the step up voltage adjustment- and just in the middle of the mppt range, say 32 volt as my panels have) ), and the tracker "seeks" there again.

I hope this got more clear now. If not, i will have to add a few current sheets. Sorry little time at moment.

I know how an MPPT works. That's not what I was asking about.

Just because it says 24V - 48V doesn't mean it's exactly those 2 settings (or 3 if it includes 36V). There's a range at each setting. 24V would be something like 22V - 29V, 48V would have a range of 42V - 58V. The range is the amount of voltage in the battery from discharged to fully charged. That's what I was asking about, not what the settings are.

Plus, you still didn't tell us which MPPT you have. You say a "chinese brand". Most are chinese brand and there loads of good ones and crap ones that are chinese brands. Calling it chinese brand doens't tell us anything. Make/Model and/or link to the product will give us loads more information. If we don't have the datasheet to look at, we can't give you solid answers or reliable suggestions on what/how to build.

If I'm building a house of "some kind of wood" and would like suggestions on support structure. This doesn't tell anything as the wood could be oak, cedar, pine, fir, or balsa. Each has a completely different structural ability and if I used balsa wood, would be better off building out of straw.
Get my point? It makes a difference as to what the device is you are building with to get the correct responses.

Korishan said:

I know how an MPPT works. That's not what I was asking about.

Just because it says 24V - 48V doesn't mean it's exactly those 2 settings (or 3 if it includes 36V). There's a range at each setting. 24V would be something like 22V - 29V, 48V would have a range of 42V - 58V. The range is the amount of voltage in the battery from discharged to fully charged. That's what I was asking about, not what the settings are.

Plus, you still didn't tell us which MPPT you have. You say a "chinese brand". Most are chinese brand and there loads of good ones and crap ones that are chinese brands. Calling it chinese brand doens't tell us anything. Make/Model and/or link to the product will give us loads more information. If we don't have the datasheet to look at, we can't give you solid answers or reliable suggestions on what/how to build.

If I'm building a house of "some kind of wood" and would like suggestions on support structure. This doesn't tell anything as the wood could be oak, cedar, pine, fir, or balsa. Each has a completely different structural ability and if I used balsa wood, would be better off building out of straw.
Get my point? It makes a difference as to what the device is you are building with to get the correct responses.

Click to expand...

Korishan, I'm sorry about the unnecessary MPPT explanation. I just didnt know what to clarify.
I agree to the range explanation you give.

The inverter i bought is here (at moment, its not the same but a lookalike) as 600 Watts
https://www.ebay.de/itm/500-600-100...60VDC-reine-Sinuswelle-mppt-SAW2/122906322941
I see it in the "crap" region , but it does properly work until now.



With every respect to you any any others, the information I gave where intended to the ones interested what i do, how i do it and why i do it. I am not fishing for advice, for the moment i know what i want and how to do it. That may change, of course .

(But, ideas of support and improvement are never rejected , or if others want to pick up the idea, thats why i am here )

If i am allowed to ask (herein this thread) then it would be the questions:
- how and where to buy used 18650 in Germany

- what cheap ( <200$) inverters ( about 500 Wp) are known to be able to work with panels and batteries and have an external control input to switch load/supply of the battery and allow to adjust the power of both from external control. That would be a replacement to what i build by my own, because i believe there is nothing of that in the market. Sure I would prefer to buy if i can.

- is there an "Batrium-alike" BMS system below, say, 200 $ (preferable less) for a full system of Hardware and Software for an 2kWh LI-Ion size. This would change my approach.

Thanks to all.

So, if I understand correctly, you are going to use a controllable DC/DC converter to throttle the power from the batteryto the Grid-Tie inverter as a way of regulatingthe power that it generates andfeeds into your housegrid? Neat idea!

watts-on said:

So, if I understand correctly, you are going to use a controllable DC/DC converter to throttle the power from the batteryto the Grid-Tie inverter as a way of regulatingthe power that it generates andfeeds into your housegrid? Neat idea!

Click to expand...

Thats dead on track.

After a detailed examination and test of the cheapo BMS samples i bought, i found that the work fine regarding cell-exact under/ overvoltage detection and Cut-out, but they are lacking balancing circuitry.
Now, my fault, its wasn't mentioned in the ads. Checking the ones having balancer showed the are really more expensive.

In trying to figure out a way to get forward with my experiments and tests the different xs/xp configuration of my modules got lots of second thoughts, how to balance them as efficient as possible, without much effort and without cancelling the advantages of the modular concept.

I found a logical looking, but very strange solution using the said BMS without balancer. Since it so much different from what i knew of systems (and what i see here being talked about), i will try to put my thoughts together and put that as a request to the specialists into the BMS Threads.

Ok, with some problems i figured out what U/I behaviour is necessary to successfully fake a MPPT behaviour to input to my grid tie inverter.

I used a cheapo (<4 Euro) chinese breakout converter of (Chinese )150 Watts.
I can adjust the current/Power simply by adjusting the external applied current limit of the converter (The converter itsself doesnt have one), in my case from 0.5 Amps up to 5 Amps on 6s, i didn't go higher, with an efficiency of 90%. That is more or less what i would expect from it, the thermal problem can be dealt with easily. I have a bigger converter in delivery, just to have a second choice.

More or less it could have been predicted how to do it, in some manner the converter is a bit more strikt (or dump) on its "MPPT" expectations as expected. Anyway, my way to an external power control of a battery connected simple grid tie is wide open now.

Last thing i need is a cheap S0 measurement to limit my grid tie input to house usage.

Real life puts away so much time, and am busy building on my first banks.

In parallel i read so much info here, i see that there is a vaste field of boards and solution which could be used in my low cost approach, and i do not know them all....

Here a shematic circuit plan of the banks and balancing Strategy i want to make. Sorry i stole the circuit plan from above and reworked it, i will soon introduce a Circuit CAD for me.





My banks are 6s8p, plan shows only 6s3p. Protecting BMS and balancing are done differently. Resistors are due to an excel shet i made, not 0,27 R but 0,63 R.
In this case i can connect battery which match in a cell voltage +-0.1 V and total Voltage +-0.1 V and still stay below about 2 Amps on any case.
This differs as well regarding the number oif Packs you have, and the number of Cells i series.

The banks will be connected to a central Balancing as follows:




Plans shows connections for 4 Banks, only one (No 2) is schown. I hope its clear.

So far, so good.

Ultra cheap building does not only mean Secondlife batterys.
It does mean for me, until now, as well secondlife PV Panels.

And now, for housing the general control, i will have a second life casing... its first life was as professional testing equipment.

Handy, neat, and holes everywhere where i do not need them....

Ill see how nice i can fit everything in I will need.

Cherry67 said:

The inverter i bought is here (at moment, its not the same but a lookalike) as 600 Watts
https://www.ebay.de/itm/500-600-100...60VDC-reine-Sinuswelle-mppt-SAW2/122906322941
I see it in the "crap" region , but it does properly work until now.

Click to expand...

That Grid Tie Unit sounds dangerous.


Because the prongs are exposed and energized coming from the inverter you risk electrocution yourself. DON'T buy these cheap type of inverters. Get one that allows you to plug in a normal power bar (non-grid tie) or a proper grid tie that connects directly to your circuit breaker and house hold wiring without any plugs. The risks are too high.

Fractions of a second can still producenasty burns or fire. And should the islanding feature fail .... :angel: .

Bubba said:

That Grid Tie Unit sounds dangerous.


Because the prongs are exposed and energized coming from the inverter you risk electrocution yourself. DON'T buy these cheap type of inverters. Get one that allows you to plug in a normal power bar (non-grid tie) or a proper grid tie that connects directly to your circuit breaker and house hold wiring without any plugs. The risks are too high.

Fractions of a second can still producenasty burns or fire. And should the islanding feature fail .... :angel: .

Click to expand...

Yes, it may sound alike, but its not (so much).

I will not electrocute me today, i tried about 45 years ago with 600 V AC when i built my first amateur radio power transmitter.
Believe me, i learned my lesson.

The converter follows german rules not to supply when it does not detect grid, i tested that. So, even if you try, you wont be successful .

This little converter seems to be better that its chinese reputation.

Cherry67 said:

Yes, it may sound alike, but its not (so much).

I will not electrocute me today, i tried about 45 years ago with 600 V AC when i built my first amateur radio power transmitter.
Believe me, i learned my lesson.

The converter follows german rules not to supply when it does not detect grid, i tested that. So, even if you try, you wont be successful .

This little converter seems to be better that its chinese reputation.

Click to expand...

That's cool. Hey do German meters spin (count) backwards? I mean if you produce more than is currently being used will the meter go in reverse?

I know that here in Canada, the units must pass tests proving the anti-ilanding but mostly that they provide pure enough sinewave and tracking.
They also are now requiring a second meter in parts of the country because of digital metering.... and money grab.

Bubba said:

That's cool. Hey do German meters spin (count) backwards? I mean if you produce more than is currently being used will the meter go in reverse?

I know that here in Canada, the units must pass tests proving the anti-ilanding but mostly that they provide pure enough sinewave and tracking.
They also are now requiring a second meter in parts of the country because of digital metering.... and money grab.

Click to expand...

Hi Bubba, yes, the old mechanical ones (Ferraris counters) do spin backwards, only a few have mechanical locks.

In future we possibly will have a rollout to electronic ones (cost free), which are common al over USA, Canada, Mexico and wherever. I know this round meters because my company is building these, and the current coils as well.

But, this rollout here has been delayed several times already, actual time schedule is in the next 2 years.
My interpretation is, they are quite content with the mechanical ones, because they are more accurate. As long as there is no trouble with the household (paying bills...) or its a low volume usage, the want to delay it obviously alap.


But, as soon as you make any sort of PV they are eager to change to an electronic one, and charge you for it....

---

A while has gone by, and i have a few stories to tell, it will be

- my harvesting style (Done)
- coming result of my harvesting campaign on real bad packs
- Imax B6AC usage
- Opus usage (new for me) (Done)
- Resistance measurement

As these themes are quite different, i should put them into the apropriate group here. If theat would be the wrong way, our friendly moderators here are welcome for any advice.

All the considerations above have been messed up by reality.

I have got Bike packs with 36 Volt which are to utterly perfect sealed, that it would be a shame/mess/hughe effort to dismount them.

And by using them "as is" i have to face alle the issues i had carefully avoided by my design.

And i WILL. Stay tuned.