Portuguese Powerwall in Poland
- Thread starter Frnandu Martiński
- Start date
Frnandu Martiński
Member
- Joined
- Mar 5, 2017
- Messages
- 101
Here's what I've been up to since last update:
I decided to buy unused old stock samsung 29E to reach the 80p which means I'll have ~10kwh powerwall.
These cells costed me ~0.80 EUR/cell, so that gives ~75 EUR /1kwh, that's quite good considering it's cells with original capacity.
My laptop processing results are 40% good cells, which at around 1EUR per laptop pack gives ~37EUR/1kwh.
So basically I come to the conclusion that paying 2x more per 1kwh is worth since the laptop batteries are of unknown state-of-life, and these are theoretically new, and don't have to spend time opening the laptop packs.
I have at the moment ~5kwh from laptop cells, and 5kwh from unused old stock, which means that total cost of cells for my 10kwh is ~600 EUR.
That's quite good I think. Let's see how long they last, hoping for at least 5 years if I treat them nicely.
So moving forward with this I soldered all the cells:
prepared the bus bars:
had to buy a torch to solder the lugs:
I'm quite happy with the end result:
Then I started building some kind of holders to keep the packs in the wall.
I kind of followed Pete's design,but did itwith just wood andaluminium, so it's kind of DIY of the DIY
It's an incredible feeling to see all the 14 packs up in the wall.
When I started this endeavour a year ago and open up my first laptop packI didn't dream that I would come this far, I was thinking maybe a 1kwh powerpack of some kind for charging small electronics, but a house powerwall seemed to me like something out of my reach. Thanks to all youtube videos of Pete's/Daniel's/Joe's and others, and with the help of the forum, I slowly become confident it will be possible even for me to do it
This DIY community is awesome!
I bought a ABB shunt trip, the same as Wim and Danielso I'm hoping it will work nicely with Batrium's Watchmon (or at least Daniel will come up with a way to make it work nice, I'm counting on him).
So the current state is that I'm soldering the glass fuses:
I'm using the 1.5amp version, hoping that voltage drop at small loads per cell will not be relevant.
Still have to connect the longmons.
In next few weeks I'll have them ready to connect them up to the PIP.
I'm thinking to test each packs capacity to try and make sure they're within very close capacity to each other, but testing something like 200ah packs with an icharger 5amp discharge will take almost 2 days per pack, and I don't know if even my icharger 106b will be able to measure/show above 99,999mah.
Otherwise I'll just connect them to Batrium and see if they are somewhat close enough that will not cause to much bypass at top.
Feeling quite excited because I start to see the light at the end of tunnel, after so many hours of working on this.
Also thinking about purchase another 3 solar panels to have at least 1.5kWp. Maybe even more if winter sun hours are not enoughto charge 10kwh powerwall given 2 days of sunnywinterweather.
I'm also impatiently waiting for Daniel's new soft for the Batrium/MPP, since I want to controll PIP AC charging from raspberry ( will probably need to use custom firmware fromhttp://forums.aeva.asn.au/).
Basically I want to charge a bit the powerwall (if SOC it low)during nightly cheap tariff IF weather forecast for next day doesn't predict too much sun hours.
This way I hope to never again have to pay peak tariff from grid even when a few winter days without sun.
Also I tested emoncms for calculating KWH from power samples, and given they're 10seconds it's a bit off compared to SDM120 kwh.
So I'm thinking still on how to precisely calculate kwh for a few things: 1) PCM solar charging 2) battery IN/OUT.
Daniel says that with influx queries some kind of value is possible, but not the best way. I'll look into that.
And also the new shunt from batrium maybe it will be able to calculate kwh besides the ah for daily charge/discharge.
That used in conjunction with the SDM120 meters, would allow to have daily kWh values available for all of these:
I decided to buy unused old stock samsung 29E to reach the 80p which means I'll have ~10kwh powerwall.
These cells costed me ~0.80 EUR/cell, so that gives ~75 EUR /1kwh, that's quite good considering it's cells with original capacity.
My laptop processing results are 40% good cells, which at around 1EUR per laptop pack gives ~37EUR/1kwh.
So basically I come to the conclusion that paying 2x more per 1kwh is worth since the laptop batteries are of unknown state-of-life, and these are theoretically new, and don't have to spend time opening the laptop packs.
I have at the moment ~5kwh from laptop cells, and 5kwh from unused old stock, which means that total cost of cells for my 10kwh is ~600 EUR.
That's quite good I think. Let's see how long they last, hoping for at least 5 years if I treat them nicely.
So moving forward with this I soldered all the cells:
prepared the bus bars:
had to buy a torch to solder the lugs:
I'm quite happy with the end result:
Then I started building some kind of holders to keep the packs in the wall.
I kind of followed Pete's design,but did itwith just wood andaluminium, so it's kind of DIY of the DIY
It's an incredible feeling to see all the 14 packs up in the wall.
When I started this endeavour a year ago and open up my first laptop packI didn't dream that I would come this far, I was thinking maybe a 1kwh powerpack of some kind for charging small electronics, but a house powerwall seemed to me like something out of my reach. Thanks to all youtube videos of Pete's/Daniel's/Joe's and others, and with the help of the forum, I slowly become confident it will be possible even for me to do it
This DIY community is awesome!
I bought a ABB shunt trip, the same as Wim and Danielso I'm hoping it will work nicely with Batrium's Watchmon (or at least Daniel will come up with a way to make it work nice, I'm counting on him
).So the current state is that I'm soldering the glass fuses:
I'm using the 1.5amp version, hoping that voltage drop at small loads per cell will not be relevant.
Still have to connect the longmons.
In next few weeks I'll have them ready to connect them up to the PIP.
I'm thinking to test each packs capacity to try and make sure they're within very close capacity to each other, but testing something like 200ah packs with an icharger 5amp discharge will take almost 2 days per pack, and I don't know if even my icharger 106b will be able to measure/show above 99,999mah.
Otherwise I'll just connect them to Batrium and see if they are somewhat close enough that will not cause to much bypass at top.
Feeling quite excited because I start to see the light at the end of tunnel, after so many hours of working on this.
Also thinking about purchase another 3 solar panels to have at least 1.5kWp. Maybe even more if winter sun hours are not enoughto charge 10kwh powerwall given 2 days of sunnywinterweather.
I'm also impatiently waiting for Daniel's new soft for the Batrium/MPP, since I want to controll PIP AC charging from raspberry ( will probably need to use custom firmware fromhttp://forums.aeva.asn.au/).
Basically I want to charge a bit the powerwall (if SOC it low)during nightly cheap tariff IF weather forecast for next day doesn't predict too much sun hours.
This way I hope to never again have to pay peak tariff from grid even when a few winter days without sun.
Also I tested emoncms for calculating KWH from power samples, and given they're 10seconds it's a bit off compared to SDM120 kwh.
So I'm thinking still on how to precisely calculate kwh for a few things: 1) PCM solar charging 2) battery IN/OUT.
Daniel says that with influx queries some kind of value is possible, but not the best way. I'll look into that.
And also the new shunt from batrium maybe it will be able to calculate kwh besides the ah for daily charge/discharge.
That used in conjunction with the SDM120 meters, would allow to have daily kWh values available for all of these:
Korishan
Administrator
- Joined
- Jan 7, 2017
- Messages
- 7,538
Very nice setup. I will have to say, OCD was calmed by those pictures Very cleanly done
I do wonder, though. Did you solder the lugs on the copper busbars? There's a lot of space visible in the one pic of the closeup of the lugs; and looks like you vise pressed them instead of crimping. This could lead to problems down the road as the metals shrink/contract over time. But then I see you have shrink wrap on them already. So, I'm hoping they are really solid now
Very cleanly doneI do wonder, though. Did you solder the lugs on the copper busbars? There's a lot of space visible in the one pic of the closeup of the lugs; and looks like you vise pressed them instead of crimping. This could lead to problems down the road as the metals shrink/contract over time. But then I see you have shrink wrap on them already. So, I'm hoping they are really solid now
Frnandu Martiński
Member
- Joined
- Mar 5, 2017
- Messages
- 101
Korishan said:I do wonder, though. Did you solder the lugs on the copper busbars? There's a lot of space visible in the one pic of the closeup of the lugs; and looks like you vise pressed them instead of crimping. This could lead to problems down the road as the metals shrink/contract over time. But then I see you have shrink wrap on them already. So, I'm hoping they are really solid now
Yes, I vise pressed them, don't have a crimp yet. And then I soldered them with the torch in the picture. The closeup pic was before the soldering.
daromer
Moderator
- Joined
- Oct 8, 2016
- Messages
- 5,665
I have done several tests with the shunt and it tops out at 5-8A when tripping. Or atleast minimum that
So Im waiting answer from Batriums engineer in terms of what their most-fet output can or will handle. So when I get that I will be posting the "Ultimate way to wire your shunt to batrium" up in close collaboration with Jaron directly. Currently I have only tested it with the relay.
And next gen of Watchmon will have even better features for it.
So Im waiting answer from Batriums engineer in terms of what their most-fet output can or will handle. So when I get that I will be posting the "Ultimate way to wire your shunt to batrium" up in close collaboration with Jaron directly. Currently I have only tested it with the relay.
And next gen of Watchmon will have even better features for it.
wim
Member
- Joined
- Apr 6, 2017
- Messages
- 556
The ABB shunt trip you (we) have there works already with the watchmon... no problem using it right now until Danielis ready withthe ultimate way to hook it up...
I have, for now,made a middle connection on my 48vsetup to get the 24v neededfor theshunt.
A relay of the expansion board is selected to be disabled in case of a critical event (so the relay is activated in normal use ) because the relay's on the expansion board have only a no contacti have a secondairy(12v)relay with a nc contact to activate the shunt.
Thesecondairy relay isalso activated in normal use and drops of in case of a critical event the nc contact closes andactivates the shunt.
The ABB shunt only draws current the few ms when tripping, there is no current after or before tripping ( only during the event )
Only drawback is there are 2 relay's activated all the time in normal use...
The new Batrium software has a "Pulse" setting that makes it possible to eliminate the extra relay, but i kind of like the way itis now...the advantage is in case the batrium stops working (ie apower failure ) the relaydrops of and activate the shunt... a extra failsafe.
atleast in my head
So, until Daniel finds the perfect way to use the shunt trip i leave it this way.... far more safer then no shunt tripin the systemat all.
I have, for now,made a middle connection on my 48vsetup to get the 24v neededfor theshunt.
A relay of the expansion board is selected to be disabled in case of a critical event (so the relay is activated in normal use ) because the relay's on the expansion board have only a no contacti have a secondairy(12v)relay with a nc contact to activate the shunt.
Thesecondairy relay isalso activated in normal use and drops of in case of a critical event the nc contact closes andactivates the shunt.
The ABB shunt only draws current the few ms when tripping, there is no current after or before tripping ( only during the event )
Only drawback is there are 2 relay's activated all the time in normal use...
The new Batrium software has a "Pulse" setting that makes it possible to eliminate the extra relay, but i kind of like the way itis now...the advantage is in case the batrium stops working (ie apower failure ) the relaydrops of and activate the shunt... a extra failsafe.
atleast in my head
So, until Daniel finds the perfect way to use the shunt trip i leave it this way.... far more safer then no shunt tripin the systemat all.
Frnandu Martiński
Member
- Joined
- Mar 5, 2017
- Messages
- 101
Can you link which exactly relay you're using?
wim
Member
- Joined
- Apr 6, 2017
- Messages
- 556
eachrelay with a nc contact will do just fine...
i used one a had laying around here.
something like this is perfect http://www.benl.ebay.be/itm/MY4NJ-D...102072?hash=item3d1c051c78:g:F~EAAOSw8lpZFHcT
i used one a had laying around here.
something like this is perfect http://www.benl.ebay.be/itm/MY4NJ-D...102072?hash=item3d1c051c78:g:F~EAAOSw8lpZFHcT
daromer
Moderator
- Joined
- Oct 8, 2016
- Messages
- 5,665
You dont need any relay extra and you dont need to run it in middle either if you dont want to
You can use the built in relay on the expansion board and there is no problems what so ever running it on 48V directly. I have tested this and it runs fine for the short period as such. Though an extra resistor doesnt hurt to divide the voltage down a slight bit. That makes the connection easy.
It takes 10-14mS for it to trip and no problem there either
The way i plan for is where you dont even add any extra wires from the battery but power it from the incomming power on the board kind of But lets see about that.
You can use the built in relay on the expansion board and there is no problems what so ever running it on 48V directly. I have tested this and it runs fine for the short period as such. Though an extra resistor doesnt hurt to divide the voltage down a slight bit. That makes the connection easy.
It takes 10-14mS for it to trip and no problem there either
The way i plan for is where you dont even add any extra wires from the battery but power it from the incomming power on the board kind of
But lets see about that.Frnandu Martiński
Member
- Joined
- Mar 5, 2017
- Messages
- 101
Thanks guys for help with the shunt trip. I'll try it out once I have the packs ready.
Meanwhile I tested first 80p pack for capacity, and icharger 106b was able to discharge at around 4amp giving a result of 198ah, it took almost 48h.
I'm aware that:
1) such low amperage will probably result in higher capacity that normal day usage, i.e. loads of 1500w for 1h (water boiler)
2) This was from 4.2v to 3.0v, but to extend life, I will rather not go outside 4.1-3.3v, but it's good to know how much there is if needed for emergencies.
Given that, I tested it mainly to check pack's capacity consistency.
Testing all other 13 packs still will take almost a month. Is it worth to confirm packs equal capacity? I used all my cells above 2000mah to fill 14s80p, so there was not equal amount of cells for each 100mah intervals. That's why I'm not confident that packs capacity will be equal enought.
Anyway, let's see after a few more packs if capacity keeps close enough.
Also one problem I found is that after soldering 1 or 2 cells per packs are at 0v or close to 0v. Does this mean I applied too much heat and destroyed them?
This is a bit annoying since now I have to fill the holes left from those bad cells with >1900mah cells and will probablby result in not so much balanced capacity among packs.
Do you guys also have this happen to you?
I really try to be quick with the solder iron, but still am getting bad cells after soldering
Meanwhile I tested first 80p pack for capacity, and icharger 106b was able to discharge at around 4amp giving a result of 198ah, it took almost 48h.
I'm aware that:
1) such low amperage will probably result in higher capacity that normal day usage, i.e. loads of 1500w for 1h (water boiler)
2) This was from 4.2v to 3.0v, but to extend life, I will rather not go outside 4.1-3.3v, but it's good to know how much there is if needed for emergencies.
Given that, I tested it mainly to check pack's capacity consistency.
Testing all other 13 packs still will take almost a month. Is it worth to confirm packs equal capacity? I used all my cells above 2000mah to fill 14s80p, so there was not equal amount of cells for each 100mah intervals. That's why I'm not confident that packs capacity will be equal enought.
Anyway, let's see after a few more packs if capacity keeps close enough.
Also one problem I found is that after soldering 1 or 2 cells per packs are at 0v or close to 0v. Does this mean I applied too much heat and destroyed them?
This is a bit annoying since now I have to fill the holes left from those bad cells with >1900mah cells and will probablby result in not so much balanced capacity among packs.
Do you guys also have this happen to you?
I really try to be quick with the solder iron, but still am getting bad cells after soldering
jdeadman
Member
- Joined
- Mar 28, 2017
- Messages
- 967
In run my packs with an assortment of capacities. So I take all my cells and sort them into increments of 100mAh. Then I start taking batteries in the method of one from each pile for each cell in my pack. then go back and take another cell from each pile for all th cells. so in the end ech pack should have really close capacity.
Now from what I am hearing you are wanting to build your packs as you get batteries and that can cause some imbalance pack to pack. But that's what a BMS is for.
As for the cells going to zero you might have damaged the internal protection in the cell. try and dent in the top on one of the packs to see if it gives you voltage and then yes you got it too hot. Also a meter might not read the power if it's got lots of flux from the soldering left on the cells. try cleaning them and try measuring it again.
Now from what I am hearing you are wanting to build your packs as you get batteries and that can cause some imbalance pack to pack. But that's what a BMS is for.
As for the cells going to zero you might have damaged the internal protection in the cell. try and dent in the top on one of the packs to see if it gives you voltage and then yes you got it too hot. Also a meter might not read the power if it's got lots of flux from the soldering left on the cells. try cleaning them and try measuring it again.
Frnandu Martiński
Member
- Joined
- Mar 5, 2017
- Messages
- 101
jdeadman said:
No, actually I did it like you did: sorted all cells among 14 packs using 100mah slots.
But if you have for example 15 cells of 2000>mah <2099mah, do you use the 1 cell that remains?
I did, so tha's why I'm not too sure about balance. I couldn't afford to not use all the remaining Modulo of 14 cells.
jdeadman said:
Some of those cells do give a voltage if I pressed hard on top, which means that probably CID has popped, probably because of too much heat during soldering.
If that's the case I try to just press a bit on top (not between the holes), if it starts working good, if not I just don't use it, since using cells with a CID reset doesn't seem to me a good idea.
I do clean the flux with a brush after soldering, so that's not a problem.
shonalex
Member
- Joined
- Oct 6, 2017
- Messages
- 39
Wow, such an awesome thread. I'm subscribing and anxiously watching. You're doing exactly what I want to do. I almost have the cells "numbers" for a 14s80p system, but the testing is taking forever. It will still be a while before my cells are all tested and I'm ready to throw them into packs of equal mah sets.
Frnandu Martiński
Member
- Joined
- Mar 5, 2017
- Messages
- 101
Finally free weekend to do substantial work on the packs, here's a few pics:
I'm a bit concerned about the negative side wire, since it's bare copper wire and it's 0.55mm, so might be a risk of corrusion after some time.
After fusing, longmon mounting:
I was eager to replace my 7s15p test pack, since it only has around 700Wh.
So as soon as I finished 7 of 80p packs, I decided to put them up on the wall.
No sun today for charging
I'm a bit concerned about the negative side wire, since it's bare copper wire and it's 0.55mm, so might be a risk of corrusion after some time.
After fusing, longmon mounting:
I was eager to replace my 7s15p test pack, since it only has around 700Wh.
So as soon as I finished 7 of 80p packs, I decided to put them up on the wall.
No sun today for charging
Korishan
Administrator
- Joined
- Jan 7, 2017
- Messages
- 7,538
Your cells are looking better than they did before. However, you need to revisit the solder joints and remelt them (specifically the ones shown in the top pic). They are still slightly on the cold side. The solder shouldn't look bumpy when done. It should almost be reflective. And definitely shouldn't have a dimple where the wire goes into the pool. Are you fluxing the wire before you solder? Or how about the cells too?
Frnandu Martiński
Member
- Joined
- Mar 5, 2017
- Messages
- 101
From watching Pete's video on glass fuse spot welding I realised that it might be actually a good idea to use the glass fuse legs for the neg side, instead of bare copper (risk of corrosion).
So for my next 7 packs I soldered them on the neg side.
Now important question is if I should redo the first 7 packs (which used 0.55m bare copper wire) with the same fuse-leg wire?
I have no idea what kind of material are these fuse legs made of...
Probably is not copper, so maybe different resistance to each cell on neg side between first group of 7 packs and last group of 7 packs.
Will this make current going more to one of the groups?
I'll be watching closely.
If I notice such unbalance after some time, I'll have to redo 7 packs neg side
Connecting all of 14 packs and seeing for the first time on multimeter 53 volts was a big smile moment.
Then turning the break on and hoping nothing explodes
I've been now getting familiar with the PIP, trying to tweak Daniel's code to make it work for my needs, i.e. turning on and off AC charging during off-peak tariff and setting charging amperage.
Next I'll probably make a load test to check out how much usable kWH I can get from 4.1 to 3.3 using for example a 1kW load.
And then connect it to House breaker box, and finally start using stored energy.
Almost no sun for the last days here in Poland, so I'll be thinking about getting more panels.
I have currently 3x260 chinese sunlink poli with angle of 13degrees connected in series to PCM60x (25meter 6mm2 cable), and during rainy days I only get peak 60W, with normal level at 30-40W.
Does this seem ok, or should I expect more?
What are you guys getting for peak wattage during rainy or very cloudy days from your panels (how many?)?
So for my next 7 packs I soldered them on the neg side.
Now important question is if I should redo the first 7 packs (which used 0.55m bare copper wire) with the same fuse-leg wire?
I have no idea what kind of material are these fuse legs made of...
Probably is not copper, so maybe different resistance to each cell on neg side between first group of 7 packs and last group of 7 packs.
Will this make current going more to one of the groups?
I'll be watching closely.
If I notice such unbalance after some time, I'll have to redo 7 packs neg side
Connecting all of 14 packs and seeing for the first time on multimeter 53 volts was a big smile moment.
Then turning the break on and hoping nothing explodes
I've been now getting familiar with the PIP, trying to tweak Daniel's code to make it work for my needs, i.e. turning on and off AC charging during off-peak tariff and setting charging amperage.
Next I'll probably make a load test to check out how much usable kWH I can get from 4.1 to 3.3 using for example a 1kW load.
And then connect it to House breaker box, and finally start using stored energy.
Almost no sun for the last days here in Poland, so I'll be thinking about getting more panels.
I have currently 3x260 chinese sunlink poli with angle of 13degrees connected in series to PCM60x (25meter 6mm2 cable), and during rainy days I only get peak 60W, with normal level at 30-40W.
Does this seem ok, or should I expect more?
What are you guys getting for peak wattage during rainy or very cloudy days from your panels (how many?)?
Frnandu Martiński
Member
- Joined
- Mar 5, 2017
- Messages
- 101
Everything is finally connected (well except for the ABB shunt trip and the ATS).
But PIP transfering from battery to grid seems to work very well.
If the battery is low and there are peak usages it will without a glitch transfer to grid, and once the peak load is gone, will transfer back to battery after a few minutes.
That's perfect.
Yesterday I raised the angle of the panels to around 60 degrees, and I immediatly noticed an increase output from ~350W -> ~600W.
I'm now hoping for much better results in kWh solar gathering, even if only with these 3 panels.
But PIP transfering from battery to grid seems to work very well.
If the battery is low and there are peak usages it will without a glitch transfer to grid, and once the peak load is gone, will transfer back to battery after a few minutes.
That's perfect.
Yesterday I raised the angle of the panels to around 60 degrees, and I immediatly noticed an increase output from ~350W -> ~600W.
I'm now hoping for much better results in kWh solar gathering, even if only with these 3 panels.
