diy solar

diy solar

LFP for EV's projected to be less than $56 per kWh within 6 months

Checkthisout said:
Skinny girls never do much for you anyway.

Ask me how I know? Well, I'm running 24 volts wishing it was 48.🥺

Also, trying not to buy 16 cells from 18650 right now....
Click to expand...
To be honest I see more options for 24V charge controllers etc (e.g. Victron solar charge controllers, mini inverters, ...).

Expanding is potentially cheaper on 24V due to only needing to add 8 cells at a time, but it's true that *if* the majority of your cost goes into the cells, then 48V is *not* a significant cost increase for the other components (breakers, fuse, etc), therefore basically you only pay for the extra cells.

And depending how you look at it, thinner cables (if you compare same level of power outputs).

But yeah, *if* I'd have to run my 48VDC to my house from the garage, I'd probably run some 120mm2-240mm2 of Aluminum Cable or something like that.
 
silverstone said:
To be honest I see more options for 24V charge controllers etc (e.g. Victron solar charge controllers, mini inverters, ...).

Expanding is potentially cheaper on 24V due to only needing to add 8 cells at a time, but it's true that *if* the majority of your cost goes into the cells, then 48V is *not* a significant cost increase for the other components (breakers, fuse, etc), therefore basically you only pay for the extra cells.

And depending how you look at it, thinner cables (if you compare same level of power outputs).

But yeah, *if* I'd have to run my 48VDC to my house from the garage, I'd probably run some 120mm2-240mm2 of Aluminum Cable or something like that.
Click to expand...

It's better cost-wise without question.

So once voltage is higher like 96 volts, any work you do on your battery pack is now "live" work and you need arc flash gear and HV gloves and tools etc.
 
Checkthisout said:
It's better cost-wise without question.

So once voltage is higher like 96 volts, any work you do on your battery pack is now "live" work and you need arc flash gear and HV gloves and tools etc.
Click to expand...
Well it's technically still low voltage (< 1000VAC/1500VDC) for all standards and purposes. Even ELV (< 120 VDC).

But yeah, for Touch Protection ABSOLUTELY WEAR GLOVES.

Concerning Arc Flash I partially disagree: 48 VDC in itself is already quite dangerous and should be stated as such.

By the way, some manufacturers of battery cells rate(d?) the short-circuit current of their battery on a sample test of say 1 to 4 cells in series. Already with 16 cells that short-circuit current is much higher (because all of a sudden the cabling between cells, chargers, inverters, becomes [in proportion] smaller).

So, while the short-circuit current might increase a bit (not much) compared to 48VDC systems, the real danger is the arc voltage and energy which of course will go up ...
 
Checkthisout said:
It's better cost-wise without question.

So once voltage is higher like 96 volts, any work you do on your battery pack is now "live" work and you need arc flash gear and HV gloves and tools etc.
Click to expand...
Why even bother with 96 volts?
Wouldn't it be more efficient to jump up to 80s (LiFePo4) or higher?
You could fit 105ah cells 40 wide on a 5' shelf, 12" deep to fit 2 rows for 80s. ~26kWh per pack.
 
42OhmsPA said:
Why even bother with 96 volts?
Wouldn't it be more efficient to jump up to 80s (LiFePo4) or higher?
You could fit 105ah cells 40 wide on a 5' shelf, 12" deep to fit 2 rows for 80s. ~26kWh per pack.
Click to expand...
Isolation, BMS with 80 channels (!), Availability of Breakers, etc ?

Not sure what's actual nowadays, but when I worked on an industrial battery I think off-the-shelf BMS solution was like 16s-32s.

Then of course nothing prevents you from building 1 battery module of 16s-32s and put it in series with other modules.

Only problem being (also in your 80s example) is: one faulty cell (or module) and maintenance becomes a nightmare.

It's not as easy as just turning off the faulty module, you lose the complete battery in such instance. You have no redundancy. Also probably balancing will become an issue (unless you have several balancers working at the same time ... think 2A balancer nowadays is already struggling a day or so to top balance a 16s battery, with 80s you'd have to take it off grid for a week to get it properly top balanced ...).

And, if you want to keep FET-based BMS, you'd need at least a 600 V rated MOSFET. These are not very efficient (not really a problem since you don't switche them on/off a lot), but the on-conduction losses are defitively MUCH more than the 60V (?) MOSFETs used in JK BMS for 16s for instance. Thus cooling becomes a problem.

Just a few examples. I'm sure there's plenty more.
 
PV arrays are commonly strung to 400-500 volts instead of the common per panel voltage of 50 volts. This is because PV wires are long and the cost for heavy gauge wire would be excessive. This is the same issue with batteries, we just accept it due to short cables. But it sure would be nice in some cases to put the batteries somewhere else than the inverter and that's just not feasible at 48 volts.

EVs are operating 400 to 800 volts even though the wiring would be very short. It doesn't take much to make higher voltage a winning idea.

We already have to deal with lots of voltages above 60 volts. Grid, load, PV. A 450 V PV array is a serious safety issue. If we can handle that, then the battery would be similar.

In some ways, having exposed 48V battery wiring engenders a certain comfort with exposed conductors that is not all that safe even at those low voltages (witness Will arcing his Ruixu stack). If everything was actually insulated and treated as high voltage, that might lead to a more uniform treatment. For example, someone inside an inverter touches what they think is a battery lug and it is AC. If everything is dangerous, then there isn't this "safe" over here, "unsafe" over there division that can get confusing to some.

High voltage makes things more efficient inside the inverter, too. Capacitors would be smaller, conductors smaller, MOSFETs smaller, etc. This applies to the AC inverter, but also to the PV to battery circuits, a smaller voltage change.

I'd skip 100 volts and go to 200 volts, 64s cells. That is 1/4 the amps in the wire. For the same losses, the wire can be 16 times smaller. You could wire this with ordinary house wiring like you use for AC.

192 V is a thing for solar, so not entirely new territory:

buy 192V 100AH 200AH 300AH Lithium Ion Battery High Voltage Lifepo4 Battery Pack 50AH 500AH,192V 100AH 200AH 300AH Lithium Ion Battery High Voltage Lifepo4 Battery Pack 50AH 500AH suppliers,manufacturers,factories

We are best 192V 100AH 200AH 300AH Lithium Ion Battery High Voltage Lifepo4 Battery Pack 50AH 500AH suppliers,we supply best 192V lithium battery for sale.
www.greensunpv.com www.greensunpv.com

Mike C.
 
42OhmsPA said:
Why even bother with 96 volts?
Wouldn't it be more efficient to jump up to 80s (LiFePo4) or higher?
You could fit 105ah cells 40 wide on a 5' shelf, 12" deep to fit 2 rows for 80s. ~26kWh per pack.
Click to expand...

A lightning strike to the array would only show up as a slight amperage increase into the batteries.😆
 
42OhmsPA said:
Why even bother with 96 volts?
Wouldn't it be more efficient to jump up to 80s (LiFePo4)
Click to expand...
Why 250Vdc, let go up to 108S like in EV with LFP cells, so 350Vdc nominal.

By the way, 48V seem like the hard stop for ''safe'' to touch with hands because anything above 60V could kill you.
So, yeah, it's 48V and the next step is 400V or even 800V like EVs. Nothing between, there is no reason.
 
mciholas said:
PV arrays are commonly strung to 400-500 volts instead of the common per panel voltage of 50 volts. This is because PV wires are long and the cost for heavy gauge wire would be excessive. This is the same issue with batteries, we just accept it due to short cables. But it sure would be nice in some cases to put the batteries somewhere else than the inverter and that's just not feasible at 48 volts.

EVs are operating 400 to 800 volts even though the wiring would be very short. It doesn't take much to make higher voltage a winning idea.

We already have to deal with lots of voltages above 60 volts. Grid, load, PV. A 450 V PV array is a serious safety issue. If we can handle that, then the battery would be similar.

In some ways, having exposed 48V battery wiring engenders a certain comfort with exposed conductors that is not all that safe even at those low voltages (witness Will arcing his Ruixu stack). If everything was actually insulated and treated as high voltage, that might lead to a more uniform treatment. For example, someone inside an inverter touches what they think is a battery lug and it is AC. If everything is dangerous, then there isn't this "safe" over here, "unsafe" over there division that can get confusing to some.

High voltage makes things more efficient inside the inverter, too. Capacitors would be smaller, conductors smaller, MOSFETs smaller, etc. This applies to the AC inverter, but also to the PV to battery circuits, a smaller voltage change.

I'd skip 100 volts and go to 200 volts, 64s cells. That is 1/4 the amps in the wire. For the same losses, the wire can be 16 times smaller. You could wire this with ordinary house wiring like you use for AC.

192 V is a thing for solar, so not entirely new territory:

buy 192V 100AH 200AH 300AH Lithium Ion Battery High Voltage Lifepo4 Battery Pack 50AH 500AH,192V 100AH 200AH 300AH Lithium Ion Battery High Voltage Lifepo4 Battery Pack 50AH 500AH suppliers,manufacturers,factories

We are best 192V 100AH 200AH 300AH Lithium Ion Battery High Voltage Lifepo4 Battery Pack 50AH 500AH suppliers,we supply best 192V lithium battery for sale.
www.greensunpv.com www.greensunpv.com

Mike C.
Click to expand...
Agree with the first part of your post and 48V making us potentially complacent. Arcing (also touch voltage ... heck I could even feel 15VDC in a lab once) is for sure an issue.

The 400V-800V of EVs are also dictated by the motors power vs speed and voltage vs speed characteristic. You don't want a 48V battery and double boost converter or isolated transformer if you can save it with so little room to play with.

Higher voltage. You don't plan to go above 1500 VDC, are you :) ?

But remember the first principle of Power Electronics (or Economics for that matter): Nothing is Free, everything is a Compromise. If you increase voltage you also increase components voltage ratings, less availability, more conduction (and switching losses), particularly for the BMS.

Capacitors are not really space efficient above a certain point (IIRC ~ 250-300 VDC for electrolytic caps). And of course ESR increases, thus you are limited by power dissipation also.
 
yabert said:
Why 250Vdc, let go up to 108S like in EV with LFP cells, so 350Vdc nominal.

By the way, 48V seem like the hard stop for ''safe'' to touch with hands because anything above 60V could kill you.
So, yeah, it's 48V and the next step is 400V or even 800V like EVs. Nothing between, there is no reason.
Click to expand...
48 V (not sure if DC) can also kill you. I'm pretty sure I saw an IEC standard with touch voltage vs exposure time.

Granted in salty environment and with wet hands, cannot remember if touch voltage or step voltage though.

If you touch for long enough it can definitively hurt or worse ...
 
silverstone said:
Just a few examples. I'm sure there's plenty more.
Click to expand...
All the issues you are discussing are being dealt with in the EV space. There are ways to BMS high voltage packs. There are ways to service them (isolate sections, like RSD in panels). The semiconductor industry is turning out EV electronics at a furious pace, so solar can ride those coattails. Lots of development there, for example:

400V & 800V high-voltage battery system design resources | TI.com

View the TI 400V & 800V high-voltage battery system block diagram, product recommendations, reference designs and start designing.
www.ti.com

Here is a 16 cell stackable BMS chip for EV high voltage packs:


You build the pack with 16s sections that all communicate to a central BMS controller.

silverstone said:
And, if you want to keep FET-based BMS, you'd need at least a 600 V rated MOSFET. These are not very efficient (not really a problem since you don't switche them on/off a lot), but the on-conduction losses are defitively MUCH more than the 60V (?) MOSFETs used in JK BMS for 16s for instance. Thus cooling becomes a problem.
Click to expand...
No, it really doesn't since the current is 1/4 at 200 V versus at 50 V. The MOSFET can be 16 times higher resistance and develop the same heat for the same power delivery.

Thermal problems scale as the square of current and linearly in voltage, so higher voltage is more efficient.

Mike C.
 
silverstone said:
48 V (not sure if DC) can also kill you. I'm pretty sure I saw an IEC standard with touch voltage vs exposure time.

Granted in salty environment and with wet hands, cannot remember if touch voltage or step voltage though.

If you touch for long enough it can definitively hurt or worse ...
Click to expand...
Sure, I fully agree.
Wet hands by sweat can definitely cause problem.
I would be better to write: 48V seem like the hard stop for ''safe'' to touch with dry hands ;)
 
silverstone said:
But remember the first principle of Power Electronics (or Economics for that matter): Nothing is Free, everything is a Compromise. If you increase voltage you also increase components voltage ratings, less availability, more conduction (and switching losses), particularly for the BMS.
Click to expand...
BMS is less losses with higher voltage because the current dropped.

silverstone said:
Capacitors are not really space efficient above a certain point
Click to expand...
Higher voltage is a BIG win for capacitors because the size of the cap scales linearly with voltage but the stored energy scales as the voltage squared. Double the voltage, you get half the capacitance, but twice the energy in the same size.

Inverters would be a lot smaller, lighter, and more efficient if they had 200 V DC instead of 50 V to work with. Then they would be cheaper, too, and probably have less idle usage.

Mike C.
 
mciholas said:
All the issues you are discussing are being dealt with in the EV space. There are ways to BMS high voltage packs. There are ways to service them (isolate sections, like RSD in panels). The semiconductor industry is turning out EV electronics at a furious pace, so solar can ride those coattails. Lots of development there, for example:

400V & 800V high-voltage battery system design resources | TI.com

View the TI 400V & 800V high-voltage battery system block diagram, product recommendations, reference designs and start designing.
www.ti.com

Here is a 16 cell stackable BMS chip for EV high voltage packs:


You build the pack with 16s sections that all communicate to a central BMS controller.


No, it really doesn't since the current is 1/4 at 200 V versus at 50 V. The MOSFET can be 16 times higher resistance and develop the same heat for the same power delivery.

Thermal problems scale as the square of current and linearly in voltage, so higher voltage is more efficient.

Mike C.
Click to expand...
That's also what I was saying with the "Battery Module" idea of 16s-32s each.

About the FETs ... Well, if resistance vs blocking voltage was linear, yeah ... Not much linear with semiconductor devices.

Quick comparison: 60V, ~40A, 10mOhm
https://www.digikey.dk/en/products/filter/transistors/fets-mosfets/single-fets-mosfets/278?s=N4IgjCBcoGwAwA4qgMZQGYEMA2BnApgDQgD2UA2iAKxwDsCMMIxATAMy21UAszI7nbmz604bAJxwIxbtyoxxVPuwTja4kS3pUlrbnBhtd-CWt56w8JsRphuDZd0tg4jy3HP8nccda8u1Pll5X0cWJ2NwsAQ2YT02GCoNeMTPcPsEFkd4Flc9GBjpEFoYdSoi8IK2LJkqBCkarxKxZXkYaNbGXKCEBG9WqgSiuXrw1pi2NKTucSm%2B3uV4KhYK9qdPGCdNxpYYLVVFrXFVjkDWTZokc%2BWwHcYwHmVaIav%2BZ-bk4oKHvLfuC6eg1oOy4HDibyBaRKuWMwVi4K031oT20ry02k%2B6IYv32tFkT3Ekwq9TEfhY9T2yNYCFE7WUNIMOwZdXpklurKkkVU8gR3KSyjZkwFYGimOO0VhdWeOPEuRarFllwFbBFkUJjDR4kUkgF2uGNJKCK1Ur4bDgHkQpvNhNeZrsZqt0T2VpY5V%2BZtdBitbBixjNQh9fAKNE%2BwZcQYQNF%2BtI8Ige8Dj6iKwLg-OIwJ9fk4u2TwMMQa1hKpxT6NBEvnZxBi9GL1YWVY4ByrNG4xYUrdbfGxdj4klJvYeJQHs0asx4nkUM2Mil6RRc5tdfHncF2S-NUleLionGLLj2luIe%2BDa4KCDndB8nxFqc4S639CDW2qd8vxhFIqcd7sdMPH8ev9kQNiE2SxYjvewVTvS4-BFM9w1-XowMPfQWC0JcUPJdCVxYK8UIdZCV0ggjULnJwHlwh52HQ3IVnQ7cuHQwQ-BAwxjBKHh8PAexW13Ght1%2BB5YjkJceAMIoWKA4pdneJd4BVTxgTAbNZPYeVwEMAxfnsThGBEIZK3Uy5PhoehNlk2lPGiFYvUPTIXFebS8V3SM5DY-Z-xABQrDnRRnk3cQYlKR9LCECN-nqZQOC4ZNtITVgosGEAAF1iAABwAFygEAAGV0oAJwASwAOwAcxAABfVgDCQaAQDQSAsDwIhSAoTy4AAAgANSDKguojPr0w67r0164aQHqPqUpADKsoAVSKgr0oAeXQABZfBMFwABXPL8Aq4gAFoslq%2Br8q25qyEgSglCS8rKs85AQAKgATLKDpcIoZsgEA%2BHSgBPVK9u%2Bza0DuoA

Quick comparison: 600V, 10-12A, 280mOhm-750mOhm

Cost can be considered similar.

But the higher voltage will result in ~1x-3x MORE losses.

Please note that this was just a quick search, by no means exhaustive :) .
 
Last edited:
mciholas said:
BMS is less losses with higher voltage because the current dropped.


Higher voltage is a BIG win for capacitors because the size of the cap scales linearly with voltage but the stored energy scales as the voltage squared. Double the voltage, you get half the capacitance, but twice the energy in the same size.

Inverters would be a lot smaller, lighter, and more efficient if they had 200 V DC instead of 50 V to work with. Then they would be cheaper, too, and probably have less idle usage.

Mike C.
Click to expand...
I don't think so (for capacitors). Cannot speak about your argument (energy !), but for use in an inverter you typically have to consider:
- Current Ripple in your inverter determines RMS current in the capacitor
- Capacitor RMS current ^ 2 x ESR (plus dielectric losses) determines your losses and your heating - You will have a maximum value stated in the datasheet
- You might also need to consider voltage ripple (yes, that's tied to energy too, since it will be around the "nominal" value)

For inverters: you get less current, but your filtering becomes more important, especially inductors. Thus more inductors (for current smoothing) and/or capacitors are needed. You don't want to have power fluctuations neither excessive ripple (for your battery).

Current ripple is essentially (dependent upon topology) proportional to DC-link voltage and inversely proportional to Inductance Value and Switching Frequency.

Voltage ripple is essentially (dependent upon topology) proportional to DC-link voltage and inversely proportional to Inductance Value and Switching Frequency and Capacitance Value.

So, again, nothing is free. Want smaller filter? Increase switching frequency and losses. Want less ripple? increase inductor size. Want less voltage ripple? Increase your capacitor bank.
 
silverstone said:
Quick comparison: 60V, ~40A, 10mOhm

Quick comparison: 600V, 10-12A, 280mOhm-750mOhm
Click to expand...
10 times the voltage rating means 100 times less current losses for any given resistance.

So the 280 to 750 mOhm Rds is actually LESS power loss than the 10 mOhm part when you up the voltage 10 times and reduce the current 10 times.

You also happened to pick a poor MOSFET. Consider:


600 V, 100 mOhm, 30 A.

That part is 10 times the resistance but that will result in 10 times less heat than the 10 mOhm part delivering the same POWER.

There are a lot of low resistance high voltage MOSFETs these days, lots of work in power systems is in that area.

If we could increase inverter battery voltage to 200 volts, we'd have cheaper more efficient inverters.

silverstone said:
But the higher voltage will result in ~1x-3x MORE losses.
Click to expand...
No, much less loss as my example above demonstrates.

If your theory was correct, EVs would be low voltage. They are not precisely for the reasons I have stated.

Mike C.
 
You must log in or register to reply here.
Back
Top