diy solar

diy solar

High-voltage DC recommendations

DoaJC_Blogger

New Member
Joined
Sep 30, 2022
Messages
4
I would like to build a high-voltage DC solar setup with 170/340 volts instead of 120/240 volt AC. I think the best option is to have a 400-volt battery bank and use a pair of DC buck converters in series. However, I have 3 issues.
  1. I need a large buck converter. I'll probably have to build this from components.
  2. I need switches and circuit breakers that can handle DC.
  3. I need to build a custom high-voltage BMS.
I don't know if I'll need to build an MPPT since they have high-voltage ones.
 
170/340V is exactly peak voltage of 120/240Vrms AC.

What kind of loads to you plan to power?

A number of consumer inverter backup solutions use a "400V" battery which is actually 48V (or some multiple) and boost converter. Efficiency losses, but it does limit short-circuit current.

REC BMS supports up to 128s

Some primarily AC knife switches support DC. Some are rated up to about 250V, while Square-D heavy-duty switches are 600VDC.

Weigh cost/benefit against off the shelf 48VDC to AC systems.
 
I ran 240v dc for a while, my biggest problem was switches and circuit breakers.
Pretty easy to get a continuous arc going, then total destruction with flames and fumes follows within seconds.
Anything plastic nearby just burns and collapses.
Terrifying stuff.

The other issue is individually monitoring a very large number of series connected cells.
Anything is possible, but maybe less than practical.
Right now I run 100v dc with thirty lithium cells and a home brew inverter.
That is manageable and works well.
 
I want to power just about everything around my house and yard except stuff with induction motors or transformers, of course. I made a YouTube video showing that most devices already support 170 volts DC, even some motorized power tools.

I think 170VDC would be too high for many 120VAC appliances, 2x the power into resistive loads. Likely 120VDC would be better.
If you pulse the 170VDC or swap polarity with some dead time (MSW) I think that would work better.
 
I think 170VDC would be too high for many 120VAC appliances, 2x the power into resistive loads. Likely 120VDC would be better.
If you pulse the 170VDC or swap polarity with some dead time (MSW) I think that would work better.
I did notice that motorized stuff like a reciprocating saw and lawn mower ran a bit faster but I didn't think it would be a problem. Do you think that's a problem? I want the power to be smooth so I'd rather lower the voltage to 120 instead of pulsing it.
 
Motors may be OK, so long as they don't overheat or tear themselves apart.
The low rider guys put 72V into 12V hydraulic pumps, just briefly.

Probably speed controls won't work. I assume they are a triac dimmer, only turn off when polarity reverses.

As was mentioned, your biggest problem will be switches burning up. It takes far more to interrupt 120V or 170V DC than AC.
That can be a problem for control switches, circuit breakers, thermostats, over-temperature limit switches.
If you could toggle polarity without dead time, power would be fairly smooth.
The ideal smooth power is 3-phase sine waves for motors.

I've put a capacitor across contacts to reduce burning, or series RC. That helped with things like a car battery powered bicycle when I was a kid, and it is what we do as as "snubber" so inductive loads don't blow out transistors.
But no guarantee they won't fail anyway and start a fire.
 
Besides switches the equipment fuses are another potential problem. Most power supplies like the one PC, television or playstation have some fusing on primary side. With DC feed the protection probably doesn’t work as intented.
 
High voltage dc fuses are available, but they are physically large and not exactly cheap.
And as Matti says, conventional ac fuses can sometimes generate a continuous arc.

Its a very sobering thought to realize that arc welders generally operate between about 20v and 60v dc.
 
^And some people die from electrocution from welding voltages, with extenuating circumstances (sweat, open wounds, etc). Found some case studies of that when i googled osha SELV (<60vdc).

So 120vac is 170v peak to peak but we don’t call it 170vac, we call it 120 because averaged out over time that’s the ‘dc equivalency’ number. So feeding things constant 170vdc is actually more than they normally get from 120vac. If you rectify 120vac to dc you will get 170vdc, but i believe the 120vac source would not be able to sustain 170vdc on the output at the same power level as it can supply 120vac because it can only create 170vdc at the ac ‘peaks’. This is a wild ass guess but if you put a 1500w 120vac circuit through a rectifier it would only do some small number like 400w at 170vdc before that 170 started to sag. But it should be able to do 120vdc at nearly the full 1500w, minus the rectifier losses.

Somebody tell me if im wrong. Im ok getting outed for wrongness as long as i actually learn right in the process!!
 
120v RMS will do the exact same work as 120v dc into a resistive load.
Two identical light bulbs would have identical brightness.
The ac is constantly varying up and down from 0v to 170v, the dc is obviously constant.
 
What do you guys think is the best way to step down DC voltage from a battery bank? Should I look up how to build a buck converter with a large thyristor, and if so, does anyone have advice about that, or should I look for an existing high-voltage buck converter?
 
Might be good to start with off-the-shelf supply.
What output voltage do you want? Universal switching AC power supplies (100 ~ 240V) would accept 360V nominal peak, close to 400V at high line voltage.

Not sure how much margin above utility specs they tolerate. Back in the 1980's, HP (known by some as "High Price") introduced the lowest cost CRT terminal in the world. When used in India, they burned up, because voltages were wilder than in the U.S. I'll bet the recall wiped out profits for that product.

There are probably expensive industrial supplies available somewhere. Expensive.

Buck converter is the way to go. Basic topology can be found in Wikipedia.
There is simple math for step-down ratio as a function of pulse on time vs. period. Somewhat less simple to determine inductor value so that ratio works for a range of load current. Below that load, output voltage would rise to input voltage, so pulse with has to be reduced (or skip pulses). There are PWM controller ICs. Check out Analog Devices in particular, and TI. Download LTSpice, which can simulate SMPS.

Any of these is for lower voltages, but maybe you could divide down with resistors for sense lines, and driver higher voltage external power transistors.
 
If I may also chime in, I would love to add another 5kw to my 10kw of LFP battery from BYD. I find there are quite a few server type LFP that have the perfect form factor to mount it flat on the wall a-la tesla style. Its only 7 inch thick and so with a steel backing plate, and a cover would add about 3 or 4 inch max. Which is perfect for me.

However the Sonny Boy Storage expects a high dc voltage input on the second terminal block for DC energy input. And, most solutions out there are 48v or less. Are there any worthy and safe options? If it were in the range of 350v-500v DC range would be good. The inputs can take 10 amps and so I could get about 4kw energy rate from or to it. SMA battery specs are:
1681695412890.png

Appreciate any pointers.
 
And, most solutions out there are 48v or less.
There is a guy selling 24v BYD battery in Santa Rosa, CA.

BYD LFP Lithium Battery 24 volt - 4.2 Kw/hr modules - New with internal BMS - 163 amp/hrs - 2 units available -
Good for mobile or off-grid applications. Second Life Units - really good price.
$500 each. ( PRICE DROP ) Direct Santa Rosa pickup.
 
However the Sonny Boy Storage expects a high dc voltage input on the second terminal block for DC energy input. And, most solutions out there are 48v or less. Are there any worthy and safe options? If it were in the range of 350v-500v DC range would be good. The inputs can take 10 amps and so I could get about 4kw energy rate from or to it.

Do you know the language/commands of communication with BMS?
Once we know what BMS to use (or have a protocol converter), more options available. Including DIY, or changing BMS of server rack batteries.
 
I find docs saying which batteries work with SBS, but would like to find BMS for DIY.
I don't find a description of protocol for SBS and compatible batteries. For SI, there may be some commands listed.
REC makes a high voltage BMS, but don't know what inverters that is for.

REC BMS for SI, manual lists RS-485 commands in a table. It also shows CAN bus connector but doesn't describe commands.

Master BMS for high voltage batteries gives RS-485 and some CAN commands or data format.


SMA SBS, I see instructions on wiring to battery but not the protocol


Battery Box also lists that it's compatible with SBS

 
Back
Top