New system 24 or 48 volt?

ChrisOlson

Re: New system 24 or 48 volt?

On the 24V system I ran the panels at ~60 Vmp (2 in series). On the 48V system I'm running them at ~90 Vmp (3 in series).

It still worked OK on the 24V system though. In the morning when the bank voltage was low it would put out 90-95 amps. Then as the controller heated up the amps would drop due to controller de-rate to control its internal temps. But the bank voltage would also come up by that time so the actual watts output to the bank remained fairly constant. It worked perfectly fine.

On 48V the controller does have more capacity though, as far as watts output because it will pretty easily handle 80 amps with the bank at 60 volts. The peak output on 24V for it is about 2.7 kW, depending on bank voltage and how hot the controller gets.

I just wanted to point out that you can put 3.5 or 4 kW of panel capacity on it for 24V, like we did, and it still works fine because you get a "flatter" power curve for the whole day instead of a peak at solar noon like we get now on 48V. 3.5 kW of installed capacity on 24V, vs 48V with the same installed capacity, will only yield more kWh on 48V on the exceptional solar day. And that's simply because the panels rarely see conditions where they can produce the full 3.5 kW. Their peak is usually closer to around 2.7-2.8 kW, which ends up being a perfect match for the Classic 150 on a 24V system @ 3.5 kW installed capacity.
--
Chris

NorthGuy

Re: New system 24 or 48 volt?

I've built a 48V system and I have absolutely no regrets about this. And I've never heard of anyone who has.

If you go through this forum, you'll find many cases of people who built 24V systems and then want to switch to 48V, but it's too hard because they have to replace their 24V equipment.

Numerous pluses are: less parallel batteries, more efficient use of MPPT controllers, less losses in the cables, much easier to expand.

I don't think 48V is such a high voltage that it causes any real problems.

Cariboocoot

Re: New system 24 or 48 volt?

NorthGuy wrote: »

I don't think 48V is such a high voltage that it causes any real problems.

Only in two ways: the possible Vmp/Voc difference causing problems V max in on the controller under certain circumstances (which can usually be dealt with with some careful planning) and the choice of disconnects and fuses/breakers of which 24 Volt system have a wider and less expensive selection to pick from.

Otherwise it is simply an issue of whether or not you need that much power stored.
No system is really easy to expand, regardless of Voltage.

niel

Re: New system 24 or 48 volt?

ultimately it will be op's choice on what to get, but i think we are all agreed he needs more than 6 pvs. if you opt for the 48v battery system then the pvs will need to increase by strings of 3pvs whereas the 24v battery system can have the 2 pvs in its strings allowing only 8 pvs to be needed there. there's allot of fence sitting with the system you propose in that it's not as easy to say what works best for you. if it were me and i don't want to worry if it can be handled or expanded upon if need be then i'd go 9 pvs in a 3x3 arrangement into a classic 150 outputting 48v so you can arrange the batteries to have fewer paralleled strings. there is a difference though between 48v and 24v in voltage drops and power surge handling capabilities and you may see less mismatching of the batteries leading to possible longer life. your choice on what inverter to go with.

mikeo

Re: New system 24 or 48 volt?

I agree and since his installer is comfortable with and supports Xantrex, that is the way he probably should go. Also his gut feeling that it should be 48 volts is what I would also agree with. I would also support either 8 at 24 volts or 9 panels at 48 volts. The only thing I couldn't recommend is that many small batteries in a series/parallel arrangement. Too many cells to keep on top of and too many terminals to keep clean. Just let one cell become bad or one terminal to become corroded and this can bring the whole battery bank down.

niel

Re: New system 24 or 48 volt?

mikeo,
i'm glad you brought up the batteries as looking back he does not already own the batteries as i assumed he did. if given a choice on the batteries you probably should look into L16s be they 6v or 2v to have higher amperage batteries to begin with and you could thus avoid having so many paralleled strings of batteries to get out of balance.

ChrisOlson

Re: New system 24 or 48 volt?

Cariboocoot wrote: »

Only in two ways: the possible Vmp/Voc difference causing problems

Yep. It really depends on what panels you choose for that. If you use the ~30 Vmp panels you'll get better efficiency from the controller on 24V with those at 60 Vmp than you will by wiring the same panels three in series for 90 Vmp on 48V. If you wire those panels four in series for 120 Vmp on 48V the efficiency from the controller will be less yet.

Already BTDT.

If I could've bought a XW6024 I would've never switched away from 24V. The 24V system is much simpler, safer and easier to maintain. And with decent MPPT controllers available these days to cut the power losses we used to have to deal with from high-amperage arrays and wind turbines on low voltage systems, I see no advantage to a 48V system unless you need more than 4 kW in inverter size.
--
Chris

northerner

Re: New system 24 or 48 volt?

ChrisOlson wrote: »

Yep. It really depends on what panels you choose for that. If you use the ~30 Vmp panels you'll get better efficiency from the controller on 24V with those at 60 Vmp than you will by wiring the same panels three in series for 90 Vmp on 48V. If you wire those panels four in series for 120 Vmp on 48V the efficiency from the controller will be less yet.

Chris

You would also have to factor in the wiring loss, but of course would be dependent on how far away the array is. I am running 90 volts with my Midnite CC on a 24 volt system, and find that I'm getting good power and charging current into the batteries. I figure that the loss in efficiency of the CC would be comparable to the wire loss with the lesser voltage. I'm not so concerned about the CC loss as the little bit of extra heat helps to keep my place warm anyway, and in summer I have plenty of excess power.

ChrisOlson

Re: New system 24 or 48 volt?

northerner wrote: »

You would also have to factor in the wiring loss

That can easily be compensated for on 24V by burying the next size bigger wire. The amps are only reduced by 33% on 48V. And typically, looking at a 3 kW array on 24V you're only dealing with 50 amps, which is easily managed over quite long distances at 60 volts. Our wire run is 108 feet from the controller to the array combiner and we buried #4 THHN in conduit way back when because we weren't using MPPT at that time. Not a lot of difference in cost between #4 and #8 wire when you buy it on the 250 foot spool. I think #4 is $1.32/ft and #8 is $1.04/ft here off the bulk spool at the local Menards.
--
Chris

northerner

Re: New system 24 or 48 volt?

Wiring loss won't be much of a factor for you with #4, even at that distance. I notice that many others make their runs with #10, albeit, their runs are usually shorter as well.

I'm currently using #10, with a short run of about 35 feet or so, from the combiner box into the house with my 1.4 kw array. The combiner box if very close to the solar panels. When I plan to expand the array, I will make separate runs of #10 in for each array of 6 panels.

westbranch

Re: New system 24 or 48 volt?

the only thing I have not heard from the OP is his PATTERN of use... We need to know not only the total day use but the period of the day that all adults and kids are inside the house and using power... think dinner time, mom cooking dad feet up watching the Sat football game and kids on the Nintendo or surfing the web... so how big is that daily peak draw?

ChrisOlson

Re: New system 24 or 48 volt?

westbranch wrote: »

the only thing I have not heard from the OP is his PATTERN of use... We need to know not only the total day use but the period of the day that all adults and kids are inside the house and using power... think dinner time, mom cooking dad feet up watching the Sat football game and kids on the Nintendo or surfing the web... so how big is that daily peak draw?

The way I look at it, even if it's over 4,000 watts you can throw Gen Support at it for an hour, or whatever it takes. A 4,000 watt generator (a little small to get full charger capacity from a XW4024/4548) plus 4,000 watts from the inverter for an hour - you can run 7 kW peak load with no problem with some reserve left.

That's one thing the Magnum inverters can't do, and why I'd cough up the extra 200 bucks for the XW. So IMHO, as long as the OP has a generator of reasonable capacity, I think peak loads are covered, regardless of which system voltage he chooses.
--
Chris

ChrisOlson

Re: New system 24 or 48 volt?

northerner wrote: »

Wiring loss won't be much of a factor for you with #4, even at that distance. I notice that many others make their runs with #10, albeit, their runs are usually shorter as well.

#10 is fine for 30 amps on a fairly short run, and even over 30 amps for short times if you use a high temp wire like THHN.

The way I've always looked at it - you spend thousands on solar panels or putting up a wind turbine. Why try to save $100 on wire when you hook it up? If the acceptable losses on the run figure out to AWG 8 - put in AWG 6 instead. You never know when you might add some more panels or throw a bigger turbine on that tower. Nothing worse than having to re-trench in wire because you buried the minimum spec the first time. :grr
--
Chris

westbranch

Re: New system 24 or 48 volt?

Chris, I agree gen support would work for him with the XW.
I wonder if there is a HOA in that little village? Sound like they may be there for the quiet and peace... might nix a gen running

vtmaps

Re: New system 24 or 48 volt?

ChrisOlson wrote: »

If you use the ~30 Vmp panels you'll get better efficiency from the controller on 24V with those at 60 Vmp than you will by wiring the same panels three in series for 90 Vmp on 48V.

Really?!?! I've never looked at the efficiency curves for the xantrex CC. You're saying that its more efficient to drop 60 volts to 29-30 volts than to drop 90 volts to 58-60 volts. Is it a significant efficiency difference?
--vtMaps

ChrisOlson

Re: New system 24 or 48 volt?

vtmaps wrote: »

Really?!?! I've never looked at the efficiency curves for the xantrex CC. You're saying that its more efficient to drop 60 volts to 29-30 volts than to drop 90 volts to 58-60 volts. Is it a significant efficiency difference?

Yep. It's enough to make the fan run steady in the Classic on the 48V system. It used to cycle on and off on the 24V system with the same installed capacity. Since that fan is turned on by internal temp, I can only assume the Classic is running hotter on the 48V system. Normally, heat means power from someplace is being dissipated. And that's probably why the Classic can't handle as many amps on 48V as it can on 24V.
--
Chris

BB.

Re: New system 24 or 48 volt?

From what I understand/guess... Higher input voltages "charge" the stray capacitance on the FETs with more stored/wasted energy (Field Effect Transistors). The charging/discharging of these capacitors is a "loss of efficiency" (Energy in Capacitor=1/2 * CV²).

So, running with a higher input voltage could certainly reduce the controller's overall efficiency.

-Bill

inetdog

Re: New system 24 or 48 volt?

BB. wrote: »

From what I understand/guess... Higher input voltages "charge" the stray capacitance on the FETs with more stored/wasted energy (Field Effect Transistors). The charging/discharging of these capacitors is a "loss of efficiency" (Energy in Capacitor=1/2 * CV²).

This would be a bigger and bigger problem as the switching frequency is increased to allow the use of smaller coils.

NorthGuy

Re: New system 24 or 48 volt?

BB. wrote: »

From what I understand/guess... Higher input voltages "charge" the stray capacitance on the FETs with more stored/wasted energy (Field Effect Transistors). The charging/discharging of these capacitors is a "loss of efficiency" (Energy in Capacitor=1/2 * CV²).

So, running with a higher input voltage could certainly reduce the controller's overall efficiency.

-Bill

IMHO, the FET switching losses are proportional to voltage to be dropped multipled by switched current.

If it's the same input voltage, then voltage drop is higher in 24V system, so 48V system will be more efficient. But even with lower input voltage (60V vs 90V), 48V system will come slightly ahead.

For example for 4kW array:

48V: (90 - 55) = 45V * 44A = 1980VA
24V: (60 - 27) = 33V * 67A = 2211VA

Based on this, 48V should have less switching losses.

Other losses are mostly because of resistance of inductors and capacitors (ESR), diode dissipation etc. These only depend on current. Since current in 48V is smaller, these should be smaller too.

Based on this, I would guess 48V with 3-panel strings should be more efficient than 24V with 2-panel strings.

inetdog

Re: New system 24 or 48 volt?

NorthGuy wrote: »

IMHO, the FET switching losses are proportional to voltage to be dropped multipled by switched current.

This seems to be based on the idea that the resistive loss in the FETs during the switching transition time is the primary loss mechanism.

The FET losses which switched off will be zero and while switched on will be proportional to the current raised to some power greater than one and less than or equal to two.
But during the fractions of a second that they are turning on, the losses will be proportional to the open circuit voltage across them times the current, more or less, for a linear change in current between on and off.
Those both would give higher losses for the higher current, but are only two of several estimating mechanisms. And the the second one makes what I think is a mistaken assumption:
I think that one reason your analysis may give wrong results is that it assumes that the length of the switching time of the MOSFET is constant, were in reality it is proportional to the voltage to which the gate was charged. This causes the power loss during switching to be proportional to V x V x I instead of just V x I.

With your numbers, that gives us 89,100 (skipping the units for now) for the 48 volt case and 72,963 for the 24 volt case. Not enough to make as much of a difference as Chris reports, but definitely favoring the lower voltage drop configuration.
If the gate voltage is actually being switched between Vin and 0, on the other hand, we then get the numbers as:
356,400 for 90 volts and 241,200 for 60 volts. A nearly 50% increase in losses for the higher voltage operation.
Where is the power FET engineer when we need him?

NorthGuy

Re: New system 24 or 48 volt?

inetdog wrote: »

I think that one reason your analysis may give wrong results is that it assumes that the length of the switching time of the MOSFET is constant, were in reality it is proportional to the voltage to which the gate was charged. This causes the power loss during switching to be proportional to V x V x I instead of just V x I.

The gate cannot go more than 15-20V from the source without destroying the FET. The gate is usually charged by FET driver chip, which always charge it at the same voltage with the same current, which should create the same speed. Better drivers produce more current and thus switch faster, but the same driver should produce the same speed.

My knowledge of FETs doesn't go far enough to tell whether drain-source current or voltage have any influence on switching speed. My guess is that there is some influence, but I have no clue how signifacant it might be.