# DC-only equipment with solar augmentation

Registered Users Posts: 9 ✭✭
We have some equipment that runs on straight 48VDC. It's currently powered by a DC-UPS system with a medium-sized battery bank. The UPS is on AC power now, and has separate connections for the load and the batteries (similar to a solar charge controller). In the event of a grid outage, the batteries are sized to carry the load for a bit over 24 hours.

The DC-only equipment has a continuous load of about 200 watts, which adds up considering it's running 24x7.  What we'd like to do is add a couple of solar panels in the range of 300-500 watts total to augment the AC power and reduce the continuous power draw from the grid. There is no inverter, because the load is 100% DC. If we connect up a solar charge controller, would we just connect the load side of the solar controller, or both the load and battery?
Tagged:

• Registered Users Posts: 221 ✭✭✭
We can plug your numbers into some formulas to figure out what you need.  Assuming your load is 200W and running 24hr per day, that's 200W X 24hr = 4800Whr.  Assuming that around the winter solstice, you only get 3 sunhours per day, that means you need
4800Whr/3 sunhours = 1600W of panels.  Divide that 1600W of panels by 85% for efficency losses, and you get 1600W/0.85= 1882W of panels.

Right now grid-tie panels in the 240-260W range are dirt-cheap.  If you got 8-9 240W panels that would give you about 1920W-2160W.  Wire them 2S4P, or 3S3P into a medium capacity charge controller.  The wiring would be the batteries to the charge controller, and then the panels to the controller, in that order.  For that many parallel solar arrays, you'll need fuses or breakers (a combiner box) to protect yourself against short-circuits.
System 1) 15 Renogy 300w + 4 250W Astronergy panels,  Midnight 200 CC, 8 Trojan L16 bat., Schneider XW6848 NA inverter, AC-Delco 6000w gen.
System 2) 8 YingLi 250W panels, Midnight 200CC, three 8V Rolls batteries, Schneider Conext 4024 inverter (workshop)
• Registered Users Posts: 9 ✭✭
I'm not concerned about sizing, I want to know the proper wiring for this dual-controller situation.

The solar doesn't have to provide the entire power source; only augment the AC controller. For us, 500 watts is more than enough to provide the functionality we need. I want the solar to reduce the AC load when the sun is up (we get about 5 hours of solar during the solstice at our latitude), plus provide a little boost when the AC is out for more than 24 hours.

So again, the question is the proper wiring when you have two charge controllers; one getting power from AC and a second one getting power from a couple of solar panels.
My suggestions:
• Point to point wiring. Charge1 to battery bus, Charge2 to battery bus... Do not do daisy chaining (Charge2 to Charge1 to Battery Bus)--The controllers can crosstalk more easily and possibly confuse each other.
• Assuming you have a negative grounded battery bus... Each wire that leaves the battery bus should have a properly rated fuse/breaker for AWG of wiring (I suggest breakers--Not much more, and you have a nice On/Off switch for servicing). See NEC simplified table below.
• You can connect solar panels directly to battery bus if they are ~1% or less of battery capacity (basically safely float bank)--Typically Vmp~72 to ~76 Vmp-array.
• If you have an MPPT charge controller, the recommended minimum Vmp-array should be at least 1.3x Vbatt-chargihng (i.e., 60 volts charging * 1.3 volts = 78 Volts Vmp-array minimum).
• If you have 2% or greater of bank AH capacity with solar--You should have a "real" solar charge controller. If the DC loads "go away" for some reason (broken loads, loads down for servicing) to prevent overcharging.
• If you have a PWM solar charge controller, you should derate the charge controller (and make Imp-array) at 80% of controller rating (i.e., 40 amp controller * 0.80 = 32 Amps Imp-array)--This is probably "overkill", but it is what NEC says to do.
• If have an MPPT controller... They will run at 100% of rated output (even if array can supply more power)--Safely and reliably.
• Whatever you decide the charge controller output current will be--Highly suggest that you "up rate" your wiring/fuses/breakers by 1.25x (1/0.80). Battery charging can easily be for many hours at "rated" charging current (0.8/1.25x is NEC continuous current derating). If you have a 32 amp Imp-array * 1.25 = 40 amps. If you have a 40 amp mppt controller * 1.25 = 50 amp minimum rated branch circuit/fusing/breaker.
• For accurate and best charging on a 48 volt bank, you want a maximum voltage drop of 0.2 to 0.4 volt drop from battery terminals to charge controller Vbatt output. Charge controller wiring should be short & heavy to keep voltage drop down. See simple Voltage drop calculator below (not this calculator is one-way run of a pair of cables--Others are round trip length).
• The voltage drop from solar array to charge controller/Vbatt-bus should be around 1% to 3% maximum drop (typical).
• When you put 2 or more charge controllers in parallel, the one with the "highest voltage" setting wins. For example, your solar charge controller may be set to 59.2 Volts Absorb and 55.2 volts Float. And the AC charger could be set to 58.4 Volts Absorb and 54.4 volts Float. That way the "solar charger" should "win" most of the time (and supply more energy to bank). You can of course crank up the solar to 60.0 volts charging/float if the bank is loaded 100% of the time--And the high voltage setpoint is just there to stop severe over charging by solar.
• And lastly--When connecting solar charge controllers--Connect battery bank first and disconnect battery bank last. If you have an energizes solar array connected to the Vpanel of the charge controller, and no battery connection, the controller can boot up "confused" or in some cases be damaged without the battery bus connected.
If flooded cell batteries (above suggested voltage setpoints are for FLA)--Watch water levels--Too much water usage (once or month or more often fill), may be over charginging. Less than once every 6 months, might be under charging. Other battery chemistries will have their own setpoints.

And monttor battery bus voltage daily/weekly/monthly as you gain more confidence the system is running correctly.

https://lugsdirect.com/WireCurrentAmpacitiesNEC-Table-301-16.htm
https://www.calculator.net/voltage-drop-calculator.html

Is the above what you are looking for?

-Bill
Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
• Registered Users Posts: 9 ✭✭
Sounds close. I'm thinking if I put a silicon diode between the AC-powered UPS and the load, and just tie the load output from the solar controller directly to the load, that will provide a ~~ 0.7 volt bias in favor of the solar. That way, if the solar has the power, it will be running the load.

I don't think I need to do the same on the battery side, as 99% of the time, the batteries will be in float. I think I just need to adjust the output of the solar controller to not fight the AC UPS.
Placing a diode in between the charging source and the battery bank--Does not really fix much, and could makes things not work (most charge controller require Vbatt on their charging terminals--no battery, no power. If you have a "bidirectional" power line--Such as the UPS--Either no charging current, or no load current.

You could put two diodes in parallel but reversed to each other--0.7 volt drop either direction). And you can even go possibly with Schottky diodes which are closer to 0.2 volt drop.

And the standard diodes get close to 1.0 amps when heavily loaded (relatively speaking). The diodes are an energy loss (0.7 volts * current)--For a 48 volt @ 200 Watt load, it is not that much"
• 200 Watts / 48 volts = 4.2 amps
• 0.7 volts * 4.2 Amps = 2.9 Watts
So, you should have a heat sink--And most power diodes--The "tab" or body is connected to one of the terminals--So you would have to do some electrical insulation so you don't have ~48 volt energized piece of sheet metal (or use insulators under tabs/body/bolts).

If you figure out what the UPS charger is set to float (and absorb?) and just set the solar charger a 0.2 to 0.5 volts above that--You should not have any problems (unless the UPS is really picky about battery voltages).

-Bill
Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
• Registered Users Posts: 9 ✭✭
No. I was not suggesting to put the diode between either charger and the batteries. Both of those would be directly connected.

What I am suggesting is to put the diode between the actual load, and the load connection on the AC charge controller, and leave the solar controller connected directly to the load. That would keep the AC charge controller out of the mix while the solar charge controller was getting sun. I think this "should" ensure that the solar powered the load while the sun is out. That would reduce our AC power usage by somewhere between 5 and 9 hours, depending on the time of year.

As insurance, I would increase the float voltage on the solar charge controller by approximately 1/10 (or maybe a little more). The AC charge controller is currently floating at 54.6 volts, so if I set the solar controller to float at 54.7 or 54.8, I think we will be good.
A diode between the load and the battery, solar connected to load side?

If so, solar charge controllers are not really designed to "regulated" (i.e., hold a steady 14.x volts) without the battery to "buffer" the voltage (in reality, in most DC power systems, the battery's ability to both take and supply current is what stabilize the bus voltage). If you "decouple" the solar charge controller from the battery bank--You could end up with lots of "voltage spikes" from the solar charge controller. (i.e., the solar charge controller passes current from the array, and if the solar current exceeds the load, then the voltage rises quickly--because the diode "blocks" current from going to "charging" the battery--The "buffering").

How each solar controller brand/model/type (PWM or MPPT) behaves with a blocking diode to the battery bank--I cannot say A will work and B will not... It is just not a common installation.

The use of two diodes installed in "reverse" to each other potentially could ameliorate the problem... The "spikes" will be be limited to Vbatt+Vdiode drop (and voltage drop is limited to Vbatt-Vdiode drop). Standard diode drops are ~0.7 to 1.0 drops--Schottky can get down to 0.2 volt drops.

Another possible issue with diodes--If your system is involved with radio transmitter/receiver--Diodes are quite noisy "electrically". I have this with simple DC rectiviers for computer power supplies (close to FCC A/B limits) and others have reported that when they turn on their LED lights, the TV loss reception. And has been reported as an issue with aircraft:

https://www.icao.int/safety/acp/ACPWGF/ACP-WG-F-31/ACP-WGF31-IP01 LED lighting.doc

Not saying this will be an issue for you--But if you have RFI (radio frequency interference issues), that the Diode in the battery line (and even LED lighting) can be an issue.

I would certainly suggest using the "standard" method of connecting solar (array=>charge controller=>battery bank... And simply program the solar to be slightly higher voltage and the AC power to be slightly less.

If you had enough solar for "day time" power--You certainly could setup a little voltage controlled switch... Battery > 14.4 volts (i.e., over AC charger setpoint and under solar set point), AC charger off. Battery bus falls below 12.8 volts (battery is beginning to be discharged), AC charger turns on. Maybe with a 5 minute delay so they don't "beat" against each other. When the battery voltage > 14.4 volts the next day, then AC charging off and solar takes over until battery bus voltage falls below 12.8 (again, use 5 or X minute delay to avoid turning AC off before sun is bright enough to sustain your loads without taking some energy from the battery bank). The details may change (depending on AC charger's settings and solar's performance).

Or, if possible, set AC charger to 13.3 volts float, and solar to 13.8 volts float... then they both share the loads to the "best" of solar's ability.

Anyway--Just some thoughts on the project.

-Bill
Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
• Registered Users Posts: 9 ✭✭
BB. said:
A diode between the load and the battery, solar connected to load side?
-Bill
No.

The AC-UPS has two connections. battery & load

The solar charge controller has two connections: battery & load

I would connect both to the batteries directly, but set the solar for a slightly higher float voltage (~~ .1 or .2 volts). This would encourage the solar to win the charge battle when sun is available. If it doesn't have the power, charge would come from the AC-UPS.

On the load side, I'm not sure if I would connect both directly. I'm thinking I should put a diode in there to ensure that power comes from the solar charge controller, but that may be moot if it's already providing the bulk of the power to the batteries.

Maybe I need to draw a diagram?

I am not quite sure about the AC-UPS having two connections... Battery & Load--Unless you mean DC Battery and 120/xxx VAC load?

The "Load Output" if DC--Possibly has an LVD (low voltage discount) to prevent over discharging the battery bank (especially critical for Li Ion type batteries). But that is a "weird function" for a UPS (mostly focused on providing battery backed AC power--Not managing a DC Loads battery bus--But it is possible).

You can post a link to the UPS for us to take a look at...

For solar charge controllers... All should have a Battery Connection--And that is as adversized, direct connect to the battery bus through a fuse/circuit breaker (recommended) for safety (circuit breaker or switch--Nice to have when servicing system--Lets you cut power while working on controller--Note that you want to make battery first, then solar, and break solar (if the sun is up) then battery connection--The controller needs battery power to boot correctly and not be damaged by high solar panel voltage--Damage is not common, but can happen).

The "Load" connection--That is usually used by the solar charge controller to turn on/off (for example) outdoor lighting on in evening, then turn off..., or turn off loads if the battery is below 10.5 (or other) voltage (LVD--Low Voltage Disconnect). Generally, the LVDs cannot manage large currents/large surge currents (such as an AC inverter powering the heavy starting loads of an induction motor). If there are heavy currents or high surge current, then the DC loads/AC inverter DC inputs should be tied directly to the battery bus via a fuse/breaker (and not through the LVD).

The issue with placing a diode between the battery bank and the solar charge controller--The controller may not boot if it is not directly connected to battery bank (through diode, the battery may look "dead" and controller will not boot, or boot correctly).

Playing with the charge/float voltages should work fine to steer the source of the charging current to favor solar when the sun is up...

-Bill
Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
• Registered Users Posts: 9 ✭✭
edited March 2021 #11
BB. said:
I am not quite sure about the AC-UPS having two connections... Battery & Load--Unless you mean DC Battery and 120/xxx VAC load?

The "Load Output" if DC--Possibly has an LVD (low voltage discount) to prevent over discharging the battery bank (especially critical for Li Ion type batteries). But that is a "weird function" for a UPS (mostly focused on providing battery backed AC power--Not managing a DC Loads battery bus--But it is possible).

You can post a link to the UPS for us to take a look at...

For solar charge controllers... All should have a Battery Connection--And that is as adversized, direct connect to the battery bus through a fuse/circuit breaker (recommended) for safety (circuit breaker or switch--Nice to have when servicing system--Lets you cut power while working on controller--Note that you want to make battery first, then solar, and break solar (if the sun is up) then battery connection--The controller needs battery power to boot correctly and not be damaged by high solar panel voltage--Damage is not common, but can happen).

The "Load" connection--That is usually used by the solar charge controller to turn on/off (for example) outdoor lighting on in evening, then turn off..., or turn off loads if the battery is below 10.5 (or other) voltage (LVD--Low Voltage Disconnect). Generally, the LVDs cannot manage large currents/large surge currents (such as an AC inverter powering the heavy starting loads of an induction motor). If there are heavy currents or high surge current, then the DC loads/AC inverter DC inputs should be tied directly to the battery bus via a fuse/breaker (and not through the LVD).

The issue with placing a diode between the battery bank and the solar charge controller--The controller may not boot if it is not directly connected to battery bank (through diode, the battery may look "dead" and controller will not boot, or boot correctly).

Playing with the charge/float voltages should work fine to steer the source of the charging current to favor solar when the sun is up...

-Bill
Once again.

I was not proposing to put a diode between the battery bank and the solar charge controller.

I am proposing to put a diode between the load and the UPS controller.

The only AC involved here is the power input to the UPS.

The AC-UPS is connected to 3 things:

1. AC - input only
2. DC out - to battery bank
3. DC out - to load
The UPS is just like a solar charge controller, EXCEPT that it gets power from the grid instead of solar panels. We have no AC loads at all; all the loads run on 48VDC. It even has an LVD like many solar charge controllers.

Here's a link to the unit we are using.
OK, the UPS is a AC in to DC output for UPS'ing DC loads.

Still not sure where the Diode suggested location is... If DC Load is connected to DC Out of the UPS, then the diode is between the Solar Charge controller and the DC Battery/UPS/LVD bus... That would mean that the solar charge controller would have to output >diode drop (0.2-1.0+ Volts)--And would

1) reduce current from solar charge controller and/or require +1 volt on the solar charge controller to overcome the diode drop and
2) at night when there is no solar power, there would be no DC Bus power to the solar charge controller (diode block current flow back to the solar charge controller's keep alive/boot circuits--And it would turn off until the sun rose again--And possibly cause the solar charge controller to not boot or have a "confused" boot as there is no battery voltage in the morning.

As I understand the system--I would connect the solar charge controller directly to the battery bank/battery bus--NOT to the LVD+Load output bus.

If the LVD ever does activate (disconnect load from Vbatt)--The solar charge controller (if daytime) will most likely cause a large voltage surge to the LVD bus/DC loads and potentially wipe the loads out (if electronic loads). Solar charge controllers are not "true voltage regulators"... And they can pass Varray voltage directly to their Vbatt output if the battery has been disconnected.

If you have a PWM controller--Literally an On/Off transistor (switch) between array and Vbatt output... No battery, then the PWM is outputting Varray and zero voltage. There is no battery to "buffer" the current spikes into stable voltage.

If you have a MPPT controller--The DC to DC (typically) Buck converter again uses the battery bank to even out the pulsed current into a stable "average" Vbatt voltage. If battery is disconnected (LVD trips), again, most likely a wildly varying DC bus voltage from the MPPT controller.

There are many different possible fault modes (LVD disconnect, battery over/under voltage, under/over charging, loss of AC power, etc.)... It is impossible to predict how every solar charge controller on the market will react to those failures. But keeping the system connections "vanilla" (i.e., solar charge controller connected to Vbatt bus, not LVD bus)--Less corner conditions to worry about/test.

Placing the Solar Charge controller on the Vbatt-bus is the way they are designed to operate. Just keep the solar charge controller voltage a bit higher than the UPS charging voltage and you should be fine. Even if LVD trips, the solar charge controller still sees the battey bank and will keep it charged to the best of its ability (amount of sun, amount of loads, charging current from DC UPS, etc.).

About the only major issue you would have to worry about is possibly excessive charging current to the battery bank (Solar+AC+no loads)... Check the battery specifications--But less 13% or less rate of charge should not be an issue (Deep Cycle FLA batteries are generally rated for 10% minimum rate of charge for best life--But you are really looking at float service batteries in this application--UPS--So check their requirements).

If you have the potential for lightning--Then the suggestion of using surge suppressors on the DC solar side is recommended.

-Bill
Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
• Registered Users Posts: 9 ✭✭
Here's the basic block diagram. I'm not certain that I need an isolation diode. If I did put one in, it would be between the load and the UPS controller (shown in blue). However, if the float voltage on the solar controller was a tenth or two volts higher than the AC controller, that might be all that's needed.

edited March 2021 #14
The upper right cable from solar LVD to load is not needed.
The LVD from the UPS controller to load is enough.
Having both connections are redundant.
If this was a telephony system, the solar would have its own battery bank and the AC charger another bank. With both solar and AC power going to the DC loads. And there would be a series diode in each line to load. That way, if one system failed the other would continue to supply current to the DC loads.
And we probably would not use LVD. Telephone central offices are critical resources and want the CO to go until the batteries are dead and/or backup genset ran out of fuel.
If it was US Coast Guard, there would be three independent power sources.
And with redundant systems, what happens if one LVD fails or shuts down while the other continues to supply current (failure needs to be detected).
Bill
Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
• Registered Users Posts: 203 ✭✭✭
I've done this with just a switching power supply connected to the solar panel. Generally panels should be wired to produce over 100V. You do have to avoid switchers that have voltage doubler inputs.  Isolate battery from from load.
• Registered Users Posts: 9 ✭✭
Yah. This is for a wireless internet installation. All of the equipment runs on 48VDC. There are zero AC loads at all. We use the LVD with alarms at set thresholds. If the power is out beyond the battery backup, then we plug in a small genset. We know when the LVD is going to trigger, and time accordingly.

The whole "to LVD or to not LVD" is a somewhat religious thing. We have compatriots with these installations on mountain tops, and many of them go without LVD because it is sometimes impossible to get to the equipment during bad weather. They'd rather run the batteries dead than have the LVD cut off service. Most of our installations are accessible such that we can bring a genset in when needed. More critical sites have a genset onsite, which kicks on automatically when needed. In those cases, the battery backup only needs to run a few minutes.
• Solar Expert Posts: 9,511 ✭✭✭✭✭
I wish the words Diode and Solar never met.  Diodes "solve" nothing except night time panel backfeed - which was solved with modern charge controllers in the 1975's

The big problem I see, is what the 48VDC equipment can withstand during the solar charging cycle, which can reach 60VDC for battery conditioning.  Can your gear tolerate that ?
Powerfab top of pole PV mount | Listeroid 6/1 w/st5 gen head | XW6048 inverter/chgr | Iota 48V/15A charger | Morningstar 60A MPPT | 48V, 800A NiFe Battery (in series)| 15, Evergreen 205w "12V" PV array on pole | Midnight ePanel | Grundfos 10 SO5-9 with 3 wire Franklin Electric motor (1/2hp 240V 1ph ) on a timer for 3 hr noontime run - Runs off PV ||
|| Midnight Classic 200 | 10, Evergreen 200w in a 160VOC array ||
|| VEC1093 12V Charger | Maha C401 aa/aaa Charger | SureSine | Sunsaver MPPT 15A

solar: http://tinyurl.com/LMR-Solar
gen: http://tinyurl.com/LMR-Lister ,

• Registered Users Posts: 9 ✭✭
Yeah. 60V is not an issue. All of it is listed as "48VDC", but if you look at the specs, they all have a wide tolerance.
• Registered Users Posts: 203 ✭✭✭
Battery voltage is the wildcard here.  Diode selection of power source only works well with voltages being within a couple volts and the battery always wants to be the lowest.  One way to achieve that is with a buck converter on the battery. I see charging as almost an afterthought being a simple maintainer off line voltage.  Solar panel supplies as much power as it can like a linear current booster with a voltage limited power supply supplementing line voltage power supply. .