MPPT charge controller and DC distribution

throwingshade · February 2019

This seems like a simple question, and maybe it is, but I can't seem to find it clearly answered anywhere.

Let's assume an off-grid cabin. Let's say I've designed for a controller/inverter combo box to handle the 48 volt solar installation. So we're talking about a straightforward setup in terms of distributing inverter AC power throughout the cabin.

But, what if I want to also distribute DC power? The combo boxes don't include a robust DC Load output like a few of the discrete charge controllers I've seen. So, I would have to tap right at the battery bank; I could hook in an LVD and a DC-DC down-converter, which would smooth everything out, and then use fat wire to run DC through the cabin.

Which would be fine, but wouldn't this confuse the MPPT charge controller? The absolute last thing I'd want to risk is battery damage. If I'm drawing variable chunks of current directly from the batteries without the charge controller knowing anything about it, won't it confuse the heck out of the controller? Won't it sense those voltage drops and jiggle things that don't need jiggling, or am I worrying needlessly?

(BTW, the whole point of all this is efficiency. The easiest approach is to just convert inverter AC power to DC, but that seems woefully inefficient; if I could use a DC-DC converter, I'd only be incurring a single conversion loss, whereas converting to AC and then converting to DC seems... a little extravagant.)

Any help appreciated.

Estragon · February 2019

What are your DC loads, and how far away from the batteries are they?

706jim · February 2019

Trust me: Eat the inverter losses and FORGET DC in the cabin. Do you have any idea of the line loss incurred in DC wiring? I speak from experience. I built my place in 1992 and have two complete wiring systems with switches and outlets; one each for AC and DC. The DC is never used. I noted 1 volt drop with a 500ma load.

Estragon · February 2019

I have DC lighting that will stay lit even after the AC inverters let the magic smoke out, so I wouldn't DC rule out entirely. Right tool for the right job?

Photowhit · February 2019

I personally said good bye to DC loads when I moved to a higher voltage system 14 years ago. I can understand the desire, but don't think their would be much gain, particularly when you will have losses with DC to DC converter and need to run heavier wire to prevent losses in voltage during delivery.

That said, you won't confuse charge controllers in general. Both Schneider and Midnite have shunt based measurement systems so the DC load would be taken into account. Outback might be a bit confused, but only if your loads were very high and I think they reset once float is reached. Other charge controllers, will just ignore things an make choices based on the battery voltage as they have since the dawn of charge controllers.

Tecnodave · February 2019

Throwingshade,

Yes you can do what you are proposing. When you connect the DC loads to the battery system do use circuit breakers rated for DC use. If your solar controller uses a shunt at the battery connect all loads so power will go through that shunt. Only one end of the shunt is connected to the battery and all loads and input power connect to the shunt. Any good controller will see all loads and charge properly. I have the most expereince with MidNite , MagnaSine, Trace Engineering ,Schneider Electric units with a bit of expereince with Outback. All of these units will have no problems with this setup. Chinese controllers sometimes have a dedicated DC output for loads but they do not correctly charge batteries. Stay away from EP Solar Tracer units. They do not correctly sense battery voltage and will leave you undercharged. One of the most important things to look for in solar system controllers is the ability to program the controller to the battery bank in use. All the good ones allow you to program the exact parameters that the battery manufacturer recommends. This is the most important factor in reliability of the system. A lot of the Chinese controllers are set for automobile batteries or RV deep cycle batteries and will not properly charge true deep cycle batteries.

On the subject of batteries, consider this. The typical automobile battery's job is done in 15 seconds or less, after that it's going along for the ride, the alternator in the vehicle does all the work. In solar once the sun sets its all on the battery. It gets no help for the next 12-15 hours or so unless you want to be running a generator frequently. Very minimum for solar is traction batteries, the golf cart batteries. They run you all over the course with no other power, they are 6 volt or 8 volt. 12 volt is not used in traction use. The common "GC-2" golf cart battery is readily available and does very well in solar use. About $100 for each 220 a.h. battery, you need two for 12 volt, 4 for 24 volt system, etc.The next less expensive traction battery is the "L-16 " traction battery commonly used in floor sweepers as used in grocery stores to scrub the floors clean. Again not available in 12 volt.
6 volt 450 amp hour L-16 are about $250 for the less expensive Johnson control units sold under many names, Interstate being one of them. The Rolls Surette batteries are more expensive at about $500 per battery. I have 4 Interstate L-16 in my 24 volt system that are 6 years old and 4 Rolls Surette L-16 that are nearly 14 years old in daily use. I'm totally off grid for many years. Do have a generator but my system is way over engineered so I never have to charge with generator. I do use a MagnaSine MS4024AE 120/240 volt 4 kW inverter but most of my loads are run from a Samlex DC-DC converter (SDC-30). which is way more efficient than an inverter. My referigeration is 24 volt DC , way more efficient than any AC powered referigerator , 46 watts at 12 volts or 24 volts, about 250-300 watt hours a day verses 1 kW / for an energy star small referigerator.

my solar gear is mostly MidNite, I have two MidNite Classic 150 MPPT controllers and two MidNite Kid MPPT controllers, plus some others, two MorningStar controllers and older Trace PWM controllers.

I bought several EP Solar Tracer 3210 and 4215 controllers and junked them. They were destroying my batteries, never fully charging them, requiring generator use to top off the batteries, horribly inefficient. Not recommended for anything bur the trash can, absolutely no manufacturers support.

MidNite Solar has stellar support that I have not seen anywhere else, you get real help, not sales people.

I am biased, I'm an engineer and do like well engineered products with manufacturers support if I should need that.

David

Tecnodave · February 2019

I cannot edit my posts from my iPad so I will toss this in....The L-16 battery format is used in traction batteries as well as true solar deep cycle batteries, My interstate units are Traction use and my Surette batteries are true "RE" renewable energy batteries I have Surette S-530 batteries, no longer available. The S-550 line has replaced them. The same great service in support is available at Rolls Surette, they truly care about their customers.
NAWS does carry Surette batteries and does have good pricing on them, I have just installed 4 Surette S-550 L-16 in my latest installation. Entire system purchased at solar-electric.com.
david

BB. · February 2019

And just an FYI. NAWS (Northern Arizona Wind & Sun or solar-electric.com) is the host for our forum here.

The rest of us here (including me, the moderator) are all volunteering our time and experience and have no other business ties with NAWS (other than as the occasional customer).

-Bill

throwingshade · February 2019

Estragon said:
What are your DC loads, and how far away from the batteries are they?

I'd just like to power a some 12V led lighting, a few brushless 12VDC computer-type fans, maybe an outside camera or two, maybe a charger or two - no more than 120 watts total, and nothing with an inductive component.

The batteries are actually going to be in a tiny walled off room in the cabin (the summertime heat would otherwise shorten battery life significantly). The maximum wire travel would be about 70 feet.

If I can run DC at all, I'll probably run 48v to a midpoint and put the DC-DC converter there, to shorten 12V wire runs.

throwingshade · February 2019

706jim said:

Trust me: Eat the inverter losses and FORGET DC in the cabin.

I hear you about the hassle and the voltage drop problem. But part of the reason I like low voltage, in addition to theoretical efficiency benefits, is that 12 volt lighting and fan installations don't have much electrical code to adhere to. Low voltage stuff gives you a freedom to do things quickly and efficiently that high voltage AC installation regulations make very complex and picky.

I'm building where inspections are not going to happen, so I don't have to follow code. But I do plan to take pictures documenting the construction, in preparation for eventually qualifying for some kind of insurance (if I can find any).

It's no problem for me to keep the AC stuff I need up to code. But it would be more fun to use low voltage alternatives for my lighting and peripheral schemes.

throwingshade · February 2019

Photowhit said:

Both Schneider and Midnite have shunt based measurement systems so the DC load would be taken into account. Outback might be a bit confused, but only if your loads were very high and I think they reset once float is reached. Other charge controllers, will just ignore things an make choices based on the battery voltage as they have since the dawn of charge controllers.

I don't think the combo boxes I've been looking at use an external shunt.

I realize there is a simple answer to that, ie, stop looking at combo boxes. But to the inexperienced they sure do look simple, clean, and in many cases they are remarkably inexpensive by comparison.

throwingshade · February 2019

Tecnodave said:

Yes you can do what you are proposing. When you connect the DC loads to the battery system do use circuit breakers rated for DC use. If your solar controller uses a shunt at the battery connect all loads so power will go through that shunt.

I'm pretty sure I've been looking at Chinese units (surely not all of them are flawed, are they?), and I'm pretty sure they don't use external shunts.

As for batteries, I'll probably start with 8 Trojan T-105RE batteries, and build so that if I eventually need L-16s it won't be a big problem to upgrade.

Thanks for the pointers.

throwingshade · February 2019

BB. said:

And just an FYI. NAWS (Northern Arizona Wind & Sun or solar-electric.com) is the host for our forum here.

The rest of us here (including me, the moderator) are all volunteering our time and experience and have no other business ties with NAWS (other than as the occasional customer).

-Bill

Got it - thanks!

Tecnodave · February 2019

And in the subject of disclosure, I am a MidNite beta tester for the Kid controller program, frequent contributor on their forum, and frequent customer of Northern Arizona Wind and Sun, I have no monetary ties to either company other than to be a frequent purchaser of solar products made by MidNite and sold by NAWS. Long ago I found the best manufacturer and dealer!

david

please note, my spelling is better than you see, it's iPad auto spell doing this horrible spoofing.

jonr · February 2019

Little know fact: many 120VAC LED bulbs run fine on 48V DC.

Estragon · February 2019

So ~10a@12vdc load on a ~225ah@48vdc bank? I don't see a big issue in terms of controller "confusing" or "jiggling" with or without a shunt.

Absent a shunt, the charge current as measured by the controller obviously will be inaccurate to the extent some current is going to loads. That aside, the controller should go through its normal charge cycle based on bank voltage much as it would with inverter loads. If a shunt can be used though, I'd use it.

A couple of things maybe worth noting though; first, some 12vdc loads may be picky about voltage and may not handle the 11-16v charging cycle range, so either check device ranges and/or use a fixed (not proportional) output converter. The other is some converters can apparently fail such that >48v input could appear on the output. Worth consideration if expensive devices are used on the low voltage side.

I run ~20 LED puck lights, 5v devices (eg. phones), an occasionally used 12v water pump, and occasionally used 10a outlet for portable dc fridge, all with no controller problems.

Tecnodave · February 2019

Throwingshade,

Before you put down cash$$$ , do plan your system well and post your plans. It will help others point out the inconsistencies. The batteries are the most expensive component of the system. 8 good quality GC-2 format 6 volt 225 a.h. Batteries are going to set you back maybe $900.00 It does not make any sense to use a Chinese controller to maintain these batteries. The most important thing you will need the controller to do is charge the batteries as per manufacturers specifications. This is very important! Most if not all Chinese controllers have several generic battery charge settings, They are setup for automotive starting/r.v. deep cycle batteries. The combo charging/deep cycle battery is not a true deep cycle battery. The settings for that kind of battery will leave a traction or RE (renewable energy) undercharged. If any battery has a "cold cranking amps rating" (CCA) it is not a true deep cycle, and as such not suitable for RE use. Golf cart GC-2 , and L-16 batteries do not have a CCA rating as they are not starting batteries, that are rated as so many amp-hours @ 20 hour discharge. Most Chinese controllers have no support. Period. You will find no one to answer your questions. I was using Trace C-40 PWM controllers and decided to upgrade to MPPT controllers. I made the mistake of buying EP Solar Tracer controllers. I did get more power from my array but my batteries (L-16 set @ 24 volts) were losing state of charge. I monitor my battery health with Bogart Engineering RV2025 battery monitor as neither the Trace or the EP Tracer is equipped to read external shunts.

With batteries going undercharged I swapped out the Chinese units and reinstalled the Trace controllers. Walla , my batteries came back. Calling the manufacturer I got only script answers, no real knowledgable technical persons at all. I called all the American controllers and peppered them with questions, mostly got script answers, and had to go through several layers to get real technical help. One company stood out in that department, they always had technical help available. That made my decision real easy. I called Northern Arizona Wind Sun and bought a MidNite Classic 150. Problem solved. I'm still using that set of traction L-16 batteries, now 6 years old. My other set are Rolls Surette S-530 L-16 true RE batteries which are now 13 , going on 14 years old, Each set has a MidNite Classic 150.

Lesson learned, you get what you pay for.

I have two complete solar systems side by side....100% redundancy....no fail that I cannot handle within minutes by throwing a few switches.

david

Tecnodave · February 2019

On the subject of DC to DC converters, some are voltage dividers, and will not supply proper voltage at output. I am using Samlex America SDC-30 DC to DC converters have a very clean 13.7 volt DC output with an input voltage range of 20-30 volts. Very well regulated, quiet electronically, low ripple. These units are made in Holland. They have a full line of units for 12 ,24, and 48 volt systems both with common neutral and more expensive fully isolated units.

I have never had an issue with these. I've used them on my fishing boat to run my radios GPS , and fish finders from the 24 volt battery bank.no radio interference at all.

david

littleharbor2 · February 2019

https://forum.solar-electric.com/discussion/comment/395313#Comment_395313

Are you saying that ,after setting charge points the controllers still undercharged your batteries? As far as I know the Tracer/Ep controllers have fully adjustable voltage setpoints.

Tecnodave · February 2019

Littleharbor2,

The EP Solar Tracers that I had did not have "fully adjustable set points" they had only three options which were flooded and two settings for AGM. There were presets and were not adjustable. Also the units and no calibration offsets to compensate for voltage drift as found in the MidNite Classic and Kid controllers.

As I remember, the models I had were EP Solar Tracer 3215N and 4210N. Bought them on Amazon, I can go back and verify model that I used. The voltage displayed was not in agreement with any of my voltmeters including Fluke 77, 87, 87-V, 177, wavetech units and others. I have access to a calibration lab and my meters are calibrated to NIST and MIL45662A.

David

Tecnodave · February 2019

BB,

I just tried to edit that post, no go.....edit button appears, does depress, does not allow me to edit. Tried several times.

Lenovo is coming on line now so I will log in with that and report back

david

littleharbor2 · February 2019

Did you have the remote lcd readout? I believe they enabled the user to make these adjustments.

BTW Editing seems to be working fine here.

throwingshade · February 2019

https://forum.solar-electric.com/discussion/comment/395308#Comment_395308

I plan on using DC just about exactly like you do. And I will make sure my DC-DC converter(s) have a wide input range and overvoltage protection.

I'm still a little concerned about using a controller that doesn't utilize an external shunt to take its measurements, and as far as I know, none of the inverter/controller combos do. If I'm drawing current directly from the batteries, the controller will read a voltage that just isn't quite accurate. I know it's much like an inverter load connected to the batteries, but I was kind of thinking (hoping?) that maybe the combo boxes, since they contain both controller and inverter (which can presumably communicate closely with one another), made adjustments which took inverter load into account.

Oh, and, if I may, just one more question (which probably doesn't deserve its own post): If I opt for a 220vac inverter (mainly for a couple of appliances), is it OK to use a step-down transformer (to 110vac) for lighter, "normal" AC loads? Do sine wave inverters have any trouble with step-down transformers?

throwingshade · February 2019

https://forum.solar-electric.com/discussion/comment/395313#Comment_395313

I probably will post reasonably detailed plans at some point, though it's shaping up to be a pretty vanilla rig. Still, it might help someone, I guess, and that someone might even wind up being me.🙂

BB. · February 2019

C/20 to C/8 is a reasonable set of "continuous" loads for an off grid system with FLA batteries. C/5 for short term loads (minutes), and C/2.5 (seconds to minute) as a suggested starting point.

225 AH * 48 volts * 0.85 AC inverter eff * 1/8 hour discharge = 1,148 Watts average load (for 3-4 hours to 50% state of charge)

But, if you had a load that was running 24 hours per day and you wanted 2 days of storage and 50% max discharge:

225 AH * 48 volts * 0.85 AC inverter eff (if using 120 VAC) * 1/2 days storage * 0.50 max discharge (longer battery life) * 1/24 hours per day = 95 Watt AC load (~2 amp load on 48 volt bus)

While there is certainly a limitation of max current draw from a battery bank (you don't want to draw 100's of amps from a smaller battery bank), many times it is the small loads that are on 24x7 (computer server, DVR on standby, kids leaving the TV/Desktop computer/video games/room lights running all day, etc.) that also tax the overall battery bank capacity and the supporting solar array (and enough hours of sun in a day).

For decades (100 years), the standard good charge controller simply estimated charging current and time based on battery voltage (i.e., max current until 14.75 volts "bulk charging" typically around 80-90% state of charge) and time... For example, hold 14.75 volts for 2-6 hours (2 hours if shallow discharge, 6+ hours if deeply discharged).

The very nicely thought out Midnite Classic using a separate current shunt in the negative leg of the battery bank--So the charge controller can monitor both current out and current back into the battery bank, as well as current draw (i.e., shut down charging absorb charging when current falls to ~2% of battery capacity) is great and much better for proper charging of the battery bank (Outback has something similar now?).

However, even a Battery Monitor with a current shunt connected to the battery bank can still have programming issues and drift from "actual" state of charge. You have to watch them closely to make sure they are not lying to you.

-Bill

Estragon · February 2019

Drawing DC from bank shouldn't give misleading _voltage_ reading to the controller, but will give an incorrect _current_ reading.

Current (amps) into the bank will be net of loads, but the controller only sees its output current. This will affect stats, but will only affect charging efficacy if the loads are large enough that the controller is unable to maintain absorb voltage. The controller will(should) see that voltage is under the setpoint though and stop the absorb cycle timer until voltage rises back to setpoint.

What can cause misreading voltage is too small output wire from controller to bank or flakey connections. Under charging load, this can cause a drop in voltage at the bank, so the controller thinks it's holding absorb voltage, but really isn't, which in turn causes chronic undercharging and shortened bank life.

Estragon · February 2019

I use an autotransformer to step 120vac psw inverter output up to split phase 240vac to run a water pump which works fine. In principle, I think it should be ok to step down providing the proper transformer is used.

A couple things to consider though:

- The transformer does have losses, so depending on the higher voltage appliances it may make more sense to use a step-up. That way, the transformer could stay off when the higher voltage appliances aren't in use.

- Also, many inverters have a search mode in which an otherwise idle inverter pulses every few seconds looking for a load greater than a few watts. A step-down would presumably stay on overnight so you can turn a 110v light on etc. If the transformer idle load is less than search setting, the transformer will pulse on and off making a possibly annoying buzz. If the idle load is greater than search, it will prevent the inverter from going into search mode. I don't normally need to fill pressure tanks overnight, so I can turn the step-up off overnight and let the search work as intended.

throwingshade · February 2019

https://forum.solar-electric.com/discussion/comment/395385#Comment_395385

Your Midnite Classic description certainly sounds like the way every discrete controller should be designed. If one were designing an all-in-one controller inverter, a plug-n-play box, though, one could achieve almost the same precision without using an external shunt *so long as all current into and out of the battery bank passed through the box*.

And of course, there's your last line, which suggests that there is still a lot of luck and voodoo involved with the successful administration of a battery bank. This is another reason (besides cost) that I'm going with T-105s to start; since I am, according to forum posts everywhere, likely to crap out my first set of batteries. The thought of doing so to 8 L-16s gives me the shudders.

throwingshade · February 2019

https://forum.solar-electric.com/discussion/comment/395387#Comment_395387

I think I'm finally understanding. So with the relatively limited DC current draw I'm thinking about, the fact that the controller isn't getting entirely accurate data probably won't hurt much.

Thanks!

throwingshade · February 2019

https://forum.solar-electric.com/discussion/comment/395388#Comment_395388

A step up really shouldn't present much of a different physical model to the inverter. All it sees is a winding, after all, so if a step up works for you, a step down should work for me. I realize there are transformer losses, but they are surprisingly (to me) small for a properly sized step down transformer.

There are really a couple of reasons I'm thinking about 220v. I'm going to audition a small minii-split, and they are slightly more efficient at 220volts, and, honestly, the least expensive overseas all-in-ones are often only available with 220v inverters.

And your description of the on-off inverter function is one of the underlying reasons I wanted to lay off some stuff to DC. I'd really like to separate always-on servos like cameras and motion detectors from the inverter so that I can get it to shut down as you describe. I read somewhere that an inverter can draw 20 to 60 watts just sitting there being an inverter.

Thanks again - you've been very helpful!

Estragon · February 2019

Running a mini-split regularly, I might have configured for higher voltage and stepped down too.

OTOH, I'd likely only need it for ~10days/year, so might have used a step up off 120v for it anyway. My situation may be a bit different in that I have a master/slave parallel 120v inverter stack, with the slave sleeping almost always. If I series stack for 240, I think I have both lit (at ~35w each self-consumption) as long as there's any load. Assuming single phase 220v, that wouldn't be an issue.

MPPT charge controller and DC distribution

Comments

Categories