Building a Solar System

Will a NiFe Battery Cut It

Grid Tie, Off Grid or Hybrid

It was after finding a pallet of 255 watt Chinese panels on eBay at a price that made them worth getting for experimental purposes that I began to think about a solar system. The two main types that I had seen are grid-tie and off-grid. Grid tie was super easy and not too expensive. Just run the wires from the panels to the grid tie inverter and then connect it to a 30 amp breaker in the panel. No need for an inverter load subpanel to be added. This would have to rank as one of the easiest things that I had ever installed. Except: The solar panels would have to be mounted in a good place for sunlight. This turned out to be over 300ft from the house. That meant that they would have to be placed mostly in series to produce a high voltage array so that smaller #10 wire could be used. Now I need an inverter that will work on a high voltage array. It seemed that the SMA Sunnyboy 5000 should work well for this. And it did.

Yes, I had thought about the off grid type. After all the Sunny Boy was limited in what it could supply in a power failure situation. Most Grid tie inverters do nothing in a power outage. But an off grid system would need batteries. Boy, what a beast those things are. The lead acid ones I was familiar with are Big, heavy, expensive, high maintenance, easily damaged, can only use 20 to 30 percent of its capacity... I had seen in a few forums where some have said that a newbie will destroy their first set of batteries. Well, I sure didn't like the sound of that!

The Sunnyboy worked great. It was a lot of fun watching the meter turn backwards. Now I knew that it worked! And pretty well to! Also, it was no trouble to install and operate. Nothing much to maintain. Just set back and watch the power bill come down. And it did - some. But as it turned out I was not getting credit for the meter going backwards. Further investigation showed that I needed to get on the power company's “Net Metering” program of course. I contacted the power company and explained what I was doing. They were very nice and said they would send me an email that would contain everything that I needed. The girl on the phone seemed quite happy that I wanted to get on their net metering program. Well I got it ok. There were numerous forms to fill out and send back to them and also to the state. One multi-page form had to be returned in sextuplet! Then I was to have an inspector come out and then his form had to be sent in. I sent these forms in with all the requested diagrams, drawings and pictures. That was well over a year ago. I have still heard nothing.

Well, that is ok. I have a strong distaste for things like this anyway. It looked like that if I was going to do this thing it is going to have to be off-grid. But I have a lot of big loads. Heat pump, electric dryer, electric water heater and so on. That is going to take quit a lot. Don't know if I'm up to this or not. After much research and study I find out that there is a hybrid system that will let you have the best of both. The system will operate totally off the grid, totally on the grid, or anywhere in between. I don't have to send power back to the grid if I don't want to but it can work with and supplement the grid. That sounded pretty good. So what do I need?

Selecting the components

One key to making this work is getting the right equipment. Outback has a real nice 8KW inverter that will do this very nicely. The references and recommendations I could find showed this to be about the best inverter for the job. But now I need high voltage charge controllers and Outback does not have them. However, Schneider Electric does and they are very well recommended and highly rated also. The problem is that I have an uneasy feeling about mixing two rather differently structured pieces of equipment of this type. I know it could be done but would it be more difficult requiring more improvising? I wanted to monitor and control everything from a computer but would this now be more complicated and the whole thing being less neat? Schneider's 6.8 KW inverter will do this job also and will integrate nicely into the system. There are some things I may have liked better about the Radian but the SE inverter works nicely and is also highly rated.

Deciding on a battery

Now for the battery. Oh my goodness, the battery! Boy, I really don't like these things. These FLAs either have the lifespan of a mouse or require you to send in all your money in triplicate. Or both. Then they have the ruggedness and durability of your mother's best crystal. I read everything I could find on these things and nothing made me feel better. Lithium batteries weren't very encouraging either.

Then I discovered nickel-iron batteries (NiFe). What in the world are these things? I have been in electronics etc. all my adult life and half of my childhood and have never heard of these things. What do you mean, they were invented well over 100 years ago? They must not be any good or everyone would know about them and they would be everywhere. What is this about lasting 30 or 40 years. Did someone say 80 or 100 years? There is some scuttlebutt about 5,000 to 10,000 charge cycles. Then you only have to renew the electrolyte to get it back to 0 again. Wow! Is that impressive! On top of all of that you are unlikely to ever hurt them by over charging or over discharging them. May wear down the electrolyte a little if you do it enough. And not sure about that. You can actually use 80% of the battery sacrificing some of the cycles but no damage. I have even heard 100% but surely there is a trade-off. Besides that would just not be practical.

Well, I don't know about all of this. There has to be some gotchas. Yes, I found that there are. Let me see if I can list some of the ones I have found.

1) If I thought FLA and Lithium batteries were expensive. Well These things... Oh good grief!

2) The NiFe voltage operating range is problematic for many charge controllers, inverters, converters, DC appliances etc.

3) All the inverters and charge controllers I have found are designed for lead-acid batteries firstly. Then a few other battery types that are not too different. I have heard that SMA has something that considers nickel-iron batteries.

4) Charge/discharge efficiency‎: ‎65–80% (FLA 50-95%) Wikipedia

5) Really big and heavy for the amp hour rating.

6) More voltage sag under load than FLAs.

7) Self discharge rate 20–30%/month (FLA 3–20%/month) Wikipedia

8) Some say they gas a lot and are messy. This is not my experience however. They do gas some and during conditioning they really gas! Sounds like a babbling brook too! But that is just one time. Never seen them messy though.

9) Some say they need to be watered a lot. Others say not. So far I say not also. But then I do not really know what a lot is. I kind of enjoy this few minutes I spend with them. Certainly better than I enjoy carrying out the trash!

I really wrestled with this. Back and forth. One day it was going to be FLA the next NiFe. I have had almost no real first hand experience with large batteries of any kind but I read everything I could find on the forums and anywhere else I could find information. So, my decision was being based on what other people were saying. And boy was that all over the place! I found that most of the posts though were from those who had never owned or worked with NiFe cells. I spoke, however, to some who have actually owned and operated an NiFe battery.

One who was a knowledgeable and respected professional in the field seemed to hate them. He had had only trouble with his and they were really messy for him. He had taken them out of service years ago and has been using FLA ever since. I remember thinking that they couldn't be this bad or they would not be using them all over Asia and parts of Europe.

Another person had a very informative post of his experience with his NiFe battery. However, the post was several years old and I wondered how they were doing now. I got up with him on the phone and had a real nice discussion with him. He had taken them out of service and sold half of them. Well, that may sound bad but he was not anti-NiFe. In fact he spoke well of them and mentioned things that may have been the trouble for him. Even though he had bought them here in the states and paid the going rate here they were not conditioned when he got them. They had not received their “initial charge” which is very important. More on that later. He had tried as best as he could with what he had available to him to perform this necessary conditioning. They may have been too small for his system but I seem to think the conditioning and perhaps size may have been his biggest problem.

I think the scale tipped after speaking to the then director of engineering at Schneider Electric. She was very supportive of NiFe even though their equipment is not designed with NiFe batteries in mind.

Finding and Buying Nickel-Iron Cells

After studying and calculating for quite some time I decided I needed a nominal 48 volt 800 amp hour battery to do what I wanted. Now, in FLA that is a relative small battery in capacity and physically for this type application. But not in NiFe. These things are huge! And boy are they heavy! This is 40 cells (or 38 - talk about that later) weighing 130 lbs each. That is a 5200 pound battery! That is not the bad part though. The sources here in the US get as much as nearly $30K for a set of these things. This would cost way more than everything else in the system put together and then some! If this was going to work something had to be done. It seems that these larger cells that are sold here in the US are made in China anyway.

Enter Alibaba. I searched on Alibaba's site and found huge companies in different parts of the world that were selling NiFe cells. Most were in China though. I went through the list and selected some to request quotes from. Got their quotes then contacted a few of these and established a rapport with the sales person asking about their NiFe cells, their company, who used their cells etc. I did not automatically go with the lowest quote. Most prices were still less than half the cost of the ones sold in the states.

I enjoyed meeting Alice and Serina with Henan Hengming Fengyun Power Source Co.,Ltd. They were very nice and patient with my many questions. However, I finally went with Zhuhai Ciyi Battery Co.,Ltd. You see, as I understand it, all of these companies make primarily NiCd batteries for their market. NiFe is a secondary market. The cells look identical. They use the same electrolyte, the same case and the same mechanical design. The primary difference is that they use cadmium in place of iron. Although I am sure the Fengyun cells would have been fine Ciyi Battery seemed to promote their NiFe cells more aggressively. Anna Wang who is their sales manager also very patiently put up with my many questions. In fact she did so for a few months as I learned about NiFe batteries and buying them from China. She even called me one evening on the phone to tell me something she thought was important for me to know. It made me feel like her customers are important to her.

The electrolyte usually comes in a powder and you have to mix it with distilled or deionized water and put in the cells. However, after some more of my many questions about that Anna said she would send the cells with the electrolyte already installed. This saved me the trouble of doing it here. I thought that was really nice of her.

Ordering and Shipping the NIFE Battery

Order placed June 15, 2015. Ready to ship July 21, 2015. Received September 25, 2015.

When we had discussed the details about the order and I was sure that Anna correctly understood everything I was ordering and she had answered all my questions she wanted me to confirm the order and request the Proforma Invoice from her. This is generally not a payable invoice but sort of like a trial invoice to be sure everything is correct and nothing has been left out or misunderstood. It also had their company bank information on it. So, at this time she also requested 30% deposit for them to start building the cells. I promptly sent the deposit and after they confirmed receipt the order could go to production.

I received an email from Anna on July 21 saying that my goods are ready for dispatch. She sent photos and everything. She said that after receiving the balance the battery could be shipped. She also wanted the “Importer of Record Number” which was my SSN and “Consignee Number” which was also my SSN. This information was so their “forwarder” could fill out the “ISF” for me. I was apprehensive at first about sending my SSN but this “ISF” is a form required by the US Customs and Border Protection agency and it requires this information. (A corporation would use their tax ID.) The freight “forwarder” is a person or company that organizes shipments for individuals or corporations. CiYi Battery has one and I can have mine on my end or use theirs. It seemed simpler to use theirs so I did. Regulations require that the shipment not leave China until 48 hours after the ISF is filed.

They also bought the insurance for the shipment and paid that expense themselves. A customs broker must also be secured and a Customs Power of Attorney form must be filled out and sent to the broker. The forwarder will arrange for a broker if you don't have a broker you want to use. In fact I found it simplest to let Anna and their forwarder handle everything to get the shipment to the port. They offered to arrange for the shipment to be carried by truck from the port to my destination and handle the hazmat thing and all. However, I had a trucking company I knew bring it from the port. It saved some money.

The shipping cost to get the battery from China to the port was less than $500. However, on this side there was a handling charge, ISF fee, DDC fee, Chassis fee, Customs Entry fee and the worst one was the $557 duty. There may have been other small fees but this is the worst of it. There was also a bonded warehouse involved to hold the shipment for pick-up. When the shipment was released I was able to arrange pickup.

There were also other documents involved such as a packing list, bill of lading etc. I would have liked to have known these things before hand instead of having to learn this as I went along. After a few times there is nothing to it. I suppose I spent less than $1800 to get them from Ciyi Battery to my basement.

Installing and Conditioning

The cells came in four large wooden crates with 10 cells in each crate. There was also another smaller wooden crate with the bus bars, connectors, some interconnect cables and other things. I got the cells into the basement and lined them up in two rows in the middle of the floor. This was the temporary arrangement until I could get the battery rack built. I installed the interconnect bus bars and ten feet of 4/0 cable from the battery to the E-panel. Pretty simple.

Now for this next essential step: Conditioning. Or as the Ciyi manual calls it “the initial charge”.

This is the instructions from the Ciyi Operation & maintenance manual for the initial charge:

“3.1 Initial charge

The new battery and the battery stored for more than 6 months should be charged at 0.25C 5 A for 13 hours, then discharged at 0.2C 5 A to 1.0V/cell, repeatedly the above charge and discharge for 3-5cycles. If the discharge duration isn’t less than 5 hours and the battery voltage isn’t less than 1.0V/cell, the battery can be put into operation.”

However Anna sent me an adjusted version:

What this means for my 800AH battery is that I must charge it at 160 amps for 13 hours and then discharge it to 40 volts. If it takes less than 5 hours to discharge it then do all it again.

I had tried pretty hard to get Anna to have the factory to do this but I had no success. I was now finding out why. However, she stressed that it was extremely important to do it. Where was I going to find a power source that would supply 160 amps at sufficient voltage for 13 hours. That is maybe 12kw to the battery at 100% duty cycle! Would probably pull 20KW or something! This would be something huge that would probably run on 480 volt 3 phase! I looked at all types of welders and plasma power supplies or whatever but nothing that I could find would work on 240 volt single phase that would do this.

Then I remembered an old (1940s vintage I guess) AC welder that my father had. That thing was very heavy duty and the dial went up to 200 amps. I didn't know what the duty cycle was but I wasn't going to assume that it was 100%. Also, It could not produce a high enough voltage under load to get more than about 55 amps to flow through the battery. Oh, did I mention that it was AC? I had ordered a 300 amp bridge rectifier from China on eBay. I don't know what the duty cycle of it was but I am pretty sure it is a long way from 100%. So, I mounted the rectifier on a good heat sink that I sat in a plastic pan of water with a little muffin fan on it. I took the welder out of the case and sat it on top of a table with a furnace blower blowing up through the bottom of the welder. So now I connected half of the battery at the time with some 1/0 cable. The welder could now produce more than enough current through the 20 cells. It was perfect. I set it to produce a little over 160 amps. It and the rectifier barely got warm the whole time. So this way it took twice as many hours but it worked very well.

Perhaps some of the newer welders with a big blower under them to get the duty cycle up would do it but I do not have that much product knowledge on welders.

For the resistive load I took two 14 foot 2x4s and put a steel bar across near each end of the 2x4s and ran 60 feet of 14 gauge nichrome wire back and forth 9 times to get the needed resistance to put a 160 amp load on the 20 cells and dissipate the energy. I had a piece of EMT I could slide up and down the wires to control the current. As the battery voltage came down the bar would have to be moved to keep the current around 160 amps. Other gauge wires are also available on eBay and you can use a nichrome chart to figure out what is needed. http://cecs.wright.edu/balloon/images/2/22/Nichrome_Wire_Heating_Element_Design_Basics.pdf

The battery really firmed up after the second cycle. I could tell a big difference in the voltage sag. The battery was providing over 1000 AH and still going but I still did it a third time though. The instructions did say 3-5 times not 2-5 times. This was a total of 6 13 hour intervals to get all 40 cells conditioned. The welder pulled about 53 amps at 245 volts. That is almost 13KW! So this thing ran for a total of 78 hours that is about 1,014 KWH!

It seems very likely that the reason many are not happy with the performance of their NiFi battery is because this necessary conditioning is not sufficiently done. I have seen little mention of it on the forums and it is a fair bit of trouble to come up with the equipment to do it right. Especially for larger cells.

It would be nice to use the 40 NiFe cells for the battery. However this puts the battery voltage out of range on the top end for the charge controllers and the 6848 charger/inverter. Fortunately 38 cells bring that voltage down to an acceptable level. It is still higher than the SE designers had in mind, it would seem, as this higher voltage creates an alarm condition at times. It is harmless and within established parameters of the equipment. Besides the custom software does not report the alarm.

By the way, I conditioned all 40 cells together so the two “spares” will be the same as the others.

Ciyi NI-FE battery FAQ http://www.ciyibattery.com/index/technical/id/155

Equipment Installation

There is nothing out of the ordinary for the installation of the arrays and equipment. I decided on the Midnight E-Panel. It is a nice smaller box that seems to provide everything I need and then some. The SE distribution panel was really big and I really didn't very easily have room for it. A secondary load sub panel was installed with the loads I wanted the system to carry. This is a laborious but necessary step and I could find no shortcut for it. I did install a relay for the water heater and control it with the axillary control output of the inverter. I connected the water heater to 120v to keep the inverter capacity available for other large loads. I have another relay connected so that I can turn on the 240 volt breaker in the main panel and switch the water heater to 240 volts from the grid if I wish.

System Software

Since the charge controllers, inverter, battery monitor etc. are not designed with the NiFe battery in mind I find that some customization is useful. Of course equipment parameters can be set to approximate what is needed by the NiFe battery. However, much more can be done for monitoring and control with the modbus interface and custom software.

I like being able to monitor voltage, current and power readings for all parts of the system in as close to real time as possible in a comprehensive and concise manner where all performance data can be seen with nearly a single gaze. The battery voltage indicator not only gives the total voltage for the battery but also provides a scale where the voltage per cell is also represented.

The software can decide when there is enough sun to take the house off-grid and how many amp-hours to use from the battery before putting the house back on-grid. It also controls when water is heated so as to heat water during peak sun if possible.

There is much more that can be done and much more to be learned. It is a work in progress.

Operating

Presently this system has about 6630 watts of PV panels. This seems to carry the house very well except for when there too many large loads. It will carry the heat pump very well when the sun is shining. When it is cloudy or night the battery will carry the heat pump very well too but it uses a lot of amp hours. Then there will have to be enough sun soon to recharge the battery. The system can put perhaps 250 amp hours or more back into the battery on a sunny day along with carrying the small house loads plus heat the water for the day. However, at night or cloudy days it is easy to use 400-500 amp hours in a day or night. This is more than can presently be put back in a single day. So I currently have it set so after 190 amp hours are used the system switches loads back to the grid. Then the next morning if current flow to the battery goes over 60 amps by 11:00 am and amp hours used are less than 190 the loads are switched back to the inverter. I expect that during the 4 months out of the year when I can leave the heat pump off this will look very different.

Since this is a nickel-iron battery we can discharge it much lower than an FLA. But how much would one want to discharge a NiFe battery? The Ciyi catalog says the battery is good for 3000 cycles at 50%. I wondered about other depths. I emailed Anna and according to Ciyi engineers 8500 cycles at 20% and 1200 cycles at 100% DOD are expected. So far, according to the battery monitor, I seem to get about 10 cycles per month. However I am not sure the battery monitor is correct. The chart shown here seems to approximate Ciyi's data.

The water heater is enabled at 11:00 each morning. If the load of the water heater causes current to be drawn from the battery for more than 5 seconds it goes off for 2 minutes then it tries again. This way water is heated only if there is enough sun to heat water without the battery. This is until 4:00 pm when the water heater is enabled and allowed to heat even if it has to get energy from the battery. Of course this is all test and trial at present. This works for me because I only use significant hot water for the evening shower.

About The Cells

As it turns out the 800AH cells use the same case as that used for the 1200AH cells. This means that there is considerably more electrolyte in the 800AH cells. There is 20Kg of electrolyte in the 800AH cell and 17Kg of electrolyte in the 1200 AH cell by weight according to the chart in their catalog. Anna tells me there is actually 18L or 21.6Kg of electrolyte in the cell. If by the chart the 800AH cell had the same amount by proportion as the 1200AH cell it would have 11.33 Kg instead of 20Kg. The Ciyi engineer confirmed that the 800AH cell will go much longer between electrolyte changes than the others. It also stands to reason that there should be longer times between watering.

I have also noticed that the case is large enough and the plate assembly small enough so that you can remove the cap and maneuver a small plastic tube over to the corner of the case and push the end down to the bottom so the electrolyte can be pumped out when it is time for an electrolyte change. Water can then be forced back down the tube under pressure to stir up any sediment from the bottom and then pump it out before refilling with new electrolyte.

The cell cases I have are made from MBS. The container materials listed are PP and MBS. Information from Beutilityfree.com suggest that PP (polypropylene) is not transparent enough to see the electrolyte level. MBS is much better and ABS is best for this but not as durable. However as I understand it they do not sell ABS since durability is more important for most of their customers. If I put a light behind the MBS CYNF800 cell, I can see the electrolyte level just fine on the other side.

Response to Gotchas

Thought I would respond a little to those “issues” mentioned earlier.

1) If I thought FLA and Lithium batteries were expensive. Well These things... Oh good grief!

It would seem as it turns out that they are available at prices that make them very similar in cost to good FLA batteries. I found a few companies in other parts of the world that sell these things at half the price charged for them in the US. It would be great to be able to support domestic suppliers and I am sure they provide a great service and can answer many questions and guide you every step of the way. Of course it was a domestic supplier that sold the set to my second reference. They apparently do not do the “initial charge”. The cells were not conditioned for him and they were unable to make the battery really work for him and they should have.

2) The NiFe voltage operating range is problematic for many charge controllers, inverters, converters, DC appliances etc.

There are 48 volt appliances I might like to use that do not have a high enough voltage rating in the specs. You would have to take a chance. Some may work and some may not. However, there is equipment that will work fine with this voltage range. I use a 30-75 volt DC to DC converter to get 12 volts for small peripherals. I am using an old MPPT charge controller I had to charge a 12 volt battery in another part of the house for larger loads such as the computer that runs the modbus HMI, radio equipment etc. Works quite well and surprisingly efficient too. It thinks the NIFE battery is a really big stable solar array. (Don't think I would try this with a PWM controller though)

So, this issue is not really a big deal.

Yes, but you need only one and there are a few to choose from that will work quite well. And fortunately these are the better rated ones anyway.

4) Charge/discharge efficiency‎: ‎65–80% (FLA 50-95%) Wikipedia

In practice, I don't really know how to put my finger on this anyway. On a good day the battery is in absorption or float a good bit. Then on less sunny days even when all the amp hours are not put back they seem to do fine. Would a few more panels or a slightly larger battery help? I don't really know how big of a deal this really is. Perhaps in time I will see. Besides, when an FLA gets old it is worse.

5) Really big and heavy for the amp hour rating.

For a vehicle this could be a problem but for a basement or out building not usually a big deal.

6) More voltage sag under load than FLAs.

This does not seem to really effect anything with the SE equipment. Set the inverter cutout voltage to the lowest setting. 40 volts on the 6848. It will smoothly invert throughout the voltage range. My other devices have also.

7) Self discharge rate 20–30%/month (FLA 3–20%/month) Wikipedia

I don't know if this is relevant in a solar application. The battery is being frequently recharged anyway.

Perhaps different sizes or brands of cells vary in this regard. The fact that these 800 AH cells have nearly twice the electrolyte for their capacity may be helpful. I don't really know.

As stated in the above, the larger electrolyte volume may be a factor. For watering, I use the Philadelphia Scientific Battery Watering system with their model X watering gun. This is deionized water which is mentioned as acceptable in the Ciyi battery manual. The TDS meter shows this to be to be as good as distilled water as far as conductive minerals are concerned. I have run across references that claim that deionized water will introduce less carbonate than distilled water. Something to do with storage I think but this information is a little sketchy. Anyway, this system is really easy to use. The watering gun automatically stops at the preset level. It is very fast although I like to take my time watering. I have installed a little water measurement device that tells me how much I put in each cell and then it will tell me the total for all the cells. Sort of neat I thought. Will look good in the battery log.

web counter