What battery monitor do I chose?

[mapmaker]

Am I correct in my understanding that a good battery monitor will provide me with that sort of information?

In terms of making decisions about when to run the generator, the battery monitor is quite effective. The biggest problem you will have is teaching the battery monitor when to reset to 100%.

As long as you get your batteries to 100% every week and reset your monitor, the battery monitor is quite accurate. Of course, you need to know when your batteries are at 100%, and that is not so easy with AGM batteries because you can't measure the SG. If you don't know when your batteries are at 100%, how can you teach your battery monitor to know?

You mentioned resting voltage as a way to monitor your battery's SOC. There is another way... end amps. In fact, that is how most shunt based monitors do it. The problem for you is to know the correct end amps. The only solution I know of is to follow the manufacturer's recommendation for charging protocol and hope for the best. Use your shunt based monitor to see what the end amps are when you have followed the manufacturer's recommendations, and then use that end amps to program your battery monitor.

You don't want to fall into the 99.9% trap. If your batteries are only getting to 99.9% SOC when you charge them, then after awhile you will be at 99.9% of 99.9% of 99.9% of 99.9%, etc. In other words you will be deficit charging and slowly losing capacity to sulfation. Plan to get flooded batteries some day.

--mapmaker

[Sunking]

I have slowly reached the same conclusion from all my reading up on monitors. I have AGM batteries so cannot use hydrometer. I do not want to each day disconnect my battery bank from PV and inverter, let rest for about four hours, then use a voltage tester to check for SOC. I have been told that I can fairly accurately estimate SOC from the resting voltage. E.g. For AGM batteries 51.6V=100% SOC and 49.6V=50% SOC. (Information provided by the dealer who sold me the batteries).

I believe that battery voltage shown during charging or discharging is not accurate which is why the batteries need to "rest" . The information I have been reading about battery monitors seems to claim that they use some sort of algorithm to more accurately show battery voltage in real time without the need to rest the batteries.

It is all marketing hype BS. With AGm the only useful test instrument to judge the battery health is a Battery Conductance Meter. They cost a few thousand dollars and do not really tell you much other than the internal resistance. Great tool if you have new batteries. YOu charge up your batteries, measure the Internal Resistance and you have a base line. Down the road you notice Internal Resistance is going higher. That tells you the batteries are done and it will not be long before they need replaced. Of course you already knew that because the batteries are no longer able to keep up.

The reason voltage is useless on a working system is the batteries Internal Resistance period. For example take a brand new fully charged 12 volt 100 AH battery, and apply a 500 watt load (40 amps). Before you apply the load you measure 12.6 volts on the battery post. Apply the load and the battery voltage sags to 11 volts. Well that tells your volt meter or monitor your battery is completely discharge and ready for the trash can when in reality nothing is wrong and is working perfectly. It means your battery has an internal resistance .04 Ohm's which is about normal for a 12 volt 100 AH battery.

Take the same battery. except this time it is 50% DOD with an open circuit voltage of 12.1 volts. Apply a 40 amp charge current and the battery voltage will read 13.7 volts. That would tell a meter your battery is above fully charged of 13.7 volts when in fact your battery is 50% DOD. That same .04 Ohm's Internal Resistance is the cause of the error.

Only a Hydrometer with a thermometer can give you real time SOC. Since you have AGM, you really do not have solutions other than voltage which is not telling you a whole lot unless your load currents are C/10 or less. AGM batteries have fairly low Internal Resistance, and with lightly loaded AGM you can get in the ball park with a SOC voltage reading. Bu tit doe snot take a expensive meter to do that, just a decent DMM or Panel Meter will do that.

[Sub]

Thanks

In terms of making decisions about when to run the generator, the battery monitor is quite effective. The biggest problem you will have is teaching the battery monitor when to reset to 100%.

As long as you get your batteries to 100% every week and reset your monitor, the battery monitor is quite accurate. Of course, you need to know when your batteries are at 100%, and that is not so easy with AGM batteries because you can't measure the SG. If you don't know when your batteries are at 100%, how can you teach your battery monitor to know?

You mentioned resting voltage as a way to monitor your battery's SOC. There is another way... end amps. In fact, that is how most shunt based monitors do it. The problem for you is to know the correct end amps. The only solution I know of is to follow the manufacturer's recommendation for charging protocol and hope for the best. Use your shunt based monitor to see what the end amps are when you have followed the manufacturer's recommendations, and then use that end amps to program your battery monitor.

You don't want to fall into the 99.9% trap. If your batteries are only getting to 99.9% SOC when you charge them, then after awhile you will be at 99.9% of 99.9% of 99.9% of 99.9%, etc. In other words you will be deficit charging and slowly losing capacity to sulfation. Plan to get flooded batteries some day.

--mapmaker

Thanks for that, mapmaker, and also for your previous post. I will have to take some time to understand it all!

Unfortunately we are not all that savvy on technical things, not that we are unintelligent, just that, as I am now aged 72, I did not grow up in an era where computers and related technology were everyday things. My 12 yr old grandson just "knows" this stuff, it is part of his everyday life.

Re your comment "plan to get flooded batteries some day", is not economically practical here in NZ. All deep cycle batteries are imported, resulting in AGM batteries from China being considerably cheaper than FLA. It cost us $5900 including installation, for eight 12V AGMs and the new controller. It would have cost me over twice that to install FLAs of similar capacity. A good quality 6V 600ah FLA costs around $1700 each here! We are retired and live on a pension and that $5900 was a considerable stretch. The grid does cover our area, but our road takes a big horseshoe bend to get around a ridge, we are on the toe of the horseshoe, and the power lines cross the heel of it, some half mile away - be more than $25K to connect to it.

I have read many posts which say that buying grid power is cheaper than installing an off grid system. It seems that we pay more for electricity in NZ that the USA, so that claim may not be totally true for us. Before we moved here 4 years ago, we were paying on average $180 a month for electricity (for two adults). Friends, also an adult couple, and who are fairly careless about conserving power, pay over $300 a month!

The original install, according to the people we bought the property off, cost $13,000, and lasted 8 years and three months before we replaced the batteries and controller. That comes in at around $135 a month.
Even if we only get five years from the new AGM batteries, that is about $100 a month. We are very careful with power use, so hope to get 7 years use before replacing.

Sorry, much of above off topic! Back on - from reading your posts, and those of Sunking, I am now wondering if it would be so much easier to stay with the set up we know - using the analogue voltage meter! Although I do like the idea of having the display for a battery monitor on an wall inside the house - much more convenient than having to go outside in the dark and inclement weather to check the present set up.

John

PS - re our discussion on another topic, I am still waiting for my son in law to send me a new voltage tester/meter. He wants to send me a good quality one, has it on back order with the dealer.

[Sunking]

I have (and like) the Trimetric 2025A. They are widely available for less than $160. Do NOT get a 100 amp shunt... you want a 500 amp shunt on a system the size of yours. Doesn't cost any more than a 100 amp

I do not see how that is remotely possible. Most all Shunts work on 50 mv scale and a 100 amp shunt is basically a .0005 Ohm 5 watt resistor. A 500 amp Shunt is .0001 Ohm 25 watt resistor with 5 times the mass and materials thus escalating cost significantly.

As for 100 or 500 amp depends on what is your full scale or full plant power. I cannot see any system designed properly ever going over 100 amps. Sure you could use a 500 amp shunt on a 100 amp circuit but you loose accuracy and resolution doing so. a 5000 watt panel system operating @ 48 volt battery with a monster size 5 Kw inverter only touches 100 amps at full inverter power. It would be a vary rare occasion to use 5000 watts from a battery inverter for any length of time.

[mapmaker]

I do not see how that is remotely possible. Most all Shunts work on 50 mv scale and a 100 amp shunt is basically a .0005 Ohm 5 watt resistor. A 500 amp Shunt is .0001 Ohm 25 watt resistor with 5 times the mass and materials thus escalating cost significantly.

As for 100 or 500 amp depends on what is your full scale or full plant power. I cannot see any system designed properly ever going over 100 amps. Sure you could use a 500 amp shunt on a 100 amp circuit but you loose accuracy and resolution doing so. a 5000 watt panel system operating @ 48 volt battery with a monster size 5 Kw inverter only touches 100 amps at full inverter power. It would be a vary rare occasion to use 5000 watts from a battery inverter for any length of time.

The default shunt size for the trimetric is 500 amps. Midnite's ePanels come with a 500 amp shunt and their "whizbangJr" is designed for a 500 amp shunt.

from Midnite's web site:

The Whiz Bang Jr is a current sense module that attaches to the standard 50mv / 500A Shunt and wires into the Classic and KID charge controllers to give Amperage readings from the shunt.

You are correct that the smaller shunt will give better resolution, but the 500 amp shunt seems good enough and is fairly standard.

I have just done a quick internet search of several vendors... the 100 amp and 500 amp shunts are about the same price.

--mapmaker

[Sunking]

The default shunt size for the trimetric is 500 amps. Midnite's Panels come with a 500 amp shunt and their "whizbangJr" is designed for a 500 amp shunt.

I understand the units come with 500 amp shunts but have no idea why that high because none of their equipment is made for 500 amps.

[Mike90250]

I understand the units come with 500 amp shunts but have no idea why that high because none of their equipment is made for 500 amps.

It's to monitor the peak current from the battery to the inverters too. Not just the charge controllers.

[mapmaker]

I understand the units come with 500 amp shunts but have no idea why that high because none of their equipment is made for 500 amps.

I think at least part of the reason is temperature... the 500 amp shunts heat up less. There's a fellow from Australia (hasn't posted here in quite awhile, posts occasionally at NAWS) who claims that shunts make very good precision fuses.

--mapmaker

[ryan.turner]

I think at least part of the reason is temperature... the 500 amp shunts heat up less. There's a fellow from Australia (hasn't posted here in quite awhile, posts occasionally at NAWS) who claims that shunts make very good precision fuses.

--mapmaker

Operating temp for shunts are pretty high -- 100A or 500A. In this case, these affordable shunts operate at -40 to 60C. Note that as ambient temp increases, the current rating decreases significantly.

Higher current shunts have lower resistances. The lower the shunt resistance, the lower the voltage drop. This doesn't really matter, though. For my circuit I'm using INA226, which can take -81.9175 to 81.92 mV. These are all 50 mV shunts; so the limit isn't the current monitoring circuit.

As for accuracy of the 100 vs 500, the stated accuracy is ±0.25% max for both specific models (1.25 A for the 500A, 0.25 A for the 100A). The voltage drop in the leads should be accounted for too or your values will be off.

The price difference between 100 and 500A shunts is very small. Considering that 2500kW at 12v is 208 amps, a 200A shunt would be too small. So, a few extra dollars makes it where 1 product can fit more potential users. The price difference today in quantity 1 is $6 from 100A to 500A.

The shunt's rated currently is not supposed to be continuous; for the one I linked, continuous operating current rating is actually 2/3s its value. For operating above 25c its even further devalued according to the doc. My ambient temp this weekend around peak was 38c; from this, the max continuous operating current is 380 A. This is still plenty of capacity, but if I'd used a 100A shunt, the max continuous current would be down to 76A -- too little to drive a 1000 watt 12v inverter!

[russ]

I'll release it all open source and make as post about it in a few nights.

Hi Ryan - Welcome to Solar Panel Talk!

Russ

[bcroe]

I think at least part of the reason is temperature... the 500 amp shunts heat up less. There's a fellow from Australia (hasn't posted here in quite awhile, posts occasionally at NAWS) who claims that shunts make very good precision fuses. --mapmaker

Those must be cheaply made shunts. The ones we used in the DC power industry were made of
big strips of constantine soldered into brass blocks. One was a hefty handful. Not much danger
of overheating, and not to be bought for a couple bucks either. Bruce Roe

[Sunking]

Considering that 2500kW at 12v is 208 amps, a 200A shunt would be too small.

What fool would run 2500 watt panels on a 12 volt battery? Anyone with common sense would run either 24 which is still low of a voltage at 2500 watts, or better yet 48 volt battery where the current is a reasonable 45 amps. Once you get above 50 amps things start getting risky even for a pro. Not to mention the huge expense of large copper cables.

[ryan.turner]

What fool would run 2500 watt panels on a 12 volt battery? Anyone with common sense would run either 24 which is still low of a voltage at 2500 watts, or better yet 48 volt battery where the current is a reasonable 45 amps. Once you get above 50 amps things start getting risky even for a pro. Not to mention the huge expense of large copper cables.

I'm not advising to do one thing or another, but products like that do exist. The best seller amazon high wattage inverters (which happen to be MSW and cheapos) are 12v. Higher quality, Xantrex makes some XPower high wattage (3000, for example) that operate on 12v.

[Colt45]

I also recommend making your own but not knowing your abilities it is hard to point you in that direction. As the above post has stated a system that can monitor the entire picture is a better solution.

My recommendation is to look into the Arduino platform if you are new to working with micro controllers. I am also a fan of the "BeagleBone Black", but it does more and requires more coding on your part. If building things is something you like check it out. It's not hard (not fast and you have a lot to learn, but not hard).

[ryan.turner]

I also recommend making your own but not knowing your abilities it is hard to point you in that direction. As the above post has stated a system that can monitor the entire picture is a better solution.

My recommendation is to look into the Arduino platform if you are new to working with micro controllers. I am also a fan of the "BeagleBone Black", but it does more and requires more coding on your part. If building things is something you like check it out. It's not hard (not fast and you have a lot to learn, but not hard).

Thanks! I'm a software engineer who has done some embedded work in the past. I chose raspberry pi just because I already have a few on hand, they speak i2c natively, and I prefer working with the linux libs rather than arduino libs. The fewer parts the better.

What do you mean monitor the entire picture? What more is there that some basic sensors can read besides voltage across the battery & current through it? The rest is in the software. The prior discussion about the inaccuracies of some specific implementations is fine; in the end all of the systems I've seen so far are based on this same hardware.