Author Topic: Lead-acid batteries - SoC vs voltage  (Read 10069 times)

0 Members and 1 Guest are viewing this topic.

Online johnd

  • Forecaster
  • *****
  • Posts: 2707
    • www.weatherstations.co.uk
Lead-acid batteries - SoC vs voltage
« on: October 26, 2011, 04:39:05 AM »
Does anyone here have any good insight (ie solid personal experience or access to definitive literature) on how the open-circuit voltage (OCV) of flooded deep-cycle lead-acid batteries varies with their SoC (State of Charge). I'm finding that a lot of the data on the web is either confused or very variable - even from apparently authoritative sources such as battery makers - or both.

This isn't just an academic nicety - running a remote solar-powered AWS site in more northerly parts it becomes quite important in winter to know the OCV at which the battery is starting to become seriously depleted so that the system can cut out to prevent damage to the battery.

I think there's reasonably general agreement that an OCV of around 12.7v (obviously measured when there's no solar charging happening) represents around 100% SoC. But it's at the other (0-10% SoC) end of the scale that there's a lot of disagreement. Some sources seem to think that about 11.8v is the lowest voltage that the battery should be allowed to go to, while others talk about 11.5v and even discharge down to 10.5v still allowing a deep-cycle battery to recover adequately. So it's particularly at the bottom end where any further insight would be useful.

I've currently got the cellular modem set to drop out at 12.0v and the Envoy/WLIP PSU drops at 11.5v (the idea being that the modem takes most power so that if it drops at significantly higher voltage then it will leave quite a significant reserve of battery power for the 0.5Whr consumption of the Envoy/WLIP to carry on. So logging should be able to continue uninterrupted even when the live long-distance data feed drops out.)

So the question here is whether my choice of eg 12.0v (representing I'm guessing 20-25% SoC) and 11.5v (5-10% SoC) are sensible numbers to choose or if anyone has a better suggestion?

NB This is with a leak-proof battery. It's not AGM-type but TBH I can't remember whether it's just a leak-proof enclosure or a gel type.

(Sorry if this seems rather an esoteric question, but it does actually become quite important in the middle of winter when you're trying to maintain a flow of live data against often very limited solar power generation.)
« Last Edit: October 26, 2011, 05:01:58 AM by johnd »
Prodata Weather Systems
Prodata's dedicated Davis EnviroMonitor website
UK Davis Premier Dealer - All Davis stations, accessories and spares
Littleport, Ely, Cambs UK

Offline SLOweather

  • Administrator
  • Forecaster
  • *****
  • Posts: 3404
    • http://www.sloweather.com
Re: Lead-acid batteries - SoC vs voltage
« Reply #1 on: October 26, 2011, 11:46:41 AM »
Hi John,

First, I wouldn't even mess with trying to roll my own low voltage disconnect. When I was designing and building small scale solar systems for municipal utility telemetry sites, I usually included a commercial device similar to this:



Per the spec sheet:

 
Quote
Continuously monitors voltage level of a 12 volt power input (terminal 86)
 If voltage is above 13.25 - power switch will turn on to supply up to 15 amps on the output terminal to power the loads
 If voltage drops below 12.8 - time delay is started
 If voltage remains below 12.8 until the timer expires—power switch will turn off, disconnecting the power to the loads
 If voltage drops below 11.8 with the timer running - power switch is shut off immediately
 Any time input voltage increases to above 13.25—power switch will turn on to supply power to the loads and timer will reset

So, there are the voltages that the manufacturer uses.

Second, if you really want to do your own research and disconnect, I'd do an experiment.

Get a brand new gel-cell or AGM battery and fully charge it. Apply a load sufficient to draw a current equal to 1/20 or so of the rated capacity. A lamp would be fine.

Measure the current over time and do a running calculation of the amp-hours until the battery is depleted to 50%. That's my guideline for maximizing battery life and the time between battery replacements.

Measure the battery voltage at that time with the load on. Disconnect the load and wait a few minutes and remeasure the battery voltage. It might be interesting to measure the voltage recovery fairly often until it levels off and graph it.
« Last Edit: October 26, 2011, 12:11:19 PM by SLOweather »

Online johnd

  • Forecaster
  • *****
  • Posts: 2707
    • www.weatherstations.co.uk
Re: Lead-acid batteries - SoC vs voltage
« Reply #2 on: October 26, 2011, 01:05:08 PM »
Chris, sorry I should have provided a few more background details in the post:

This is an installation which is already up and running.

(Actually the data is here: www.weatherlink.com/user/elydev )

Last winter there just wasn't enough solar to keep it running without drop-outs until spring, so I've made some upgrades over the summer but obviously the season of minimal solar is approaching again and so I'm aiming to monitor solar, voltages etc more closely this year.

The upgrades have been to double the panel power to 60W (battery is still around 85Ah IIRC) and to put in an MPPT solar regulator, which hopefully should help a little. (The regulator obviously provides overvoltage etc protection on its own.)

I have also already split the power supply to the Envoy/WLIP (5v derived from the 12v battery supply via a Dimension Engineering DE-SW050 high-efficiency converter) and cellular modem (12v) and run each supply through a separate adjustable voltage drop-out switch, which I managed to find available commercially and quite cheaply here in the UK. Currently as per my initial post the modem is set to drop at 12.0v (IIRC) with the logger not until about 0.5v lower (eg 11.5v). This is all installed and has been operating fine over the summer, which is good but unsurprising.

So, given the time of year, I'm just starting to look back at the details of this set-up and wondering whether I've chosen the voltage dropout points as well as I could. Of course, time will tell over the winter but any new background insights would also be a useful extra input.

The overall problem here at 52N (with often quite persistently overcast conditions in midwinter) is that to specify a solar system that could provide guaranteed power and with something in reserve too would be too costly. In dollar terms, you might be looking at having to spend say $1500 on the solar PSU to be sure of getting a continuous feed of live data back from a $500 weather station. So instead we have to try to be more clever about how to make a $500 solar PSU go as far as possible. (Of course another answer would be to add a wind turbine into the mix, but that would cost a separate/additional $500+ and would add further practical complications of its own.)
« Last Edit: October 26, 2011, 01:06:41 PM by johnd »
Prodata Weather Systems
Prodata's dedicated Davis EnviroMonitor website
UK Davis Premier Dealer - All Davis stations, accessories and spares
Littleport, Ely, Cambs UK

Offline Bushman

  • Forecaster
  • *****
  • Posts: 6881
    • Eagle Bay Weather
Re: Lead-acid batteries - SoC vs voltage
« Reply #3 on: October 26, 2011, 01:06:59 PM »
Why not just add batteries?

Online johnd

  • Forecaster
  • *****
  • Posts: 2707
    • www.weatherstations.co.uk
Re: Lead-acid batteries - SoC vs voltage
« Reply #4 on: October 26, 2011, 01:20:45 PM »
Why not just add batteries?

Yeah, it's a good point and it may well be that the 85Ah that I currently use could usefully be increased, maybe eg doubled.

But with any solar PSU like this there's a balance between generating capacity (is solar panel wattage) and storage capacity. Whatever (realistic) battery capacity you put in, it's going to run out before too long if there's not enough charging capacity. And once it's run out then a 200Ah battery is not going to reacquire charge any faster than a 100Ah one when charging capacity is limiting. Plus, good deep-cycle batteries are not cheap themselves and so that's a factor also.

But I don't disagree - the balance I currently have between solar and battery capacity may not be optimal - not easy though to work out what the optimal balance should be.

Overall, I will surprised if - for a given level of cost - it's possible to keep this system feeding live data continuously throughout the winter. But I would like (i) to keep the uptime as high as possible; and (ii) to keep the system logging even if the live data feed drops out at times. And using the optimal voltage drop-out points is going to be key to achieving this AFAICS.

Prodata Weather Systems
Prodata's dedicated Davis EnviroMonitor website
UK Davis Premier Dealer - All Davis stations, accessories and spares
Littleport, Ely, Cambs UK

Offline Bushman

  • Forecaster
  • *****
  • Posts: 6881
    • Eagle Bay Weather
Re: Lead-acid batteries - SoC vs voltage
« Reply #5 on: October 26, 2011, 01:26:30 PM »
Over here, a deep cycle batt is about a buck/amp hr.  No need for AGM or spiral in your application.  I would think that a larger capacity  would allow for longer discharge periods.  I would also look at a smaller bank of say 100 amp hr batts with auto switching between them.

Online johnd

  • Forecaster
  • *****
  • Posts: 2707
    • www.weatherstations.co.uk
Re: Lead-acid batteries - SoC vs voltage
« Reply #6 on: October 26, 2011, 01:44:07 PM »
Over here, a deep cycle batt is about a buck/amp hr.

Here it's more like $1.50/Ah but sure not too different and not a huge cost for eg 100Ah but not inconsequential either.

Obviously if you double the capacity then you double the runtime before you have to recharge. But what I'm saying is that once the batteries are significantly discharged then the limiting factor becomes the panel power not the battery capacity.

To put some numbers on this, the installation probably consumes say 10-20Ah daily. So even a fully charged 100Ah battery is only going to last a matter of days extra, whereas the winter period lasts eg 90-120 days. So battery capacity is only one factor.
Prodata Weather Systems
Prodata's dedicated Davis EnviroMonitor website
UK Davis Premier Dealer - All Davis stations, accessories and spares
Littleport, Ely, Cambs UK

Offline SLOweather

  • Administrator
  • Forecaster
  • *****
  • Posts: 3404
    • http://www.sloweather.com
Re: Lead-acid batteries - SoC vs voltage
« Reply #7 on: October 26, 2011, 02:10:58 PM »
In an interesting coincidence, I'm going through a similar design exercise with the 5 station ranch project, and I recall a previous thread where we discussed panel tilt angles and such.

I've never cut a design so close as to have to worry about such fine tuning of the voltage dropouts, and you are using WLIP, which limits my other recommendations. I think your choices are as good as any.

Forgive me if this has been answered before, but have you measured the current draw of everything as it was in use so the daily amp-hour need is known?

Are there any power reducing features on the cell modem you have missed? Sometimes there is a setting or jumper that kills the LEDs, which can save considerable power.

Does the WLIP have a programmable upload time, and if so, will the application be OK with a longer time between transmissions? The cellular transmitter is likely the biggest single power drain.

(One of the things we programmed into the WeatherElement data hub is the ability to change the update rate remotely from the server.)

I had some other ideas, but the system design negates them. It does give me some ideas for possible WeatherElement enhancements.

I'm just gonna say empirically, I think you are OK. I measured my solar powered WE station at about 7AH per day. Your site is not that different, although I think maybe the WLIP might draw a little more than my data hub. 60 watts on the panel is an optimistic 5 peak amps. Even just with 2 hours of sun, and 1.1 fudge factor for battery charging, you should be good to charge on a moderately sunny day.

Draining an 85 AH battery to 25% uses about 63 AH. At an estimated 7AH per day, that should run the station for an optimistic 9 days in the dark (negating derating for cold temperatures).

 But, only time will tell... :)

One last thought for next year. Surrounding the panel with some shiny polished metal angled to reflect sunlight onto it would increase its output inexpensively.

Offline SLOweather

  • Administrator
  • Forecaster
  • *****
  • Posts: 3404
    • http://www.sloweather.com
Re: Lead-acid batteries - SoC vs voltage
« Reply #8 on: October 26, 2011, 02:19:45 PM »
To put some numbers on this, the installation probably consumes say 10-20Ah daily.

This post came in while I was writing my last reply. This is the number that you really need to nail down to know how to size the solar system. I spent a lot of time measuring my hub hardware and the cellular router before deploying the solar mountaintop site. And I've been doing the same for the ranch, for all of the solar powered hardware.

Don't trust what the spec or data sheets say. They usually overstate/round up, but sometimmes they are low. And it's key to measure everything assembled and in normal use. Modem current goes up during transmit. In my case, switch current goes up depending on how many ports are in use.

RE: the modem, read the receive current and the transmit current and then calculate an average current draw based on the duty cycle.

Offline Flag

  • Forecaster
  • *****
  • Posts: 609
Re: Lead-acid batteries - SoC vs voltage
« Reply #9 on: October 26, 2011, 07:03:47 PM »
This type of WLIP scenario requires the biggest capacity (and separate) as the backup to maintain the WLIP/data. This power supply/battery does nothing else except provide backup when all the other devices have depleted power and have shutdown or been disconnected via LVD etc (from their own separate power/battery/panel)

Big balancing act, how much power is enough? how many batteries are enough? BUT at the end of the day the thing is to still have that logger running and data stored for download when things come back up. Even the archive interval has to be linked to the expected length of power outage.

Even with a few hours of sun can be enough to raise the charge to switch things back on, upload then die again but generally this re-instate voltage is considerably higher than the cutout voltage.

  
« Last Edit: October 26, 2011, 07:05:19 PM by Flag »

Online johnd

  • Forecaster
  • *****
  • Posts: 2707
    • www.weatherstations.co.uk
Re: Lead-acid batteries - SoC vs voltage
« Reply #10 on: October 27, 2011, 09:28:03 AM »
To put some numbers on this, the installation probably consumes say 10-20Ah daily.

On reflection I've probably overstated this figure a little - it's maybe more in the range 7-10Ah daily, but the same principle obviously applies.

Quote
RE: the modem, read the receive current and the transmit current and then calculate an average current draw based on the duty cycle.

This is the trickier part. It's quite easy to log voltage inexpensively which is why that's what I've been focusing on voltage measurements on the battery - an EL-USB-3 Voltage Data Logger (around $60 here) lets me log collect 5-min voltage data for extended periods - not online admittedly but long as I visit the site at intervals I can collect the data easily enough.

In a way, I guess I'm going through this design/evaluation process pragmatically and in real time. I've guessed some suitable sizing parameters and I'm now following the voltage profile on the battery week by week. Where I came in with this thread was in trying to interpret the voltage figures as the battery gets progressively depleted and in trying to judge the point at which the battery might be in danger of damage.

But back to your point: It would be great if I could think of a way of monitoring actual current consumption in the same way that I'm logging voltages, but I can't identify a simple/inexpensive way of doing so. (Actually I have experimented with a low voltage current clamp linked to a millivolt logger but it's not working out very well as yet.) I'm sure that there must be a better way of doing it and that I just haven't identified the optimum hardware yet, but I'm not an electronics engineer and limited to scouring the web in my amateur way for ideas.

The problem is that the current drain by the modem isn't constant but varies according to the transmission state of the modem and I don't have any good handle on the timing of the duty cycle - the current consumption spikes up substantially during transmission but too quickly really to get a fix on by eye and with simple equipment. I possibly need something like a memory scope (that is inexpensive and will run off 12v so that I can use it as the field site :lol:). Probably the duty cycle is very reproducible once you can get a fix on it - just a question of identifying the size and timing of the consumption spikes and then you could probably get quite a good estimate of the overall power consumption. Any other ideas would be very welcome.

And just to comment on the broader question: What I'm trying to do with this site is to relay genuinely live data. OK it's currently provisioned with a WLIP logger so 'live' can't be better than 1-minute updates, but its purpose is not to be powered down then up again at intervals (or at least I'm sure that it's worth cycling a cellular modem at short intervals of less than 1-2 minutes because of the time it takes to re-register on the network). I totally appreciate that data at eg 15 or 60-min intervals may be perfectly adequate for many remote data applications - I'm just interested in what might be possible for genuinely live data. From a Davis POV presumably the Connect product may be suitable for interval data once it appears (still looks like 4-6 months away AFAICS) and of course there are other excellent alternatives ;)
« Last Edit: October 27, 2011, 10:02:07 AM by johnd »
Prodata Weather Systems
Prodata's dedicated Davis EnviroMonitor website
UK Davis Premier Dealer - All Davis stations, accessories and spares
Littleport, Ely, Cambs UK

Offline SLOweather

  • Administrator
  • Forecaster
  • *****
  • Posts: 3404
    • http://www.sloweather.com
Re: Lead-acid batteries - SoC vs voltage
« Reply #11 on: October 27, 2011, 12:02:32 PM »
Sorry I wasn't clearer, John. I meant to do all of that measurement and calculation as a part of the design or review process, not read it live remotely and try to do something with the information reactively. I learned that after many design cycles at my past employment.

The next time you go out to the site, take a DMM and make a few readings and you will have a much better idea of what you are working with. If you can buy or borrow a "peak hold" feature, that will make reading the transmit current of the modem that much easier. Or, since you are using WLIP, you must have an Ethernet jack on the modem or router. Plug in a laptop and do something like upload a file that will key the xmitter long enough to get a good current reading.

My WE data hub sends a long enough data stream, (about 350 bytes) that on the bench I was able to get a good xmit current reading off the BlueTree router, as well as a good timing of the xmit cycle with a stopwatch.

If you need faster than one minute updates, I've tested WeatherElement hubs as fast as every 10 seconds on VP2 consoles. There shouldn't be any reason they wouldn't go faster. I suppose I should test that sometime just to know what the top limit is.

Of course, faster updates = more current draw = bigger solar power system. :) And more records in the database, too. Every 10 seconds would be 8,640 per day or 3,153,600 per year!

Offline Weather Display

  • Forecaster
  • *****
  • Posts: 2611
    • West Coast Road Weather Data
Re: Lead-acid batteries - SoC vs voltage
« Reply #12 on: October 27, 2011, 11:53:12 PM »
to make the deep cycle batteries last, I would not let them go below 40% SOC
and also if they are lead acid, they will need a period of rapid charge to stop them from stratifying out
also a equalization charge (i.e sustained just over charging) to get them all up around 100% SOC every few months is also good too
Brian
info@weather-display.com
http://www.weather-display.com

Online johnd

  • Forecaster
  • *****
  • Posts: 2707
    • www.weatherstations.co.uk
Re: Lead-acid batteries - SoC vs voltage
« Reply #13 on: October 29, 2011, 04:16:16 PM »
to make the deep cycle batteries last, I would not let them go below 40% SOC

That's interesting. Aren't deep-cycle batteries designed to be cycled deeply (by which I'd assume down to 20% SoC if not lower)? Maybe I've misunderstood how they're intended to be used?

The problem about charging with the set-up at a remote site is that it's quite difficult or at least pretty inconvenient to do other than allow the combination of solar input and MPPT regulator to do their thing and, ultimately, if all else fails,to up the specification of the solar PSU. Charging is routinely at up to 14v - limited by the regulator. Anything much beyond say once per year maintenance wouldn't be very convenient.

As per the start of this thread, I'm logging the battery voltage and now that October is wearing on and daylight hours heading towards a minimum I'm taking more interest in the details.

Thus far, the solar PSU is holding up OK. It's still the case (at least up until the last voltage download I did a few days ago) that on sunny days the battery is recharging to 12.8-12.9v (reading taken after dusk when charging has finished for the day), which I take to be essentially 100% SoC. And by dawn (after probably 14 hours of no charge) the voltage is typically down by 0.2v which is maybe a 10-15% fraction of total battery capacity. But this is then recouped most days - remember I'm talking about the current month of October here, the situation will deteriorate towards Christmas -  to restore 12.8-12.9v provided there's been a total solar energy input of say 6MJ/sqm or more (the VP2 is a Plus model and so I can get actual on-site TSE figures).  

Of course some days are seriously overcast and the TSE is no more than 3MJ/sqm. Under these latter conditions there's no more than a 0.1v recharge of the battery (in addition of course to powering the cellular data feed during daylight hours). So there's certainly the potential for the battery to deplete by a net 5-10% daily if there's an extended run of overcast days (this is England remember!) and we've not yet got close to the shortest day.

If anyone's interested I'll maintain an occasional update on this thread of observed voltage vs TSE as winter wears on. I think that I may be able to predict that if we get a run of 10 days with an average of no more than 3-4MJ/sqm TSE daily then we'll get to the point where the live feed over the cellular network will drop out, but we'll see.

Edit (Sunday 30th): The past 7 days battery & weather data downloaded and reviewed today: Charge/discharge pattern essentially unchanged; lowest dawn battery voltage still 12.5v (14th, 18th and 28th Oct) but yesterday, for example, 6MJ/sqm TSE still restored the battery to 12.9v by dusk. But I'm afraid that we'll see an increasing number of days like Thursday 27th when only 1.8 MJ/sqm TSE was measured and contributed precisely 0.0v to the battery voltage by the end of the day (but at least we didn't end the daylight hours worse off than when we started!).
« Last Edit: October 31, 2011, 10:09:52 AM by johnd »
Prodata Weather Systems
Prodata's dedicated Davis EnviroMonitor website
UK Davis Premier Dealer - All Davis stations, accessories and spares
Littleport, Ely, Cambs UK

Online johnd

  • Forecaster
  • *****
  • Posts: 2707
    • www.weatherstations.co.uk
Re: Lead-acid batteries - SoC vs voltage
« Reply #14 on: November 06, 2011, 01:43:23 PM »
Not sure whether this is of any interest to anyone here, but here's a picture of the control panel for the remote VP2 installation referred to in this thread. This uses a solar panel PSU and uploads data from an IP logger via a cellular modem. There's a 12v supply for the modem and a 5v supply for the Envoy, each fitted with automatic low-voltage cut-outs. Probably looks a little Heath-Robinson and more complex than it needs to be but that's partly because I'm looking to log voltages at different points in the circuitry (for troubleshooting as and when the solar gives up in midwinter) - it's surprising how needing to find attachment points for logger probes ends up complicating all the wiring.

Edit: Just added a second image that shows the control panel in context (inside a wooden shed to provide some shelter - for me and the equipment!). Battery is lower right. The cellular modem is on the desk below the panel and there's a Vue console to the right used for spot readings and data backup.
« Last Edit: November 06, 2011, 01:52:50 PM by johnd »
Prodata Weather Systems
Prodata's dedicated Davis EnviroMonitor website
UK Davis Premier Dealer - All Davis stations, accessories and spares
Littleport, Ely, Cambs UK

Offline Weather Display

  • Forecaster
  • *****
  • Posts: 2611
    • West Coast Road Weather Data
Re: Lead-acid batteries - SoC vs voltage
« Reply #15 on: November 06, 2011, 01:50:58 PM »
Quote
That's interesting. Aren't deep-cycle batteries designed to be cycled deeply (by which I'd assume down to 20% SoC if not lower)? Maybe I've misunderstood how they're intended to be used?

we use a SMA (leaders in the field) Sunny Island
and its default setting is to start up a generator at 40% SOC
Experts I have talked to say that batteries should not be allowed to go below that
otherwise you start getting ionisation of the plates (lead acid type battery)
Short periods might be OK (a few hours or so) though
Brian
info@weather-display.com
http://www.weather-display.com

Offline SLOweather

  • Administrator
  • Forecaster
  • *****
  • Posts: 3404
    • http://www.sloweather.com
Re: Lead-acid batteries - SoC vs voltage
« Reply #16 on: November 06, 2011, 01:57:21 PM »
There are probably 2 basic modes or uses for "deep-cycle" batteries.

In occasional "emergency" use, like in a UPS, or as another emergency supply during a power failure, you could probably discharge the battery down to 10% and still be able to use it a few more times.

However, in anything that requires regular cycling, like a daily charge/discharge cycle in solar use, max battery life is attained by not dropping below 50% or so before recharging.

In either case, I replace them after 3 years, 4 years max.

Offline Weather Display

  • Forecaster
  • *****
  • Posts: 2611
    • West Coast Road Weather Data
Re: Lead-acid batteries - SoC vs voltage
« Reply #17 on: November 06, 2011, 04:51:18 PM »
yeah, we use 45% as the cut in for the generator instead of the default 40%, to help make the batteries last longer
Brian
info@weather-display.com
http://www.weather-display.com

 

anything