Author Topic: Battery strings (monitoring of)  (Read 6856 times)

0 Members and 1 Guest are viewing this topic.

rossw

  • Hero Member
  • *****
  • Posts: 834
  • Country: au
Battery strings (monitoring of)
« on: April 04, 2007, 01:15:46 AM »
We (the players: Zubbly, Tom, DonC, Jam and myself) were having an interesting and animated discussion about dumpload controllers on irc recently.


One train of thought was to use a standard 12V dumpload controller across only one battery in the 60V string. This had the (clear) advantage of not needing to modify the controller and being easy.


There is a problem though, in that you are sampling only one battery in the string and with non-identical characteristics, you risk damaging the whole string because of one battery failure. Proponents of the "12V" scheme pointed out that even brand new strings have different voltages across each battery, so is not really an issue.


The problem though (and what I was asked to write up to share with everyone else) is what happens if your "reference" battery fails. Lets take a fairly typical "worst case" failure - where one cell fails short-circuit.


The battery now shows only 10V, and charging will happen in earnest.... on and on, as the reference battery will never reach its "high point". Result: the rest of the string gets boiled/cooked/ruined!


Using a simple voltage divider (two resistors) across the whole string, in the example cited, would show a string dropping from nominal 60V to 58V. Sure, it'd charge a little harder, but it would be a mild overcharge and not a completely destructive result.


The bigger the (serier) string, the worse the effect becomes.


So while there is an argument that monitoring just one 12V battery in a string is easy, scalable, has standard bits etc, beware of the downside. In this case, the chance of the failing battery being the sense battery is 20%, but if that 20% happened, it's a 100% chance you'll cook the other 4 batteries!


Food for thought!

« Last Edit: April 04, 2007, 01:15:46 AM by (unknown) »

wdyasq

  • Hero Member
  • *****
  • Posts: 1324
Corin Sparrow diary
« Reply #1 on: April 03, 2007, 08:42:35 PM »
http://www.los-gatos.ca.us/davidbu/sparrow/s_diary.html


Tells an interesting story of battery charging, equalizing and killing in a short time...


Worth the read as it costs nothing but your time.


Ron

« Last Edit: April 03, 2007, 08:42:35 PM by (unknown) »
"I like the Honey, but kill the bees"

billymc

  • Newbie
  • *
  • Posts: 16
Re: Battery strings (monitoring of)
« Reply #2 on: April 03, 2007, 10:50:59 PM »
Hey all,

Man am I glad to see this discussion.  The dumpload control is the last piece of my puzzle.  I was looking at the Oatley controller that WindStuffEd has (k220a) and was wondering if I could hook it up as 24 volt across the two center batteries in a 48 volt string of 4 batteries. I would like to set my mill to furl at about 500 watts and the controller says its for 300 watts or more!?


I did find a schematic for a homebuilt 48 volt controller on a discussion board at the Backshed board by Ross Wheeler.  It looks doable but not a lot of instructions for someone with my limited electronics ability.


Thanks

billymc

« Last Edit: April 03, 2007, 10:50:59 PM by (unknown) »

rossw

  • Hero Member
  • *****
  • Posts: 834
  • Country: au
Re: Battery strings (monitoring of)
« Reply #3 on: April 03, 2007, 10:57:04 PM »
lol  That's me.


Is this one any better/clearer? It was just a quick knock-up design, a few guys made PCBs for it, I'm sure they'd share their layout.


It can be made easily on veroboard too. None of the components are critical, and most opamps will work quite happily. Use what you have in the junkbox - or if you don't have one, hunt up your local "electronics and radio enthusiast" - they'd probably be more than happy to help out with parts, construction and testing.

« Last Edit: April 03, 2007, 10:57:04 PM by (unknown) »

billymc

  • Newbie
  • *
  • Posts: 16
Re: Battery strings (monitoring of)
« Reply #4 on: April 03, 2007, 11:24:02 PM »
OK, Thats funny!  

That's the one! I was online checking parts, making a list last night.  Was going to head off to the local electronics house in a few days.  Wish me luck.


billymc

« Last Edit: April 03, 2007, 11:24:02 PM by (unknown) »

Flux

  • Super Hero Member Plus
  • *******
  • Posts: 6275
Re: Battery strings (monitoring of)
« Reply #5 on: April 04, 2007, 01:20:59 AM »
This all seems rather crazy to me. Sensing from one battery is asking for trouble.


At 60V I also question how you are intending to dump the load. Asking a controller intended for 12v to switch at 60v would be a tall order without knowing what its switching components were.


Flux

« Last Edit: April 04, 2007, 01:20:59 AM by (unknown) »

rossw

  • Hero Member
  • *****
  • Posts: 834
  • Country: au
Re: Battery strings (monitoring of)
« Reply #6 on: April 04, 2007, 01:28:11 AM »
Trivial if its using a relay.


If it's using powerFETs, as long as they are sufficiently high voltage and the load is resistive (not too much inductive component) it should be fine.


Obviously one doesn't use the same value resistors as a 12V installation though!

« Last Edit: April 04, 2007, 01:28:11 AM by (unknown) »

SamoaPower

  • Sr. Member
  • ****
  • Posts: 417
Re: Battery strings (monitoring of)
« Reply #7 on: April 04, 2007, 02:23:23 AM »
I agree with Flux on this one, it's nuts!


To sense at a portion of the string makes no sense.


But, even more basically, this type of bang-bang battery charge controller went out ages ago. Many seem to forget the primary job of a controller is to provide an adequate charge algorithm for the batteries, not to just protect the wind turbine from overspeed.


A simple bang-bang controller that switches at a fixed voltage will leave the batteries in a state of chronic undercharge which is what kills most of them.


Much better is one of the PWM controllers with scaled sensing if necessary and FET drive on the output to accomodate the necessary voltage/current.


Think batteries first!

« Last Edit: April 04, 2007, 02:23:23 AM by (unknown) »

Flux

  • Super Hero Member Plus
  • *******
  • Posts: 6275
Re: Battery strings (monitoring of)
« Reply #8 on: April 04, 2007, 03:08:35 AM »
I agree about fets if they have enough volt rating, ok if you check. I don't agree that switching significant amps at 60vdc with conventional relays is trivial, and beyond 100v you need several contacts in series to break the circuit at all with conventional relays.


You should at least be using RC snubbers at 60v and that may not be enough.


Flux

« Last Edit: April 04, 2007, 03:08:35 AM by (unknown) »

billymc

  • Newbie
  • *
  • Posts: 16
Re: Battery strings (monitoring of)
« Reply #9 on: April 04, 2007, 03:19:20 AM »
I would agree with you and flux, if I were counting on the system as my primary source of power.  I would invest in a Xantrex or TriStar in a heartbeat if my lights counted on my system. I am only playing with wind power as a hobby since I'm in town and can't really put up a `real' system.  I also like to build, scrap and scrounge as much as possible.  There is a certain since of pride in a home built system.  Also, if I bought off the shelf components, what would I do in my garage all day?


I will be looking for someone to submit plans for a pwm controller maybe based on the pic chip so I can spend some more quality time in my garage.


PS: I would like to thank all those who take the time to submit drawings and info that makes my shop time what it is.


billymc

« Last Edit: April 04, 2007, 03:19:20 AM by (unknown) »

zubbly

  • Sr. Member
  • ****
  • Posts: 371
Re: Battery strings (monitoring of)
« Reply #10 on: April 04, 2007, 06:59:05 AM »
hey Ross!


thanks for posting about our recent discussions. my proposed ideas brought out some great thinking and also helped to bring to light possible problems with using a 12 volt controler tied to one 12 volt batt in a series string of 12 volt batts.


story time now :) LOL


until now (approx 4 years) i have done without any form of battery charge controler. daily monitoring and manually turning loads on and off have been my normal way of keeping the batts at reasonable charge levels.  talk about a pain in the butt!

most of what i do is for fun and seeing what i can make from a pile of salvaged junk. practically everything from my gennys and controls have come from salvaged dumpster goodies.


this winter i have been using my 7.5hp conversion to feed 3 phace AC to my heater bank i have installed in my oil fired furnace for additional heat. ballpark estimate, saved maybe 20% on fuel oil bills. largest problem i noticed was that very often the genny was running just under the 100 volt cut in and was basically loosing everything being produced under that. i used every conceivable heater configuration i could dream up including capacitors to bring the genny on line to small voltage sencing relays to control the load relays to the elements.


my solution, and i am sure gives me more captured energy in the long run, was to charge batteries and have them dump DC direct to the elements via a dump load controler.


a simple charge/dump controler that used to be available was the "mikes" charge/dump controler that was available as a build it yourself kit. it was basic, simple, adjustable cut in and cut out points, and was 100% better than no control at all.


so what i did was put the batteries i have ( a super mismatched set consisting of 2 6 volt trojans, a wallymart 12 volt starting-marine duty, and some desulphated car batts that seem to be working ok which i got from an auto repair facility). i totally agree with anyone that says that is a bad battery bank, and it is, but its what i got and it does work and does the job for now.


i did make 2 attempts in the past to build a mikes controler, but both did not work. the looks of them would even have scared Frankenstein. my electronic skills lack desire. the third one however does work.


so what i tried first was setting the batts up as a 48 volt bank. the genny is in a 1 star configuration but seemed to slightly stall under load unless wind was very strong. simple solution was to add another 12 volt battery and raise the cut in rpm, and it also brought up the kw rating of the element bank.  all elements (6) are now connected paralell, and when the contoler dumps, i get a 32 amp input to the elements with about an average of 68 volts DC. almost a 2200 watt load.


the reason i tapped off of only 1 12 volt batt was because it was a simple soluton at the time, and does work extremely well for dumping. i have the trip points set at 14.6 volt and 12.6 volt. the dump mode often stays on quite a long time when i have good wind outside.


the controler does however not carry any of the dump load. the only load on the controler is a tiny 1 pole N/O-N/C 12 volt relay which is just used to control the coil in a 3 pole 40 amp contactor. the 3 poles of the contactor have been ganged together to be used as 1 pole. so far so good.  DanB has suggested i find one of the old mercury displacement relays sometimes available on e-bay which would handle the dc voltage and current much better in the long run. i am sure i will eventually burn the contacts in the present one in use (but i got lots of them).


i daily take part in discussions on the irc chat line (i just go there to drive them all nuts ;) and with the knowlege and expertise of many of them there, have helped to come up with some re-design strategy so that this simple controler can be made with additional parts, truely compatible with 12-24 and 48 volt battery banks. it will not have PWM or any fancy charge settings. it will be a basic charge/dump controler able to work on all 3 of the common voltage banks, and be fully adjustable  for the charge and dump settings you determine.


i did some research locally and on e-bay, and found that the cheap controlers do not even have setable trip points (some locked at 13.8 and 11 volt) to the more elaborate (cadillac) of units and also cadillac pricing. price ranges seem to be from $23 to well over $500.


this is what has brought me to the idea that a revival with 12-24 and 48 volt capability of the original  "Mike's controler may suit some of the needs by people out there who need a basic controler at a very affordable price. when all the bugs have been ironed out and i make another unit and test for sometime, i may consider having these units available for sale. personally at this time, a cadillac unit is beyond my budget and think many others are the same.


so far, the unit has been operating for almost 4 weeks without a glitch and does seem to work very well. that crappy battery bank of mine has seen all batts stay within .2 of a volt of each other after resting over night. to get an equalizing charge, i simply turn it off.


anyone with ideas, views, opinions good or bad, please spit it out. i would rather have possible problems brought to light now rather than later.


zubbly

« Last Edit: April 04, 2007, 06:59:05 AM by (unknown) »

Opera House

  • Sr. Member
  • ****
  • Posts: 261
Re: Battery strings (monitoring of)
« Reply #11 on: April 04, 2007, 07:51:48 AM »
At least that is not another 555 circuit.  That circuit when tripped stays on for 10 seconds and is totally on or off.  I posted a TL494 shunt regulator that regulates by PWM at about 100 Hz.  
« Last Edit: April 04, 2007, 07:51:48 AM by (unknown) »

fungus

  • Sr. Member
  • ****
  • Posts: 448
Re: Battery strings (monitoring of)
« Reply #12 on: April 04, 2007, 07:54:02 AM »
emm.. the 555 circuit does -not- stay on for 10 seconds. It has a timing period of approx 1.1 seconds.. which can easily be changed by changing resistor values.
« Last Edit: April 04, 2007, 07:54:02 AM by (unknown) »

TomW

  • Super Hero Member Plus
  • *******
  • Posts: 5130
  • Country: us
The Mike Controller....
« Reply #13 on: April 04, 2007, 08:15:10 AM »
Hey Zubbly;


I thought it might help if everyone could eyeball that original circuit.


Here is a photo:






I modified mine for 24 volt use by changing that 3.3K resistor that is across the "ALT" switch to a 10K. Then the trip points are 1/4 of the desired voltage due to the 3:1 ratio of that divider formed by the original pair of 3.3K resistors. No long term tests on it yet but it did pass the tests described in mikes original instructions. Oh, if you build this remember those fets act as switches in the - [negative] lead of your fan or relay.


Cheers.


TomW

« Last Edit: April 04, 2007, 08:15:10 AM by (unknown) »

ghurd

  • Moderator
  • Super Hero Member Plus
  • *****
  • Posts: 8059
Re: Battery strings (monitoring of)
« Reply #14 on: April 04, 2007, 08:18:40 AM »
Hi Zubbly,


I will have a dump controller kit out soon.  The PCBs are giving me fits.

Very Cheap and simple from all common parts, except one. Nothing earth shaking. Samoa's "Bang-Bang" type.

For 12 or 24V, adjustable threshold voltage, hysteresis adjustable during assembly.

Aimed toward the 50 to 200W people, who often have no controller.

Currently the last pic in my files.


Thought you might like to know.

G-

« Last Edit: April 04, 2007, 08:18:40 AM by (unknown) »
www.ghurd.info<<<-----Information on my Controller

Nando

  • Hero Member
  • *****
  • Posts: 1058
Re: Battery strings (monitoring of)
« Reply #15 on: April 04, 2007, 11:39:50 AM »
INDEED, rather crazy and illogical to me as well.


One needs to understand the basic electrical principles for batteries to define if such step is practical and warranted.


Differences in manufacturing, materials, and battery placement in the battery bank, as well as, how the power is extracted from each battery in reference to the battery connections ( paralleling and in series) defines the charge/discharge cycle of each battery in the bank.


Nando

« Last Edit: April 04, 2007, 11:39:50 AM by (unknown) »

Opera House

  • Sr. Member
  • ****
  • Posts: 261
Re: Battery strings (monitoring of)
« Reply #16 on: April 04, 2007, 02:19:34 PM »
My comment was to the Wheeler op amp circuit that charged up a 3.3uF through the diode and discharged through a 4.7M resistor creating an off delay.   555 circuits have been known to work, but they take too many parts and are difficult to diagnose by a neophyte.  They have been around for more than 30 years, but I don't take any 555 design seriously.
« Last Edit: April 04, 2007, 02:19:34 PM by (unknown) »

Bruce S

  • Administrator
  • Super Hero Member Plus
  • *****
  • Posts: 5375
  • Country: us
  • USA
Re: Battery strings (monitoring of)
« Reply #17 on: April 04, 2007, 02:30:37 PM »
Just to add what Ghurd is saying. I took part of his idea and ran with it for my output from the HF PDU, and it's doing just fine.

Looks ugly as all get out but runs along just fine. linda set it a forget it like:--)


Cheers

Bruce S

« Last Edit: April 04, 2007, 02:30:37 PM by (unknown) »
A kind word often goes unsaid BUT never goes unheard

Bruce S

  • Administrator
  • Super Hero Member Plus
  • *****
  • Posts: 5375
  • Country: us
  • USA
Re: Battery strings (monitoring of)
« Reply #18 on: April 04, 2007, 02:31:57 PM »
oops linda has nothing to do with it. I don't tpye for a living ;--) should'a said kinda set it and forget it.

B-(

« Last Edit: April 04, 2007, 02:31:57 PM by (unknown) »
A kind word often goes unsaid BUT never goes unheard

Bruce S

  • Administrator
  • Super Hero Member Plus
  • *****
  • Posts: 5375
  • Country: us
  • USA
Re: Corin Sparrow diary
« Reply #19 on: April 04, 2007, 04:29:32 PM »
Ron;

   AND he's back to updating the diary. I sent him and email a long while back asking if it was okay to "borrow" his idea of using the NiCds to charge the "wet-cell" bank.

Never did get a reply so I did it anyway. AND still works.

The inverter is slightly higher in power , but was free and well.... as the story goes is still working...

The NiCd bank is as large as I EVER want to get it and as already proved itself with the last STL power out age.


Bruce S

« Last Edit: April 04, 2007, 04:29:32 PM by (unknown) »
A kind word often goes unsaid BUT never goes unheard

BigBreaker

  • Sr. Member
  • ****
  • Posts: 302
Re: Battery strings (monitoring of)
« Reply #20 on: April 17, 2007, 08:03:48 AM »
At university I took a lab in digital design ten years ago.  We used discrete parts (ICs and such) to build some really amazing things.  That was about the time when things were changing though (certainly earlier in industry - this is the education/hobby world).


The solution to a lot of problems for us turned out to be a PROM (basically a look up table) that took a bunch of state inputs and then output a dozen or more gating signals or processed outputs to run the project.  Wiring ICs turned into programming PROMs.


Now with microcontrollers the whole project is programmed and sits on a single piece of silicon.  They have built in oscillators, D/As and PWMs.


I agree that 555s are "lightweights" but they still have a place.  A decent multimeter has a frequency sensing mode.  That mode can verify that the 555 is working.  The parts are easy to scrounge.

« Last Edit: April 17, 2007, 08:03:48 AM by (unknown) »

dinges

  • Hero Member
  • *****
  • Posts: 1294
  • Country: nl
Re: Battery strings (monitoring of)
« Reply #21 on: April 17, 2007, 08:23:00 AM »
"A simple bang-bang controller that switches at a fixed voltage will leave the batteries in a state of chronic undercharge which is what kills most of them"


When the trip point is correctly set the batteries will not be undercharged. The exact trip point also depends on the type of battery. A datasheet of the battery should give all the information one needs. Generally, for trickle charging, I use 13.85V as set point; for intermittent charging 14.40-14.50V (gell cells).


In cars, the voltage regulator in the alternator is a simple bang-bang regulator. My last (cheap!) battery lasted 6 years. Can't be that bad then, I suppose.

« Last Edit: April 17, 2007, 08:23:00 AM by (unknown) »
“Research is what I'm doing when I don't know what I'm doing.” (W. von Braun)

SamoaPower

  • Sr. Member
  • ****
  • Posts: 417
Re: Battery strings (monitoring of)
« Reply #22 on: April 17, 2007, 12:06:03 PM »
Sorry Peter, can't agree.

When charging with a simple switch controller, the first time it reaches the upper trip point, the SOC will be at 80-90%. To do it right, it then needs to remain at that voltage for some period of time, 1-3 hours typically, to bring it up the rest of the way before switching off. This is the absorption phase.


With a bang-bang type, it will take many, many cycles to bring it up close to 100% because each cycle adds so little. Batteries left even a few percent shy of 100% will have a shorter life (chronic undercharge).


An auto alternator regulator is not a bang-bang type. It modulates field current to stay at fairly high voltage to top off the battery, around 13.8-14.0. This is higher than a typical float.

« Last Edit: April 17, 2007, 12:06:03 PM by (unknown) »

SamoaPower

  • Sr. Member
  • ****
  • Posts: 417
Re: Battery strings (monitoring of)
« Reply #23 on: April 17, 2007, 12:06:36 PM »
Sorry Peter, can't agree.

When charging with a simple switch controller, the first time it reaches the upper trip point, the SOC will be at 80-90%. To do it right, it then needs to remain at that voltage for some period of time, 1-3 hours typically, to bring it up the rest of the way before switching off. This is the absorption phase.


With a bang-bang type, it will take many, many cycles to bring it up close to 100% because each cycle adds so little. Batteries left even a few percent shy of 100% will have a shorter life (chronic undercharge).


An auto alternator regulator is not a bang-bang type. It modulates field current to stay at fairly high voltage to top off the battery, around 13.8-14.0. This is higher than a typical float.

« Last Edit: April 17, 2007, 12:06:36 PM by (unknown) »

dinges

  • Hero Member
  • *****
  • Posts: 1294
  • Country: nl
Re: Battery strings (monitoring of)
« Reply #24 on: April 17, 2007, 02:29:10 PM »
An auto alternator modulates field-current to compensate for varying RPMs of the motor, NOT for the goal of PWM-ing the output voltage.


That way, despite hugely varying RPMs, the output voltage remains constant (14.0V as you say).

« Last Edit: April 17, 2007, 02:29:10 PM by (unknown) »
“Research is what I'm doing when I don't know what I'm doing.” (W. von Braun)

dinges

  • Hero Member
  • *****
  • Posts: 1294
  • Country: nl
Re: Battery strings (monitoring of)
« Reply #25 on: April 17, 2007, 02:33:06 PM »
You may be right in that it still is not bang-bang, I realize now. It's more like a float-charger, constantly charging/topping it up. BTW, I thought the voltage was higher than 14.0V, 14.4V. At least, that's what I measured in my car.
« Last Edit: April 17, 2007, 02:33:06 PM by (unknown) »
“Research is what I'm doing when I don't know what I'm doing.” (W. von Braun)

SamoaPower

  • Sr. Member
  • ****
  • Posts: 417
Re: Battery strings (monitoring of)
« Reply #26 on: April 17, 2007, 05:41:53 PM »
The auto regulators I'm aware of are not PWM at all, but linear types.


I believe they are temperature compensated so that in colder weather the voltage will be higher. I usually measure less than 14.0 volts on mine, here in the tropics.


Way back when, I used a bang-bang controller on my RE system and only got about three years out of a set of Trojan T-105s. After learning what a better charge algorithm is, I now get six years plus.


Even so, I still think the popular three-step PWM controllers are pretty crude and don't do it properly. I have a design for what I consider the right way, but, unfortunately, it's still way down on the project list. One of these days.

« Last Edit: April 17, 2007, 05:41:53 PM by (unknown) »

dinges

  • Hero Member
  • *****
  • Posts: 1294
  • Country: nl
Re: Battery strings (monitoring of)
« Reply #27 on: April 17, 2007, 11:12:04 PM »
Yes, that was what I was saying: auto alternators are NOT PWM but linear.

« Last Edit: April 17, 2007, 11:12:04 PM by (unknown) »
“Research is what I'm doing when I don't know what I'm doing.” (W. von Braun)

solarengineer

  • Full Member
  • ***
  • Posts: 136
Re: Battery strings (monitoring of)
« Reply #28 on: May 30, 2007, 07:14:35 AM »
totally NOT trivial at all!!


Ross What you need to do is scale the 60V to a nomial value you can work with.

so a voltage divider that would divide by 5 will give you 12V when the input is 60V

A typical divider I use all the time is a 4.7K resistor in series with a 1K resistor.


the 4.7K to Batt+ and the 1K to Gnd.... the center tap or junction between them is your output. This is a little more than divide by 5 but will work for you.


This monitors the bank as a whole, sensing 1 is asking for absolute disaster..

Been there, done that.....lost the rig and more than once.


Jamie

« Last Edit: May 30, 2007, 07:14:35 AM by (unknown) »

rossw

  • Hero Member
  • *****
  • Posts: 834
  • Country: au
Re: Battery strings (monitoring of)
« Reply #29 on: May 31, 2007, 05:21:29 AM »
I think you're confusing me with someone else.


I was NOT an advocate of monitoring only one cell at all, and made much the same comments.

« Last Edit: May 31, 2007, 05:21:29 AM by (unknown) »

TomW

  • Super Hero Member Plus
  • *******
  • Posts: 5130
  • Country: us
Re: Battery strings (monitoring of)
« Reply #30 on: May 31, 2007, 09:46:01 AM »
Ross;


Uh, hmmm, first time that has happened on this forum. In fact, I remember the initial IRC discussion and you were quite vigorously opposed to monitoring one of X. Zubbly seems to be doing it successfully on his mismatched, used starting battery bank, however.




Cheers.


TomW

« Last Edit: May 31, 2007, 09:46:01 AM by (unknown) »