Anotherpower.com Forum
Renewable Energy Questions/Discussion => Automation, Controls, Inverters, MPPT, etc => Topic started by: frackers on July 01, 2012, 07:41:36 pm
-
Greetings all
Just landed on my feet!! Tonight I get to unload 21 12v 100amp-hr UPS batteries from my trailer having picked them up yesterday. They aren't very old (couple of years I think) so should have plenty of their 10 year life left. Perfect timing as the 8 golf cart style batteries I bought new 4 years ago have just about died :(
These AGM batteries came out of an 80KVA UPS that was operational until last week so they all have a reasonable state of charge (all 13.2<>13.3 volts) apart from one which is sitting at 12.95V.
I don't have a great deal of experience with AGMs so the question is, when I'm doing a *Bulk Charge*, what should I limit the charge voltage to or should I allow them to sit at whatever voltage they like provided the charge current is within spec. I realise that once I get to *Absorption Charge* I should limit the voltage and just keep an eye on current, similarly *Float Charge* should be constant current with an eye that the voltage doesn't get too high.
I want to make sure I characterise these batteries before I wreck them!! I'll be using a charge counter to track where in the charge cycle they are, the problem being they are AGM so I can't check the SG to see what the starting point is. What I *don't* want to do is just go out and buy a controller where I have no knowledge of the algorithms being used - I trust my own software development skills better than that!!
This is my starting point for moving from wind to solar power so I'd like to get it right - there is another few KW of Oztules panels about to take shape :)
-
I am no expert but I know a lot of batteries of a lot of types seem to depend on the manufacturer.
Who made them? Maybe they have some charge guidelines we could find
-
Like Wolv said, there is a lot of difference between manufacturers of similar chemistry batteries on how they like them charged.
Must be differences in how they are made.
Flooded batteries seem to be the same but SLA, VRLA and AGM manufacturers seem to vary a lot on how they think you should do it.
Tom
-
These are Century UPZ 12.100 VRLA batteries. The 2 page spec sheet has most info but not charging voltage - only float.
-
For reference
Spec sheet obtained from here
(http://www.anotherpower.com/gallery2/main.php?g2_view=core.DownloadItem&g2_itemId=20854&g2_serialNumber=1) (http://www.cyb.com.au/files/All_PDFs/UPZ12-100_Century_Spec%20Sheet.pdf)
-
I've lived with this same type awhile now. Hope you got them for free or near free :)
Float may be the only charging scheme permissible, but note the C/5 max charge rate makes for rapid bulk charging.
The 10-year warranty batteries I have suggest a high-rate float at 0.15 or 0.20V increase from normal float is allowable ONLY if another UPS 'outage' use period is scheduled that will draw the battery SOC down and damage the plate chemistry worse than cooking them at a higher charge rate. Looks like 13.56V at 30°C (86.0 F) is it.. and accurate temperature compensation is MANDATORY.
http://tinyurl.com/7tudlec <--- "Effects of AC Ripple Current on VRLA Battery Life"
http://tinyurl.com/7oce4fz <--- "Thermal Runaway in VRLA Batteries"
^^ You can't dump 'wild' rectified three phase into the 10-year VRLA/Gel batteries without consequences. These are thoroughbred pedigreed batteries made to be maintained by lab quality chargers.
There are hybrid SLAs etc. that use some qualities of different designs that can accommodate automotive spec charging, the 14.4 or 14.7V and no temp compensation, but then they are designed around a shorter warranty and no special equipment etc.
So if you're feeding wind power you get to invent a constant voltage - constant current charger AND filter the AC crap out of the DC charge lines - And you need to fine tune it for accurate temperature compensation.
-
Thanks Dang
Some very useful information there and stuff I'll be able to use in my controller one way or another!! I can monitor volts, current, temperature and state of charge (from charge counting) so now I have the start of a set of tables I can construct suitable charge/divert algorithms.
Since its pretty easy to do, I might split the battery strings a bit so I can monitor current and temperature in separate blocks (I run a 24 volt system).
With the current turbine its unlikely I'll be able to push more than C/20 (50 amps) into the bank and max discharge is about C/5. Assuming I get some solar onto it and start moving loads off grid then those numbers will change somewhat but so far I don't expect to suffer any thermal effects although the ripple current effects are something to think about...
Cheers
-
Good score on those AGMs are there any left? ;)
you will injoy the power 2kw of solar gives, my system is keeping up with the house needs on pv alone in these frosty days.
-
"This is my starting point for moving from wind to solar power so I'd like to get it right - there is another few KW of Oztules panels about to take shape :)"
............... you do realize that your social life will suffer making panels.... they use up a lot of time.
But once you get up into the 3-4kw of solar, you will sadly lose any interest in wind made power..... except to see it turning.
Even today, only a handful of days past the shortest day of the year, mine made over 20000 watt hours today.. 20kwh. Both my 4 meter mills are in the shed...... to re-pot the magnets to stop them from deteriorating ......initially,...... but I just can't get excited about putting them back up, as I have too much power all the time now..... very sad really :'(
When they do go back up..... they will only be shorted I suspect and spinning very slowly...... just for memories............
When you can't use the power...... you can't use the power.
................oztules
-
Just reading the pdf for these batteries again and I see they can be used on their side - that would make getting to the terminals easy especially as I was thinking of having a 2 tier rack of them to save on floor space.
I wonder if they will also be OK on their short side - I can pack more side by side then!! Anyone any ideas or is it best to just have them right way up...
Cheers
-
Just reading the pdf for these batteries again and I see they can be used on their side - that would make getting to the terminals easy especially as I was thinking of having a 2 tier rack of them to save on floor space.
Here's mine, and generally being on their side makes lots of things easier I think.
(http://house.albury.net.au/23may2010/100_4463.JPG)
-
Ross,
That installation is much too clean ::)
You're going to have to do something about that!
Kevin
-
Ross,
That installation is much too clean ::)
You're going to have to do something about that!
Kevin
Yep, Ross does nice work. Don't tell him I said that tho.
Tom
-
Ross does nice work. Don't tell him I said that tho.
Aww, shucks guys....
There's an easy answer. I only post pics of the stuff that turns out reasonable :)
-
"You can't dump 'wild' rectified three phase into the 10-year VRLA/Gel batteries without consequences"
I don't think the ripple they are talking about is related to rectified 3-phase.
To me, it looks like they are talking more about unfiltered rectified single phase.
Not sure how poorly a system would have to be designed to have 5A of ripple current on a 100AH battery from rectified 3-phase charging.
I'll leave that for Ross' calculator.
Could put a cap on the DC side of the rectifiers, asuming the wire run between the rectifiers & cap to the battery was fairly long (like tower to battery distance). That would keep the ripple down to "almost none"?
G-
-
"You can't dump 'wild' rectified three phase into the 10-year VRLA/Gel batteries without consequences"
I don't think the ripple they are talking about is related to rectified 3-phase.
To me, it looks like they are talking more about unfiltered rectified single phase.
Thats what I thought - and charging at 50amp (on rare occasions) max into 1000 amp-hrs means the ripple per cell is already so low as to not be easily measured.
Mind you, providing a large cap between the rectifier and the battery bank is not likely to be very effective seeing the internal resistance of the bank is so low that to get a suitable time constant would require kilo-farads !!
-
Mind you, providing a large cap between the rectifier and the battery bank is not likely to be very effective seeing the internal resistance of the bank is so low that to get a suitable time constant would require kilo-farads !!
Which is why in many high-current applications, I've seen the use of chokes in series rather than caps in parallel. Far more "bang for the buck".
-
Talking about someone elses $5,000 or $7,000 10-year-plus-life battery bank the final word is do no harm not let's get busy taking shortcuts.
The paper on current ripple does make some assumptions - like even with high-grade temperature control there's still a possible 3% life reduction through slight mismatches, that they say in print. Let's amplify those errors with site conditions, like not enough or uneven cooling in the battery room, erratic charge-dump cycles during high-wind events, multiple charge sources happening at different calibrations, non-calibrated VOMs letting values drift, series-parallel interconnections playing with impedance or sp strings not equalized and who knows what else.
Another assumption the papers make is supplying 'constant frequency', not a turbine bumping against cut-in 15 or 30Hz rectified (for hours at a time) from the first or first/second phases breaking charging potential with pulse currents. Oh yeah, and thats after maybe a full day charging from solar plus earlier wind that has the battery string warmed up. Do we talk about partial shorted coils innocently being run for months with the resultant voltage offsets hammering at the battery chemistry? I'm sure there is other conditional stuff that would love to pile on unnoticed.
On inductor current smoothing, since y'all mentioned it :)
Where to best insert inductors? AC Reactor, AC side Chokes or DC Link Choke, the DC side is cheaper since there is only one after the rectifiers but would not help protect the diodes, if possible, from surges from nearby lightning, etc.. After all it is connected to a tower, it'd be another passive payer of protection. On the AC side at what size inductor could diode conduction overlap and would that be a problem? If doing an LC circuit analysis with adding caps what would be the best values, design around 24 - 48V and 16 - 32 amps system peak or just worry about cut-in smoothing? Would an AC side inject harmful losses or resonance back into star/wye or delta connected with air-core stator coils?
Attached photo is $50 worth (shipped!) of Allen-Bradley and Fanuc inductor & reactor I sniped from fleece-bay to play with for current smoothing & dump loads on the Dan's 10' mill when it gets flown :) The reactor is rated 130A and .13 mH, the 3-gang choke is 4.125" O.D. wound with 4AWG at 12mH 150A (I think). These are salvage from older robotic assembly controls.
[attach=1]
-
Just to clarify, I was referring to quite large, DC chokes specifically for ripple-reduction, not high-frequency switching noise and lightning protection. As such, they'd be a major liability on the AC side of any turbine, and only suitable for use after the rectifier. They were also quite large, in the order of 100 lbs or more, had soft-iron cores and lots of turns of heavy copper. No idea their inductance.
-
"This is my starting point for moving from wind to solar power so I'd like to get it right - there is another few KW of Oztules panels about to take shape :)"
............... you do realize that your social life will suffer making panels.... they use up a lot of time.
................oztules
Is "fred480v" on fleabay the best reseller these days? He does seem to have a regular supply of tabbed cells without too many dings on them (judging from the pics) and kits with the interconnect ribbon etc. He is also selling EVA roll ends - maybe not the best way to get it!!
Just got to work out the best way of getting the delivery costs down to NZ - perhaps surface (container) shipping or getting together with a few mates and getting 10KW in one hit.
-
freds cells are very good for quick panels. I have noticed that a percentage of them (after a year or so) get fine fracture lines in them. It does not seem to affect the output, but it does create tempory heat spots. If you use normal window glass it will probably cause cracking. On the tempered stuff i used, it was no problem, and seems to have stopped.
By far the best cells i have used (but slow because they are virgin cells ) come from China.
They are 6x6 7.5Awp ones.
They take a lot longer, but if you have the time, i think they are better longer term. A lot of my panels are 2 years on, and no change except for a few panels with freds second hand cells..... the tiny brown tell tale signs of internal fractures perhaps....... They have had a lot of handling by the time he and then we get them, so it is not unexpected, just a little sad perhaps....
The ones i got from China are from these folks. all were spec.
http://www.aliexpress.com/store/product/Promotion-price-A-grade-3-87W-156mm-polycrystalline-solar-cell-16-efficiency-Factory-original-package/505815_529127899.html
This is free shipping, about 33c per watt for over 5.8kw for little less than 2000 bucks..... but a lot of work.
or if you want less power then
http://www.aliexpress.com/store/product/Solar-Cell-156mm-156mm-6-polycrystalline-3-70W-A-grade-two-busbar-promotion-price-Free-Shipping/505815_549917575.html
Check the feedback, but I could not fault them or the cells..
For EVA perhaps this :
http://www.aliexpress.com/product-fm/416117721-Free-Shipping-Solar-Cells-Panel-EVA-film-sheet-Encapsulant-1M-x100M-NEW-wholesalers.html
or this:
http://www.ebay.com.au/itm/Solar-Cells-Panel-EVA-film-sheet-Encapsulant-1M-x-50M-0-5MM-/160733168893?pt=AU_Solar&hash=item256c7184fd
or this for less..... remember to check the width, these are both 1000mm wide:
http://www.aliexpress.com/product-fm/416120817-Free-Shipping-Solar-Cells-Panel-EVA-film-sheet-Encapsulant-1M-x40M-NEW-wholesalers.html
If you can spend the time, i think these chinese ones are amongst the best. No breakages, and no sign of internal fractures showing up.
2 bus bars are easier than three......
If your using window glass, not hardened, spot heat will cause you heart burn if the cells have hot spots.... so chinese new ones are best for this i think.
If you have iron free hardened glass, then perhaps freds for convenience, but i would go china now, and spend the time......
If you have to buy glass, or frames or whatever, panels are down to a dollar a watt in Aust now (CE approved), so do some home work first.
.................oztules
-
Thats what I thought - and charging at 50amp (on rare occasions) max into 1000 amp-hrs means the ripple per cell is already so low as to not be easily measured.
Mind you, providing a large cap between the rectifier and the battery bank is not likely to be very effective seeing the internal resistance of the bank is so low that to get a suitable time constant would require kilo-farads !!
The ripple is already low. ("so low as to not be easily measured")
The 'R' in RC is the wire run.
Only need to smooth out the ripple voltage a bit.
Not sure of the L in the coils. Been way too long since...
All combined, might almost think of it as RLC low pass filter?
It is one of those things I can't get from my head to the keyboard.
I tested it several years ago.
Small 3-ph iron core PMA but only used ONE phase.
1000uF? 1500uF? cap on the rectifier at the PMA.
Long run of wire to the battery.
Cap increased the output at the same RPM by 7% if I recall correctly, which is what I expected.
I still believe it is the change from "rectified unfiltered single phase", the ripple current in the resistance of the wire, and the i^2*R wire loss when the wave is in a higher amplitude portion of the current flowing.
The cap smoothed out the current flow (less ripple current), which reduced the average wire loss.
At least thats the best I can explain whats in my head.
I couldn't make Flux understand it either.
I just know the charging current increased, like I expected it would.
G-
-
"I still believe it is the change from "rectified unfiltered single phase", the ripple current in the resistance of the wire, and the i^2*R wire loss when the wave is in a higher amplitude portion of the current flowing.
The cap smoothed out the current flow (less ripple current), which reduced the average wire loss.
At least thats the best I can explain whats in my head."
From our experiments at the backshed, it has a lot more to do with how the armature reactance changes with a positive power factor. A reasonable capacitance across the output will change ( increase) the flux available to do work, a lagging current will affect the field adversely.
ie a cap allows the flux to increase, a lagging current causes the field to decrease. On long salient pole iron cores, this effect is quite marked. Too much does not help though. There is a 30-40 page dissertation on the backshed regarding this called visual effects of capacitors if I recall it correctly.
I appreciate the current peaks being less with caps also, but from experiments conducted, it seems to be the reactance and the effects on the magnetic "timing" (teeth to pole) thats the real difference. It advances or retards. Herb of NZ actually used a timing light to see the effects in real time.
...........oztules
-
I wish they'd get that down to what is correct!
By the time I spent an hour sorting out an idea on one page, the next page said that part was incorrect. LOL
I am talking about a simple cap on the DC side of the bridge, and the bridge at the PMA, and a long wire run.
G-
-
Yep, it's a jumbled mess...... thats why I was not keen to do a direct link....... it takes a brave man to read all the way through.
Interesting how we think one thing and do another........ I knew you said after the rectifier...... but somehow was thinking before the rec... typed away...... then re read your lines, and thought after the cap and current smoothing....more typing then didn't put the two together Ghurd..... sorry about that.
..............oztules