Daily charging parameters

SolarMusher · July 2016

Vic said:

The Classic CC does have in the latest FW versions, the ability to Skip Charge days, where the number of days to be skipped is entered by the user. On Skip days, the CC begins in FloatMPPT, and stays in Float all day, until the day after the Skip number has been satisfied. We Skip three days on both of the larger systems here, and recharge on the fourth day ... FWIW. Vic

Vic, what would happen on the 5th day if the 4th day was rainy/cloudly and the classic can't reach absorb/float?
Will it start in bulk?
I've never used skip day but want to give it a try this summer, actually mine (750Ah) are only 5% discharged every morning and that's not great.
A+
Erik

Vic · July 2016

Erik,

Am using 1849 Classic Firmware, and if day #4 was a poor Solar input day, and Absorb did NOT finish, Day #5 will be the first day Skipped in the next interval ... ie, the Classic will not try to get the full charge cycle done, it just follows this fixed Skip/Charge sequence as entered in the Advanced menu.

This is a bit too bad for folks with fairly frequent storms. Have not tried any of the more recent FW versions to see it there is a provision for trying to get the full charge done, regardless of how many days it requires.

Skip days is a big help in assuring that the batteries here, really DO get cycled moderately deeply, at least on occasion. AND, in the Summer, it obviously reduces the heating of the battery, as there are fewer charge cycles, and so on.

Have Fun, Erik. Vic

Vic · July 2016

herodotus said:

Hey Vic,

Yeah, I have a hydrovolt hydrometer that I have been using to measure SGs. I did the whole bank a couple of times at the start and now I'm testing a couple of test cells each time. I've got my panels switched off for the moment so I can get a good deep cycle on the batteries. The FNDC battery monitor will probably wind up reading about 85% SOC, but I suspect the SGs will say otherwise - presumably because my batteries are actually lower capacity right now than they will be once they are "broken in".

Cheers,,

Hi hero..,

The Hydrovolt does a good job, and is very easy to read SG values. Just make certain to thoroughly clean it after each measuring session, with four or five rinses with Distilled Water (IMO). There is a fairly large amount of surface area exposed to electrolyte, and one should work to completely rinse the insides of the Hydrovolt.

It is a very good idea that you are using test cells to quickly indicate the nominal SG of the battery bank. Saves time, and is still a good indicator, often encouraging one to monitor SGs relatively often -- a good thing.

Also a good idea to force some deeper discharges by switching off the PVs (we used to do this, as well, before the Classic's Skip Day function was implemented).

Have fun and good luck with the new system/batteries. Vic

herodotus · July 2016

Vic said:
Hi hero..,

The Hydrovolt does a good job, and is very easy to read SG values. Just make certain to thoroughly clean it after each measuring session, with four or five rinses with Distilled Water (IMO). There is a fairly large amount of surface area exposed to electrolyte, and one should work to completely rinse the insides of the Hydrovolt.

I confess that I'm a bit lazy/cheap about this - I rinse with our spring water rather than distilled. It's relatively soft water and I figure that the residues left behind after drying are going to be small enough to be irrelevant.

Vic · July 2016

Hello hero..,

The main reason for careful rinsing of Hydros, is to try to reduce/avoid that pesky sticky residue that can build up on the inside surfaces of Hydrometers. This residue allows air bubbles to stick to the inner parts of these instruments, and causes erroneous readings -- often reading higher SG than actual.

Any hard deposits from minerals might (do not know from experience) make this residue more difficult to rinse out, and so on ... have only used real Steam Distilled Water, forever.

One person on this Forum had wondered about saving the spent rinse water, as replenishment water for adding back to batteries. Perhaps not a bad conservation step, but, with batteries being a bit expensive, have always tossed this spent rinse water.

FWIW, Cautious Vic

zoneblue · July 2016

Sounds like you have this mostly sorted, but just to explicitly spell out that many people on this forum have found that certain Rolls L16 batterys can be notoriously tough to break in. See posts by Chris Olsen and blackcherry.They reported that the SG takes huge absorbs and lots EQ to get up in the first 6 months, then it comes right. This process is important to form the plates so id definitely be preserving with it. As others have said, well watered, you dont need to fear over absorbing a new battery, as the end of day terminates the absorb one way or another.

herodotus · July 2016

zoneblue said:

Sounds like you have this mostly sorted, but just to explicitly spell out that many people on this forum have found that certain Rolls L16 batterys can be notoriously tough to break in. See posts by Chris Olsen and blackcherry.They reported that the SG takes huge absorbs and lots EQ to get up in the first 6 months, then it comes right. This process is important to form the plates so id definitely be preserving with it. As others have said, well watered, you dont need to fear over absorbing a new battery, as the end of day terminates the absorb one way or another.

Interesting - that's really useful, thanks. I think what I'm settling on now is a cycle of turning off the panels for a day, to get a good discharge, followed by a really good deep absorb charge the next day - at least 4 hours in absorb until the current comes right down to a 0.5% or so. At some point we should get some more substantial loads sorted out that will eliminate the need for turning off the panels. Once the batteries are broken in I might think about toning down the absorb charges and only doing a full charge a bit less frequently.

Cheers,

Vic · July 2016

herodotus said:

zoneblue said:

Sounds like you have this mostly sorted, but just to explicitly spell out that many people on this forum have found that certain Rolls L16 batterys can be notoriously tough to break in. See posts by Chris Olsen and blackcherry.They reported that the SG takes huge absorbs and lots EQ to get up in the first 6 months, then it comes right. This process is important to form the plates so id definitely be preserving with it. As others have said, well watered, you dont need to fear over absorbing a new battery, as the end of day terminates the absorb one way or another.

Interesting - that's really useful, thanks. I think what I'm settling on now is a cycle of turning off the panels for a day, to get a good discharge, followed by a really good deep absorb charge the next day - at least 4 hours in absorb until the current comes right down to a 0.5% or so. At some point we should get some more substantial loads sorted out that will eliminate the need for turning off the panels. Once the batteries are broken in I might think about toning down the absorb charges and only doing a full charge a bit less frequently.

Cheers,

Couple of thingsl
hero.., you will probably have some difficulty reaching 0.5% of C, especially at the recently-elevated Absorb voltages ... and EA behavior will probably change as the bank gets more cycles.

zoneblue, certainly DO recall ' blackcherry04's experience with the Surrette L-16 2V batteries, getting the SGs up after heavy-cycling. BUT, in my experience, and that of some neighbors that have recently replaced L-16s with Surrette L-16s, there have been NO protracted issues with cycling and getting batteries fully-charged, relatively easily ... not to be argumentative. AND, do not recall seeing similar references, recently, on this or other Forums ... just sayin'.

FWIW, Vic

herodotus · July 2016

Vic said:

Couple of thingsl
hero.., you will probably have some difficulty reaching 0.5% of C, especially at the recently-elevated Absorb voltages ... and EA behavior will probably change as the bank gets more cycles.

Actually, following Rolls's recommended charge voltages and absorb times, I don't think I will have trouble getting to ~0.5% EA. My peak charge current is about 45A at the moment. Following the standard formula 0.42 * 470 / 45 gives me a bit over 4 hours. After 2-3 hours of absorb I'm seeing currents of < 4A going into the bank. Maybe I won't quite get down to 0.5%, but it won't be far off.

I'm also noticing that the SGs of the bank don't easily go above about 1.255 unless I really boil the batteries - which seems consistent with the idea that they need a bit of breaking in. To be honest, at this point, I've heard so many different opinions that I'm going to stick (broadly) to what Rolls themselves recommend - which includes:

Tolerably regular deep cycling to 50%
Long(ish) absorb times at 60V
Expecting it to take ~50 cycles for the batteries to be broken in

As time goes by and the batteries settle down, I will look at trimming back the absorb times and experiment with working at a lower average SOC more of the time with a view to extending the life of the batteries, but I've found nothing in the Rolls documentation to back that up, so to start with I'm going to do it "by the book".

herodotus · July 2016

Just a quick follow up - I went for the full 4 hour absorb today at 60V. Return amps reached 4ish (0.85%) after about 3.5 hours, and slowly crawled down to 3.8 in the last half our or so. I'm not sure if it would have got any lower if I'd carried on, but it was certainly dropping very slowly by that point - down only 0.7A over the last 1.5 hours. By the time it went it to float SG on my test cells was hovering at around 1.255 - but as I mentioned before, so far I've been unable to get it much higher than that without really boiling the cells. I'm going to keep cycling the bank like this for the next few weeks, and I'll give it a heavy EQ charge once a month to try and drive the SGs right up.

MarkC · July 2016

herodotus said:

vtmaps said:

herodotus said:
In practice for reasons that aren't entirely clear to me yet (still some poking around to do on this) we're actually struggling to pull more than 2.5k from the panels at the moment
You have 3750 watts of array. With a 0.77 derate you should expect to see 2887 watts.

Yeah. And the 0.77 derate is obviously just an average for use in specifying a system. At the moment I'm hunting around for explanations of where *specifically* I'm losing all of that extra right now. I can account for some of it with temperature (although lacking an IR thermometer I'm not sure exactly how much I should really expect to be losing there), panel tilt (I'm at 40 degrees and solar normal is about 13 here at this time of year) and a bit of cable loss (~1.5%), but unless my cell temperatures are ~75C (nominal is 47C at 20C ambient with a light breeze) there's something else going on... but that's another question that I'm working on separately!

I might have missed, but did you ever reconcile your measured production?

If I did my math correctly, the angle correction would reasonably lose about 10% of the surface for irradiance. I've measured panel temps with an IR gun and during hot summer temps (90 oF), the temp rise at PV surface is between 25 and 30 oF (- 17-20 oC). Seems the .77 derate would be close???

I'm grid tied, so my MPPT inverter always tries for the max "watts". However I do see lower wattage production that I would expect (based on power reduction due to the PV temperature and mounting angle) - especially during the hottest months. I have seen (on the other hand), above rated wattages during fall/spring days with bright clouds. I've not washed my panels for over 2 years

, so might be part of it.

herodotus · July 2016

So another update:

Solar output: I disconnected my array->CC cabling from the CC and the array, bonded the ends together in a loop and measured the resistance of the whole lot. It barely registered on either of my multimeters - certainly less than 0.1 ohm in total. So I'm pretty sure it's not that. I also turned each string on separately and looked at the readings on the CC, and any differences in output get lost in the normal fluctuations of each string. Tomorrow I'm going to try popping the diode box covers off each panel and measuring the voltages of every panel during bulk charging, but I've seen nothing to suggest that I'm going to find anything enlightening by doing so.

A friend who lives not that far away and has a similar Outback setup (with a smaller array) is typically seeing his MPPT bulk settling on 76-78V rather than 70-71V, so I remain convinced that my problem is that the voltage is being pulled down for some reason. I think I'm going to have to try and find someone with an IR thermometer to measure the temperature of the panels so I can quantify how big a discrepancy there really is.
Battery charging: I gave the batteries a deep discharge down to below 1.200 SG and then did a charge with a 4 hour absorb (@60V before temp compensation). The SGs only reached about 1.220-1.230 despite current into the bank (which I'm not using to terminate the charge right now) dropping down to below 5A :-/ In the end I used the last of the sun to do a short EQ at 63.6V and got the SGs back up to 1.245ish. Today I'm trying a 5.5hr absorb at 60.5V (before temp compensation). Although I don't want to be hitting 100% every day, I still think that a normal charge (perhaps only done every other day) should be getting me above 1.260.

Cheers,

Wheelman55 · July 2016

H...the IR Thermometers are pretty inexpensive. Why not just go out and buy one?

You likely know this, but just in case, you measure the panel temps from the bottom (back) side of the panels.

https://www.amazon.com/s/ref=nb_sb_ss_c_1_9?url=search-alias%3Daps&field-keywords=ir+thermometer&sprefix=undefined%2Caps%2C274

Dave Angelini · July 2016

I did not see anyone in this long thread mention running loads that will cycle the battery and avoid all of this. A mini-split heat pump cooling and heating occasionally over night and 2% end amps with an occasional test cell SG and you are done.

herodotus · July 2016

Dave Angelini said:

I did not see anyone in this long thread mention running loads that will cycle the battery and avoid all of this. A mini-split heat pump cooling and heating occasionally over night and 2% end amps with an occasional test cell SG and you are done.

Well, I am thinking about adding some water heating to help fix this problem, but it's only part of the story. As I mentioned above, I've done a good deep discharge, and if I'd stopped the subsequent charge even at 1% end amps I would have been *way* off fully charged. Even going well below 1% and absorbing for 4 hours my batteries aren't getting anywhere near fully charged after a reasonably deep discharge. I'm having to charge at high voltages (>60) for long times (>5hrs) to get anywhere near 1.260 SG.

STOP PRESS: Just thought to double check voltages with my multimeter and it turns out that the FM80's battery reading was a full 0.5V higher than what it was actually putting into the batteries. The PV voltage was dead on what my multimeter read, but the battery side was completely out. I guess that would probably explain a fair amount of the problems I've been having with getting things fully charged! Have adjusted the FM80 battery voltage calibration now. Will have to wait and see what happens as I go round another charge cycle to see if it has improved things.

Vic · July 2016

hero..,

Some of the apparent poor voltage calibration could possibly be from voltage drop in the cable from the CC's output to battery connection.

The CC might be very close to the DC breaker and the run to the battery, or battery buss might be very short. And the cable may well be the largest that the various terminals can accept, but please re-torque all connections, as cable cold-flows after a period of time, and so on.

Your DMM can measure the voltage drop in all of your cables, if the test leads are conveniently long, and so on.

FWIW, Vic

herodotus · July 2016

Vic said:

hero..,
Some of the apparent poor voltage calibration could possibly be from voltage drop in the cable from the CC's output to battery connection.

I don't think so, since I've had the cover off the CC and was measuring the voltage at the output terminals of the CC itself - and there it disagreed with my DMM by 0.5V... whereas testing the input terminals from the PV gave exactly the same reading as the CC was reporting.

Vic · July 2016

Hi hero,

OK, had read the previous post as saying that "at the battery" the voltage was different.

0.5 volt error is significant. Am surprised that the CC's output V metering would be that far off.

Have you tried measuring the voltage drop from the CC battery + terminal to the battery +, just for kicks? With the CC wiring compartment cover of the CC off, you could just measure the CC bat +, to bat -- terminals, and immediately measure the battery voltage, ideally during early Absorb, where the current is high, and the CC is regulating voltage. Sometimes one can be surprised at the accumulated drops in cable and breakers and connections at a nominally high current.

And using any standard DMM to measure the resistance, as you noted above, is very imprecise, as contact resistance in the meter jacks, lead plugs, R in the leads and in the connection of the lead to the test lead prod, itself are probably higher than the R of the cable under test. Also, there is a spec of the meter, where it needs +/-- several "counts" just for the meter and its conversion process, etc.

You would want a MilliOhmeter, or, easier, is to run a known current of 10 - 25 Amps DC through the cable under test, and measure the voltage drop ... as you might guess.

FWIW, Vic

mike95490 · July 2016

extra panels now=less generator time in winter
This is a time to exercise the breakers in your PV Combiner box. Switch all but one array off, let it stabilize for 5 min, then measure voltage and amps. Go to next string and compare them. If they are all the same, you likely don't have a high resistance issue on one of them. Get an IR thermometer and scan/shoot each splice/connection. Is one warmer than the others ? Bingo.
Switch the array off one morning, and let the batteries run down, Wait till just before solar noon the next day, Power up one string at a time and test it, batteries are low and will demand max current.

herodotus · July 2016

Vic said:
0.5 volt error is significant. Am surprised that the CC's output V metering would be that far off.

Yeah, I was pretty surprised myself! But the fact that the DMM agreed with the PV input reading on the FM80 means that one side of the FM80 must be wrong by that amount, unless something truly bizarre is happening!

Have you tried measuring the voltage drop from the CC battery + terminal to the battery +, just for kicks? With the CC wiring compartment cover of the CC off, you could just measure the CC bat +, to bat -- terminals, and immediately measure the battery voltage, ideally during early Absorb, where the current is high, and the CC is regulating voltage. Sometimes one can be surprised at the accumulated drops in cable and breakers and connections at a nominally high current.

Will give this a go tomorrow. When I was measuring today there wasn't a lot of current flowing so there wasn't much point. I'd be a bit surprised if it was huge since it's a very short run of 16mm2 cable, and in reality I'm rarely pushing more than 45A into the batteries right now.

And using any standard DMM to measure the resistance, as you noted above, is very imprecise, as contact resistance in the meter jacks, lead plugs, R in the leads and in the connection of the lead to the test lead prod, itself are probably higher than the R of the cable under test. Also, there is a spec of the meter, where it needs +/-- several "counts" just for the meter and its conversion process, etc.
You would want a MilliOhmeter, or, easier, is to run a known current of 10 - 25 Amps DC through the cable under test, and measure the voltage drop ... as you might guess.

Yeah, I'm aware it's not super accurate to measure resistance like this. The meter doesn't register any resistance more than it does when touching the two leads together, and this is replicated across all three meters. I have also tested the voltage when the system is under load, and the voltage at the panels is about 1V higher than at the CC when the CC is registering an input of ~20A, which would suggest a resistance of < 0.1 ohm as well.

herodotus · July 2016

Hey Vic,

So you where right to be sceptical

I measured everything under load again. The round trip from the CC to the batteries costs me about 180mV at 43A. A fair amount of that is the breaker; I don't see any evidence that there are bad connections, I guess the rest must just be the cabling itself - and in fact that's consistent with what I'd expect from a 2m round trip of 16mm2 cable. Perhaps I should consider replacing it with 35mm2; when I specced it I was really only thinking about losses (a few watts here or there doesn't make much difference) rather than about the ability of the CC to measure the battery voltage with precision over a range of different currents. Really there ought to be a way for the FM80 to use the value measured by the FNDC directly from the battery for all of its charging logic - and that value should itself having a calibration option. Sadly, neither of these things is true

On rechecking the array cabling, I'm losing more like 2.4V on my cabling from the array to the CC when the current is 35A - which is still < 0.1ohm but is more loss than I thought I was getting.

Measurements in the 50-70V range are obviously complicated by the fact that I don't know whether to trust any of my meters' accuracy down to fractions of a volt in that range. I've ordered a reasonably well respected voltage reference from ebay so I can calibrate, but it may be a while before it gets here.

For the time being I've maxed out the FM80s battery voltage calibration at -0.5V since I'm definitely 0.2V lower at high current thanks to losses and most likely more (but won't know for sure until I get my voltage reference).

herodotus · July 2016

Another quick update for anyone interested. I've been running 5 hour absorbs over the last couple of days, since this is what Rolls' formula would recommend officially based on a charge current of 40A. I'm also running with my voltage calibration on the FM80 maxed out at -0.5V, so the effective absorb voltage is going to be that much higher as well.

During these charges my end amps clearly bottoms out at 3.5A, which is about 0.7%. If the batteries aren't particularly discharged at the beginning of the absorb (as they weren't today), I hit this value quite some way before the absorb timer terminates the charge, as would be expected. My SGs this evening are back to 1.270ish, so those long absorbs at a higher voltage are obviously making a difference.

From here, I'm going to try another reasonably deep discharge (panels off tomorrow), and then see what happens to my SGs if I charge up to an end amps reading of 3.6A (also interested to see how long it takes to get to that value).

IR thermometer on order (it's only getting hotter here!). When all my shiny measuring tools arrive I might be able to shed some clearer light on what is actually happening here...

Cheers,

herodotus · July 2016

Another update. I wrote to Rolls technical support for some advice on breaking in batteries, and I've now had a response. Steve's recommendation based on my system details was, broadly, to do longer absorbs without pushing up the voltage (and to drop the voltage for the inverter charger, not that I ever really use that). He also recommended sticking to a lower EQ voltage (61) "as too much EQ can be a bad thing". Overall "[t]he Batteries are going to respond to a lower charging voltage for a longer period of time better than hitting them with higher charging voltages for shorter periods of times."

The final point was "Yes running them, or cycling them early will help break in overall performance faster... but is not always necessary."

So I think it sounds like I'm doing broadly the right thing. When I get my voltage reference I should be able to configure the system exactly, and I might need to replace my CC->battery cables with thicker ones to help the CC maintain battery voltage more accurately, but in principle I think setting up the longer absorb times + lower end amps is probably the right answer, and I won't worry *too* much about the cycling - I'll just make sure I give them a decent discharge periodically.

Vic · July 2016

Hi hero..,

Thanks for the Updates. Good that you communicated with Surrette.

That suggested EQ voltage is surprisingly low -- Surrette used to recommend a MINIMUM of 61.9 V for their batteries.

Agree about the forced cycling ... have usually just cycled the batteries normally, and have not ever seen any issues. Generally it is good to try to cycle below about 90% SOC on each cycle, even it that means that it might require several days for the bank to become discharged by that much.

Then there is the strategy that avoids fully-charging Flooded batteries, cycling them between about 75-ish percent and 90-ish percent SOC, and fully-charging them within several weeks, and perhaps EQing somewhat more frequently than might normally be expected.

Would like to hear Surrette's comments on this approach ... will e-mail them.

Have fun with the new system. You are paying careful attention to the batteries, which is a great way to start.\Thanks, Vic

Vic · July 2016

Should have mentioned (but you appear to have seen the above post, so will not Edit it).

There may be a considerable voltage dorp on the CC to battery connection, in the DC breaker in the circuit.

Out of curiosity, what is the manufacturer, and model number of that breaker? It is difficult to see this info on an installed system, so you may not know. You could measure the drop across this breaker, but, again, it may be quite difficult for you to access the terminals without tearing into the system.

FWIW, Vic

herodotus · July 2016

Vic said:

Should have mentioned (but you appear to have seen the above post, so will not Edit it).
There may be a considerable voltage dorp on the CC to battery connection, in the DC breaker in the circuit.
Out of curiosity, what is the manufacturer, and model number of that breaker? It is difficult to see this info on an installed system, so you may not know. You could measure the drop across this breaker, but, again, it may be quite difficult for you to access the terminals without tearing into the system.
FWIW, Vic

Hi Vic,

It's not actually *that* bad. I measured it the other day when I was looking at all the different voltages in the system. The total loss from the CC to the batteries when the charge current is about 40A is about 180mV. Of that, about 70 was the breaker, and the rest was the cable (16mm2 to the positive terminal of the battery and to the battery negative bus, and 35mm2 from the battery negative bus to the battery negative terminal). I'm thinking of replacing the 16mm2 with 35mm2 to keep the battery charging voltage as close as possible to what the CC thinks it is, particularly if I can get a bit more output from the panels when it's cooler/when I fix whatever is limiting it at the moment.

Lucian

WaterWheel · July 2016

At this point I'm still wondering if the better higher SGs are a result more of the .5v higher voltage from the calibrated meter or the longer Absorb time.

Anybody have an opinion on which is making the larger difference?

SolarMusher · July 2016

Longer absorb time.

WaterWheel · July 2016

My thought as well, I recently went to a different charge controller but SGs are very low by the end of the Absorb cycle. Trying a longer absorb time myself.

Dave Angelini · July 2016

Why not just do this the way the battery manufacturer wants you to do? Use their bulk absorb / voltage and adjust the time until the SG is correct? There seems to be alot of overthinking these days?

Daily charging parameters

Comments

Categories