What approach are you folks using when trying to arrive at a proper EA value? I'm watching my system today, and I know my batteries have had enough, specific gravity is to spec.
Problem is the EA reading is jumping anywhere between 3.5 - 4.3 amps. With such a "large" variance in the value from second to second, what would you suggest I use for EA? I've had the system for years, but thought I'd once again look at using EA to terminate absorb after loading the latest firmware, but it doesn't appear that EA is any more accurate than it was years ago. I've just been using a timed terminate, and it works fine, but would really like to revisit EA.
I guess the real question should be, why does the EA value jump around so much?
I'd really like to know why that value is not a lot more accurate, it's kind of useless at the moment with the 8/10's of an amp swing I saw, who know's, it could have been more. That was the biggest difference I saw during the 5 minutes I was watching.
I think the reason it is jittering is because it is in Absorb and the voltage is regulating on and off.
This is most likely because the input PV voltage related to battery voltage may be a high ratio and it is hard to keep it absolutely steady.
You might need to set the ending amps slightly higher ?
Are you using the Whizbang Junior ? Maybe the low pass filtering needs to be a longer time constant in the Classic for EA ?
Hi Bob, yes I'm using a wbj. Does that low pass filtering you speak of have a setting somewhere, not familiar with it. I'll have a look through the manual as well.
Muskoka,
We, also, notice a bit more Jitter than we would like.
Several thoughts; We care most about EA, near the end of Abosrb. At that point in the recharge, the batteries are at a relatively high impedance part of the cycle. This makes the jitter greater. Noise from the inverter (ripple currents, cause ripple voltage at the battery terminals), can cause quite a lot of jitter, particularly, if one is running some Opportunity loads, near the end of Absorb. Changes in load current on the inverter (if you are using one), causes changes in voltage drop in the cable and breaker on the CC output, will cause changes in the EA value, and so on.
The exact details of your system are not clear (inverter, Vin to the Classic, etc). Some of those details affect the minute detail of the behavior of the system.
One other thing that we have noticed, is, that for our aged FLA batteries, the proper EA value -- where the battery charge current is no longer descending, varies, depending upon the DOD in the previous discharge. Have begun to think, that we could do well if CCs had a Rate Of Change in battery charge current setting, to end Absorb. This is really what we all are watching for, when we watch EA during the last 10 - 30 minutes of an Absorb stage ...
All, just opinions, FWIW, Vic
Quote from: Muskoka on March 03, 2020, 02:54:55 PM
Hi Bob, yes I'm using a wbj. Does that low pass filtering you speak of have a setting somewhere, not familiar with it. I'll have a look through the manual as well.
Sorry, there is no setting for a filter for the WB Jr. current.
The EA current has to be below the setting for 1 minute. For now, can you raise that EA setting just a wee bit so that it can stay there for 1 minute or more ? Maybe the setting now is low enough and can stand to be a bit higher ?
Also, the Absorb timer can be set so that it is just a bit longer than you see it take for the ending amps to reach the level it is at now. That way, it is also timed.
Thanks Vic and Bob for the responses.
Sure, I can set the EA to a number that "will work", but it's not the optimal number. What I was trying to do yesterday was find the "optimal number" for EA, by setting the absorb time to a very low number, and see how low the EA would go. Right now I don't know if the true "batteries are full" EA number is closer to 3.5 or 4.5 amps as that's the range it was bouncing around.
So, to be safe, I'll set EA somewhere in the neighborhood of 5 amps, and see how it goes.
Quote from: Muskoka on March 04, 2020, 06:45:45 AM
Thanks Vic and Bob for the responses.
Sure, I can set the EA to a number that "will work", but it's not the optimal number. What I was trying to do yesterday was find the "optimal number" for EA, by setting the absorb time to a very low number, and see how low the EA would go. Right now I don't know if the true "batteries are full" EA number is closer to 3.5 or 4.5 amps as that's the range it was bouncing around.
So, to be safe, I'll set EA somewhere in the neighborhood of 5 amps, and see how it goes.
If you want to get a good idea for EA setting. Log the WBjr amps with the LA for a day with the Absorb timer set a little higher than normal. Export the LA data to an MS Excel file and then graph the WBjr data over time and you will get a nice curve showing the bulk charge and then show it ramping down in Absorb and you will see the the graph level off. It will also show the range of the noise. I usually set the graph to have 1 amp minor lines. Yes it’s noisy, I usually see about 1 amp of jitter. Especially if my laptop is plugged in to the inventor and is charging. I’ll try to find one of my graph and post it when I’m on my laptop which is where I have them all saved.
Re-reading my post I incorrectly said I set absorb time to a low number, I knew what I meant, but it "reads" wrong, to the readers here. I have absorb set to 6 hrs, I'm normally in float within 3 hrs, so 6 hrs will work fine for trying to find a EA number.
Gathering, reading the data is not really the issue here. The issue is the amount of jitter. You can't accurately set a EA value when it's jumping around + - a full amp every second.
Does it really matter in the long run if EA's set at say 4 amps or perhaps 5 amps, maybe not, but your still just "picking" a number instead of a precise value. Like I stated, mine was jumping from 3.5 to 4.3 amps, and that's just what I noticed, it could have been more. The frequency of logging doesn't allow for "precise" readings, so there could have been a much greater variance than just the 8/10's of an amp that I saw.
So to safely ensure my system goes to float using EA, I'll have to set my EA somewhere close to 5 amps, even though I saw readings as low as 3.5, but I can't "trust" that number, due to the jitter. That's all I'm saying. Is it going to make much of a difference over the life of the batteries, probably not, because you have to constantly adjust these values as the system ages anyways, but it would be nice if the data / readings didn't fluctuate so much.
It probably really doesn't matter all that much because like Vic mentioned the actual end amps setting may change depending on what was going on with the batteries.
Plus when batteries are close to full they will get to the end amps point pretty quickly even if it is off a bit.
And lead acid batteries are very forgiving .
The way I see it they are not precision devices and don't we all know people who have not done much of anything special and they get a long life out of their batteries ? Sure the opposite could be true - people ruin their batteries but that is usually either total neglect or the batteries are cycled so low they just don't have a long life .
Ross W has monitoring on his individual 2v cells and there are devices sort of like bms that can keep them all in balance - that is probably the best way to get real accuracy and longevity .
Larry
Muskoka, could you tell me, if possible, about how long you see the jitter rise above the EA set point current ?
Is it seconds ? 10's of seconds ? Minutes ?
It may be that what needs to be done is to make it so that that 60 second or 90 second EA timer (I can't remember which one it is), just needs to 'wait' a few seconds when that current rises above the set point level before it resets the timer ?
That shouldn't be too hard to do. Then, the current would still reflect the real jitter but wouldn't stop it from ending the Absorb cycle.
boB
Bob, I've only been looking at this the last couple days. I'm still trying to find a "happy" EA value to set. I haven't actually arrived at a solid figure yet.
Tuesday was a good solar day. If I remember correctly I had EA set to 2 amps, artificially low. I was looking for the point at which the amp's flat lined. It got as low as 3.5 to 4.3 amps. That's the widest range I witnessed while monitoring Grahams app.
So, I don't think I have an answer to your question, yet. It never got to the EA value I had set, so it "never did" rise above it, it was "always above" the artificially low value I had set. I would need to set it to some "close" EA value, one that my system could actually achieve, to give you a better answer.
Your last sentence is all I'm concerned about, going to float prematurely, because of the jitter, leaving the batteries not quite full.
Glen
Quote from: boB on March 05, 2020, 03:32:16 PM
Muskoka, could you tell me, if possible, about how long you see the jitter rise above the EA set point current ?
Is it seconds ? 10's of seconds ? Minutes ?
It may be that what needs to be done is to make it so that that 60 second or 90 second EA timer (I can't remember which one it is), just needs to 'wait' a few seconds when that current rises above the set point level before it resets the timer ?
That shouldn't be too hard to do. Then, the current would still reflect the real jitter but wouldn't stop it from ending the Absorb cycle.
boB
boB
Since the LA is recording data about every 2 seconds. You can get a value above your EA setting in one or two reading that makes it start the EA count all over. So I would say it’s seconds. The only way I have found to avoid that is to see what the top value of the jitter is after the absorb curve levels off and set EA above that value. All my graphs show a jitter value around one amp.
It’s pretty easy to see if you graph the WBjr amp reading and have a good vertical scale of at least 1 amp resolution.
And as Vic has said the EA value will change with varying loads.
Thanks guys.
I will look into seeing how much is involved in possibly making a quick change.
Here's an example of the WBjr Amp data from the LA. Pretty easy to see the jitter once it drops to Float. The data is with Loads on.
(//)
Quote from: Resthome on March 08, 2020, 03:32:24 PM
Here's an example of the WBjr Amp data from the LA. Pretty easy to see the jitter once it drops to Float. The data is with Loads on.
(//)
Thanks John !
Thanks guys !
Since Graham and Matt have gotten the raspberry pi and Classic all synced together and all the modbus data now available on MQTT it is easier to play around with the data in Node Red . There are lots of graphing options in Node Red along with writing functions etc.
I would like to ask - what would be the ideal way to chart looking for the correct ending amps point ?
Which data to graph or overlay in a graph ?
Would there be a formula to find the exact ending amp point according to the manufacturer data - since they give a percent change over time ?
sample graphs just so show the node red works ( created by Matt)
(//)
Larry