Axpert settings for LiFePO4

[Adri]

My AXPERT 1000VA/24V works perfectly. I have replaced the Trojan T-105s with new AKKUTECH LiFePO4 batteries (2 in parallel).
I set the bulk setting (program 26) to highest possible value = 29.2V (instead of 32V). 

It charges well on PV and "utility" (Heishkom) but at about 27.5V it switches to "absorption" (program  27) and charge current drops to almost nothing. This means in theory that I will never be able to reach max.  capacity of 2 x 100Ah. Why does that happen? Shouldn't it keep charging until close to or equal 29.2V?

What should I set program 27 to?

Thanks.

 

 

[Adri]

Someone might be interested in my monitoring of my batteries : to show true batt. volts and charge/discharge current.

I use a Mooshimeter I bought on the 'net.  It can display 2 functions simultaneously : my case batt. volts and true charge/discharge current.  I connected 3 input plugs to (a) B+ terminal of Axpert, (b) B- terminal on Axpert and (c)  B-  terminal at the batteries, so creating a small volt-drop. It's then displayed in the 100mV range which is exactly proportional to the current [in both directions].  By quick calibration I measure about 5mV/A.

It's that simple and I carry my phone/display anywhere in the house.  It has logging functions I haven't mastered yet!

If someone needs help/more info, let me know.

 

[Coulomb]

8 hours ago, Adri said:

I set the bulk setting (program 26) to highest possible value = 29.2V (instead of 32V).

I assume you mean that your Trojans would have preferred 32 V (at least for equalisation), but instead you used 29.2 because that's as high as the inverter would go.

But now you have an LFP battery; it is a totally different chemistry and should not be charged to 29.2 V (presumably 3.65 V per cell). It's not obvious what you are meant to charge to from what's written on the battery label; "cut off" seems to imply an  "absolute never exceed" value. So you need something less than 28.4 V. Perhaps the battery has a BMS that is actually cutting the battery off when the charge voltage reaches some 27.5 V. Though that's a slightly low value for 8S LFP; I would suggest 27.6 V. Or ideally, find a datasheet for your battery and follow the manufacturer's recommendations.

It may be that the imperfect MPPT tracking is overshooting the bulk charge voltage set-point and upsetting the battery's BMS (if it has one). By using a more realistic bulk charge voltage, it might start behaving. Use a slightly lower value for the float charge voltage (setting 27); say 27.4 V, though you could go down to 26.9 V for longer battery life.

[Adri]

Thanks for the reply, Coulomb.

The LFPs are fitted with BMS.

I have now set the bulk to 28.4V and float to 26.9V.

I'll see what happens now and give feedback. 

(My batteries are in different enclosure, no displays.)

 

AT-SPEC-LiFePO4 Battery 25.6V100Ah SPEC.pdf

[Adri]

Details from spec.sheet of AKKUTECH 25.6V  100Ah batteries :

Nominal voltage       25.6V

Energy               2560Whrs

Charge Mode         Constant charge at 0.2C to 28.4V

Charge current        20A

Max. charge current          50A

Charge cut-off            28.4V  +/- 0.2V

Max. continuous current             50A

Discharge cut-off         21.6V +/-  0.8V

 

(reduced info from spec.sheet)

 

 

[Coulomb]

4 hours ago, Adri said:

Charge Mode         Constant charge at 0.2C to 28.4V

...

Charge cut-off            28.4V  +/- 0.2V

It's pretty weird that the recommended charge voltage is equal to the charge "cut-off" voltage. Or it might even be 0.2 V too high, per the specs.

But maybe "charge cut-off" doesn't mean what I expect (i.e. the voltage at which the BMS will cut off the battery to prevent overcharging).

Also, 0.2C recommended charge current seems low for an LFP battery.

[Adri]

I am as confused as you are, Coulomb.   Perhaps it's a language/translation problem?

I will be in touch with the supplier next week and will post new information.

Thanks.

 

[Adri]

20191216

I have copied the most important details from the spec.sheet because I can't attach the .pdf for some reason.  I cannot get it to charge above 27.5V.  I am thinking that as Coulomb says that the BMS says no at 27.5V and shuts off charging, on PV or mains.  My Mooshi never shows higher than 27.5V.    I can't speak to supplier yet as it's another holiday here (Mon. 16 Dec.).

 

Akkutech spec1..ods

[PJJ]

On 2019/12/13 at 2:49 PM, Coulomb said:

Also, 0.2C recommended charge current seems low for an LFP battery.

I also noticed that on the cells from https://lithiumbatteriessa.co.za/ as well.

The recommended charge current on the spec sheet for those cells are also 0.2C 

[Coulomb]

1 hour ago, Adri said:

I cannot get it to charge above 27.5V.

What does the LC Display on the inverter say that the battery voltage is just before it switches to float?

When the current drops and you are assuming it is in float mode, is the charge LED (middle LED, green) on solid?

[Adri]

13 hours ago, Coulomb said:

What does the LC Display on the inverter say that the battery voltage is just before it switches to float?

 

27.5V  - same as Mooshi.

 

13 hours ago, Coulomb said:

When the current drops and you are assuming it is in float mode, is the charge LED (middle LED, green) on solid?

Flashes.  That's why I assume it's in float.

[Gerrie]

On 2019/12/08 at 2:32 PM, Adri said:

I use a Mooshimeter I bought on the 'net

Where did you buy this mooshi?

[Adri]

2 hours ago, Gerrie said:

Where did you buy this mooshi?

Google for suppliers, e.g. SparkFun.

 

[Coulomb]

Quote

27.5V  - same as Mooshi.

Flashes.  That's why I assume it's in float.

Then it looks like the Premature Float Bug.

Does your model have support for bulk/absorb stage time, e.g. does setting 32? If so, you could set that to say 90 minutes to try and get more charge into your battery.

Otherwise, there is nothing you can do to make the inverter-charger behave; you'll need an external MPPT solar charge controller, and not use the PV-in on the inverter-charger.

Edited December 18, 2019 by Coulomb

[Adri]

No, the "MK" model has those features to set time but mine ("MKS") does not.

The  manual is 3.5MB (.pdf) so I have split it into two posts.

 

Axpert MKS-MKS PLUS 1-5KVA Manual on 20140306-pages-1-15.pdf

[Coulomb]

5 hours ago, Adri said:

No, the "MK" model has those features to set time but mine ("MKS") does not.

Right. I can't see a solution other than the external MPPT. Sadly.

[Adri]

I fiddled much in the last week.    I now have bulk set 29.2V (the max. my Axpert allows me to go)  and float 29.1V and it's charging perfectly, on either PV or mains. And batt's are still room temp.       So I'm thinking it wasn't the BMS triggering the previous 27.5V cutting to float.

Possibly the label, spec.sheet indicate the 28.4V limit to have the batteries remain in "safe area". But then is the 100Ah capacity based  on 28.4V?

Any ideas?

 

[ojeysky]

On 2019/12/26 at 12:44 PM, Adri said:

I fiddled much in the last week.    I now have bulk set 29.2V (the max. my Axpert allows me to go)  and float 29.1V and it's charging perfectly, on either PV or mains. And batt's are still room temp.       So I'm thinking it wasn't the BMS triggering the previous 27.5V cutting to float.

Possibly the label, spec.sheet indicate the 28.4V limit to have the batteries remain in "safe area". But then is the 100Ah capacity based  on 28.4V?

Any ideas?

 

I use Axpert Variant as well, but when i set mine at 29.2 the BMS cuts in at 27.5 but it occasionally will spike to 29.1v and then go way down to 26.9v. I am wondering if you are experiencing similar. I feel 29.2 is quite high for lithium, so i have set my bulk to 27.6 and float (even though float is not applicable in lithium) to 27.2v don't know if this will be sufficient for my 200AH LifePo4 battery.

[Coulomb]

3 hours ago, ojeysky said:

i have set my bulk to 27.6 and float (even though float is not applicable in lithium) to 27.2v don't know if this will be sufficient for my 200AH LifePo4 battery.

Those should be fine, for an 8S LiFePO₄ of any capacity.

Edited February 6, 2020 by Coulomb

[ojeysky]

2 hours ago, Coulomb said:

Those should be fine, for an 8S LiFePO₄ of any capacity.

Thanks @Coulomb but what i have noticed is that once it gets to 27.6v on the inverter after a while the battery voltage drops to 26.9 and the inverter float status gets indicated and it does not leave float thereafter and charging current goes to zero. I am concerned that it seem my battery is not getting adequately charged (i like it to be up to 90% full). I have just tweaked the settings to 27.8v for both bulk and float and now the battery voltage indicator on the battery stays between 27.4 to 27.6v even though the voltage reading on the inverter is between 27.6v and 27.8v. It seem there is si 0.2/0.3v difference between the meter on the inverter and that of the battery

Is my new settings of 27.8v(bulk and float) still fine or will i be killing my battery earlier than latter?

If i disconnect the charging source, battery voltage drops rapidly from 27.6 to 26.6 hence it seem i am yet to achieve a 90% charge on my cells?

[weber]

Coulomb asked me to look in here, as he is rather busy at the moment.

I believe Coulomb has already correctly explained what's going on here, namely the infamous Voltronic Premature Float Bug, which is in every Voltronic inverter firmware we've ever looked at. Instead of going to float when the battery voltage is near the absorb voltage (setting 26) and the current has fallen below one fifth of the maximum charge current (setting 02), it goes to float when the battery voltage is near the float voltage (setting 27)!! and the current has fallen below one fifth of the maximum charge current (setting 02).

@Adri If you can find us a firmware update file for your Axpert MKS 1K-24, we may be able to patch it to fix the bug, as we have done for several other Axpert models. But no promises.

@ojeysky What model is your Axpert inverter? We may already have patched firmware for it. See the Firmware section here.

Failing that, there are two ways of avoiding the bug:

1. (As you have discovered) Set the float voltage (setting 27) to what the absorb voltage setting should be, i.e. don't have a float mode. This is a terrible thing to do to lithium batteries. Unlike lead-acid, they age most rapidly when held at a high voltage.

2. Don't let the charge current fall below 1/5th of the max charge current setting (setting 02). That means you either have to keep the charge current high (until the absorb voltage is reached), or you have to set the max charge current setting (setting 02) low. For the Axpert MKS 1K-24 I suggest putting setting 02 to 20 A. For more powerful Axpert models I suggest setting it to a little less than what your PV array is capable of, but no more than 40 A for a 200 Ah battery (0.2C).

What were your settings 02 and 11? And what charge currents do you normally see from your solar panels?

Does the problem occur when you charge from utility?

Coulomb also gave correct absorb and float voltages for 8S of LFP cells, namely 27.6 V to 28.4 V for absorb and 26.9 V for float. I suspect you will need to use 28.4 V for setting 26 for two reasons:

(a) so that the series cells will get their SoCs balanced by the BMS, and

(b) to keep the charge current high all the way to absorb, to avoid the premature float bug.

But you will still be hit by the premature float bug on overcast days.

There's another issue. Current sharing. When you have two batteries in parallel you have to ensure you have the same resistance in the link between their positive terminals as between their negative terminals, and the takeoff cables must come from the positive terminal of one battery and the negative terminal of the other (diagonal takeoff), with the same lug-stacking order (e.g. takeoff lugs on top). And the batteries should be stacked beside each other and have the same thermal environment, as internal resistance goes down as temperature goes up, and so a small current imbalance can become a larger one.

Edited February 7, 2020 by weber

[ojeysky]

5 hours ago, weber said:

What model is your Axpert inverter? We may already have patched firmware for it. See the Firmware section here.

It's an Axpert variant(clone?) specifically felicity IVCM3024 here is the link to the inverter. Will I still be able to do the update?

 

5 hours ago, weber said:

(As you have discovered) Set the float voltage (setting 27) to what the absorb voltage setting should be, i.e. don't have a float mode. This is a terrible thing to do to lithium batteries. Unlike lead-acid, they age most rapidly when held at a high voltage.

Noted with thanks

 

5 hours ago, weber said:

Don't let the charge current fall below 1/5th of the max charge current setting (setting 02). That means you either have to keep the charge current high (until the absorb voltage is reached), or you have to set the max charge current setting (setting 02) low.

I am off-grid so i may not be able to always ensure that charge current remains high as it depends on the weather condition but my PV array(2.25kw) on a sunny day does up to 50A

 

5 hours ago, weber said:

For more powerful Axpert models I suggest setting it to a little less than what your PV array is capable of, but no more than 40 A for a 200 Ah battery (0.2C).

Okay I have set the max charge to 40A

 

5 hours ago, weber said:

What were your settings 02 and 11? And what charge currents do you normally see from your solar panels?

Settings 02 was on default of 60A but i have just changed it to 40A.

Settings 11 is at default of 30A but its not applicable as i am offgrid

5 hours ago, weber said:

I suspect you will need to use 28.4 V for setting

Done and i have reduced the float to 26.9v as well

 

5 hours ago, weber said:

When you have two batteries in parallel you have to

This is not applicable I only have one 200AH

Thanks a lot for the support @weber

Edited February 7, 2020 by ojeysky

[ojeysky]

5 hours ago, weber said:

And what charge currents do you normally see from your solar panels?

@weber here is my battery charge current log, and here is my PV harvest log. I started using LFP in the last 3 weeks, so you may only correlated figures within the last 2-3 weeks

Edited February 7, 2020 by ojeysky

[weber]

@ojeysky Sorry to have got your hopes up in regard to firmware. We only have patched versions for the 5 kVA 48 V models. So it's the same offer as for Adri. If you can find us a firmware update, we'll look at it, but no promises. I can't tell whether the IVCM3024 is a rebadged Voltronic or a copy made unethically by another company (a clone), but since we don't have a firmware update, it doesn't matter.

I also said the wrong thing about the current setting (setting 02). I suggest putting it a little lower than the maximum that ends up going into the battery (after loads). Based on your battery charge current log, it looks like 30 A would be the best compromise between missing out on a little charge when the sky is clear and avoiding premature float when it is cloudy.

Please let us know how this goes. It is quite possible that I'm completely wrong. The explanation for the strange behaviour may lie elsewhere.

Edited February 7, 2020 by weber

[ojeysky]

4 hours ago, weber said:

Based on your battery charge current log, it looks like 30 A would be the best compromise between missing out on a little charge when the sky is clear and avoiding premature float when it is cloudy.

It was quite sunny today and I didn't hit float all through the day. Do you think I should still change the charge current to 30A or I should just maintain the 40A currently set?

 

4 hours ago, weber said:

If you can find us a firmware update, we'll look at it, but no promises

I will look for it. Though I bought the inverter to serve as stop gap while I receive replacement part for my PIP-3024GK which suddenly threw up error 32. I know the float bug affects that as well, do you by chance have the patch for those? as it will be useful when I finally get to put my GK back online.

Edit: It's almost 6pm and now the inverter has a float indicator.

Thanks @weber

Edited February 7, 2020 by ojeysky