hello all.
I came here searching for answers about a strange way of work of this inverter that brought a friend of mine.

we have 15s litium batteries that is supposed to be charged @61v, so we set both bulk and float charge to 61v from the display.
the strange thing is that when on charging the voltage goes to 58v it stops to charge. I tink that there is some offset to calibrate and i read something here but i like to discuss about whit someone that have much more experience.

ty in advance 

 

i add some information:

firmware is 79.02 and i think that it is a clone but don't now the original reference.

On 2024/05/17 at 9:40 PM, deg87 said:

have 15s litium batteries that is supposed to be charged @61v,

15 X 3.6V Max would be 54V and at 3.65V absolute maximum voltage per cell you'd be looking at another 0.75V, so 54.75V maximum charge Voltage

On 2024/05/17 at 9:40 PM, deg87 said:

strange thing is that when on charging the voltage goes to 58v it stops to charge

I am guessing the BMS is preventing the further charging of the battery, since it probably is either, A: allready full and the Voltage is too high, or B: it is not happy with the excessive Voltage from the charging source

 

So, you should set your charge Voltage to no more than 54V and the float would probably be ok at, say 3.4V/Cell, thus 51V...

sorry but why you state that cell's have nominal voltage of 3.6?

the correct nominal voltage is 3.7v and the full charge is for theese batteries 4.2v.

this is not an opinion is the measured and declared data of these batteries.

maybe you assume is lifepo but it is not, it is a very large package of 18650 cilindrical batteries put in series and parallel to react a 15s and many ampere.

to avoid misunderstanding i manage to charge those before put under this inverter un to 61v (that is 4.05 for cell).

other clue is that i built a 7s version too on other system and i can charge it at 4.1 @ cell whithiut any problem and a scope of testing i tried to react 4.2v too and it goes like a sharm, so battery are cleare capable of doing it.

 

the bms is in both cases daly smart version and i think that is not the culprit but maybe i miss some settings?

the only i can find in bms settings is the nominal voltage and the max and min voltage to kick in protection, and the settings are the same for 7s and 15s but on 15s as i said it is impossible to go out of 58.4 v

 

Edited May 20, 20242 yr by deg87

1 hour ago, deg87 said:

the correct nominal voltage is 3.7v and the full charge is for theese batteries 4.2v.

I have not seen a Cobalt based Lithium battery in 15S configuration being sold for Solar purposes, 13S, yes, but not 15S... give me more info on the battery, make and technical specs...

Normal would be LiFePO4 based Lithium batteries which come in 15S and 16S flavours, 48V nominal and 51.2V nominal... but if that is not what you have, then you had better provide more details so I can understand what you have and what its parameters are.... and LiFePO4 cells are nominal 3.2V and 3.65V absolute max, probably best not to push above 3.6V and if you value a longer life, maybe limit it to 3.45 or 3.5V per cell... but apparently this does not apply to you battery....

1 hour ago, deg87 said:

the only i can find in bms settings is the nominal voltage and the max and min voltage to kick in protection, and the settings are the same for 7s and 15s but on 15s as i said it is impossible to go out of 58.4 v

Ok, re-reading your lot, I don't know the Daly BMS, but I suspect like most it would be happier with a 13S setup for Cobalt based Lithium cells and thus no more than 13 X 4.2V peak Voltage... 54.6V, maybe 14 cells at somewhat less than 4.2V... but I think if you research the Daly BMS, they may well have a combined Voltage limit, which you have reached at 58V...

i spoke whit bms dealer and he confirm that whit this settings the battery can be charged up to 63v, infact i have charged it up to61.5v whit dumb charger and no problems occurs. 

You are talking whit the battery manufacure because it'me🙃.

i got a lot of lithium battery pack that have the spec i tell (3.7v nominal voltage and 3 to 4.2 range). i got theese batteries in 5s block so it is unconfortable to cut away cells, it is reused battery to be much clear.

 

as many nows the mayor cost in theese sistem is the battery so i got theese for like nothing and have to use in a smart way. i manage to use it in other application in 7s config (but cuttin cells apart) and use it daily whit good result staing in the voltage range (for cell) as mentioned before.

58 minutes ago, deg87 said:

You are talking whit the battery manufacure because it'me🙃.

Then you should not be asking for help, if you are the manufacturer... 😉

Have you considered that the inverter may not be designed for that kind of Voltage range? most of these designs started with 24 X 2V Lead acid cells, which probably don't god much above 57.6V peak Voltage...

Either way, the norm for Cobalt based Lithium batteries for stationary storage applications seems to be 13S setups and that is probably what you will have to aim for...

Since, in your own words, you got the cells for nothing, now is time to spend a little bit to reconfigure them into 13S with many parallel cells batteries and that will still be cheaper than buying batteries, like most of us do... (I bought 16 X 304Ah LiFePO4's and connected them to a 150A JK BMS, which is more than plenty for the 5kW inverter I have at the moment)

in fact my theory is that this inverter does not react 61v despite of what it avail me to set up from display. 

in symply word my question is this:

is this inverter able to raise it's declared voltage?

my bet is no observing the result of the test i conduct.

so return to the original point.

i have to modify this inverter bus<->battery ratio that seems to be like 8 now to something like 7:1 i ask help about this if some one have exerience about change battery level voltage and how he\she do it.

i teorify a way but maybe there is other and better way

 

11 hours ago, deg87 said:

in fact my theory is that this inverter does not react 61v despite of what it avail me to set up from display. 

It's not a Voltronic Axpert inverter. It's a PowMr, which started out as Axpert/Infini clones. This one can export to the grid, so it's more like an Infini than an Axpert.

It's quite possible that they copied specifications from an Infini or Axpert model, then did their own minor modifications, and were not able to live up to the specifications that they copied.

11 hours ago, deg87 said:

i have to modify this inverter bus<->battery ratio that seems to be like 8 now to something like 7:1 i ask help about this if some one have experience about change battery level voltage and how he\she do it.

The only way would be to change the turns ratio of the inverter, or use a 5 kW transformer from another model. This is quite difficult. Merely unsoldering the old transformer without butchering the PCB would be a huge challenge. You would also have to modify the firmware so that it was aware of the new turns ratio. The MOSFETs may need a higher voltage rating as well.

Sorry, I don't think that this is practical at all. Modifying the battery to be 13S would be much more practical.

 

 

3 hours ago, Coulomb said:

Sorry, I don't think that this is practical at all. Modifying the battery to be 13S would be much more practical.

I agree if it is a classical case installation. Let me say that i build my sistem whit the goal to save money.

So the first thing is to not invest too much money(that anyway i have'nt) because the recovery time will be long.

other point is about it is practical or non practical. The batteries are 40 single unit to modify one by one, losing like 10/15% cutting away cells is other problem, and last but not least the bms 15s that i already have.

 

so the batteries has to stay as they are, there are other reason i think to stay on highter voltage on battery side but maybe we discuss it later 🙂

3 hours ago, Coulomb said:

It's quite possible that they copied specifications from an Infini or Axpert model, then did their own minor modifications, and were not able to live up to the specifications that they copied.

Yes i think the same, they copied a firmware that allow to charge up to 61V but the hardware they made/clone simply is not able to go that hight.

i have to say that using a software to query the menu it allows me to put Voltage even highter like say 65V 🤪 but other strange thing is that anyway the max you can save is 61V (the max on the display too) highter value are rejected.

This make me think that in some way the firmware is able to work whit a 61V charge battery and so i take one battery that i charged in other way and put it on.

The result is that it work whitout any error or warning and the log show the bus bump to 480v.

i have observed that until the bus go under 465V (discharging battery to 58.4) the pv input is somehow interdict.

All point me to the fact that all work good on software side, exept for the internal bus protection that lock solar (and maybe grid too) until return in safe level.

3 hours ago, Coulomb said:

The only way would be to change the turns ratio of the inverter, or use a 5 kW transformer from another model. This is quite difficult. Merely unsoldering the old transformer without butchering the PCB would be a huge challenge. You would also have to modify the firmware so that it was aware of the new turns ratio. The MOSFETs may need a higher voltage rating as well.

I tink that the max bus voltage protection is there for a reason so of course the only way is to change the ratio between batteries and bus. 

Desoldering and other working around electronic for me is daily job so my limit is as i sayd budget and specific know about inverters.

my idea is to little change the ratio make it to stay under 460V, so i think that if the change is little maybe the software simply react regolating his output drive signal offsetting a bit. i think that because i can set the value i need and this poor thing tries to obbey, what i see on graph is a output fixed to max 58.4v that clear decrease gradually the current.

component on low voltage side have already work whit 61v so i think there is nothing that can't hold 61v, the only part i can think now is the capacitor that maybe can be 63V in this case they will be replaced whit 80 or 100v,

the low voltage section if i understand is the dumbest part of the thing, correct me if i'm wrong, but from what i see untill now it simply straight convert low dc to high dc in a fixed manner (bump to 480v bus can be a clue of this)

 

so the main question is:

it is the part to modify the two transformer usually sitting between the heat sinks? 

if are they the way is to unwind the low voltage side counting the turns dividind by 58.4, then multiply to say 63 and rewind same as original but whit the correct amount of turn to give finally what i need.

18 hours ago, deg87 said:

my idea is to little change the ratio make it to stay under 460V, so i think that if the change is little maybe the software simply react regolating his output drive signal offsetting a bit. i think that because i can set the value i need and this poor thing tries to obey, what i see on graph is a output fixed to max 58.4v that clear decrease gradually the current.

It looks like they may be using 1:8 transformers but firmware designed for 1:7.5 or some other value. You charged the battery to 61V but it may have sagged to closer to 60V under load. Thus 480/60 = 8.0 ratio. A 7.5 ratio would have a bus voltage of 458V, which should allow solar charging as long as your panels are below that voltage. The high PV voltage solar charger can only boost, not buck. So yes, if you can modify the transformers too about 1:7.5, it might work. If I'm right about the firmware, then it might even work better than now. There are several places in Voltronic firmware that assume a fixed turns ratio. 

18 hours ago, deg87 said:

the low voltage section if i understand is the dumbest part of the thing, correct me if i'm wrong, but from what i see untill now it simply straight convert low dc to high dc in a fixed manner

Well yes, transformers are very fixed in their voltage ratios. 

 

18 hours ago, deg87 said:

if are they the way is to unwind the low voltage side counting the turns dividind by 58.4, then multiply to say 63 and rewind same as original but whit the correct amount of turn to give finally what i need. 

Err, you want to reduce the turns ratio. I don't think that you want to mess with the primary, it is centre tapped (for models with two transformers) and uses either very thick wires or copper bands. So you would take off a few turns from the secondary. I have no idea how many turns it will have. You would have to do both transformers the same, I assume. The secondaries will be connected in series, like in a Max. 

It will be interesting to see how this goes. I think one person has already done a transformer transplant, likely a single transformer. I can't quickly find the post. This post shows a burned transformer being removed. You can see the very heavy bars connected to the primary, and the litz wire that I think is the secondary. 

@Coulomb you more or less confirm what i guess.

of course i mean to act on the two secondaries 🫠(because i imagine i will found 2 transformer).

the plan is before disassemble all the things to measure voltages in work condition (maybe welding thin wires in strategical place to avoid me to die bad😆).

once i measure all the relevant data carefully detach the transformers and then on bench give on primary like 12 or 24 AC to se what ratio roughtly i got on (i guess) the two twin secondary windings.

once i have transformer original ratio i can then unwind the secondary counting the turn and then calculate how turns in needed to multiply the output of a factor like 1.05 1.1 (next i do math seriusly).

if i see that rebuild in a good and safe way the transformers is not in mine abilities i will pass it to a specialist that lukly should be not so difficult to find and non to so expensive too.

 

once i will open the inverter i check all the component and see if there is something too near the limit or irreasonably cheap. many manufactures i notice that optimize price reducing for example the mosfet quality or specs. so i plan somehow to pimp a bit all this to ensure that after it will be at least not less safe thas originally it is.

Edited May 22, 20242 yr by deg87

after a long time i come back on this inverter that now is mine.
i managed to remove the two transformers and then unwind them to finally understand how they are made..

i end whit the conclusion to rise the internal turn that was 8 to 9 but reducing a bit the size (10%) and now one of the transformer is done.

unfortunatly i come to a doubt because it is the first time i try a thing like this... the doubt:

does the wire i use to wind the transformer bee insulated from itself? i mean all the turn must never meet the turn made before??

i ask because i do it this way, meaning insulating every single turn, but when i come to unwind the second transformer i notice that not all the turns are insulated from the others.

 

So now that i have already remount the first transformer my big question is:

does every single wind bee insulated? 

it does?

it does not affect?

it affect the transformer?

 

Sorry but it is clear i need a bit of help here

ty to any that try to help

On 2024/11/21 at 2:39 AM, deg87 said:

i end whit the conclusion to rise the internal turn that was 8 to 9 but reducing a bit the size (10%) and now one of the transformer is done.

Are  you saying that you had to reduce the cross sectional area of the wire, or the number of  turns?

Reducing the cross sectional area will affect the power rating.

Changing the turns ratio will change the voltage at the load end (either the bus end or the battery end, depending on whether the battery is discharging or charging), and this will confuse the firmware, which has a fixed ratio in mind. I don't know what the effects would be.

On 2024/11/21 at 2:39 AM, deg87 said:

does every single winding need to be insulated? 

Yes!

There is something like one to several volts induced for every turn. If a part of the winding shorts to another winding, then isolation is broken. If one turn shorts to another turn  on the same winding, then there is a shorted turn, and that will cause massive currents to flow, probably damaging the driving transistors and/or other components.

Normally winding wire is varnished; the varnish is amazingly resilient even when the wire is bent fairly sharply. But the low voltage winding of these transformers has to handle very high current, and not a lot of turns, so the "wire" is more like a bar or rod. But it still needs insulation.

Rewinding a transformer is pretty hard core.

i ended whit fault 09...

It can be seen as an half fail i guess. 

My problem is that when i have unwind the transformer i havent notice the fact that is not insulated on any turns but only one time every circa 3 turns.

so i tryed to undestand how it is made and i'm not sure 100%. i update some pictures to show

this is how i think was the core wind insulated (not sure)

 

and this is how the transformer is winded 

internal and external winding is made of copper sheet insulated at any turns and core winding is made of 4x 2,5mmq bare copper wire as seen in previous image.

my guess is that the core has a real turn count of 3, so i ended whit the only choise to ad a turn going to 4.

to have the space to make an additional turn i have to reduce the height of the wires but elevate the number of wires. i ended up whit 6 bare wire obtaining 6x 1,5mmq.

i rewind in the same way, i mean:

1 a complete layer not insulated,

2 then insulate layer

3 a complete layer not insulated,

4 then insulate layer

5 a complete layer not insulated,

6 then insulate layer

7 a complete layer not insulated,

8 then insulate layer

9 rewind the original external in the same way as originally was leaving of course the internal one untouched

was difficult but possible in the end.

all amazing no smoke or big fails but..

 

this 09 fault code and now i have to debug

i mean i would be nice to use this inverter

 

17 hours ago, deg87 said:

My problem is that when i have unwind the transformer i haven't notice the fact that is not insulated on any turns but only one time every circa 3 turns.

That's Litz wire; many strands of thin insulated (varnished) copper wire, to reduce the skin effect. Those bits of red tape are possibly just to hold the bundles together for ease of winding. Maybe they can't get that weird wire to wind by machine.

Check with your multimeter. Copper doesn't stay shiny like that for more than a few minutes before tarnishing, unless it's covered with varnish. This happens to be clear varnish, or it's so thin that it appears to be transparent.

You really should replace the winding (if necessary) with similar wire. Otherwise, your transformer will get very hot (again?), and efficiency will be poor.

I would expect that you would need the exact same number of turns either side of the centre tap, if that's indeed what it is.

 

i see... 

as you said the wire is coated so effective turns is 8.

you point me on correct way and i also understand that i messed up the things.

what i have done in the end is run this inverter on 4:1 ratio i guess because i have wind 4 turns instead of 8. 

the fact is that i messed up the battery side whit the bus side so instead make more turns i have to do less turns...

this is good because is enought to reuse the original internal winding subtracting a turn and so come to 7 turns vs 8 original turns.

 

the problem now is that i blew up the mosfet of bus side so now i have to fix those before rebuild again transformers and try again...

 

 

On 2024/12/02 at 7:22 AM, deg87 said:

this is good because is enough to reuse the original internal winding subtracting a turn and so come to 7 turns vs 8 original turns.

I repeat, I strongly recommend sticking to the original turns ratio. The firmware will not be able to properly control bus voltage or battery charging otherwise.