Merged LiFeYPO4 1000Ah Winston prismatic cells and all electric galley...

[newhaul]

Originally Posted by Catapault

How long did it take you to Top Balance the new 400AH winstons ?

I've been running for 2 days now, at 3.55v on the PSU. Current started at 4.250 A, down to 4.080 A now. Cell voltage is only up to 3.320v so far, and climbing at glacial pace. (0.001mv. every 4 hours or so). (Note: Cell voltages being measure with a calibrated Fluke Multimeter at the terminals.)

I appreciate its 1600AH of bank atm, (4 cells wired in paralell), but expected a bit faster movement of the Cell voltage over the time period, so wanted to check others experience with same cells.

So let's get this right for now you have all 4 cells in parallel making 1600ah at top of 3.65v
Currently at 3.32v well you need to charge with a lot more than 4amps.
My recommendation would be to rewire in series making it a 400ah 12v nom bank .
Charge to approx 14 volts. Then break it down and rewire in parallel and start the top balance.
400 ah needs at least 40 amps to do a reasonable time charge. That's a .10C charge rate.

The way you currently have it I doubt it would be charged by Christmas.

Actually with your current setup it should take about 3 weeks to charge.
Rewire as a 12v bank and charge with as big a charger you have available. To 14 volts
3xample a 40 amp charger will take about 6 to 8 hours from where the cells are now. Then top balance a5 3.55 should only take about a day

[s/v Jedi]

Originally Posted by Catapault

How long did it take you to Top Balance the new 400AH winstons ?

I've been running for 2 days now, at 3.55v on the PSU. Current started at 4.250 A, down to 4.080 A now. Cell voltage is only up to 3.320v so far, and climbing at glacial pace. (0.001mv. every 4 hours or so). (Note: Cell voltages being measure with a calibrated Fluke Multimeter at the terminals.)

I appreciate its 1600AH of bank atm, (4 cells wired in paralell), but expected a bit faster movement of the Cell voltage over the time period, so wanted to check others experience with same cells.

The problem you have is that your power supply is measuring voltage at it’s own terminals instead of the battery terminals. Some good PSU’s have separate voltage sense wires so that they can supply maximum current for longer.

You can do this manually… but need to keep an eye on it of course. To do this, connect your Fluke to the cell terminals to measure voltage. Now increasing PSU output voltage until either it maxes out on current, or the Fluke shows the voltage is up to 3.55V.

From experience I recommend you do the step balance system from Rod Collins (MarineHowTo). For this you first balance at 3.4V which is 95% of the work. This also allows you to keep a safe distance to max. cell voltage while adjusting PSU output higher while measuring at cell terminals.

After they balance at 3.4V you then balance at 3.5V which only takes a short while compared to the first step.

With Winston cells from Julia Yu, they are already balanced so alternatively you can just connect them in series, hookup a BMS or cell monitor and charge as a complete 12V (or 24V, 48V) battery

[Catapault]

Originally Posted by s/v Jedi

The problem you have is that your power supply is measuring voltage at it’s own terminals instead of the battery terminals. Some good PSU’s have separate voltage sense wires so that they can supply maximum current for longer.

You can do this manually… but need to keep an eye on it of course. To do this, connect your Fluke to the cell terminals to measure voltage. Now increasing PSU output voltage until either it maxes out on current, or the Fluke shows the voltage is up to 3.55V.

From experience I recommend you do the step balance system from Rod Collins (MarineHowTo). For this you first balance at 3.4V which is 95% of the work. This also allows you to keep a safe distance to max. cell voltage while adjusting PSU output higher while measuring at cell terminals.

After they balance at 3.4V you then balance at 3.5V which only takes a short while compared to the first step.

With Winston cells from Julia Yu, they are already balanced so alternatively you can just connect them in series, hookup a BMS or cell monitor and charge as a complete 12V (or 24V, 48V) battery

Perhaps I was unclear in this note:
"(Note: Cell voltages being measure with a calibrated Fluke Multimeter at the terminals.)". - These are the battery Terminals, not the PSU terminals.

Agree the readings on the cheap PSU's themselves suck.

All voltages were set at Battery terminals, or at the end of the PSU cables (When setting the PSU voltage when disconnected from battery) using a Callibrated Fluke meter.

I did start doing the stepped balance at 3.4v. but that was even slower, with 1.4A going into the cells, and neither current nor measured cell voltage budged for 6 hours. (not budged as in, same reading on the fluke, down to the millivolt)

I agree the batteries looked pretty balanced on arrival. Consecutive serial numbers, 4 resting voltages at exactly the same reading, down to the mv.

The PSU I'm using is a cheap 10A one, still, thought it would do the job. maybe not.

However I'm still curious on my origional question, how long did yours take to balance on arrival, and which method did you use ? Do you happen to recall the resting cell voltage when they showed up ?

[wholybee]

Originally Posted by Catapault

Perhaps I was unclear in this note:
"(Note: Cell voltages being measure with a calibrated Fluke Multimeter at the terminals.)". - These are the battery Terminals, not the PSU terminals.

Agree the readings on the cheap PSU's themselves suck.

All voltages were set at Battery terminals, or at the end of the PSU cables (When setting the PSU voltage when disconnected from battery) using a Callibrated Fluke meter.

I did start doing the stepped balance at 3.4v. but that was even slower, with 1.4A going into the cells, and neither current nor measured cell voltage budged for 6 hours. (not budged as in, same reading on the fluke, down to the millivolt)

I agree the batteries looked pretty balanced on arrival. Consecutive serial numbers, 4 resting voltages at exactly the same reading, down to the mv.

The PSU I'm using is a cheap 10A one, still, thought it would do the job. maybe not.

However I'm still curious on my origional question, how long did yours take to balance on arrival, and which method did you use ? Do you happen to recall the resting cell voltage when they showed up ?

It could take weeks with a 10A supply the way you are doing it-depending on the state of charge. Note that how balanced it "looks" doesn't mean anything at all unless the voltage of the cells are over 3.4V or so. Below that voltage, the cells can be very far out of balance, and still be the exact same voltage.

The _fastest_ way to balance them is to connect them in series with the BMS in place. Set the BMS Battery HVC to around 14.2V. Set the Cell HVC to 3.5. Then charge normally with a regular charger until the BMS disconnects.

You can measure the voltages of the cells at this point to see how balanced they are. Maybe you don't need to do anything more.

That will get the cells 95+% charged. Then, take it all apart, place them in parallel, and proceed with balancing with the 10A supply. Don't use alligator clips or the thin wires the supply probably came with. Using heavier wires with ring terminals will balance much faster. And make sure you set the PS voltage with the battery disconnected. The voltage will drop when you connect it. That's ok, the voltage will rise as the cells charge.

As Jedi mentioned, you can increase the voltage of the PS to max out current, but ONLY do this if you have a separate meter connected to the paralleled cells and are watching it without taking any breaks. When the cells are full, the voltage will rise very sharply and quickly. A 5 minute break of not watching them and your cells could be at 4.0V per cell. Many people have ruined cells this way, so I recommend patience and not taking that shortcut.

I only balance to 3.5V. That is well within the steep part of the curve, which is all you need. You might get <1% more capacity by going higher, but also the balance circuit of the BMS should handle that tiny bit over time anyway.

[s/v Jedi]

Originally Posted by Catapault

Perhaps I was unclear in this note:
"(Note: Cell voltages being measure with a calibrated Fluke Multimeter at the terminals.)". - These are the battery Terminals, not the PSU terminals.

Agree the readings on the cheap PSU's themselves suck.

All voltages were set at Battery terminals, or at the end of the PSU cables (When setting the PSU voltage when disconnected from battery) using a Callibrated Fluke meter.

I did start doing the stepped balance at 3.4v. but that was even slower, with 1.4A going into the cells, and neither current nor measured cell voltage budged for 6 hours. (not budged as in, same reading on the fluke, down to the millivolt)

I agree the batteries looked pretty balanced on arrival. Consecutive serial numbers, 4 resting voltages at exactly the same reading, down to the mv.

The PSU I'm using is a cheap 10A one, still, thought it would do the job. maybe not.

However I'm still curious on my origional question, how long did yours take to balance on arrival, and which method did you use ? Do you happen to recall the resting cell voltage when they showed up ?

Please re-read my post. I know you have the Fluke at the cell terminals and this is good. But you need to adjust the power supply voltage up until the Fluke measures the voltage (3.5) even though the power supply is at a higher setting. Expensive power supplies have voltage sense leads to do this automatically, but yours doesn’t (mine don’t either but I have a 30A unit).

[CatNewBee]

Originally Posted by Catapault

Nice Job !!.



Where in the circuit are the new shunts loctated ? The photo seems to show one between two of the RBS's ?



Did this replace your 1000A Victoron Shunt, or doing a different Job ?

Between the RBS is only the 300A shunt, like all shunts it has 2 small screws to attach the probe wires, the circuits are in the box, 2 small wires are the probes to each shunt, you measure tiny voltage differences (mV), there is no current flowing.

If you look at the pcb, there a 3 wire pairs, 1 pair per shunt and one for the power supply of the circuit and ground reference. The positive voltage probes of the INA226 are attached to the high side of the respective shunt, so you can measure not only A and count Ah, but also Voltages, Watts and count Wh precisely.

(well I count mAs and mWs for the aggregation, then divide the result by 3600000 to get Ah an Wh)

[Catapault]

Originally Posted by CatNewBee

Between the RBS is only the 300A shunt, like all shunts it has 2 small screws to attach the probe wires, the circuits are in the box, 2 small wires are the probes to each shunt, you measure tiny voltage differences (mV), there is no current flowing.

If you look at the pcb, there a 3 wire pairs, 1 pair per shunt and one for the power supply of the circuit and ground reference. The positive voltage probes of the INA226 are attached to the high side of the respective shunt, so you can measure not only A and count Ah, but also Voltages, Watts and count Wh precisely.

(well I count mAs and mWs for the aggregation, then divide the result by 3600000 to get Ah an Wh)

Hi Cat,

Thanks for the info.

What I meant was, where in your overall circuit diagram, did you insert these new Shunts ?

- Shunts are normally on the negative size, between battery and everything, but I got the impression from your post that these new shunts were measuring specific busses only, rather than everything ? And if they are between the RBS's, then they are on the high side ?

Just clarifying if I'm misunderstanding what they are measuring is all.

[Catapault]

Originally Posted by s/v Jedi

Please re-read my post. I know you have the Fluke at the cell terminals and this is good. But you need to adjust the power supply voltage up until the Fluke measures the voltage (3.5) even though the power supply is at a higher setting. Expensive power supplies have voltage sense leads to do this automatically, but yours doesn’t (mine don’t either but I have a 30A unit).

Yup, got that part. All good.


Still curious how long yours took.

What I'm really asking is, info from others with SAME cells, to understand what their experience is with the SoC is of the cells when they are delivered. (If available). i.e. emperical evidence rather than a 'somewhere around here' guess I see online elsewhere. Or trying to calculate it from cell voltage (hard). NP if not avail, was just wondering.

[s/v Jedi]

Originally Posted by Catapault

Yup, got that part. All good.


Still curious how long yours took.

What I'm really asking is, info from others with SAME cells, to understand what their experience is with the SoC is of the cells when they are delivered. (If available). i.e. emperical evidence rather than a 'somewhere around here' guess I see online elsewhere. Or trying to calculate it from cell voltage (hard). NP if not avail, was just wondering.

400Ah Winston cells, right? Mine were at 30% SOC when delivered. They were sequential serial numbers and balanced within a couple mV.

I forgot how long the balancing took, but I used a 30A power supply so much faster than a 10A unit.

[PaulCrawhorn]

If the isolated resting voltage at the posts is 3.32V that can be considered a "healthy Full" for all intents and purposes, including capacity and internal resistance testing benchmarks.

Use verified-accurate measuring instruments at the posts, ignore anything else.

For normal cycling, even a little lower is fine. Any higher than that is just surface charge, not adding useful energy / capacity utilization, more stressful than conducive to long lifespan, and is actually tending to push imbalances.

Yes to get there, the charging voltage (as opposed to isolated at rest) needs to get up to 3.45-3.50V depending on C-rate available.

The only reason to get it higher would be the result of a top balance procedure, ideally with gear allowing adjustment of setpoints, maybe to isolated resting 3.34V

A delta of 30mV even 50mV is fine, with top balancing less is more - you really do not want cells to sit at high SoC for too long if you care about longevity.

Good quality cells, not pushed too high up the voltage/SoC shoulders in use, may only need a rebalance every few months or even years. Being OCD about it will actually exacerbate the issue you think you are preventing.

Personally I am not a fan of top balancing for that reason, the mid-region point, that good cells are delivered balanced at, is best, and a bottom balance second best, but that is not what most gear is designed for.

[newhaul]

Originally Posted by PaulCrawhorn

If the isolated resting voltage at the posts is 3.32V that can be considered a "healthy Full" for all intents and purposes, including capacity and internal resistance testing benchmarks.

Use verified-accurate measuring instruments at the posts, ignore anything else.

For normal cycling, even a little lower is fine. Any higher than that is just surface charge, not adding useful energy / capacity utilization, more stressful than conducive to long lifespan, and is actually tending to push imbalances.

Yes to get there, the charging voltage (as opposed to isolated at rest) needs to get up to 3.45-3.50V depending on C-rate available.

The only reason to get it higher would be the result of a top balance procedure, ideally with gear allowing adjustment of setpoints, maybe to isolated resting 3.34V

A delta of 30mV even 50mV is fine, with top balancing less is more - you really do not want cells to sit at high SoC for too long if you care about longevity.

Good quality cells, not pushed too high up the voltage/SoC shoulders in use, may only need a rebalance every few months or even years. Being OCD about it will actually exacerbate the issue you think you are preventing.

Personally I am not a fan of top balancing for that reason, the mid-region point, that good cells are delivered balanced at, is best, and a bottom balance second best, but that is not what most gear is designed for.

What is being discussed here is a specific controlled precommishing top balancing

[Catapault]

Originally Posted by s/v Jedi

400Ah Winston cells, right? Mine were at 30% SOC when delivered. They were sequential serial numbers and balanced within a couple mV.

I forgot how long the balancing took, but I used a 30A power supply so much faster than a 10A unit.

Perfect, thank you, and yes, 400Ah Winston's. That was my main question just for a comparison. Mine also sequential serial numbers, and all read the same voltage down to .001 V I was expecting them to come at a higher SoC, but that's fine, and yours sound like delivered similar.


Also I found my charging problem, and was none of the above.

The variable 10A PSU I'm using has issues. It intermittantly craters the voltage, and cycles, plus it just point blank refuses to go over 4.3A, regardless of voltage settings etc. This issue tested with non battery loads as well.

The technical term is 'U/S' ! (several other terms spring to mind). Seems it didn't survive being onboard and chose 'partial failure' instead of total.

Note: YES, it should be slow with a 10A supply,(as others mentiond) no issue there and I was aware and had the time. ( I understand the math etc). But this was 2-3x slower than it should of been, so was puzzling.

Time for a new benchtop supply, then I'll try again next week, As I need to do this again in 6 months, (when I get the additional 8 cells), will prob get something decent.

Moral of the story "pay more attention, HEF is bad".... !

PS: I miss my bench test equipment. Right now I'm missing my bench multimeter that logs, which would of shown this up right away on a graph ...

[s/v Jedi]

Originally Posted by Catapault

Perfect, thank you, and yes, 400Ah Winston's. That was my main question just for a comparison. Mine also sequential serial numbers, and all read the same voltage down to .001 V I was expecting them to come at a higher SoC, but that's fine, and yours sound like delivered similar.


Also I found my charging problem, and was none of the above.

The variable 10A PSU I'm using has issues. It intermittantly craters the voltage, and cycles, plus it just point blank refuses to go over 4.3A, regardless of voltage settings etc. This issue tested with non battery loads as well.

The technical term is 'U/S' ! (several other terms spring to mind). Seems it didn't survive being onboard and chose 'partial failure' instead of total.

Note: YES, it should be slow with a 10A supply,(as others mentiond) no issue there and I was aware and had the time. ( I understand the math etc). But this was 2-3x slower than it should of been, so was puzzling.

Time for a new benchtop supply, then I'll try again next week, As I need to do this again in 6 months, (when I get the additional 8 cells), will prob get something decent.

Moral of the story "pay more attention, HEF is bad".... !

PS: I miss my bench test equipment. Right now I'm missing my bench multimeter that logs, which would of shown this up right away on a graph ...

I have this one and can recommend it:
https://volteq.com/volteq-power-supp...rotection.html

[Catapault]

Originally Posted by s/v Jedi

I have this one and can recommend it:
https://volteq.com/volteq-power-supp...rotection.html

Thanks for the recomendation.

[CatNewBee]

Just use the Victron gear to initial charge to a safe voltage starting low, like 13.6V, 13.8V, 14.0V, 14.4V, 14.6V and then balance a little over it.. It can be done in 4-6 hours.