Designing yet another BMS (yaBMS)

[ikanode]

Originally Posted by s/v Jedi

Software. So what do we need to do. Below I compiled a list of questions, please shoot at them

1. Temperature range.

2. Voltage range for soft signaling to stop chargers

3. Voltage range for hard protection

Is it possible to make these user settable? Best would be in some sort of user interface. 2nd best would be "defines" at the top of the code so that expert users can change and recompile.

The BMS isn't very useful if you pick parameters. It could only match your situation and not mine. (I tend to pick very conservation settings--but when I need more Ah, I change them temporarily to be more aggressive.)

[s/v Jedi]

Originally Posted by ikanode

Is it possible to make these user settable? Best would be in some sort of user interface. 2nd best would be "defines" at the top of the code so that expert users can change and recompile.

The BMS isn't very useful if you pick parameters. It could only match your situation and not mine. (I tend to pick very conservation settings--but when I need more Ah, I change them temporarily to be more aggressive.)

Yes I will do that. I just want to discuss ranges because I suspect there will be different ideas about them from members here as well as from equipment manufacturers (chargers, regulators etc.)

[tanglewood]

Originally Posted by s/v Jedi

It does. The advantage is that you know what you’re doing because it is just the inverter behind the contactor, not the whole boat.
But like tanglewood wrote: it can ruin the BMS contactor so would be good to prevent that. It probably means Ineed to measure voltage on the boat’s main DC bus as well

I don't know exactly what parameters companies like MG use to make, or not make the final connection, but they pre-charge with a current-limited source of something like 6 or 10A. But then I don't know if they wait for some minimum voltage differential before closing, or if it's just timed and then they close no matter what, or if they will never close if some max is never reached.


The contactor data sheets typically show survivable current vs voltage. Calculating the initial surge current at this point is beyond my rusty EE training, and probably requires knowledge of the inverter input circuit that is not readily observable. So assumptions will have to be made. One datasheet I recall talking about how pre-charging to 80% voltage significantly reduced arcing and extended contactor life.



I agree there is an argument that this circuitry belongs in the inverter, or whatever capacitive load is at play. But the problem is seriously exacerbated by the low internal resistance of LFP batteries vs LA, since that's probably the largest inrush limiter. And why is it the inverters job to protect an external switch? Shouldn't the switch protect itself? It can be argued many ways, and the practical solution is pre-charge to reduce the voltage differential before contact closure.


Now, all this said, I haven't done any pre-charging in my home brew BMS, and instead elected (at least for now) to deal with it manually. But for anything resembling an actual product, I think pre-charge is important. And the issue should at least be understood, even for home brewers.

[tanglewood]

On your temp sensors, have you found the sensors/probes to be electrically isolated from their contact point with the battery cell? I have encountered some where one side or the other of the thermistor is connected to the case, and if the case is brought in contact with the battery terminal, which is arguably the best sense point, you have a short circuit back at your sense circuit for all but one cell in the series string. Making matters worse, I found most data sheets were silent on what was electrically isolated and what was not. It's the same problem as voltage sensing, trying in some to way have electrically isolated sensors.

[s/v Jedi]

Originally Posted by tanglewood

On your temp sensors, have you found the sensors/probes to be electrically isolated from their contact point with the battery cell? I have encountered some where one side or the other of the thermistor is connected to the case, and if the case is brought in contact with the battery terminal, which is arguably the best sense point, you have a short circuit back at your sense circuit for all but one cell in the series string. Making matters worse, I found most data sheets were silent on what was electrically isolated and what was not. It's the same problem as voltage sensing, trying in some to way have electrically isolated sensors.

Yes there is galvanic isolation... it’s heat conductive only. The sensor housing is plastic, will be covered by epoxy on the top and by a silpad where it contacts the cell. A silpad is a heat conducting, silicone kind of double sided tape. It is designed to do this. It will be stuck anywhere on the cell casing, not a terminal.

For balancing leads I am going to larger wire diameter. I will provide connectors that accept active balancers, cell monitors and even balancing cell chargers. Will be interesting to see what the BMS thinks of them as they interfere with cell voltages
The reason for this wiring is that I only want one set of leads to the battery for a heat installation. I can also provide the transient surge protection on that wiring so that everything that connects to it is protected.

[Steve_C]

Originally Posted by s/v Jedi

For balancing leads I am going to larger wire diameter. I will provide connectors that accept active balancers, cell monitors and even balancing cell chargers. Will be interesting to see what the BMS thinks of them as they interfere with cell voltages

You may need to enter a pre-determined balance mode for a short period and then recheck cell voltage after a short rest period. I don't see how you can actively balance and check for balance at the same time?

[s/v Jedi]

Originally Posted by Steve_C

You may need to enter a pre-determined balance mode for a short period and then recheck cell voltage after a short rest period. I don't see how you can actively balance and check for balance at the same time?

Worst that can happen is that the battery is disconnected from the main busbars. For normal, manual balancing, this is what is done anyway.

These active balancers can only balance with 1A or so. I think it will not be enough to matter for the measurement and if it does, it should make it better because it takes power from the highest cell and feeds it to the lowest. I bought one, it only cost $10 or so.
A bigger balance charger may be a different situation, but I don’t see those in use for house banks. Same with 48V, I don’t think many boats have that for house bank... for propulsion you need a different BMS, this one is for house banks that cruisers currently run even without BMS or they go overkill to get the features they need (like external contactor).

[ikanode]

Originally Posted by Steve_C

You may need to enter a pre-determined balance mode for a short period and then recheck cell voltage after a short rest period. I don't see how you can actively balance and check for balance at the same time?

One BMS that I have (by ElectroDacus) has these default settings:
Balance on 6s
Balance off 1s.

I think charging continues and that it's just the balancing power that switches on/off. It's only the cell Δv that matters for balancing and not the absolute voltages.

[s/v Jedi]

Originally Posted by ikanode

One BMS that I have (by ElectroDacus) has these default settings:
Balance on 6s
Balance off 1s.

I think charging continues and that it's just the balancing power that switches on/off. It's only the cell Δv that matters for balancing and not the absolute voltages.

I believe that is passive balancing with bypass resistors? yaBMS will not do any balancing; it’s for house batteries.

p.s. you guys saw the 280Ah cells for less than $100 each? Alibaba.

[s/v Jedi]

More progress today:

I have a Pro Mini operational incl. a USB FTDI interface to program it from the Arduino IDE. I’m gonna modify the Pro Mini, by removing it’s voltage regulator and power LED. This all to minimize power consumption, which should go down to uA level.

Also, I have the DS3231NS real time clock chip working, incl. alarms that generate external interrupts to wake up the uC that is in power down sleep mode.
I found that the breakout board also has an EEPROM and temperature sensor aboard. Nice, now we can check BMS temperature as well and have a safe spot to store log entries (the module has a 2032 lithium button cell for backup and the uC only has access to the EEPROM via I2C communication).
The board does have a power led... may be it can be switched off.

I think I’m ready to start coding and building a prototype PCB

[Steve_C]

Originally Posted by s/v Jedi

p.s. you guys saw the 280Ah cells for less than $100 each? Alibaba.

I did look at these on Alibaba after I saw your post. There are at least 3 or 4 different listings for these all in the "Under $100" price range. This certainly bodes well for a significant price drop across the board from all the suppliers. I am in a position to hold off for a while, so I will certainly wait to see how it develops!

By the time I am ready to order, you may have the BMS perfected!

[s/v Jedi]

Originally Posted by Steve_C

I did look at these on Alibaba after I saw your post. There are at least 3 or 4 different listings for these all in the "Under $100" price range. This certainly bodes well for a significant price drop across the board from all the suppliers. I am in a position to hold off for a while, so I will certainly wait to see how it develops!

By the time I am ready to order, you may have the BMS perfected!

Talk is that these prices are related to the Coronavirus... but I think that is a marketing trick. I guess we’ll find out. The ones I linked to, the people who ordered them got the genuine deal, new and up to spec. They test and match the cells, make a video of them doing that with a card with your name visible on top of one of the cells. After some (minor) shipping damage, they improved packaging dramatically. “Good customer service” is the verdict. This is all on that forum run by Will Prowse.

Some of the listings are used ones and one buyer could confirm that.

[ikanode]

re: Alibaba 280Ah cells

I have no expertise re: these cells but I suggest reading the second blog post at:
<https://batteryhookup.com/blogs/news> titled:
"Read before you buy off-brand lithium batteries."

According to the post, initial capacity tests can be misleading...

[s/v Jedi]

Originally Posted by ikanode

re: Alibaba 280Ah cells

I have no expertise re: these cells but I suggest reading the second blog post at:
<https://batteryhookup.com/blogs/news> titled:
"Read before you buy off-brand lithium batteries."

According to the post, initial capacity tests can be misleading...

These are from Ganfeng Battery, one of the largest lithium cell producers in the world. Check out their video: Jiangxi Ganfeng battery technology co., LTD. -- official website -- | electric vehicle battery | new energy | lithium iron phosphate battery pack | power battery pack | battery management system | battery pack | wind-solar energy storage system

[s/v Jedi]

defining the ranges for temperatures and voltages. Please comment

// LiFePO4 charge temperature range is 4C to 57C

#define tempChargeErrorLow 4            // disconnect charge bus <= 4C
#define tempChargeWarningLow 7          // signal charge sources to stop charging <= 7C

#define tempChargeErrorHigh 57          // disconnect charge bus >= 57Cs
#define tempChargeWarningHigh 54        // signal charge sources to stop charging >= 54C

// LiFePO4 discharge temperature range is -20C to 71C

#define tempDischargeErrorLow -20       // disconnect discharge bus <= -20C
#define tempDischargeWarningLow -17     // signal discharge loads to stop discharging <= 17C

#define tempDischargeErrorHigh 71       // disconnect discharge bus >= 71C
#define tempDischargeWarningHigh 68     // signal discharge loads to stop discharging >= 68C

// LiFePO4 max charge voltage 3.65

#define voltErrorHigh 3.65              // disconnect charge bus >= 3.65V
#define voltWarningHigh 3.50            // signal charge sources to stop charging >= 3.50V

// LiFePO4 min discharge voltage 2.5

#define voltErrorLow 3.00               // disconnect discharge bus <= 3.00V
#define voltWarningLow 3.15             // signal discharge loads to stop discharging <= 3.15V