New SBMS0 Firmware soon to be released.

[Dacian Todea (electrodacus)]

For the past few weeks (since the end of January).  I started working on the SBMS0 new firmware.

I added quite a few new features as you can see some of them below.

1) Monitoring

 Not much changes here other than cell bar-graph is now green in normal range and will change to yellow above or below UV and OV and red it it gets above OVLK or below UVLK.

I also added two more temperature sensors and you need to select betwen seeing the IT1 and ET3 (what was available before) and the IT2 and ET4. ET3 and ET4 are the same external temperature sensor just measured by two separate ADC's and the IT1 is the ISL94203 (BMS IC internal temperature) and IT2 (newly added) STM32 micro-controller internal temperature.

I added two new graphs the Battery voltage graph (data was there but was not displayed before) and the SOC graph (at least one person asked for this) and it took a bit of extra RAM that is fairly limited.

The Diversion is now added just need more testing and also all the wiring drawings corresponding to the state for dual PV array and diversion was added.

And on the same subject the Diversion graph will work displaying current and power for diversion. Still needs to be cleaned up by some of the things I have now for diagnostics

There will also be DMPPT functionality for diversion that needs a bit of testing also.

2) EXT IOx Settings had a huge upgrade and will require quite a bit of testing.

There are two extra EXT IOx available EXT IO1 and IO2 and they are the I2C data and clock pins and can be used.

Main reason for this two extra IOx was so that I can do Digital max power point tracking for diversion so select between two restive heating elements to get the max power point.

There is a table showing all the settings at a glance but to set things you enter each EXT IOx separate menu.

Here is the separate menu for EXT IO4 as an example. It is set as type 1 so Charging but there are much more settings now. 

There is a Minimum and Max temperature limit for each EXT IOx and with T sensor you select one of the 4 temperature sensors that will be used to enforce the limits in this case is IT2 (the internal temperature of the microcontroller).

I set here as 20C so below the actual temperature so you can see the OTP (over temperature protection) as a red bliking ON/OFF at 1 second flag in the bottom left corner.

Other settings include Current limit , SOC (SOC was available before) and now you can set 1 to 30 days for when the SOC limit is overridden and allow full charge of the battery. Before it was just an option for charging to full every day and you can get the same by setting 1 day.

Also you can see the red flag in the EXT IOx table. But in the bottom left corner can be see form anywhere including monitoring page.

3) This is what I worked for the past few days and is now almost done.

It is a test for the internal resistance of the cells and a very useful test to see if you have good connection between cells.

You just need to turn OFF your loads enter this menu and push OK to start the test that takes about 20 seconds and makes 3 tests than then averages in one result.

This is one of my A123 System battery 20Ah pouch cells and it is about 14 years old.

This first test is at 500mA 0.025C and this is a bit to low to get accurate results as there is just around 1mV drop per cell at this low current so not enough resolution even with me oversampling the voltage by averaging 10 quick measurements is still a bit to low. 

Here is at 1A so 0.05C charge rate. Test is done by switching ON/OFF solar charging but you need DSSR20 or DSSR50 or something that can as fast react to ON/OFF.

Here is with 2A so 0.1C and this is about the minimum I will recommend for best results

Here is with 4A 0.2C and that is best as there are about 10mV per cell so much more to work with for calculating the internal resistance. The graph in the middle shows with magenta the portions in the graph where the 10 measurements for voltage are taken for charge enabled and the same for charge disabled to measure the delta in voltage for each cell. That graph in middle has data at 70ms interval so about 5 seconds for the entire width of the graph where the other two graphs are at standard 1 second per data point 60 seconds one minute total per graph. So in the middle is about a 12x zoom

You can see this is page 1 of 2 and on the second page I want to implement something similar for the cell balancing wire resistance as that will save a lot of time diagnosing bad cell sense wire connections. So likely another few days to do that.

Here is anther test at same 4A as above but results are slightly different each time the test is run. Still fairly close each time especially when you have 10mV or more delta to work with.

The % value is with red only when above 40% but this is not a great battery after 14 years. I remember it was around 1mOhm per cell when new and maybe 1.5mOhm max with the connections between cells when new but both cell degradation and likely connection oxidation  got to this. I'm very curious to see what value I get for my large 8S2P EVE pack but it will need to wait a week to 10 days to finalize most things before installing the new firmware V6 on that SBMS0.  I will need at least two weeks of testing on that before I can start shipping with this new firmware.

In the Diagnostic this is all it was before but even this was updated.  I added temperature compensation for current measurement to make it even more accurate for low currents  and I added the option to use the offset it makes at power up or a factory calibrated offset that I will set when testing.

And last thing in the diagnostic menu was a way to reset the LCD but evolved in me implementing a graphics benchmark.

It comes with a warning as there will be fast moving images with up to 1.1 milion pixels per second in some benchmarks (a total of 14 benchmarks)

Here is drawing small rectangles at 888kpx/second

And here is a ray-tracing test up to around 270kpx/s.

This tests do not affect the SBMS0 functionality he will take care of battery and everything as normal.

Unfortunately this is only tested and will only work on the SBMS0 v0.3e hardware and will not work with the older SBMS0 version or any other older SBMS models. It will take way to much time to port all this to the older models and each of them will need to be tested separately. This will likely take a full year or work and I just do not have the time to do that.

As for all of you that have the new SBMS0 and want to upgrade to firmware V6 from V5 I will need to make a tutorial and maybe offer some programmers with cables so you can easily do that.

Let me know what you think and if there is something you wanted that I did not implement. If it is easy there is still time to implement. I do not have much Flash or RAM left is now with V6 almost fully utilized.

[Sasa Duric]

Thank you Dacian for all the changes you have done! Now, as you ask for the wishes, here it is... I know you already advised to have external thermostat to keep the batteries at 15°C, which I did. But having all those extra EXIOs and not being able to use it to trigger external heater to me is kinda ignorant. The logic that blocks charging under certain level can be used to trigger the heater. From all the improvements you did, I would appreciate the most if you add another seting type, or two in the EXIO menu for the heating and cooling if that does not cost much of the resources. I bet it cost less memory than those nice graphics for the display ray tracing test and it is certainly more useful as the batteries are much more expensive than a display or a whole SBMS0. Currently, I have another external esp32 running esphome for which I can freely adjust the battery heating, cooling and hysterisys. It works great, but I really think it should be a BMS feature.

[Bjorn Teani]

Looks like great work Dacian ! 

I think it will be time for me to buy a new SBMS0 so I can mesure diversion, test my batteries and have the old one as backup !

One wish : being able to manually enter historical data so you can transfert them from one SBMS0 to another and keep track of how much watts you gathered over the years.

[sailingharry]

Dacian,

Your continual improvement is awesome!

Three thoughts:

*  Your pretty graphic with the squares showing Battery, PV, Load, and Diversion -- since not all users implement Diversion, can that block be optionally disabled?  Adding unused information distracts.

*  Your IR testing is slick, and I'm looking forward to playing with that.  It would be useful (although I have pencil, paper, and a calculator so I can do it myself...), if you could total it -- a "Battery IR" result.  And, since you have a measurement point on the output of the shunt, it would be interesting to get a "System IR" that includes the entire system from the negative connection to the output of the shunt (I'd actually like the output of the fuse, but there aren't any measurement points there). I've already done my own calculation of "System IR" that even includes the fuse, since in my mind that is the number that impacts short circuit current (for fuse ratings).

*  It looks like you've set up high/low charge temps (Type 1) independent of high/low disconnect temps (Type 5).  YAAAY!  (Actually, it looks like you've allowed independent temps for each Type).  I've implemented this on my boat by setting the low temp disconnect at some crazy low number (-20F/-30C?), and then setting my charge sources (they have temp controls) to something like 40F/10C.  This past winter we had an especially cold snap (for us -- balmy for you!) of about 2 weeks below freezing, and my battery only fell to 75% with various incidental uses.

None of these comments are critical or negatively impact the use of the SBMS0 -- the system rocks.  But they would be nice if possible.

Keep up the good work!

[Dacian Todea (electrodacus)]

Sasa,

The Ray tracing uses no significant amount of recurses else I will not have included it. The LCD has his own memory (each pixel has a 16byte memory for the color RGB565) so when I write a pixel to screen it will remain there no need to have a buffer for that. The LCD has much more memory than my micro controller  320x240 * 2byte =153.6KB of video memory vs just 32KB of RAM on my microcontroller.

Most of the RAM on the microcontroller is used up by the graphs on page 4 of the monitoring those use over 80% of available RAM thus very little is left for everything else.

 

That said your request is not significant at all in terms of resources I just did not found it useful in the past when others have asked for the same thing. 

While you think 4 now 6 EXT IOx are plenty I find they are fairly limited to waste one of them for battery heating element control.

That said if you want to do that  (with the new firmware)  you can use an EXT IOx set as type 2 (load) to have a heating element ON until your set temperature limit say +15C is reached then it will turn OFF the heater until temperature drops 3C below that limit so +12C then heater will be re enabled.

But I will still suggest you use a $5 electronic thermostat that comes with his own thermometer probe that way you save an EXT IOx pin and the only external temperature sensor of the SBMS0.

Me I will be using the extra EXT IOx for DMPPT for water heating and the external temperature sensor to monitor the water temperature to not exceed +80C.  And yes I can do the same with an external thermostat for the water as I do now but using the SBMS0 external temperature to monitor the water temperature also gives me a more interesting data log as I can know the amount of energy delivered to water and the energy loss from water to environment.

 

Bjorn,

If you have the new SBMS0 model (the one with the 16pin green connector at the back instead of the older one with just 10pin white side connector) then you can write the new firmware when available and the historical data will be preserved.

sailingharry,

* That will normally just show zero for diversion if unused. Will you prefer that diversion bloc just be completely gone leaving a black spot in that place ?

*I was thinking of the total battery IR will be useful information or just extra distracting information :)  

My measurement points are limited from battery- terminal to battery+ terminal so nothing outside of that. The cell resistance includes the bussbar between the cells and connection point between the cell terminal and bussbar.  So the main use for this is to see if you have bad connections between cells.

*Yes each EXT IOx has his own independent low and high temperature limits. The one that was before controlled by ISL94203 is now disabled so that this independent ones are to be used.

So was your charging disabled for 2 weeks ? and you only used 25% of your battery in that time ? Seems like a very unusual use case if my understanding is correct.

The default values as you see now are just temporary I will need to think on what default's I should use before release of V6 firmware version. But for type5 I will likely leave the -40C and +80C limits. Of course anyobe can change that to what they like but they need to remember to use the Store device settings else they will reset to default after a power cycle. 

[sailingharry]

sailingharry,

* That will normally just show zero for diversion if unused. Will you prefer that diversion bloc just be completely gone leaving a black spot in that place ?

*I was thinking of the total battery IR will be useful information or just extra distracting information :)  

My measurement points are limited from battery- terminal to battery+ terminal so nothing outside of that. The cell resistance includes the bussbar between the cells and connection point between the cell terminal and bussbar.  So the main use for this is to see if you have bad connections between cells.

*Yes each EXT IOx has his own independent low and high temperature limits. The one that was before controlled by ISL94203 is now disabled so that this independent ones are to be used.

So was your charging disabled for 2 weeks ? and you only used 25% of your battery in that time ? Seems like a very unusual use case if my understanding is correct.

The default values as you see now are just temporary I will need to think on what default's I should use before release of V6 firmware version. But for type5 I will likely leave the -40C and +80C limits. Of course anyobe can change that to what they like but they need to remember to use the Store device settings else they will reset to default after a power cycle. 

Dacian,

*  Making the diversion block just a blank spot (assuming it's cleaned up with no connection lines, etc) would be nicer-- but this really is just icing on the cake, not a big deal.  I assume that the "no diversion" portion of your users is large enough to make this useful -- but I may be mistaken.  My GPS has a couple of permanently displayed "errors" for things that don't exist (SIRIUS antenna, for instance) that distract from "real" errors like "GPS signal lost."

*  Understand on the Battery IR.  Really looking forward to playing with that.

*  My use case is unusual, but not completely uncommon.  In winter (during layup), I have several very low loads (monitoring, lights when working on the boat, occasional inverter power for tools, etc).  Use under 20Ah/day (12V system).   This is probably typical of boats, RV's, summer cabins, etc during winter.  In the summer, when actually using the boat, my daily consumption is closer to 300Ah/day!

[Dacian Todea (electrodacus)]

Short update.

I implemented the cell balancing wire voltage drop.

Here is normally connected to the 20Ah 8s battery with a short ribbon cable maybe around half the length of what I include with the SBMS0 

Here is with a simulated fault by adding a 110mOhm resistor in line with wire #3 the one between the cell1+ and cell2-

Still acceptable as if you had a longer sense wire in that place.

Easy to see that issue is between cell 1 and 2 as both show a higher voltage drop than the other's

While it is not a pass it still OK to use.

 

Here is the 110mOhm now replace with 1Kohm and it will look and work normal except if cell balancing is enabled.

Same fairly simple to see where the issue is.

And here is an update to the cell internal resistance. I increased the number of measurement points to increase a bit the accuracy.

Still needs at least 10mV drop per cell to get below 10% measurement tolerance. 

[Peter]

That looks awesome! Really appreciate the work your putting into the firmware!

- Love the new function to independently set temperature for charge and discharge!!!! thumps up!

- Diagnosis for cells and sensing lines looks very cool, also the time limits for full charge.

- What is the purpose of having External temperature measured by different ADCs?

- What is the hardware setup requirements for using DMPPT? I have DualPV setup with the DSSR20s, would i need DSSR50? Is this the case where you input the resistance of the heaters? But how will SBMS0 know which voltage the panels currently deliver? Currently i monitor those individually with separate Esp32, but that cannot switch anything. 

cheers

[Dacian Todea (electrodacus)]

Peter,

- What is the purpose of having External temperature measured by different ADCs?

The ISL94203 has two temperature sensor inputs. On the SBMS60,SBMS40, SBMS100, SBMS120 the second one was used internally to measure the temperature on the board mostly to protect the mosfets.

Since that was not needed for SBMS0 instead on leaving that temperature sensor pin disconnected I connected together to the the other input so now both temperature senor inputs measure the same single external temperature on the SBMS0.  Maybe if there is ever a new hardware revision of the SBMS0 I will expose that second input on a separate connector to have two external temperature sensors.

- What is the hardware setup requirements for using DMPPT? I have DualPV setup with the DSSR20s, would i need DSSR50? Is this the case where you input the resistance of the heaters? But how will SBMS0 know which voltage the panels currently deliver?

No you do not need the DSSR50 you can use your existing setup. There is a new menu called diversion settings replacing the old DMPPT menu as some settings where moved in the new EXT IOx settings like max current in dual PV setting "MPV" as seen default 10A for the type 6 and I introduced a new one is the ratio between the two arrays default set here 2.00 meaning the large array is 2x larger than the small array.

 

So here is the menu where you input the resistance for heaters. If they are set to 0.0499 Ohm (minimum value) they are considered OFF or not set and above 0.0500 Ohm they are considered ON

If you had a single PV array or want to divert just one of the two arrays to heating You will set just the first one Div1R1 if you wanted max power point with single array diverted then Div1R1 for small heating resistor and Div1R2 for large heating resistor (resistors could also be equal).

Then for dual PV array diversion and with DMPPT you will have 4 resistor values that you will need to set the Div6R1 and Div6R2 will be for the large array controlled by the type 6.  

As for voltage measurements to calculate current and power and and to be able to do the DMPPT logic the ADC2 and ADC3 are used.

The ADC2 is for the small array and the ADC3 for the large array voltage.

Since SBMS0 knows the array voltage and it knows the heating element or elements resistance it will know to calculate the current and from there also power.

Then for max power point it can control in this example EXT IO1 and IO2 set as type 3 and type 4 that control R1 and R2 from one or both arrays.

For the EXT IO1 and IO2 I will need to make an adapter with two sets of optoisolators  that you will need. This will have two inputs IO1 and IO2 and 4 outputs so that you can simultaneously control R1 and R2 from both the small and the large array. 

So when battery is fully charged and both arrays are diverted you can turn ON/OFF both R1's with EXT IO1 set as type 3 measure the current and calculate power (using ADC2 and the resistance settings) then turn that OFF IO1 and turn ON IO2 the R2's will now be connected measure current and power with ADC3 and then connect all of them and measure power and decide witch of the 3 variants provide most power for those solar conditions then have that selected until there is a large enough change in solar.

It is all implemented just needs to be tested and modified in the next weeks.  Then I will also need to almost fully rewrite the user manual as there are quite a few changes with this new firmware version. And also I need to modify the computer app the one the comes on the Wyse mini computers.

[Dacian Todea (electrodacus)]

It took a bit more than I expected but from today the V6.0 firmware is installed on my battery for testing. There was already a small change I needed to make this morning to correct the dual PV array current limit setting. But other than that it works fine up to now. I also need to do more wiring to test the DMPPT for diversion as I was not setup for that up to now. That also needs quite a bit of testing since is completely new function.

It is a bit of a mess of wires and photos are a bit overexposed due to bad lighting but here are a few tests.

The cell wire resistance result is perfect (not sure if cable is standard length 1.8m (6ft) or maybe a bit shorter as I may have used an end of roll cable that was not long enough to sell).

In any case all wires are same length so result is withing 2.6% delta  77.9mV highest cell sense wire drop. 

Cell resistance is a different story and was a bit surprising but it can be explained

Cell's resistance is very good better than I expected should be about 0.260mOhm per cell as seen from cell 1 (two cells in parallel 0.130mOhm)

Cell 1 is the only one where cell sense wires are directly across the cell terminals and do not include the bussbar voltage drop.

Bellow you can see a photo of the battery and connections and my explanation for the results

The cell 5 that reads the highest resistance includes the resistance of those long connection (two) at the back about 18cm each made with 16mm^2 copper cable (marked as AWG6)

All the other connections between cells is made with shorter about 6cm cables.

According to spec one meter long section of this cable is 1.15mOhm so a 18cm long section should be 0.207mOhm but since there are two in parallel is half around 0.1035Ohm 

So if we remove that from the measurement we get 0.257 - 0.1035 = 0.1535mOhm for the cells and cell connection resistance.

All the other cells have about 3x shorter connections so maybe 0.03Ohm can be attributed to that  so since they are 0.18 to 0.19 they will be 0.15 to 0.16mOhm 

So cells are very close in internal resistance no issues but the connections in particular the long ones between cells 4 and 5 included in the measurement of cell 5 are very significant almost doubling the cell total resistance and thus the fail result.

So while the test says fail as long as you understand where the issue comes from and you are OK with that then all is fine. I can not make those connections shorter but I could use 25mm^2 or maybe even 50mm^2 copper cables to reduce that. To me is not an issue with only around 120A max charge and discharge current. 

At the moment I switch to v6.0b I have about 30khours and 7MWh used from DC loads (this excludes the diversion to heating that was not logged by SBMS0 with v5 firmware) and out of 7MWh  4MWh went trough battery.

Since battery is around 13kWh capacity  4000kWh/13kWh = 308 full cycles equivalent over the past 3 years or so.

I have not touched the cell connections since it was installed and torqued to spec. I only changed the circuit breaker a few months ago as that 120A was to close to limit and was getting to hot (was a bit smelly but did not failed or tripped with no reason). This is way larger than it should be but the v6.0 has electronic over-current protection  (that I will need to test).  It should be able to trip the circuit breaker if type 5 is ever set due to any reasons including UVLOK, OVLOK, overcurrent or over or under temperature.

The monitor page is not that different on how it looks but there are a huge amount of changes under. Almost the entire software was changed from v5 to v6

I did also a bit of reverse engineering and you do not even need to power cycle if you change things like overvoltage limit in parameter settings. As soon as you make the changes and save parameter settings the settings will take effect without the need to power cycle.

I also added in diagnostic the option to calibrate the cell voltages and temperature if you have precise enough equipment to do that but for most the factory calibrations done by Intersil  after production are more than adequate for battery management.

One of you asked to remove the diversion display from page 6 monitoring and that is also done when diversion is not used as well as skipping after page 5 that is also about diversion. Again sorry for the bad photo quality.

I will continue to test this for at least the next 10 days and if all is good I will start shipping the version v6 on all SBMS0.

I also need to design a optoisolator board for the EXT O1 and O2 . For my own testing I build a prototype the purple board on the right of the SBMS0 but the final version will be smaller and more compact.

It is almost like a new device. All functions where improved in speed by significant factor 2 to 5x so it feels much more responsive. 

[Dacian Todea (electrodacus)]

Sasha, you asked by email but I will answer here in case others want to know.

The cell resistance is measured when called from the menu as it needs to turn OFF all loads and turn ON/OFF charging in order to measure the voltage drop across each cell when charging is enabled vs charging disabled in order to calculate the resistance.

The 59 bytes data available over USART/USB  and WiFi contains more data now. There are two extra temperatures sent alternating on the same place the initial two where one second is the existing ones the other second is the other set. And there are a set of 3 bytes where data changes each second starting to repeat after 30 seconds meaning another another 180bytes of data that is not changing often or never changes. Like the total uptime of the SBMS0 the HW and SW version and total life energy counters (basically all the info in About page) then the EXT IOx types are also there and the parameter settings.  I will need to make a new manual for SBMS0 and this v6 firmware as there are to many changes to just modify the existing manual  and the existing one need to remain as it is for the v5 version.

Peter, thanks. Not sure why your message is not visible for me on the forum.

[Sasa Duric]

Thanks Dacian for the answer. I mistakenly sent a question by email.
Would it easy for you to implement MQTT receive Load/Charger ON/Off commands? Right now I intercept the control signals so I can remotely disable charging/discharging which adds to my messy installation. 

[Dacian Todea (electrodacus)]

Sasha,

There is no way for the ESP32 to communicate anything back to SBMS0 main microcontroller. The STM32 and ESP32 are connected over USART but only the TX pin is used so ESP32 receives data can not communicate anything back to the STM32. 

The only way to turn OFF a Load or a Charger is either trough SBMS0 menu or trough disconnecting the EXT IOx (likely what you are doing now).

It will not be safe to allow an external device either wired or wireless to control the functionality.  

[Dave McCampbell]

Hi Dacian.  We bought our SBMS0 about 5 years ago, and it has worked perfectly ever since.  While talking with friends with LFP systems on their boats, I have been convinced that there is no better BMS.  Especially for those that have internal Mosfet BMSs.  I have been watching your recent posts re the new v6 SBMS0 firmware you are working on.  Of particular interest is the included resistance monitoring capability and extra EXTIOs.  Our onboard computer programmer, my wife Sherry, is asking if any of that can be installed on our current SBMS0?  If not, and even though we have two of the SBMS0s onboard (on a cruising sailboat you can never have too many spares!), it might make sense to buy yet another new SBMS0.

[Dacian Todea (electrodacus)]

Dave,

I think you have the v0.3 hardware so the one with 16ping green connector on the back instead of the limited v0.2 hardware version that has a 10pin white side connector for shunt and EXT IOx connections.

If you have the v0.3 version the one that looks as the currently sold one then you can just upgrade the firmware to version 6.0 in a few weeks when it will become available. It still needs some testing for the next two weeks.

To easily use the extra two EXT IOx you may need a optoisolator board that I will offer in a few months. Need to design and build a few of those.