Smart BMS

[Teh Stork]

Hello Endless-sphere,

I've been working on a better BMS and I'd like to offer some members of ES the opportunity to test the first revision. This first batch is scheduled to be done in September 2014.

In general:

• Every feature you would find on a normal BMS and then some
  Key functionality (By plugging in a SD card or a USB stick)
  Datalogging capabillities

Functionality:

• Adjustable charge voltage balance treshold
  Adjustable protection voltage (both HVC and LVC)
  Adjustable temperature cutoff (both for charging at low temperatures and discharging at high temperatures)
  Broken Wire detection, if a sense wire comes loose - the BMS will detect it

User interface:

• A computer program will aid in the programming of all the adjustable parameters, the SD card will transfer the settings to the BMS.

Datalogging capabillities:

• Histogram output of common data (Temperatures, discharge current, cell voltages)
  Total AH in, Total AH out

Pricing:

• Is not set yet, but a price of around 5-10 USD per cell is a ballpark figure.

Avatible sizes:

• 6s
  12s
  18s
  24s
  30s

An 18s BMS can be used with 13s, 14s, 15s, 16s or 17s battery - no special modifications needed, only correct wiering. This applies for every odd number of cells with a different size of BMS.

Specifications:

• Balacing current: 25mA
  No uneven draw on battery cells
  Standby current 20uA (will be tested)
  JST-XH (2,5mm pitch) connectors
  Minitarized design

Updates will be posted here as the design is finished. Mainly I'm interested in the size people are interested in. For now, the 18s version is the one that will be focused on.

 

[dnmun]

why?

we can buy well designed and well built and tested BMSs for 20S at less than $35 now.

there is no need to have a different balancing voltage and HVC and LVC from the normal specs so why even worry about designing in a feature that is not needed? or temperature, like we won't know what the temperature is.

the more ways the the device has to fail, the more likely it is to fail.

 

[Teh Stork]

This BMS is not developed for the average ES-users that willingly hack chargers and the like to achieve the wanted result. I'm not going to disclose full techincal details before it is done, but I'm sure several of the features will be appreciated.

The quoted 20s 35USD BMS won't detect a sense wire coming loose, and the creepage distances on the pcb are quite bad. Conformal coated? It will also only balance at 4,2V.

The reason I'm reaching out to the ES community is to gain feedback on my product.

EDIT: I might have been unclear, these first "engineering samples" will be delivered freely for an honest review and improvements list in return. The quoted price are for the finished product.

 

[Tench]

I for one would love to put one on my 20s 20ah pack made up of 80 5ah HK lipo cells. The ability to lower the HVC and still have a balancing facility and top charge voltage set at say 4.15v would be great. As far as I am aware no one else offers this in an end user programmable BMS.
The ablity to monitor and record accurate total AH out from hvc to lvc would give good info for establishing cycle life attained/cell degradation and how it could be improved.

 

[dnmun]

but you don't know what causes cell degradation.

i think the people who make the BMSs that we already have available know more about them than people picking up an interest in this as a hobby.

there is no way i would ever trust a home built and designed BMS versus one that had been designed and tested by people who have the engineering background and expertise it takes to do these kinds of electronic devices.

 

[Teh Stork]

Well then I should add that I've got an B.sc in Electronics and currently I'm studying to a M.sc in Electronic Design.

 

[Tench]

but you don't know what causes cell degradation.

i think the people who make the BMSs that we already have available know more about them than people picking up an interest in this as a hobby.

there is no way i would ever trust a home built and designed BMS versus one that had been designed and tested by people who have the engineering background and expertise it takes to do these kinds of electronic devices.

As you have quoted from the Battery university site before I am sure you are aware that charging to a lower voltage decreases degradation and extends the number of cycles before the same % of capacity is lost, so lets not have the "You don't know what causes degradation" line when it suits you to act dumb.

From one of your posts;

it does not hurt the cell to discharge it fully but it shortens the life if you do it constantly. there is an article at battery university about how storage at full charge and high temperature also shortens cell life.

The battery university article;
http://batteryuniversity.com/learn/article/how_to_prolong_lithium_based_batteries

Charging to a lower voltage is good for cell life, so being able to program a BMS to balance charge and cut off at a lower HVC will be beneficial.
I for one prefer to encourage development rather than be derogatory towards it, that attitude would see us all rubbing sticks together to make fire.

 

[agniusm]

Thubs up for the effort. When you say sense wire detection is that it or it will have isolation? If its just an alarm whats the point in having it?
Best of luck

 

[jonescg]

There's never any harm in trying to build a better mousetrap. I think the main thing we need to achieve with a BMS is getting the size and quiescent currents down.

 

[Teh Stork]

Thubs up for the effort. When you say sense wire detection is that it or it will have isolation? If its just an alarm whats the point in having it?
Best of luck

No isolation, no. It will detect if a sense wire comes loose, normal BMSs won't do this and HVC and LVC protection will not kick in as a result of this. An alarm will be issued, but the battery can be operated as usual (also something that is subject to adjustment via programming).

There's never any harm in trying to build a better mousetrap. I think the main thing we need to achieve with a BMS is getting the size and quiescent currents down.

It is actually coming up as very small indeed. Each 6s module measures 4 x 2 cm. Place for discharge and charge MOSFETS are not designed yet, but obviously this is dependent on wanted amperage. Big 1210 SMD (up to 80mA) discharge resitors have been swapped out by 0805 (~25mA) resistors to save space. Since the BMS main MCU is powered from the whole string (not just the bottom two-three cells) there is no uneven draw, and no need for high balancing currents.

Quiescent current are very interesting. During my initial prototyping I found that the electrolytic capacitors actually leaked (3,5uA) more than the rest of the circuit in sleep (3uA). The ASICS used are rated at 10uA max in sleep.

 

[zhivko]

Is there any progress on this?

 

[Teh Stork]

Is there any progress on this?

Some, updates will come at the end of the month - I think. No promises. Currently busy with exams.

This first batch is scheduled to be done in September 2014.

 

[riba2233]

why?

we can buy well designed and well built and tested BMSs for 20S at less than $35 now.

there is no need to have a different balancing voltage and HVC and LVC from the normal specs so why even worry about designing in a feature that is not needed? or temperature, like we won't know what the temperature is.

the more ways the the device has to fail, the more likely it is to fail.

But there is, what about 4.35 V cells? And I want to balance and charge them to, lets say, 4.3 V?

 

[circuit]

we can buy well designed and well built and tested BMSs for 20S at less than $35 now.

Where can I find such BMS?

 

[circuit]

so no bms? Currently looking for 8s BMS with CAN bus.