Chargery BMS News / Update (July-27-2020)

[Factory400]

Hi,
I just ordered a 300A-DCC. I will use this with a Chargery 16T and 16 LiFePo4 cells with 280Ah@1C.

I am a little bit confused about the maximum current it will carry. In the specs it says 300A@100V.
So I assume that my system with 48V (round the half of 100), will reduce the maximum current to 150A maximum.
My load will be at the maximum around 200A, so I have a dout if the 300A-DCC is to weak...
Can someone confirm my thoughts?

A MOSFET (or MOSFET array) that is properly rated at 300A - 100V will always pass 300A regardless of voltage. 1V, 5V, 48V - whatever up to 100V.

 

[mswoboda]

O.k. that answers my question, thx!

 

[grizzzman]

Steve, I'm wondering if you have any info (that you can release) on the next-gen Chargery BMS. specifically, The ADC there planning on using. I'm trying to see if I'm going to wait or move on with a more accurate BMS. Thanks for all your work BTW.

 

[Steve_S]

16 Bit ADC. Active Balancing, updated firmware, possible UI changes & communications.
2nd week of October Jason was finalizing the new BMS PCB Boards.
I have a "feeling" that new product should start appearing end of year, or very early 2021.

 

[tim0shel]

16 Bit ADC. Active Balancing, updated firmware, possible UI changes & communications.
2nd week of October Jason was finalizing the new BMS PCB Boards.
I have a "feeling" that new product should start appearing end of year, or very early 2021.

Hoping I can retro fit it in my setup! Will it use the 4pin RJ9? I have remotely mounted mine through different cable into a case so I hope display remains simular size and connects the same. I need to keep things simple if I ever build anything else.

 

[Steve_S]

Hoping I can retro fit it in my setup! Will it use the 4pin RJ9? I have remotely mounted mine through different cable into a case so I hope display remains simular size and connects the same. I need to keep things simple if I ever build anything else.

Unfortunately, I don't know those details, I would certainly hope the connectivity would be similar if not identical.
It would be nice if we had some "official Advanced Info". There may be suggestions / ideas that could bounce back "pre-production".
It would also be nice to get an advanced Alpha Version (again pre-production but final layout & setup) to test, & evaluate prior to full production, that would have been helpful with the DCC.... Well, I guess in Essence, I got 1st version of DCC noted 3 issues which Jason solved for the production run.

 

[grizzzman]

16 Bit ADC. Active Balancing, updated firmware, possible UI changes & communications.
2nd week of October Jason was finalizing the new BMS PCB Boards.
I have a "feeling" that new product should start appearing end of year, or very early 2021.

Thank you Steve. Looks like I will wait for the new product.

 

[Factory400]

16 Bit ADC. Active Balancing, updated firmware, possible UI changes & communications.
2nd week of October Jason was finalizing the new BMS PCB Boards.
I have a "feeling" that new product should start appearing end of year, or very early 2021.

From the perspective of a power electronics engineer........

16bit ADC says nothing about accuracy. It only describes the resolution of the output.
If an actual voltage is 5 volts - a 16 bit ADC could spit out 4.9167 or 5.12843, neither of which are accurate.

When I design an voltage measurement point, it takes extreme care to get really accurate results. In a 16bit system, the last 2-3 bits are useless noise. An accurate system relies on very good analog design with a very good internal voltage reference. For BMS purposes, I would be perfectly satisfied with a 12bit ADC and excellent front end analog design.

 

[grizzzman]

Well, I have a Mooshimeter Specs Here to decide the accurateness of the voltages. So....we will see.

 

[Bob B]

Won't be able to argue with that ....

 

[jasonhc73]

I just ordered 4 BMS16 and 4 DCC from Jason.


I have been resistant to spend this kind of money on a BMS, but I didn't realize the Chargery did active balance also. Since it does, I am happy to go.

1​	BMS16T-300 v4.0​	16S management system,​	4​	155​	620​
2​	DCC-300HB​	300A DC contactor​	4​	90​	360​
​	​	freight by UPS (1carton, 10kg)​	1​	150​	150​
​	​	SubTotal​	​	​	US$1,130.0 ​
​	​	paypal handel fee​	​	​	US$56.5 ​
​	​	Total​	​	​	US$1,186.5 ​

 

[Steve_S]

I just ordered 4 BMS16 and 4 DCC from Jason.


I have been resistant to spend this kind of money on a BMS, but I didn't realize the Chargery did active balance also. Since it does, I am happy to go.

1​	BMS16T-300 v4.0​	16S management system,​	4​	155​	620​
2​	DCC-300HB​	300A DC contactor​	4​	90​	360​
​	​	freight by UPS (1carton, 10kg)​	1​	150​	150​
​	​	SubTotal​	​	​	US$1,130.0 ​
​	​	paypal handel fee​	​	​	US$56.5 ​
​	​	Total​	​	​	US$1,186.5 ​

The Chargery does PASSIVE BALANCING not Active Balancing. Jason is developing an Active Balancing BMS.

Li-Ion BMS - White Paper - Dissipative vs. nondissipative balancing (a.k.a.: Passive vs. Active balancing)

 

[Podbuilder]

The Chargery does PASSIVE BALANCING not Active Balancing. Jason is developing an Active Balancing BMS.

Li-Ion BMS - White Paper - Dissipative vs. nondissipative balancing (a.k.a.: Passive vs. Active balancing)

Steve. You seem to have a good working knowledge of these units. The passive balancing happens as per the sepcs?

• Balance current of 1.2A per cell

 

[grizzzman]

Steve. You seem to have a good working knowledge of these units. The passive balancing happens as per the sepcs?

• Balance current of 1.2A per cell

I have tested this function., as it adjusts it goes over 1.2 amps. so.... yep.

 

[Factory400]

I finally got my first large battery system working on the test bench with a Chargery BMS - rather dissapointed with it overall.

Curious when the newer model may arrive and how it will perform. Not so sure I am will to try it though. Honestly, the majority of the low-cost BMS's that flood the market are rather marginal devices. It seems no one makes a mid-level product - either low-end and cheap or high-end and very expensive.

 

[Bob B]

I finally got my first large battery system working on the test bench with a Chargery BMS - rather dissapointed with it overall.

Curious when the newer model may arrive and how it will perform. Not so sure I am will to try it though. Honestly, the majority of the low-cost BMS's that flood the market are rather marginal devices. It seems no one makes a mid-level product - either low-end and cheap or high-end and very expensive.

Please provide some detail of what you don't like ... That way Chargery is getting the feedback they need for improvement.

 

[Factory400]

I only have a single day of observation, so my opinion is limited. The most obvious flaw is that the cell monitoring accuracy and noise rejection is not good. The numbers reported are jumping all over. Admittedly, batteries are a new disclipline for me but I do know that a battery pack sitting unloaded is not constantly fluctuating by tens of mV's as seen on Chargery. On my Keithley 6.5 digit DMM, it is easy to verify that the cells are stable down to nano volts as expected. I suspect this is a filtering issue which should be on the front end of the input and/or possibly mux noise from a switching array that feeds all inputs into a single A/D that needs more settling time. Either way, the cell monitoring on a battery can be filtered to a sub-hertz level so the noise disappears. I suspect a lot can be done in software to help this, but solving the problem in the analog domain is where I would personally begin.

On the current shunt, the long wire connection is totally unsuitable for accuracy. The shunt is generating useful information in the micro-volt range which will never make it to the A/D converter in the main box without being destroyed by noise. The far better option is to put the differential amplifier on the shunt and return a much larger voltage or a differential signal back to the main controller. This would drastically improve the current monitor accuracy instead of seeing it bounce around by 10's of amps. Yet another step up is to put the A/D converter on the shunt as well and return the data in a digital stream like SPI, I2C, CAN or whatever. It is the same components, just put in different places.

 

[Steve_S]

I can't speak to the filtering & digital aspects etc, out of my experience. Regarding the Shunt, absolutely agree... I use a Midnite Classic with a WizBangJr attached to a Deltec 500A/50mv, and it's very accurate. Then we look at Victron & other better shunt setups and the same thing as you suggest.

I'm quoting the separate notes to help Jason note each one.
Ping to @Chargery
A few good points & suggestions I think should be noted.

a battery pack sitting unloaded is not constantly fluctuating by tens of mV's as seen on Chargery. On my Keithley 6.5 digit DMM, it is easy to verify that the cells are stable down to nano volts as expected. I suspect this is a filtering issue which should be on the front end of the input and/or possibly mux noise from a switching array that feeds all inputs into a single A/D that needs more settling time. Either way, the cell monitoring on a battery can be filtered to a sub-hertz level so the noise disappears. I suspect a lot can be done in software to help this, but solving the problem in the analog domain is where I would personally begin.

On the current shunt, the long wire connection is totally unsuitable for accuracy. The shunt is generating useful information in the micro-volt range which will never make it to the A/D converter in the main box without being destroyed by noise. The far better option is to put the differential amplifier on the shunt and return a much larger voltage or a differential signal back to the main controller. This would drastically improve the current monitor accuracy instead of seeing it bounce around by 10's of amps. Yet another step up is to put the A/D converter on the shunt as well and return the data in a digital stream like SPI, I2C, CAN or whatever. It is the same components, just put in different places.

 

[Cal]

I only have a single day of observation, so my opinion is limited. The most obvious flaw is that the cell monitoring accuracy and noise rejection is not good. The numbers reported are jumping all over. Admittedly, batteries are a new disclipline for me but I do know that a battery pack sitting unloaded is not constantly fluctuating by tens of mV's as seen on Chargery.

Cell voltage accuracy is +/- 5 mV. Therefore you will see jumps of 10 mV. I believe he's using a 12-bit adc. Filtering will help. He might be upgrading to a higher resolution adc. When I'm pulling 160A from the inverter I will see a 120 Hz cell voltage ripple. Current display is much worse. I can see 20A ripple. Current absolutely needs a filter.

 

[Factory400]

A 12bit ADC can deliver 1mV accuracy if implemented well. A 16bit ADC can deliver +/-5mV accuracy if implemented poorly. Bits do not deliver accuracy, they only deliver resolution.

I am no battery expert, but my intuition tells me that 1mV accuracy (not resolution) is good for a battery management application and a 12bit ADC is suitable for that. With that said, the cost delta between 12bit and 16bit is almost nothing. The challenge is in the analog implementation and how the firmware processes the resulting data.

As for the current shunt, the common INA226 (or similar) would probably suffice for this kind of thing - it is low cost with rather good performance. It has the differential amplifier married to the the 16bit ADC for both voltage and current monitoring over SMBUS. The shunt LSB is 2.5uV which offers good performance at low-currents. That is important if you want reasonable SOC reporting on systems that have periods of low current activity. Lots of ways to skin the cat - just spit balling ideas. If this circuit is implemented on the shunt itself - you can actually take advantage of the precision and shoot the results back on SMBUS (I2C). It even has on-board over sampling and averaging to deliver a median result without seeing all the noise from inverters.