Chargery BMS16T Teardown and circuit analysis? Anyone?

[Cal]

Got to pick your battles. The noise floor isn't one of them. I've mentioned already a long time ago the design is sloppy.

 

[Factory400]

Got to pick your battles. The noise floor isn't one of them. I've mentioned already a long time ago the design is sloppy.

I am picking my battles - you are the one that has hyper focused on this particular detail of cell noise which is barely the tip of the iceberg.
When I see the tip of an iceberg, I intuitvely know that there is a massive chunk of floating ice that needs to be avoided. Not really worried about the tip of it.

Shall we move on? I am getting bored with this part of the conversation.

 

[toms]

I agree Chargery is fortunate to be getting free R&D done for them.

I have only worked with 3 BMS manufacturers, all of them modified their systems based on my feedback - i don’t think its at all uncommon.

The manufacturers are very lucky to receive feedback that makes their products better. I fully expect to get looked after (ie free upgrades) if i am providing the service of quality feedback.

If someone can dissect the Chargery and other BMS’ and use this information to create a better BMS, this is a great thing also.

 

[Maast]

I would think simple ferrite bead chokes on the voltage sense leads would clean up a lot of the noise. Maybe add some pf filter capacitors. Also use twisted pair power supply wires and a choke on those as well. Simple and easy enough that anybody could do it.

I'd like to see some thermal transfer grease between the mosfets and heatsink.

I think expecting a elegant design from a sub-$500 BMS is futile. I spend an awful lot of my time with $50,000 pieces of avionics taken apart on the repair bench yelling at it "what dipshit designed THIS?! I want to meet him just so I can kick him in the jimmy"

 

[Factory400]

I'd like to see some thermal transfer grease between the mosfets and heatsink.

I did that immediately with some good quality T-global S606C paste and PGS pads which is a massive improvement here. Sadly - the heat sink is completely enclosed inside the enclosure - so it is really only a heat spreader.

I think expecting a elegant design from a sub-$500 BMS is futile.

You are correct....my expectation were very low and this product met those expectations.

I would think simple ferrite bead chokes on the voltage sense leads would clean up a lot of the noise. Maybe add some pf filter capacitors. Also use twisted pair power supply wires and a choke on those as well. Simple and easy enough that anybody could do it.

Now that I have a few weeks where I am re-configuring this battery bank - I can do some of those things.
I still need to update the firmware, which should be first I guess. From what I have heard, it sounds like the newer firmware does a lot of averaging to stabilize the displayed data. That would have a similar impact to adding hardware noise reduction efforts.

Powering the unit from the EXT port is a good idea as others have pointed out. I have a circuit with a low voltage cutoff that I can use to power the unit direct from the battery and still have it turn of in the event of a low-voltage situation.

That is about as much as I would ever put into this system - it is just not up to the standard I would like in my system. I do not want or need space rated avionics - but this is essentially an advanced DIY project designed by people that are just not aware of the details necessary.

Like I said before - I really like the modular concept of this system but the practical implementation is dreadful and I am moving on to something else. Batrium perhaps? Design my own, maybe?

 

[joeblack5]

Factory 400 ....A little unfortunate that you did not go deeper besides very general remarks about to simple and to much noise and layout.

Having a 6Ghz scope and low noise lecroy does not mean all that much by it self.
With those and your implied credentials I would have expected a little more thorough analysis of for instance the quality and specification of MUX. Or does the second set 9 to 16 cells have a separate reference and causes more issues.
Or why the balancing seems to work within 5mV but the display readout does not and can vary with 60mv or more on cell 16.

At least that would have shown of your expertise.

Now after an easy way out it looks more like negative publicity attempt from the competition.


Johan