FlexiBMS - 0.2 HW under work - Flexible configuration and charging BMS

For what? (I didn’t got the point) I am using 15 MCU cause I want each block to independent from whole system, in case it goes bad or etc I could disconnect one block and still have working pack :slight_smile:

It would be kinda modular but not hotplug style system :slight_smile: Plus its easier to balance this way 15P pack :smiley:

1 Like

I figured you, we’re using that many to monitor individual cells and didn’t have enough pins. Not having separate 1p packs

Yep I making separate 12S packs :slight_smile:

This will be going on top of 2 floor 6 cell length group

Sorry @SimosMCmuffin for offtopic :stuck_out_tongue:

1 Like

Good for sure. Not much cheap though. I remeber I paid something like 9.5€ each for the group buy. 15 of them will cost more than the battery itself :wink: (probably not)

I have them already from private works :wink: Bought whole pallet for like 3 eur/ea

Man this is great news. Can you keep some for the next diebiems group buy?

Lets move discussion to my topic :slight_smile:

1 Like

Just for curiosity, what is your assembling process ? First LTC and MCU and power stages then you check by powering ? I’m not an Electronics Engineer, but learning to become one :smile: so if you have some advices, I take them all !

Part by part is for the WEAK :smiling_imp:.

1 Like

I start by soldering MCU with enough components to power it up so for e.g. VREG and MCU and etc if its recognized by programmer then I continue :smiley:

I personally like to assemble everything in “blocks” (at least in non-tested designs).

I started first with the voltage regulators and then confirmed that they were supplying the correct voltages, because if I hadn’t and I had started with the MCU, DAC, CAN or any other ICs that are powered from those regulators. They might have blow if the regulators were badly soldered or faulty. This approach is about making sure the very basic things are okay before putting anything else on the board.

Second thing is figuring out the best order to start placing components based on their size and location. You don’t want to put the tallest/highest things on first and trying to solder components on between high components. This is about just making the assembly progress easier and putting components on in a smart order. Although this approach is more relevant to tightly packed boards, as you might be space limited. With sparsely populated boards this isn’t as much of an issue.

Also, hot air reflow soldering might look intimidating and difficult, but it really isn’t. I personally thought this way too before I actually bought a 50 euro hot air station from ebay and started soldering on my first Quad-sided IC. I was more pissed off that I hadn’t tried it sooner, so easy. SMD components all the way from there onwards for me. I bought some soldering paste and flux from digi-key like 3 years ago and I’m still using them. Might not be the most fresh things, but still solder and work fine. A decent basic soldering iron is also recommended as you’ll have to clean up things every once in while, like solder bridges between component legs after hot air reflow with a bit too much solder paste.

Below are some pictures of my tools, which are very basic and affordable, but very capable once you learn how to solder and work them.

Generally, try to come up with projects that motivate you to make things and try to incorporate some small new things always or often so your experience and skill field expands. That’s how I’ve done it since trade school, just making stuff that interested me and trying new things.

4 Likes

Thank you so much for this amazing reply. I knew that I’ll arrive to the SMD game soon or later but you convinced me that it may be sooner than later :smile:

Oh right I forgot about Vreg side first test to make sure it works correctly

More components have been soldered and MCU has been hello worlded.

Not quite sure, if the 0402 resistors in the balancing circuits made the assembly any easier compared to the more cramped 0603 resistors in the 0.1 .

7 Likes

Arent balancing resistor bit too small? how much current are you going to drain?

The actual power bleed resistors are the 0805s on the bottom side next to the FETs, balancing current should be around 90 mA.

Resistor current: 4.2V / 51Ohm = ~82 mA LED current: (4.2V-1.7V) / 120 Ohm = ~20 mA

but the P-fets Rds(on) will add some more series resistance, so both currents will be a bit lower.

1 Like

Will it be enough to balance like >10Ah packs?

What about >50Ah packs? LoL still 2A charge though so no worries

If the cells are healthy, they won’t start drifting a lot in use, so not a lot of balancing might be necessary to keep the pack in check after the initial balancing after building the pack. If it takes a lot of balancing time at the end of the charging , then that would indicate more that some of the cells are either damaged or not-up-to-the-task compared to the rest of cells. This was the case with my first DIY 4S5P pack of LG MJ1s. One particular series cell would always be the lowest voltage and warmest out of all the series cells, but for some reason reach the balancing voltage first. I concluded that there was something going on with those particular series cells.

Even if there is need to balance during charging, as long as the pack is otherwise relatively balanced, by the time you get to balancing you’re already somewhere +95% SoC, so you’re not losing much range if you don’t let it balance and just head out for cruising.

The first working prototype will go into my 10S6P 35E pack build with nominal capacity of around 19,5 Ah, so I can give more detailed reports with it’s performance on that. Can’t see any real problems in this point in time.

What are you building :smiley:? xS+20P packs? Those are starting to be on another scale capacity wise and be more fitting for a bigger BMS as a whole when compared to compact Esk8 battery packs.

Possible to still balance over time and monitor cell voltages via CAN, USART, bluetooth or USB.

More components soldered, programming has begun, LTC6803-3 has been tested and has been confirmed to be in working order.

Pretty much only the boost converter components aren’t on yet.

9 Likes