Update
Have been using the 0.3 board to charge my pack on the start of the week and I’ve been testing it’s performance.
I started with just comparing the cell measurement results between the 0.2 and 0.3 boards. Both showed steady numbers for all the cell measurements, but the 0.2 board showed slightly lower voltages for all cells, namely about 2-3 mV lower compared to the 0.3 board. So I used my multimeter to manually check the cell voltages to compare against the boards measurements. The DMM showed identical numbers to the 0.3 board, but a doubt started to rise within in about one thing. How accurate are my DMM’s voltage measurements? I need to calibrate/verify it’s readings before I can say whether the results are good or not.
Brought the DMM to office to compare against better and more expensive equipment. So I hooked up my chinese Aneng AN8008, a Fluke 179 and a Keysight 34461A benchtop DMM to a low noise voltage source and tested couple different voltage measurements on different ranges to see how accurate the indicated measurements were. The Keysight is probably the most accurate DMM, so it’s an accurate enough reference to compare the AN8008 results against.
The measurements seemed to be within 1 least significant indicated number. All the DMMs had some noise on their indicated least significant numbers, but that’s normal as the voltage source was not a voltage reference, but just quite clean DC output. So with that test I could trust the steady DC voltage measurement were pretty spot on in the 3.000 - 4.200 V range, which was the range I was interested in due to wanting to verify the Li-Ion cell voltages.
So what is the result?
Well my DMM is showing identical numbers to one measured by the LTC6803 on the 0.3 board and I have verified that the DMM is accurate withing ±1 mV in the used voltage range, so all I can say is that I’m pretty darn happy about it’s performance! Also not bad for a 22€ DMM.
I suspect on the 0.2 board that the MCU is loading down the LTC6803’s VREG enough that it disturbs the measurement results, even though they are steady, they show consistently 2-3 mV less then the 0.3 board results. But that is really what I wanted to fix for the 0.3 iteration, by completely supplying the MCU through a separate regulation, removing it’s influence on the LTC6803’s VREG output.
Future moves and HW RC (Release Candidate)
I’m moving to a ramp-up stage on the development. I did very slight changes to the HW and with that I’m moving onto 0.4/RC1 HW (Release Candidate). I feel like the HW is now mature enough and the SW is usable enough (still under development) to start considering doing a development batch of 10 boards meant for distributing to HW testers and SW developers. I will be opening a sign-up form for testers and developers in the coming days. Link to it will be posted on this thread along with more info.
I have ordered components for a couple hundred euros for the upcoming developer build, which I I’m going to hand assemble, so I have the best/personal control over quality. I can also test the hand assembly time for larger batches, so I can compare the time/cost ratio for future production batches. The 0.4/RC1 gerbers are ready and I’m planning on ordering the boards next week. I want to test couple extra connector options for the XH and PH connectors on the board before ordering the 0.4 boards.
Other BMS news, FlexiBMS HV
I’ve been doing HW design with an American start-up company for a high voltage variant of the FlexiBMS for a bit over a month now. They contacted me through PMs on this forum, commenting about this Lite thread and expressing interested for the Lite and a HV version. I’m not on a payroll, I’m working rather for the possible future sales, we have been discussing and planning for me to give them a license allowing them to use the HW design (single-time payment) and then a small commission fee for every unit sold.
So what’s this HV version about then?
- LTC6804 battery stack monitor (newer gen. to the LTC6803)
- Master/Slave topology
- Cascadable design for chaining multiple packs together in series (they are going for 8 units -> 96S)
- True isolated connections between LTC6804’s
- Isolated external charging contactor control on Master module
- CAN-bus on Master module
What application is the HV meant for?
Higher voltage use, EVs, power storage. The real point is that you can connect multiple packs together in series for larger S-configurations and be able to measure all series cell voltages and balance them if necessary. They for now have been talking about using them in recycled Tesla model S packs for re-use and want a BMS for that.
It’s not meant for the small scale as the Lite is (3S-12S) and is not an All-In-One, but for bigger pack configurations (12S+) where you need to start designing your system as a distributed one.