TI Fuel Gauge Circuit EVM - Review

Table of contents

RoadTest: TI Fuel Gauge Circuit EVM

Author: jurist88

Creation date:

Evaluation Type: Development Boards & Tools

Did you receive all parts the manufacturer stated would be included in the package?: True

What other parts do you consider comparable to this product?: EV2400

What were the biggest problems encountered?: EV2300 is just a tool for work with BQ EVM... No problems with EV2300. The biggest problem was to get all info about BQ EVM into my head.

Detailed Review:

Intro

 

There are many great tools suggested for RoadTest by element14, however, EV2300 kit IMHO is the one of the most unusual one.

The main reason is that you don’t need this kit alone. When you want to test another BMS / Fuel Gauge / Charger / Protector EVM from TI, you need EV2300 as SMBUS link between EVM and GUI at yours PC =). In my case RoadTest additional cost was 100$ - the price of BQ40Z50 EVM.

However, I’ve been brewing plan to test one such Battery Management System SoC for a long time...  Thus, noticing opportunity to get one of tools for testing, I just applied for EV2300 RoadTest and ordered BQ40Z50EVM.
When I just started testing, I though that I'll review in deep details both EV2300 and BQ40Z50EVM... However, during testnig I understood, that BQ40Z50 is so complicated system, that it requires its separate review.

Therefore in this review I'll try to cover as much as possible about EV2300, and I'll mention only most important/interesting stuff about BQ40Z50EVM.

So, Let's start.

 


Technical Task

Before we dive into tool itselfs, let me answer you a question "why do I need BQ40Z50" and what is the Device Under Test (DUT)?

 

The main reason is that I wanted to evaluate residual charge of powerful cell pack precisely enough. When I tried just measuring Voltage and Current of cell pack with MCU, I concluded that this is totally not simple stuff, especially when you want to do it in long-term: battery is ageing, operation temperature vary a lot, discharging happens at high currents up to 50Amps ( you can read more about this in TI Battery University)... Moreover, I still needed various cell protections, which are fast enough. Therefore I've chosen BQ40Z50 as very advanced solution for this purpose.

 

I believe, that protections in BQ will work as they should (and they worked very well, many times, before I found all protections that I need to configure ). The main test purpose was to evaluate charge estimation mechanism of BQ.

The cell pack I wanted to test is for big quad-coptersadnd has folowing params:

  • Qnom = 5600mAh (@0.5C charge)
  • Vnom=14.8V (3.7V/cell, 4 cells)
  • Vcharge = 16.80 +/-0.08V (4.20 +/-0.02V/cell)
  • Vcutoff = 12.0V (3.0V/cell)

Unboxing

When received the parcels, EV2300 looked as described. However, I’ve been surprised by BQ EVM: on picture at TI.com there is clearly visible USB-mini connector; on real board it doesn’t exist! I checked revision of the board, and it was the latest. Thus, looks like TI removed it for some reasons.

 

Both kits were nicely packed into ESD-safe bags. Quite useful, because no air in the room is quite dry, and for me it is enough just to move a bit on the chair, to get 0.5cm long sparks from fingers to grounded objects (at some moment I even thought that I killed BQ EVM, when it didn’t want to start in the middle of testing ).

The first thing I’ve done – downloaded User Manuals for both boards; Each manual is very good structured, however doesn’t provide 100% of info you need; Therefore – Google =)

 

Modifications

Considering that I wanted to test EVM with currents from 20A(continous) up to 50Amps (short pulses), following modifications were done (after all protections were set correctly):

  1. Mosfets were bypassed by wire. My purpose was to test BQ at currents up to 40Amps. Si7114DN MOSFETs on EVM board allow 18.3A for 10s, and in 11.7A steady state;
  2. AWG14 wires were soldered directly to EVM PACK+, PACK- of J3 and 1N, 4P of J1 terminals. Original screw terminals accept 18AWG wire maximum size, which recommended ampacity is 16Amps.
  3. According to BQ EVM manual, wire jumpeer between SYS_PRES and GND J3 terminal pins must be done.

 

Following considerations during test must be kept in mind:

  1. Rsense resistor’s value is 1mOhm 1W . Calculation shows that it handle currents up to 31.6Amps continously (assuming air temperature is low enough).
  2. There is thermistor near power FETs – I left default 10k resistor. It is adviced to solder PRF18BA103QB1RB inside; obviously, it's so rare (www.findchips.com gave no results), that even TI wasn't able to get it and solder onto board
  3. BQ manual tells you about possibility to update BQ SoC firmware (EVM comes with BQ40Z50-R1 firmware, now there is already R2 avaiable). I decided first to test EVM with R1, while everything works just fine.

Software Installation

After all hardware stuff is finished, software side must be done. 

  1. EV2300 folder (http://www.ti.com/tool/ev2300) suggests only 32-bit version of drivers; fortunately, there is such great place as TI E2E community, where you can search for answers on common problems. In one of the forums ( https://e2e.ti.com/support/power_management/battery_management/f/180/t/539162 ) I was able to find the 64-bit version of driver, which fortunately worked on my laptop.
  2. BQ Studio (http://www.ti.com/tool/bqstudio ). There were no issues with installation.
  3. I decided to download and install Chemistry Updater package (sluc564 + sluc261), because it was mentioned in many manuals and forums. However, when attempted to install it, got message:

 

 

On this point I wasted some hours, while trying to figure out how to install it. Later I understood that I actually don't need it and there are a lot of obsolette  and misleading information from TI and forums (At least IMHO sending log-file to TI pre-defined e-mail is the only way now, how to define chemistry of your cell-pack).

4. Later, when I already found out, which chemical ID my cell pack has, next problem was, that BQ Studio didn't have this ID in its database (although, I downloaded the latest version of BQ studio). Luckily, I remembered, that sluc564 files had some chemdat files. After copying them in BQ Studio  installation folder, I found ID that I needed. Epic Win!

 

First Launch

 

At the same day, when I received parcels, I connected EV2300 and launched Battery Management Studio (BQ Studio) GUI. After selecting your EVM’s name, GUI leads to folowing screen:

 

my first reaction: WTF!? what tools should I use if I don’t want to risk with, not to get dead chip on EVM. 

However, just a bit more patience is needed in this case:) . Day later, all electrical connections were done, and BQ EVM was powered from cells. Lauched GUI detects the EVM type itselfs and shows the workspace’s action screen with all cool stuff (which is one of the main reasons for me, why first to buy Evaluation Tools, but not to create your own embedded PCBA)

User interface

My first impression: the GUI is well-structured, and intuitively understandable; Working area is divided into zones, which can be adjusted; each zone contains tabs, which can be opened by pressing on “Registers”, “Data Memory” and other icons.

One of the GUI’s epic fails (IMHO), is the Data Graph tab, where I was trying to find for a very looong time, how to start graph recording (of course, without looking into user’s manual =) ) Occasionally it was tiny green triangle button (on the pic it’s between “Remove” button and “5000” drop-down menu), which starts recording.

While opening some tabs, there were some problems with text displaying, which didn't fit into windows correctly:

I tried a few things like changing display settings (like resolution / Text size etc.). It didn't help. This part wasn't so important for me, therefore I left it as is.

Going into details...

I decided to consequently search info, which could provide me step-by-step instructions, how to set BQ from scratch to fully operating system. My main objectives were:

  1. Change series cell quantity from 3 to 4 in config
  2. Calibrate the EVM board
  3. Select the Chemistry that fits my cell pack the best. (intially I selected 0x0150 ID)

 

After long search at www.ti.com, I found nothing useful there. Then I faced “Introduction to Multi-Cell Gas Gauges” ( https://e2e.ti.com/support/power_management/battery_management/f/180/t/128277 ), which is the best I could find, and which helped to get better understanding where to start. Later, I concluded that the only and best way to fully understand how all settings in BQ Studio must be done, is reading BQ40z50-R1 technical reference (bq40z50-R1_Tech_Ref_sluubc1c) .

 

As result, I found that I need to change in “DA Configuration” CC0 bit and set it to 1 – as result, 4 cell voltages became displayed on screen.

 

Calibration was much more intuitive and user-friendly. Moreover, user’s guide provides useful information on it (like "you should bypass FETs for more precise result"). Therefore the biggest challenge was to get precise current and voltage references.

 

It is perfectly described in http://www.ti.com/tool/GPCCHEM , how to find chemistry, that match cell pack’s chemistry the best. The main idea is to log time, voltage, current and temperature of battery pack during charge, then relaxation, then slow discharge, and then relaxation again. After that you send log-file to TI, and they automatically reply you what chemical ID suits you the best .

In my report, there was: Best chemical ID : 1356 Best chemical ID max. deviation, % : 1.51

 

Later I programmed it into EVM, and tested with custom big, hot, powerful load, which allowed to change current from 0 to 50Amps with the help of potentiometer;


As result (considering, that cell packs real capacity was 100mAh higher than stated in datasheet, which I find out during chemical ID detection test), the error was around 100mAh, which for 5700mAh is about 1.7%, which is veeery satisfying.

 

Conclusions:

In total I spent about 3-5 working days to fully understand how to use dev kits, and how the BQ chip works. Some tests needed to be repeated many times, until succesful result; However, I think that spent very little time to understand, that I like BQ chip that I tested, and that it’s able to satisfy my battery evaluation requirements. Moreover, there is possibility to optimize system for room and low temperature performance, which will give even more precise result.

I should admit, the guys, who created BQ SoC, did huge and great job. Because even configuring BQ chip is quite challenging – it has hundreds of registers and data fields. And all of them ust be configured properly to allow system operate without fails.

 

There are still some small issues with provided documentation/software, which can be solved relatively fast. Yes, they don't give you 100% satisfaction of ideal product. However, in engineering you can-not be perfectionist-procrastinator which says: "It's better to be done perfectly and never, than anyhow but today" .



Finally, I'd like to say Thank You for everyone:

Thank You for TI, that they developed such a nice battery management system and evaluation tools;

Thank You for Element14 community, that they gave me opportunity to test device and write this review;

And, of course, Thank You, Reader, which was able to finish reading this text! I hope that it was useful for You.
 

Anonymous