I find the world of connectors very confusing but I thought "BEC" was usually the abbreviation for Battery Elimination Circuit. Many electronic speed controllers will include a BEC so the receiver can be powered with the same battery used to power the motor. I'm sure many BECs used that type of connector but I'd be surprised (though not very) if the connector itself is called a BEC connector.
In radio control circles, BEC does stand for battery eliminator circuit, but on a broader level, it's the same as a certain type of JST connector. So BEC is used both ways. JST is a company tradename, and they sell many types of connectors, so it's handy to note the particular type of JST connector. In the case, the connector type is RCY.
Maybe I'll order some BEC connectors and measure how much drop I get at 4 Amps. I certainly prefer the size of it.
An online calculator says that the voltage drop for 6" of 22 gage wire at 3.6V at 4 amps is 0.065V. Sounds perfectly acceptable to me. It will be interesting to know how close the calculator is to your real world test.
I received the BEC / JST RCY connectors this morning which had 22 AWG pigtails installed. I made up three test samples of approximately the correct length, I shorted the wires together at one end and powered the other end with my old HP rack mount power supply. The measured voltage drops reported below is over the combinations of the power and ground wires and the connector pins.
The drop at 2 Amps, my anticipated average running current with the robot driving around, was 100 mV. I unplugged and replugged the connector a bunch of times and I didn't see much of change in the readings.
The desire to optimize things is deeply rooted in my core so I would have liked to see a little lower drop which directly translates to an increased run time but I think this numbers are perfectly adequate for our needs. I'll run a 5.5Amp test for a couple hours just to be sure but I think the BEC is our winner.
Thanks to everyone for their feedback. I'm not a RC hobbyist so the recommendations were very helpful.
Thank you for sharing this process as you develop the S3. From my hobbyist perspective, it's fascinating to see all the steps you are taking to bring a product to life!
Thank you for sharing this process as you develop the S3. From my hobbyist perspective, it's fascinating to see all the steps you are taking to bring a product to life!
I finished assembling the back boards yesterday so I thought I would share my progress and some photos. I was able to charge a battery with all 4 boards and the 5V boost switcher appeared to be working correctly. The FTDI serial convertor isn't being recognized by my PC so I'll have to dig into that, there's probably something happening with the USB mux chip. This was good initial results, at least no fires erupted!
-Ben
Boards with the SMT parts placed ready for the reflow toaster oven:
Close up of a compete board with the through hole parts installed:
I had one funny moment last right before calling it a day. Of course you should never power up a board right at the end of the day or your risk not sleeping that night wondering what went wrong but I never seem to be able to resist. After connecting everything, the LED's immediately showed a fault, the charge status LED's flash red/blue to indicate faults. I checked all of the voltages, the battery looked good as did the USB input. After a few minutes of head scratching I realized that I had forgotten to install the bottom side thermistor! The charge chip thought the battery was over heated. A few minutes later I had it charging and slept fine last night. Far from my worst mistake. -Ben
We're getting to a point where we can formalize our schedule. After Ben is done prototyping we're ordering a prototype from the factory. My guess is that it'll be six months until we've got S3s in our grubby little hands. We'll continue to keep this post alive with the latest information.
We're getting to a point where we can formalize our schedule. After Ben is done prototyping we're ordering a prototype from the factory. My guess is that it'll be six months until we've got S3s in our grubby little hands. We'll continue to keep this post alive with the latest information.
Ken Gracey
Great! Sounds like I can add an S3 to my Christmas list!
Another small update on the back board bring up. My problem with the USB Serial Convertor just turned out to be bad USB cable, well more specifically I was using a "charge only" cable which didn't have the data lines in it. With a good cable, it works just fine. I was also able to charge at the full 2 amps by sending out the correct I2C commands to the charger chip using the TI Dev board. They have a little Windows program that lets you setup the I2C registers that works pretty well. The chip defaults to a max charge current of 500 mA unless you configure it to be higher. There's a little PICmicro on the back board which will serve as the Soft Power controller to handle this initialization and also handle power off. I have also routed the I2C connection to the Propeller in case someone wants to override the default settings. With this approach, everything will work correctly regardless of the Propeller program. I've pretty much tested everything that I can on the back board so I'll move on to building the front boards tomorrow morning. The bring up went smoother than I had anticipated.
For anyone interested, I finished spec'ing out the Soft Power Controller operation this morning. I pasted it in below and also attached a state diagram. Any feedback is welcome as always.
Thanks
-Ben
The S3 robot includes a rechargeable LiPo battery, battery charger and power path circuitry. The power control user interface includes of a soft power switch and battery charge indicator LED's. A press of the power switch turns the robot on if it is currently off. A press and hold of the switch turns the robot off if it is currently on.
Major Hardware Components:
A TI BQ24295 serves as both a battery charger and intelligent power path controller for the USB power input. It additionally contains a 5V boost power supply which supplies power for most of the other logic circuitry used in the robot.
A PIC10F200T microcontroller acts as the intelligent Soft Power Controller configuring the battery charger IC via an I2C interface and monitoring the power switch.
A push button switch drives one of the soft power controllers IOs as well as the waking the battery charger IC when fully turn off.
The battery charger is primarily controlled via an I2C bus interface. The I2C bus is also connected to the user programmable Propeller processor and EEPROM. The controller holds the Propeller in reset when it communicates with the battery charger to avoid any I2C bus conflicts. This is NOT a multi-master I2C bus. At power on, the controller configures the battery charger with default settings and then takes the Propeller out of reset. The I2C and Reset IOs should be hi-z when not communicating with the battery charger IC.
Controller State Descriptions:
While the user interface is intentionally simplistic from the users prospective in only including On and Off states, several additional states are required to fully support the desired operation.
Deep Sleep State:
The Deep Sleep State is the lowest power mode during which the battery is fully disconnected from all circuitry except for the battery charger IC. The Deep Sleep State is purely conceptual from a programming stand point since the controller is powered off. Following a press of the power switch or connection of a powered USB cable, the S3 enters the Wake State.
Wake State:
The Wake State is the same as a power on event from the controllers stand point. Following charger IC configuration, the controller reads the fault status and branches to the On State or Battery Fault State accordingly.
On State:
The On State is entered from the Wake State if no faults are detected. After the input voltage limits and the charge voltage limits are programmed, the controller takes the Propeller out of reset and the polls the power switch for a press and holds of more than 1 sec in length
Battery Fault State:
If a battery fault is detected by polling the I2C bus in the Wake State or the Off USB Powered State, the soft power controller moves to the Battery Fault State. The battery charger IC flashes the battery status LEDs Red/Blue when a charging fault is detected. The controller should stay in the Battery Fault state for 60 seconds and then enter the Go to Sleep State.
USB Power Check State:
The USB Power Check State branches to either the Off USB Powered State or Go to Sleep State based on whether USB power is present when the power switch is pressed.
Off / USB Powered State:
The state is entered when the user turns the robot off by holding the power switch while the USB cable connected. From the user standpoint, this state is the equivalent to the Deep Sleep State with the exception that Charge Indicator LEDs may be lit if charging is active. The input voltage limit is lowered to 4.3V and which has the effect of increasing the charging current.
Go to Sleep State:
The Go to Sleep State is a transitory state while moving into the Deep Sleep State during which the controller powers it-self off.
As the hardware moves forward we've also got a productive iOS programmer on board working on the iPad S3 GUI. As you can see, some basic functions of the GUI are working and it produces code which can be downloaded from a PC.
- basic interactions are implemented (adding/moving tiles, zooming, scrolling)
- few first commands are implemented: start/end (new program, new procedure), led, observe (each command has a worksheet tile, editing panel, Spin conversion, SCB conversion)
- Spin compiler is integrated and works fine for already implemented commands
- save/load/send/clear/autosave options for worksheet are implemented
- for saving/loading/sending I'm using exactly the same format as the original S2 Gui - the S2 Gui is producing a little bit specific XML using Perl library, but I'm reproducing this XML in exactly the same way (including MD5) so we can exchange SCB's with the S2 gui.
- next up are the additional commands
Pretty neat, isn't it? Since the S3 has our WiFi programming circuitry, we will of course add that into the S3 iOS GUI as one of our upcoming milestones.
Great work, Ken! Your iPad programmer is doing great work. He looks to be Polish, I know some very clever Poles who are great EEs and programmers. Super nice people too, except for the team of pickpockets who nabbed my wallet in Warszawa.
Great work, Ken! Your iPad programmer is doing great work. He looks to be Polish, I know some very clever Poles who are great EEs and programmers. Super nice people too, except for the team of pickpockets who nabbed my wallet in Warszawa.
Yes, he is. And he's a truly awesome, competent iOS programmer who we were very lucky to sign up. At some point we'll get him to appear on the forums. But, like Chip, I want them to focus on the needs for project completion right now.
Interestingly, we have 12 external developers on our team right now. This is larger than ever before. They're located in Hungary, Poland, USA (AZ, NH, CA, MI) and the UK.
I finished assembling a pair of the new front boards last night, a few pics are attached. I'm just starting my testing but the initial results look promising. I was able to program the Propeller and the bus voltages look good. Board bring ups are always fun, I'll post more details as I progress with the testing.
Pop quiz for the day. Something on ONE of these boards went pop after being powered up for about 10 mins, the other board was perfectly fine. There are lots of hints in that sentence. The first person to spot it gets a gold star. -Ben
I finished assembling a pair of the new front boards last night, a few pics are attached. I'm just starting my testing but the initial results look promising. I was able to program the Propeller and the bus voltages look good. Board bring ups are always fun, I'll post more details as I progress with the testing.
Pop quiz for the day. Something on ONE of these boards went pop after being powered up for about 10 mins, the other board was perfectly fine. There are lots of hints in that sentence. The first person to spot it gets a gold star. -Ben
It appears you can install C14 and C46 any which way you like!! The top board has the one way, the bottom board a different way.
Was it hard to find those special bi-polarized electrolytics???
And we have a winner and in record time, nice job. Those big caps were the last two parts I had installed, I must have been excited to power the board up.
-Ben
Comments
In radio control circles, BEC does stand for battery eliminator circuit, but on a broader level, it's the same as a certain type of JST connector. So BEC is used both ways. JST is a company tradename, and they sell many types of connectors, so it's handy to note the particular type of JST connector. In the case, the connector type is RCY.
http://en.wikipedia.org/wiki/DC_connector#JST_RCY_connector
As Rich points out earlier, it's breadboard friendly, unlike other types of polarized battery connectors.
An online calculator says that the voltage drop for 6" of 22 gage wire at 3.6V at 4 amps is 0.065V. Sounds perfectly acceptable to me. It will be interesting to know how close the calculator is to your real world test.
Sample A Vdrop: 0.197V @ 4 Amps, 0.245V @ 5 Amps
Sample B Vdrop: 0.199V @ 4 Amps, 0.247V @ 5Amps
Sample C Vdrop: 0.204V @ 4 Amps, 0.256V @ 5Amps
The drop at 2 Amps, my anticipated average running current with the robot driving around, was 100 mV. I unplugged and replugged the connector a bunch of times and I didn't see much of change in the readings.
The desire to optimize things is deeply rooted in my core so I would have liked to see a little lower drop which directly translates to an increased run time but I think this numbers are perfectly adequate for our needs. I'll run a 5.5Amp test for a couple hours just to be sure but I think the BEC is our winner.
Thanks to everyone for their feedback. I'm not a RC hobbyist so the recommendations were very helpful.
-Ben
I finished spec'ing out the battery today and I printed a quick mock-up just to be sure it would fit ok which it did. Pic's attached.
- Ben
BTW someone will have to tell me the secret for making the pictures show up instead of as an attachement.
Thank you for sharing this process as you develop the S3. From my hobbyist perspective, it's fascinating to see all the steps you are taking to bring a product to life!
If you use "Reply With Quote" to this post, you can see what I entered to embed the photos.
I'm glad that you find it interesting. I feel really lucky to get to work on such a fun project!
-Ben
I finished assembling the back boards yesterday so I thought I would share my progress and some photos. I was able to charge a battery with all 4 boards and the 5V boost switcher appeared to be working correctly. The FTDI serial convertor isn't being recognized by my PC so I'll have to dig into that, there's probably something happening with the USB mux chip. This was good initial results, at least no fires erupted!
-Ben
Boards with the SMT parts placed ready for the reflow toaster oven:
Close up of a compete board with the through hole parts installed:
And no smoke.
S3 by Christmas?
Jim
We're getting to a point where we can formalize our schedule. After Ben is done prototyping we're ordering a prototype from the factory. My guess is that it'll be six months until we've got S3s in our grubby little hands. We'll continue to keep this post alive with the latest information.
Ken Gracey
Another small update on the back board bring up. My problem with the USB Serial Convertor just turned out to be bad USB cable, well more specifically I was using a "charge only" cable which didn't have the data lines in it. With a good cable, it works just fine. I was also able to charge at the full 2 amps by sending out the correct I2C commands to the charger chip using the TI Dev board. They have a little Windows program that lets you setup the I2C registers that works pretty well. The chip defaults to a max charge current of 500 mA unless you configure it to be higher. There's a little PICmicro on the back board which will serve as the Soft Power controller to handle this initialization and also handle power off. I have also routed the I2C connection to the Propeller in case someone wants to override the default settings. With this approach, everything will work correctly regardless of the Propeller program. I've pretty much tested everything that I can on the back board so I'll move on to building the front boards tomorrow morning. The bring up went smoother than I had anticipated.
-Ben
Scribbler Stack Attack!
Thanks
-Ben
The S3 robot includes a rechargeable LiPo battery, battery charger and power path circuitry. The power control user interface includes of a soft power switch and battery charge indicator LED's. A press of the power switch turns the robot on if it is currently off. A press and hold of the switch turns the robot off if it is currently on.
Major Hardware Components:
A TI BQ24295 serves as both a battery charger and intelligent power path controller for the USB power input. It additionally contains a 5V boost power supply which supplies power for most of the other logic circuitry used in the robot.
A PIC10F200T microcontroller acts as the intelligent Soft Power Controller configuring the battery charger IC via an I2C interface and monitoring the power switch.
A push button switch drives one of the soft power controllers IOs as well as the waking the battery charger IC when fully turn off.
The battery charger is primarily controlled via an I2C bus interface. The I2C bus is also connected to the user programmable Propeller processor and EEPROM. The controller holds the Propeller in reset when it communicates with the battery charger to avoid any I2C bus conflicts. This is NOT a multi-master I2C bus. At power on, the controller configures the battery charger with default settings and then takes the Propeller out of reset. The I2C and Reset IOs should be hi-z when not communicating with the battery charger IC.
Controller State Descriptions:
While the user interface is intentionally simplistic from the users prospective in only including On and Off states, several additional states are required to fully support the desired operation.
Deep Sleep State:
The Deep Sleep State is the lowest power mode during which the battery is fully disconnected from all circuitry except for the battery charger IC. The Deep Sleep State is purely conceptual from a programming stand point since the controller is powered off. Following a press of the power switch or connection of a powered USB cable, the S3 enters the Wake State.
Wake State:
The Wake State is the same as a power on event from the controllers stand point. Following charger IC configuration, the controller reads the fault status and branches to the On State or Battery Fault State accordingly.
On State:
The On State is entered from the Wake State if no faults are detected. After the input voltage limits and the charge voltage limits are programmed, the controller takes the Propeller out of reset and the polls the power switch for a press and holds of more than 1 sec in length
Battery Fault State:
If a battery fault is detected by polling the I2C bus in the Wake State or the Off USB Powered State, the soft power controller moves to the Battery Fault State. The battery charger IC flashes the battery status LEDs Red/Blue when a charging fault is detected. The controller should stay in the Battery Fault state for 60 seconds and then enter the Go to Sleep State.
USB Power Check State:
The USB Power Check State branches to either the Off USB Powered State or Go to Sleep State based on whether USB power is present when the power switch is pressed.
Off / USB Powered State:
The state is entered when the user turns the robot off by holding the power switch while the USB cable connected. From the user standpoint, this state is the equivalent to the Deep Sleep State with the exception that Charge Indicator LEDs may be lit if charging is active. The input voltage limit is lowered to 4.3V and which has the effect of increasing the charging current.
Go to Sleep State:
The Go to Sleep State is a transitory state while moving into the Deep Sleep State during which the controller powers it-self off.
https://www.youtube.com/watch?v=NQ3g4M01vGw
A brief summary of his work:
- basic interactions are implemented (adding/moving tiles, zooming, scrolling)
- few first commands are implemented: start/end (new program, new procedure), led, observe (each command has a worksheet tile, editing panel, Spin conversion, SCB conversion)
- Spin compiler is integrated and works fine for already implemented commands
- save/load/send/clear/autosave options for worksheet are implemented
- for saving/loading/sending I'm using exactly the same format as the original S2 Gui - the S2 Gui is producing a little bit specific XML using Perl library, but I'm reproducing this XML in exactly the same way (including MD5) so we can exchange SCB's with the S2 gui.
- next up are the additional commands
Pretty neat, isn't it? Since the S3 has our WiFi programming circuitry, we will of course add that into the S3 iOS GUI as one of our upcoming milestones.
Ken Gracey
Not sure why the video was categorized under "comedy" on youtube.
Jim
Yes, he is. And he's a truly awesome, competent iOS programmer who we were very lucky to sign up. At some point we'll get him to appear on the forums. But, like Chip, I want them to focus on the needs for project completion right now.
Interestingly, we have 12 external developers on our team right now. This is larger than ever before. They're located in Hungary, Poland, USA (AZ, NH, CA, MI) and the UK.
Ken Gracey
Ken Gracey
-Ben
When you're ready, contact Jeff for the latest iPad downloader tool so you can also test out the WiFi programming circuitry.
What I would do for a working S3 in my hands today. . .
Ken Gracey
It appears you can install C14 and C46 any which way you like!! The top board has the one way, the bottom board a different way.
Was it hard to find those special bi-polarized electrolytics???
-Ben