-
Notifications
You must be signed in to change notification settings - Fork 6
Gibbot Board v1
ajgriesemer edited this page Aug 9, 2013
·
2 revisions
The first iteration of the Gibbot Board was intended to fully drive every aspect of the Gibbot except for the battery management circuitry. This includes:
- dsPIC33EP512MC806 microprocessor
- [AIS3280Q 3 axis accelerometer] (http://www.st.com/web/en/resource/technical/document/datasheet/CD00289364.pdf)
- [A3G425OD 3-axis digital output gyroscope] (http://www.st.com/web/en/resource/technical/document/datasheet/DM00047823.pdf)
- Stepped down supply voltage from 48V to 24V using LT1976 Buck Converter
- Stepped down supply voltage from 24V to 5V using LT1976-5 Fixed 5V Buck Converter
- Stepped down supply voltage from 5V to 3.3V using TLV1117 Linear Rectifier
- MOSFET H-bridge circuit for driving the BLDC motor using NPN transistors to drive the MOSFET gates
- ACS716 Current Sensors on each of the 3 BLDC output legs
- Inputs for 2 1800 CPR, No Index E3 Encoders from US Digital
- 5 Debug LED outputs
- 6 IR LED outputs
- User input button
- Reset button
The board was never fully implemented because of a few issues that became apparent during soldering. These are:
- Short between power and ground planes - all of the vias next to test point pads were unintentionally connected to both power and ground planes causing a short. The vias were eventually drilled out.
- Inefficient h-bridge drive circuit - The use of NPN transistors to drive the MOSFET gates resulted in a significant loss in the maximum duty cycle as detailed in the BLDC Driver page.
- Buck converter capacitors were too large - the typical application notes for the buck converter recommended 22uF and 510 uF high voltage capacitors which were 1" long and 1/2" diameter. Which were much too large for the space constraints limiting the board.
- Improper buck converter installation - the 48V to 24V buck converter circuit overheated after initial installation.
- Sensor soldering issues - both the accelerometer and the gyroscope came in very small packages. During debugging it became difficult to prove that the sensors were soldered correctly with reliable connections.
The v2 iteration of the board scaled back on functionality because of the issues that came up in the v1 boaord.