pointhi / highpower-mechaduino

Mechaduino derivative which drives up to 10A, drawn using KiCad

View this project on CADLAB.io.

HighPower-Mechaduino

KiCad based Mechaduino derivative which drives up to 10A.

This is the original Meachaduino Project: https://hackaday.io/project/11224-mechaduino

Warning: This Board is currently under development. Simply fabricating it doesn't mean it would work! See v0.1_BUILD_LOG.md for some problems found while building the first prototype.

A big thanks to the Shuttleworth Foundation for supporting the development of this Project.

Improvements

• ESD protection (especially USB)
• flyback diodes
• stronger drivers
• better silkscreen
• correct use of blocking capacitors
• fuse (for protection against polarity reversal and shorts)
• input voltage measurement
• supply of the microcontroller via the motor voltage supply
• external storage for calibration and configuration data
• Temperature sensor

Planned Electrical/Mechanical Ratings

• Input Voltage-range of about 12V-50V
• N-Mosfet Driver which is capable of about 10A
• board in the size of NEMA 23 (56x56xmm) or NEMA 24 (60x60mm)

Interfaces

• USB
• UART (3.3V)
• Step/Dir/Enable (PoStep25-32 Compatible)

various informations

• BOM can be created using KiBoM
• There exists a modified version of the Firmware for this project: ElektronikNode/Mechaduino-Firmware

License

Hardware including documentation is licensed under CERN OHL v.1.2. license

highpower-mechaduino's Issues

Consider using RPi Zero

Based on the 4EUR cost, RPi zero may be a very good choice: https://www.raspberrypi.org/products/pi-zero/ please consider that.

price optimizations

The current component price range is about 50-60€ per board (when buying for 10 pbc's). Not problematic for the prototype board, but we should optimize it.

This is a list of the most expensive parts of the board (based on the discounts you get when you buy for about 10 boards):

Component	price per board	description
AS5047D	8,19€	encoder
IPD079N06L3	5,90€	motor driver mosfets
ATSAMD21G18A	4,78€	controller
A4955GLPTR	2,84€	motor controller
RB162MM-60CT	2,11€	schottky diodes
744773122	1,96€	inductor for step down
10UF POLY ELCO	1,83€	elco for step down (very low esr)
TERM BLOCK 4POS	1,67€	step down connector
MCP16331T	1,26€	step down regulator

those few (main) components sum up to about 30,54€ per board.

programming header P1

On the programming header P1 the pin 7 is connected to the ground, and in the Arduino zero schematic it is not. I know this is not a big deal, but maybe (in some parallel universe) it could bother the programator when connected.

AS5047D supply

According to the datasheet (page 9, bottom) the pins VDD and VDD3V3 should be connected to 3.3V.

C2=1uF can be omitted in this case.

Current Reference VREF

The text saying "I_MAX=14"next to the voltage dividers for VREF (R7-R10) is wrong. Instead it is 11.6A.

U2 = U1R2/(R1+R2) != U1R2/R1
;)

Also I think that the cutoff frequency of the low pass should be a bit higher. With R9, R10=1k and C7, C8=100n the cutoff frequency is ~1.5 kHz.

For a 200 steps/rev motor turning with 2000 RPM (which is not too unrealistic I think) the step frequency (in our case: sine frequency) would be 6667 Hz. I would suggest to set C7 and C8 to 10nF instead of 100nF.

Export schematic in pdf

It would be great if you could also export the schematic as PDF so it is readable online.

5V USB power switch

Maybe it would be beater to change the switch part, so the MOSFET Q1 should be triggered by the USB and not the VDD. I think it would be beater this way because the USB voltage is constant, +5V, and VDD is not, it can be from 12V to 50V, and the gate current would not be the same. See the reference picture:

step down converter MCP16331

The collector of the bootstrap diode D5 should be connected to the BOOST pin an not to the SW pin! See the example circuit on page 2 in the datasheet.

Also I think that the output diode D2 should be directly in series with the induction L4. Otherwise the output voltage could drop due to a transient load current. This could also cause stability problems of the voltage regulator.

Free Wheeling Diodes

The free wheeling diodes D6-D9 and D10-D13 are schottky diodes so the appropriate symbol should be used.

Also the lower diodes should be connected to the MOSFET-drains instead of GND if possible.

pcb mounting holes

I started to design a housing/mounting for the HighPower-Mechaduino on a NEMA24 stepper motor and discovered that the mounting holes in the PCB are made for M4 only while they are for M5 in the motor. See this screenshot:

(The CAD model is based on measurements form an actual motor.)

See also this datasheet of a NEMA 23 motor:
http://www.omc-stepperonline.com/download/pdf/23HS16-0884S.pdf

I think it would be good to change mounting holes in the pcb to fit to M5 screws (a diameter of 5.2mm should be sufficient)

There might be some NEMA23/24 motors where the mounting holes are actually M4, but it is easier to fit an M4 screw into an M5 hole than the other way round. ;)

Shunt Resistors

I think a resistor with 4-wire-connection should be used for the shunt resistors R16 and R19. Otherwise the load current flowing throw the groundplane could lead to a wrong current feedback due to the parasitic resistance and inductance.

The connection on the drain side is already (almost) correct, but not so on the ground plane.

Current rating

I believe 6A current rating at 48V is completely sufficient for most applications the board would be used with NEMA 23 motors.

MCP16331_5V circuit

Check the datasheet for the reference circuit. If you change the schematic like in #11, you can lose the D2 diode. Otherwise move the resistors R20 and R21 in front of the diode D2.

Have you tested the board?

Hi,
Tried to find any contact info to send you a private message, but couldn't find any thus I'm writing this post.

Have you managed to test current layout of the board? I'm in need of such controller and just wondered if it's working or are there any serious issues that you're aware of?

Let me know if I can help in further development.

Cheers,
Maciej

diode D1 outdated

The diode D1 is not recommended for new design, maybe it would be best to change it. They will probably stop producing it in near future. See the manufacturer link: RB160M-60

magnetic encoder option TLV493D

Regarding enhancements mentioned in the #9 issue, possible option for magnetic encoder is TLV493D which is ~ 1 EUR. There is even a small dev kit for it TLV493D-A1B6 2 GO KIT