[smarbaga]

Hello:
Attached is a new servo controller design similar to my old "MTMC" (multi type motor controller) from about 7 years ago.
The old desighn never got finished because I had a home breakin and all my electronic stuff got ripped off
I think I will start to collect some more electronic stuff. I recently got my leg amputated and i am getting boared.
It has been about 7 or so years since i have been into this electronic stuff so I have a few questions, lots of time tho.
Nothing on this partial schematic is carved in stone so all sugestions and advice will be greatly appreciated.

1. I am not sure what kind of data will be comming in from the controlling PC software ?
2. How would I connect the usb chip to the controller chip. (handshaking, polling, I2C ...etc.?
3. I guess if i had to I could use some spare lines from the usb chip for Home, Limit, Auxillary I/O, Error Signals .. etc.
4. Lots more questions as I have not written any software for this thing yet.

When I get some hardware I thing i will proceed to get some leds flashing on the usb chip via a breadboard and go from there.
As it stands the cirrent limit chip INA282 can only handle 80 volts. I was going to use another current monitor but it was 3mm,
I have used 3mm chips before but they are to small for me to handle. so any other higher monitor chip sugestion would be good.

Attached is a pdf of my old mtmc (with bom etc) and the proposed new desigh.
Thanks in advance for your interest and input. Sincerely Garry.

[H500]

It might be interesting to take a look at Microchip's app notes on vector control. This is the state of the arts technique for motor control. It's actually not as complex as it sounds. The algorithm measures the currents and rotor position and then adjust the 3 PWMs to point the magnetic force directly in the direction of rotation. Maximum efficiency and performance is achieved because no power is wasted, pushing against the center of the rotor.

Microchip has already written the motor code and tuning interface. To use it, you just need to change the current monitoring circuit to a simpler one that connects to the ADC. It would also be good to use a dspic33 variant that has peripherals for motor control.

[smarbaga]

Hello:
Thanks H500 for your input. I will look intoyour sugestions.
one of my intents was to use trapazoidal ramping, this also poses lots of questions.
upon setup:
1. Do I have the user load a maximum speed in the controller?(non volitile ram)
2. Should the ramp rate be programable by the user.?
3. If the controller is going home should the ramp and max speed be different?
4. Should the user just select the tool type and these params be set automatically?
5. If this was on a milling machine, ramps amd max speed would be different for
material (steel, aluminum) as well as tool size
6. or are these paramaters set in the pc cnc software.
I have built lots of motor controllers but have never used or had any pc cnc software.
so i don't know exactlly what it does. guess i got to know that first to be able to handle its data.
7. I am open to sugestions how to join the usb chip to the controller chip with minimum overhead.
should i use somthing like I2C, or a bit bang connection, or something else.
I would like to make this controller as standard and as easy to use as possible.
The hardware is not carved in stone and a program can always be modified or changed.

I will attached (for the last time) a jped of my old mtmc and a jpeg rather than a pdf of the
proposed new design, I think jpegs are easier to view.

Thanks in advance for your interest and input. Sincerely Garry.

[James Newton]

7. I am open to sugestions how to join the usb chip to the controller chip with minimum overhead.
should i use somthing like I2C, or a bit bang connection, or something else.

I would recommend using a standard pin header for the TTL serial connection and then let people plug in whatever interface they want. e.g. you can connect a TTL to USB converter like
RLC_3:USB-TTL serial converter massmind

or a TTL to RS 232 connector:
RLC_1:RS232-TTL level converter in a DB9 backshell massmind

or directly wire the TTL lines to an Arduino or Rasp Pi or whatever

and perhaps provide a separate i2c connection.

[James Newton]

1. Do I have the user load a maximum speed in the controller?(non volitile ram)
2. Should the ramp rate be programable by the user.?
3. If the controller is going home should the ramp and max speed be different?
4. Should the user just select the tool type and these params be set automatically?
5. If this was on a milling machine, ramps amd max speed would be different for
material (steel, aluminum) as well as tool size
6. or are these paramaters set in the pc cnc software.
I have built lots of motor controllers but have never used or had any pc cnc software.
so i don't know exactlly what it does. guess i got to know that first to be able to handle its data.

The standard for CNC (and 3D printers) is a separate motion controller that puts out step and direction to the motor drivers. It makes sense to do this OUTSIDE the driver because the motion often needs to be coordinated between multiple motors. So in that use, your driver should listen to step and direction, use that to maintain a desired position, and then run the motor to that position.

Because the step pulses will be ramped up and down by the motion controller, you don't really need to worry much about speeds... just a simple max speed setting should be enough, I would think.

[grabik]

Look on microchip project AN1017 with dsPIC30F ,dsPIC33j ,dsPIC33EP .
link http://www.microchip.com/wwwAppNotes...pnote=en025522

LibStock - Space Vector Modulation (Sinusoidal Voltage ) Control of a PMSM Motor with Phase advance

on youtube
https://www.youtube.com/watch?v=wqtW4poaKG0

https://www.youtube.com/watch?v=ShHluVtlOSM

[H500]

Garry,

The common CNC programs such as EMC and Mach, takes care of most functionality, including speed and acceleration. It will output the step and dir signal to control the motors.

The only purpose for the serial port is to tune the servo motor feedback loop parameters. The cnc programs does not handle this. You will need to write a PC program, or use the one supplied by Microchip. If it was my design, I would omit the FTDI chip. Just include TX, RX, CTS, RTS on a header. USB to serial cables are readily available for $1-$5. No point including the circuit on every board. Motor tuning is normally only done once.

I recommend including the current measuring circuit in case you want to experiment with vector control. All you need is an quad op amp such as the mcp6024, configured non-inverting amps to boost the shunt resistors readings.

[smarbaga]

thanks grabik and H500 this makes things easier especially the usb to 232 thing. btw the ina chip in the diagram reads the shunt and will output to an adc for curent measuring.
if anyone out there knows of a different current monitoring chip that, puts out a tunable approx 0-5v, is not 3mm x 3mm ( i can work with this size , but a pain), and is higher than 80 volts pls let me know. I will redraw the cct and post it. Thanks all. garry

[smarbaga]

Thats about all the drawing for now. I am sure i will have to re-route some i/o lines due to re-mapping constrictions.
Now it is just a matter of reading and trying to understand the 500+ page data manual.
I am always open to suggestions. thanks !

[H500]

I would add two .05 ohm resistors to low side of the mosfets for current measurement. The third branch current can be calculated from the other two.

You can eliminate the 6 fet drain-source diodes. The fets already have them built in.

[smarbaga]

Hello:
It will be a while untill i get some test equipment together but if anyone wants to glue all the free software out there together and test out some

of this cct. I will gladly change the schamitic to your specs and if it all works I will draw a pc board and I can get cheap pc boards made.

Attached is some of the ic data and a jped of perhaps a starting point to this controller.

Thanks in advance for your interest and input. Sincerely Garry.

[smarbaga]

Hello:
I like drawing, but I like the hands on better, there is nothing to draw any more, looks like the cct. is really simple (kiss). If there is any body out there who is interested in this concept one can put the cpu on a bread board and scope the outputs, I know the drive cct works.we can always add bells and whistles later. So I guess I must get a solderless bread board and a pic programmer and a scope to continue.
Minimum list of what one needs:
1. Power supply ( current limit ).
2. Volt Meter.
3. Oscilloscope.
4. Solderles Bread Board.
5. Components.
6. Computer, Software.
7. DC motor with encoder.
8. Tools.

Thanks in advance for your interest and input. Sincerely Garry.