Re: Low cost Servo Controller

Bertho Boman wrote:

> From: Bertho Boman <boman@...>
>
> Hello Jon,
> I like to sign the non-disclosure and take a look at what you got.
>
> I like your operating mode and wish you best of luck.

Attached file follows :

Meanwhile, here is a non-disclosure agreement, for you to sign and
send back to me by US post.

I, __________________________ (undersigned) agree to the following

terms to protect the confidential nature of the circuit design that will be

provided to me by Jon Elson of Pico Systems, of Kirkwood, MO.

This circuit is a pulse-Width Modulated servo amplifier for

brush-type DC motors.

1) I will not disclose the details of this circuit to any third

party, unless it is for technical help in constructing and

setting up the circuit.

2) If it is necessary to engage others in the construction of this

circuit, I will require them to agree to these same terms of

non-disclosure.

3) I will not, under any circumstances, manufacture amplifiers based

on this design, or allow others to do so.


Signed
__________________________



Date
___________________________



Oh, yes, my address is :
Jon Elson
543 Lindeman Rd.
Kirkwood, MO 63122
(USA)



Well, sorry to have to go through all this, but I want to reserve the
right to make commercial sales of these servo amps in the future.
I did go through quite a bit of effort (and expense counted in blown
transistors and chips) to get to this final design. Sometime during the
course of this project, I'll give some of the history of that.

Finally, I need everyone to give me some info on the motors you plan
to use. I think there will be (probably) 3 basic versions of these
amps.

1. A low-voltage amp, good up to 80-90 Volts, and anywhere from
5-12 Amp capacity, with small filter inductors and sheet metal
heatsinks on the transisors. These will be identical to what I'm
using right now.

There is only one part we have to 'make', that is the charge pump
that delivers power to the high-side drivers and the fault sense
circuit. It involves winding 2 isolated 75-turn windings on a little
pot core, and only takes a couple minutes to do.

2. A high-voltage amp, good to 200 VDC, and up to 12 Amp capacity.
The heat sinks may still be sheet metal bolt-on units, I don't
expect the transistors to run much hotter. But, the inductors
in the output filter need to be made custom, I've run some designs
using Magnetics, Inc. design guide, but I'll have to actually get a
sample and test it to be sure it won't overheat.

3. A spindle drive amp, with all the power circuitry built off
board, with 30-50 Amp IGBT's, hand wired (with heavy
wire) on a heat-sink extrusion. The PWM frequency needs
to be turned down to 25 KHz or so (one component change).
I have some big lengths of heatsink material, so we can select
what you need. The filter inductors will definitely need to be
custom made here, pretty much a design for each application.

Note, also that you will need an external +/- 12 Volt DC power
supply,
about 1/2 A for the +, and .25 A for -, for each amplifier. One big
supply is fine. You will also need a DC power supply for the servo
power. I'm using a 15 Amp, 68 Volt unregulated supply for my setup,
but your motors determine what is appropriate. You will also need
a card cage for the 4.5 x 6" cards (standard Vector board format).
I used some 22/44 contact wire-wrap sockets for the board to plug
into. I used some 1/8 x 1/2" bus bar (armature bar) to connect
ground
and servo power to all the amps. I ran the bar between 2 rows of
pins,
and soldered 4 pins to each bar. For the motor output, I bent 4 pins
together, and stuffed them into a crimp-type male 1/4" quick
disconnect
terminal. I then soldered the terminal to the pins. There are 2 of
these terminals for each motor. Finally, you need a good DC tach
to provide velocity feedback for the servo loop. I got 3 really nice
DC tachs from C&H surplus in Pasadena, CA. Unfortunately, those
are all gone. I'm looking for some more, and if I find a suitable
source,
I'll let everyone know. The trick here is it must have minimal
ripple
at constant speed. Some small ironless-rotor DC servo motors, like
Maxon and Escap are made exactly the same as their tachs, and will
work fine.

The final gizmo to construct is the power up/down control. I use a
setup with a small relay and a big contactor to ramp the servo supply
up and down. When the CNC control commands the servos to
come active, the first relay connects a resistor to charge up the
caps
located on the amp cards. When enough time has elapsed, a simple
delay circuit closes the direct connection to the DC power supply
with the big contactor. When the control tells the servos to power
down, the small relay opens, which cuts power to the contactor, as
well as connecting a dump resistor (same actual resisitor as the
charging one) to the amps, draining the on-board cap, and also
burning off any back-emf coming from decelerating motors.
I'm having some slight problems with burning contacts on the small
control relay, so we might try a bigger one with more contact gap.

My phone at home (with the recorder) is (314) 965-5523.
I'm usually here in the mornings (central time) and evenings, and of
course
on weekend. On weekday afternoons, I am at Wash U., at (314)
935-6547.

Jon