120 Degree Phase Shift Osc.

Nov 05, 2010 468 Replies

This is the output of a three-phase generator. So, glue a bar magnet to the shaft of a small DC motor, and put three pickup coils around it. Pole pieces for the pickup coils are optional.

Use a PWM switch to run the motor, with a frequency-sensing feedback loop to generate the error signal. A '4046 can do the frequency-voltage conversion for you, off any of the phases. You might even be able to generate pulses on all the zero-crossings of the three phases, that'd give you six times the base frequency to lock the PLL to.

The OP finally made an appearance and says that the relay coils aren't being used to energize relays; they're evidently just the right size for OP to make a rotating magnetic field.

Cheers! Rich

[...]

+--------+ |ADUC7021| +--------+

:)

John Devereux

I've been known to go super crude and use a shift register, resistors on each output, sum thru an OpAmp. On-chip I've even done this at

1GHz ;-) ...Jim Thompson
| James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | Ding Dong! The Witch is Dead. Wicked Old Witch Pelosi! Hurrah!

This is "crude?" I'd have thought it rather elegant, if you want to go through the tedium of selecting the right resistor values.

Cheers! Rich

The ADUC7021 was a sort of "cheat" - a joke - but I am starting to think it is probably the best solution. It's a single chip ARM, with 4 good 12 bit DACs and an internal 1% oscillator. There's code space for a 12 bit lookup table for each DAC, though at 10Hz it could calculate the waveform in real time!

John Devereux

uld

een

Yep. If you want to get tricky you can use Don Lancaster's magic sinewaves or something similar to conjure up PWM waveforms that cancel the various harmonics, which makes the filtering even easier.

Yes. OTOH, you might even direct-drive the coils here, since they're being run off Vdd anyhow. That would save even the op-amp. It doesn't get any simpler than that.

-- Cheers, James Arthur

y

tween

s,

All he needs to do is compute a reasonable, semi-optimal sequence in advance, then output it with efficient code.

The generator costs $0.50, one part, plus filters and drivers. It's really all pretty easy.

-- Cheers, James Arthur

.
I
s

could

..

y

tween

s,

Belay that last part--the OP likely wants a clean magnetic field, which means PWM --> filters + coil drivers, as you said.

-- Cheers, James Arthur

he

nly

between

s

ips,

g
e

To approximate a 10Hz sine wave with was new mark to space ratio every

50usec, you need 500 mark-to-space ratios, and successive ratios aren't going to differ by more than one part in 637, so you want to code them to ten bits or better, and your pulse width timer wants to be clocked at at 20MHz or faster, all of which is starting to get difficult with $0.50 micro.

If you don't think about it in any detail.

-- Bill Sloman, Nijmegen

annel.

summing

lse none.

e use R2R ladders.

-+ =A0+---+

|-->PHASE 1

=A0+---+

+
+

=A0+---+

|-->PHASE 3

-+ =A0+---+

ne :-)

r

hatever,

And when I was, he had problems comprehending what I'd posted, leading to several quixotic quests on his part, trying to get me to admit that I'd screwed up because his interpretation of what I'd posted didn't make sense. It proved impossible to convince him that the problem lay with his interpretation (which couldn't possibly be wrong), rather than what I'd actually posted (which he couldn't possibly have misunderstood).

-- Bill Sloman, Nijmegen

--

  • =A0+---+

F|-->PHASE 1

  • =A0+---+

=A0

-+

-+

  • =A0+---+

F|-->PHASE 3

--

  • =A0+---+

Hewlett-Packard did it first. It's only crude if you don't use enough taps and don't get the resistor values right (including the windowing function). When you get it right it is a hardware implementation of a finite impulse response filter, and can be very give very nice results.

-- Bill Sloman, Nijmegen

Exquisitely beautiful chip and, at less than $6 in qty 100 for the Rolls-Royce version, it seems to be economical. But, then there _is_ the cost of the development tools and the cost of climbing the learning curve. You're probably right if the OP is thinking of volume, but for a one-off (or maybe even a million) I'll stick with my shift register solution and either low-pass or band-pass the 10Hz output square waves to get the required 3 phase sinusoidal outputs, depending on what he needs. So far he's been pretty sketchy about describing what he wants to do, so unless he comes forth with what he's up to, I think he's just data mining

You might want to make the LUTs 384 or 768 bytes, so it's divisible by

  1. :-)

Cheers! Rich

e:

The

only

n between

was

ships,

th

big

-

ine

h

You're assuming interrupts and a timer for PWM. That's fatness. Don't do that. Use an ATMEGA part, or a few resistor-DACs if you can't figure out how to do without those crutches.

None of those are a big deal. Like I said, it's really pretty easy. No need for that pile of FPGAs with multiple supplies, etc.

I once generated a 1.2MHz direct-sequence spreading modulation, with data, real time, in an 8051. Several million were sold.

This is easier (and the PIC's faster).

-- Cheers, James Arthur

e:

k

tion.

I am

y to

rt from

age,

uch

he

ough

e four

s.

f

ves

Hill

d it

S

-dds...

a

gic

Hz

red to

degree

3,4
0).

three

ch,

4

er,

art of

ts

on of

n PWM

ving

=A0 =A0 =A0Vcc

=A0 =A0 =A0 =A0 =A0 =A0 |

[DIODE>]-+ =A0

=A0 =A0 | =A0 =A0 =A0 =A0 =A0| =A0

-[COIL]--+ =A0 =A0

=A0 =A0 =A0 =A0 =A0 =A0| =A0

DE>]-+ =A0

=A0 =A0 | =A0 =A0 =A0 =A0 =A0| =A0

=A0 =A0 +--[COIL]--+ =A0

=A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0| =A0

--+-[DIODE>]-+ =A0

=A0| =A0 =A0 | =A0 =A0 =A0 =A0 =A0| =A0 =A0 =A0 =A0 =A0 =A0

=A0 +--[COIL]--+

|

=A0 =A0 =A0 =A0 =A0 =A0 =A0

=A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0

He just did.

-- Cheers, James Arthur

=A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0

Of course I notice, but why would I harp on it? I prefer to learn from people's strengths rather than dwell on their weaknesses, for one thing. Besides, the only thing I've seen recently is that he thinks he skis faster.

-- Cheers, James Arthur

What sort of entrance fee though? Some sort of demo or development board, toolchain, programmer, etc?

Grant.

from

four

to

degree

of

Change the subject? Irrevelant? This is sci.electronics.design, you useless twit.

John

No fair distracting him--he's busy proving a uC can't PWM a 10Hz sinewave.

"If you think you can do a thing or think you can't do a thing, you're right." --Henry Ford

-- Cheers, James Arthur

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required