Re: Quest for the simplest zero voltage switching

On Aug 14, 3:44 am, "petrus bitbyter"
<pieterkraltlaatdit...@xxxxxxxxxxxxxxxxx> wrote:
"MooseFET" <kensm...@xxxxxxxxx> schreef in berichtnews:1187053482.282841.143630@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

On Aug 13, 4:21 pm, "petrus bitbyter"
<pieterkraltlaatdit...@xxxxxxxxxxxxxxxxx> wrote:
"MooseFET" <kensm...@xxxxxxxxx> schreef in
berichtnews:1187014014.937032.70700@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

On Aug 12, 3:58 pm, "petrus bitbyter"
<pieterkraltlaatdit...@xxxxxxxxxxxxxxxxx> wrote:
"JoeyB" <joseph.bur...@xxxxxx> schreef in
berichtnews:1186924269.912470.58160@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

All,

I'd like to do proportional control on a 120VAC, 1.5kw heating
element
(capacitive load) as simple as is possible. Ideally, I'd like to
have
a single pot that I could adjust from zero to one hundred percent
duty
cycle input to a zero voltage switching controller chip. This ZVS
controller chip would then control the Triac swiching the load. I've
found several chips out there that provide zero voltage switching
but
only one that DIDN'T require a dc voltage for operation AND simply
allowed a pot to be connected to the control pins, the Motorola
CA3059. Or at least, it was the only one that had a circuit in the
data *** showing such a configuration. Unfortunately, its
discontinued and I cannot find any suitable substitute for its
operation. The MOC3081 seems like a possibility however it requires
a
digitial input control. This would mean some kind of timer circuit
(again DC would probably be required) unless I'm missing some
obvious
passive solution to providing this input. Any help from the power
circuit design pro's out there would be greatly appreciated.

The most simple zero cross switching can be obtained by buying a solid
state
relay with a build in zero cross switch. Building a timer to controll
it
will be a little bit harder. I ever build one using a classic 555 with
a
stereo potmeter and some other passives and diodes. The 555 was used
in
astable mode and set to about 10Hz. The potmeter controlled the duty
cycle
from about 1 to 100%. As the zero cross switch only switched on zero
voltage, I could controll power by the half cycle. Unfortunately I had
to
add a small power supply. The old 555 required 10-15mA at 15V and the
solid
state relay added another 5mA. Using a resistor directly from the
mains
would have made me to get rid of over 3W heat (220-230V/50Hz mains).
As
you
have only 120V mains and can use a CMOS 555, powering via a series
resistor
will not be difficult.

Nevertheless, this days I'd go for a PIC10F200 and a normal 10k lin
potmeter
to build the timer. With one or two extra components and a little
effort
you
can build the zero crossing switch into it as well.

I think if you move up that product line just a bit you will do a
little better. You want a comparitor to do the zero crossing and a
ADC to read the pot. You don't really need the PWM stuf because 60Hz
is so low you can count off the cycles in a loop.

The PIC's output isn't quite strong enough to trigger a triac so an
external (gasp) transistor would be needed.

The PIC's ADC uses the supply voltage as a reference. If the pot used
the supply as its input, the value of Vcc drops out.

Since you are counting cycles of the mains, the RC clock of the PIC is
more than good enough.

The total current draw of the control circuit would only be several mA
at the most. Power can be done like this:

R1 C1
Mains---/\/\---!!----+---->!----+----
! !
5.1V /-/ ---
Zener ^ ---
D1 ! !
--------------------+----------+-----

The ratings of R1 and D1 are mostly set by the turn on inrush. Once
the circuit is up and going most of the mains drop is on C1.

You will need a relative good 5V power supply.

I don't see why. The PIC doesn't need one, the pot doesn't need one
and the SCR trigger doesn't need one.

Maybe you're right. I tend to stay on the safe side.

The safe side does have its attractions but remember one goal is
keeping the costs down. The OP said "simplest" not "reasonably simple
and yet reliable".

[....]

If you use a triac, you need far less average current. The PIC only
needs to make the output high long enough to trigger it.

Depends heavily on the triac. Some types require 50mA, and - as you say -
the trigger pulse needs to last long enough. It has been some time now but I
had problems with trigger pulse duration. Don't remember the required pulse
length anymore.

You have to apply gate current until the device is passing much more
than the holding current. In a zero crossing switch this can force
you to pulse longer than normal. If the load is inductive matters are
extra messy. The OP is running a resistive load so we don't have that
to worry about.


I misunderstood.

It happens to us all.


You don't need a "zero crossing pulse". You only need a divided down
copy of the AC waveform. The PICs have comparitors in them. You just
need to use one comparitor to detect the zero crossing.

That small PICs have either a comparator or ADCs, not both. But speaking
about the PIC10F222, it has two ADCs so you can use one to look after the
mains phase.

You are right. Either we need to add a comparitor (perhaps just a
transistor) or we have to get very clever. First lets see how few
parts the comparitor could be:

The PIC has weak pull-ups. We can use that as the pull-up on a
comparitor. 0.7V is very close to zero when considering mains voltage
so we could:

-----------PIC pin
100K !/
AC----/\/\/----+-----! Small NPN
! !\e With large HFE at low I
1N914--- !
^ GND
!
GND

A small JFET may actually be a better way to go if you have a low
Vgs(off).

If you really want to go cheap and nasty, you could just run the
resistor to the pin of the PIC. The substrate diodes would clip the
signal.


Now for getting very clever:
BTW. Re-reading the posts I see I mentioned the PIC10F200 in my first reply.
Should be PIC10F222 as the 10F200 has no ADCs. Sorry for that.

I had assumed you meant a member of that family not that specific one.

Lets see if we can do without the ADC.

If we run a simple RC from an output to one pin of the comparitor and
the pot to the other, we can use a sort of delta-mod ADC. We have to
be careful about how we code the whole thing to be sure to get around
the loop at the same rate but I will leave that detail out and write a
bit of pseudo-code to show what I mean:


if RC < POT then
OUTPUT = HIGH
inc MeasuredPot
else
OUTPUT = LOW

LowerPart = LowerPart - MeasuredPot
if BORROW then
dec MeasuredPot

The MeasuredPot will settle at the point where the rate it causes
borrows from the subtract matches the portion of the time that the
OUTPUT has to be high to cause the filter to match the setting of the
pot.

.