Scr firing

Hello all,

We have been trying to build a prototype of a single-phase power
controller.The amount of power to be output is controlled by a PID
algorithm. The PID algorithm resides in the external flash associated
with a processor. The output from PWM module of the processor controls
the firing of the SCR. Let me add that the input to the PID algorithm
are the current temperature of the heat lamps which are controlled by
the SCR. The temperature is measured using thermocouple and acquired
using SADC.The part till the output from PWM is complete.
I am not sure how to proceed from there.We need an
optoisolator for electrical isolation.And then there is the SCR
firing circuit with a couple of SCRs connected in anti-parellel
fashion. What else shall I need to consider the SCR firing circuit.
Or do I need to change my plans and go in for Mosfet based induction
heating.Could you please shed some light on the above problem.

Thanks in advance,
Robotnik

 

Hi, R. The easiest, most straightforward way to go is to purchase a
solid state relay (SSR) that will handle your problem. SSRs have the
optocoupler and thyristor built in. Heating lamps generally provide a
more or less resistive power factor, so I'm not sure why you seem to
be focused on inverse-parallel SCRs. For most loads, a triac-based
SSR should do well.

A couple of caveats here. First, to be safe, think 2 watts power
dissipation in the SSR per amp of load current. Purchase a heat sink
accordingly, use thermal heat sink compound to mate the SSR to the
surface of the heat sink, and use forced air where appropriate. Make
sure to purchase a SSR which is "random turn-on". *DO NOT* get one
which is "zero crossing", unless your control algorithm takes that
into account.

Apart from that, you might get better advice if you described your
problem a little better. For instance, what kind of heating lamps do
you have, are you shooting for phase control, integral half cycle
control, or proportioning on/off control of your load, and of course,
what's the load power, current, and voltage?

Note also that, if this is a school project, a SSR probably won't cut
it -- you'll have to "show your work" rather than just buying a
module.

Feel free to post again.

Cheers
Chris

 

Why do you want to do this? What would define a successful project? Why not
an off the shelf solution?

 

For SCR controlled AC, you need to sense zero crossing, and add a delay for
phase modulation. For slow response like a heater, you can use PWM
techniques, but only if the PWM frequency is much lower than the AC line.
In other words, you just turn on the power for a few seconds, and then turn
it off for a few seconds. You probably don't want to use the PWM module for
this.

Induction heating is something else altogether, using RF coils to heat a
metal object. Heat lamps are more properly called radiant heating, using
infrared to heat the object.

For antiparallel SCRs, you need two isolated gate drive circuits. Pulse
transformers are the simplest and cheapest, but may not be as reliable as a
well defined gate pulse circuit, which uses a narrow high current pulse
followed by a "back porch" of gate current just over the minimum turn-on
spec. You also should be careful to avoid gate current during the quadrant
where the SCR has reverse voltage applied to it.

You can also use a single SCR connected to the + and - terminals of a full
wave bridge, so you can use a single gate circuit. However, the diodes
waste power and the SCR works twice as hard.

For any SCR gate drive, be sure the isolation is sufficient. An
optoisolator like 4N25 has isolation voltage of over 4000 volts, which is
just enough for a 480 VAC circuit. If you build gate drive power supplies,
make sure the transformers are rated for the mains voltage you are
controlling. If you use DC/DC converters, isolation voltages are often only
500V, which is just barely enough for 120 VAC. The isolation barrier must
be designed to withstand high voltage transients on the power line without
degradation. A breakdown will allow mains power to connect to your logic
circuitry, with dangerous and catastrophic results.

These concerns are another reason to use a commercially available solid
state relay.

Paul

 

Your two main choices,m ere are phase controlled power (fast
smooth and electrically noisy) or burst fire integral half
cycles (slower, simpler, lumpier and electrically quieter)

For quantity one, you best bet is an appropriate solid state
relay. The line voltage isolation is included, both for the
control signal and for the heat sink.

There are instantaneous turn on types, if you want to use a
small transformer to provide a line frequency zero crossing
reference for the micro, so you can delay the firing pulse
with software, for phase control. There are zero crossing
turn on types for burst pulse control, where you need only a
slow PWM digital output to control the average power.

There are also units available that convert a variable DC
control voltage or current into phase or burst pulse power
control, if you have a DC control output available. The
last kind are most expensive, but you will definitely spend
more designing, building and debugging your own, if you are
going to need only 1 or a few.

 

I have done PID loads of times over 25 years for motors, but its a
similar principle for controlling heating.
In all cases I used the mains as a reference and the output from the
PID was scaled to provide a pulse of between 0 and 9ms. A triac was
used in every project.
I have to say that PID is not always to convert into a real project
especially for a beginner.
I found that often a bit of fine tuinin gwas required to get the
critical damping just right.
If its not right you will get hunting of the speed or the motor will
grumble badly (not a problem with heat lamps)

www.ckp-railways.talktalk.net/pcbcad21.htm

 

Try also looking at:

http://focus.ti.com/general/docs/lit/getliterature.tsp?literatureNumber=slaa043a&fileType=pdf

Might give some ideas.

Jon