So after doing some more reading it turns out I am getting pretty confused on the heat portion of this.

Some parts generate more heat depending on the amount they are switching, and some parts generate more heat depending on how long they are "on" for. (SSR's, mosfets, transistors, and triacs). At least I thought it was this way.

Could someone shed some light on this?

 

Hi everyone,

I am new here.
I just want to back track, to add my own experience. this may be an overkill for the current project.
all of this is add ons suggestions on what is said above already and @Mahonroy s findings

I live in a 220V world, 120VAC is very similar but i may make some assumptions at some point.

I have done some research on controlling AC motors. All the controllers and designs i have tested and found works as designed in pure R loads. sadly I don;t have many of those in use other than my test bench many of our loads are RL excluding few heating etc stuff.

I am also a MAJOR fan of micro controllers we can now have an atmel controller for $1.5 (ready made PCB ready to connect few wires and go.) because needs always change and redoing the whole PCB or re soldering etc is not efficient. Sorry we live in virtual of everything, cloud of everything but some of us still try to design one job/purpose circuits.

cheap and dirty: (mostly what we need) I also mean no etching or designing PCBs etc...
your budget is $50 you can get this working below $25 with the MCU included

1)
if you don't need fine control, then use an off the shelf SSR for DC, that is suitable for your load that is random on/off then use a PVM base frequency. you can also use a zero crossing SSR (cheaper) with very similar results. you may miss a cyle here and there but that is it. SSRs are packed insulated opto isolated cheap packages.
(read below option 2)
if you keep your PWM to 6Hz you can easily get a realistic 10 step variation. PWM frequency above 50~60 Hz will only produce on or off with most of the SSRs.

2)
again go and get an SSR with potentiometer, Variable resistance, input (dirt cheap) $4~5 for 25A version
now the dirty part.
now instead of the Variable resistance use an LDR (Light dependent resistor).
connect a LED to your PWM output, and use a good amount of electrical tape to put both of them together and clock ambient light. Volaaa you have created your self an opto isolator with variable resistance output.

dont expect the output to be linear, you need to do coding and conversion on your MCU to compensate, but it is software and easy to do, change or update.

- be careful with LDR voltage depending on the SSR you pick. some has line voltage.
- be careful with LDR heat dissipation again depending on the SSR you pick.


I put it in 3 different ways of controlling an AC power. similar to @Mahonroy

But I have to keep everything still AC. Remember my pesty AC inductive loads..
And No slow control eg 1 second on then 3 seconds off etc...

1) Triac Dimmer Circuit. just chopping the wave. same as OP... cheapest and mostly abused one.

2) removing few or all cycles out.

This is on off but you switch off or on the full cycle, 60Hz we have sixty of them a second, 50Hz etc..
how many is off and how long to wait can be explored and tested using micro controllers. (this is like slow control if you try to have to much steps.
for 10 step, resolution, variation the control cycle will repeat 60Hz 6 times, 50Hz 5 times
for 5 step, resolution, variation the control cycle will repeat 60Hz 12 times, 50Hz 10 times etc

3) Doing what we do in digital DC world with AC. PVM the actual signal. yeaaay...

this is same as what @Mahonroy mentioned as option 3 but NO rectification. I need to keep both sides of the signal.
and what I am doing at the moment is similar to what @Mahonroy suggested but two of them at the same time.

these were originally design ideas of Motorola (umm Freescale, nope NXP now).

images are from my notes. (i will try to find the document.if i have time)


IGBT design



mosfet design



I have tried the IGBT ones...

and software depending an what MCU you are using....




cheers Frank

 

Forget to add.

when buying SSRs use at least double or quadruple of the actual current that the load will use.
manufacturers generally suggest %50 of the current for R loads and %30 for LR loads.

I have no idea why they dont actually use the actual load current as operational current...

cheers

 

I think it's most likely due to repeated surges.

A discussion of how to implement current limit would be cool. Commercial units typically have that as a external adjustment. Many have metering as a meter or bar graph. The bar graph in recent years.

I used a lot of controllers operating into Variacs and much higher controllers did better. I had a choice of 10 and 25 A. The controllers incorporated a VERY EXPENSIVE I2t fuse which were like $25.00 each to replace. I re-wired all to use current limit and added an extra normal fuse that cost less than a $1.00. Wiring was effectively exposed inside of a vacuum chamber and shorts occasionally happened.

Later, I convinced everyone to change to a 1 RU DC power supply and a programmable temperature controller.

 

The MOS-FET you chose has a pretty high on resistance. Do you really need a 600 volt part? If not then choosing a lower voltage FET will let you get one with a lower on resistance.

The lower on resistance will help with the heat sinking of the FET but it will not help a bit with the bridge rectifier.

 

You have 120 ish and 240 ish systems.

 

As a late reply to the OP, I came across the interesting app note from PicMicro AN958 and its companion AN954.
Outlines the details of a Triac replacement for a Adj Mech Thermostat using burst firing to prevent EMI etc. Another feature, if needed is auto shut down after a prolonged period.
It uses a small 10f or 12f Pic, The App also includes code.
Max.

 

I am at a bit of a loss to understand how the sound is transmitted through the vacuum that is inside a light globe!

Sound does not travel in a vacuum. However, sound travels through other media as well. For instance, a light bulb that is making noise: the filament may be vibrating, but you're not going to hear that because of the vacuum. However, you CAN hear it because sound also travels through the wire support structure, into the glass and then resonates out to the globe.

Just had this discussion a few days ago: Does sound travel in a vacuum? Answer - depends on that medium it's traveling through. I was once working at a building site where someone was striking the concrete floor about 75 to 100 feet away. At first you heard the strike through the concrete, followed closely behind by the sound of the hammer being transmitted through the air. For a moment assume there was no air. A pure vacuum. The sound would still travel through the concrete. Granted, without air to resonate at my end I wouldn't have been able to hear it. But if a vibration sensor were mounted directly on the concrete - such as a microphone diaphragm, there WOULD be a sound.

The take-away from this is that sound can travel through the wires and the glass structure itself. Even if there's a vacuum in the bulb. But now I have a question for you: Does a light bulb actually have a perfect vacuum? Or is there some inert gas pumped in to replace it?

 

Why not consider a magnetic amplifier. Simple reactor control.

 

Why not consider a magnetic amplifier. Simple reactor control.

GREAT IDEA! Completely forgot about those weird things. Y'know, a magnetic amplifier is exactly how they control the power to the traction motors on a freight train.